Inoculation against misinformation
New
analyses for submission
Panel analysis
Goal
The objective of this memo is to present some analysis to determine whether controlling for pre-survey outcomes increases precision of treatment effects (TE) estimates of the IAM experiment (see Wiki; paper).
Takeaways
Main takeaways
- Wide vs. Panel
Misinformation and Non-misinformation sharing rates:
For the baseline linear estimations (no covariates nor fixed effects except for user-ID fixed effects), point estimates and standard errors have the same magnitude. This is obtained from comparing column (1) of each table in the “OLS analysis” section in the wide memo to the last column of the “Prediction” subsection of the “Linear analysis” section in the long memo. However, adding post-domain fixed effects in the panel analysis increases precision. More precisely, the standard errors decrease from 0.01114 (t-stat: -10.05) to 0.01107 (t-stat: -10.27) for Misinformation sharing rates and from 0.01252 (t-stat: -7.016) to 0.01227 (t-stat -7.244) for Mon-misinformation sharing rates.
Adding covariates increases precision even further, increasing the gap between wide and panel specifications, going down to 0.009287 (t-stat: -12.06) for Misinformation for the model with demographics (but no further increases in precision for also including sharing rates and tactic sharing rates as covariates) and 0.01082 (t-stat -8.545) for Non-misinformation.
Linear panel specifications with post-domain fixed effects and demographics have larger point estimates compared to linear specifications in the wide analysis. For Misinformation sharing rates, point estimates in the former range from -0.117 to -0.1174 but only -0.1154 to -0.1163 in the latter. For Non-misinformation, point estimates range from -0.09243 to -0.0933 for panel analysis and -0.0883 to -0.0927 for the wide analysis.
In sum, the panel analysis yields more precise estimates particularly when incorporating post-domain fixed effects and demographics (but there are no further gains in precision from adding sharing tactics and sharing rates as covariates). This is the case since both standard errors slightly decrease and point estimates slightly increase compared to wide analysis. This gain in precision is consequential due to the presence of marginally statistically significant results in the original set of results.
- Treatment effects estimates from Logit vs. Linear models
When comparing estimated marginal means (last column in the linear models section) vs. difference in predicted values (penultimate column in the logit section) for all possible combinations of Post-treatment and Misinfo: Point estimates from the logit models are slightly smaller to those of the linear models across all specifications (for the different 5 sets of covariates and 5 combinations of post-type/post-ID fixed effects)
t-ratios (last column for linear models) to z-ratios (logistic models) are very similar in magnitude.
This might suggest that we are in the ascending part of the sigmoid function, where logit behaves as a linear function.
In general, linear models yield a little more precise treatment effect estimates.
Secondary takeaways
Linear model:
The inclusion of User-ID fixed effects and Post-Domain fixed effects improves the precision of estimates (the standard error for the triple interaction Emotions x Post-treatment x 1[Misinfo] is reduced by ~20%, going from 0.0162 to 0.0129).
While including Post-ID fixed effects also reduces the standard errors, most of this reduction comes from incorporating information about the post-domain, rather than the combination of post-domain and post-tactic, as seen from comparing column (3) to (5) and (6).
The specification that incorporates Post-ID fixed effects yields the largest estimate of treatment effects. However, the difference in magnitude is not sizable, compared to the specification with Post-Domain fixed effects (around 4%, going from 0.1184 to 0.1226).
Logistic model:
The treatment effects from the logistic models are smaller across the board, compared to the linear model.
The standard errors of treatment effects from the logistic model (\(\approx (\hat{se}^2_{facts} + \hat{se}^2_{emotions})^\frac{1}{2}\)) are similar in magnitude to those of the linear model.
Other takeaways
- The inclusion of covariates do not increase the precision of estimates. Minimal changes to standard errors may be due to already including User-ID fixed effects.
Experimental Design
The goal of this experiment is to study the effect of different text message courses on misinformation sharing and discernment. The figure below diagrams the experimental design. We recruited participants through Facebook ads to complete a five-day text message course, plus a pre- and post-survey, for a mobile airtime payment. Participants who clicked an ad were directed to the pre-survey on Qualtrics, where they were first randomized on whether they would see an accuracy nudge in the pre- and post-survey. Conditional on completing the pre-survey, participants were randomized into one of the text message course interventions and enrolled in the course. Participants in all course interventions, except the No-course baseline, received one text message a day for five days starting on the day they completed the pre-survey. Participants in the No-course baseline received the Combo course after the post-survey to ensure we fulfilled our recruitment promise of a text message course. On the last day of the course, participants received a link to complete the post-survey on Qualtrics. Participants who completed the post-survey were paid KSH 500 (about $4 in U.S. dollars) in mobile airtime. Seven to eleven weeks later, participants who completed the post-survey were randomized on whether they would see a prime in the text message recruiting them to a follow-up survey. Participants who completed the follow-up survey were paid KSH 350 (about $3 in U.S. dollars).
knitr::include_graphics("figures/Experimental Design.png")
Experimental Design
The pre- and post-surveys each contained 10 posts:
- 3 non-misinfo posts
- 6 misinfo posts (2 emotions, 2 reasoning, 2 combo)
- 1 attention check post
These posts were created by taking 15 facts/domains, then creating a non-misinfo version and 3 misinfo versions (emotions, reasoning, combo), for a total of 60 posts (plus 2 attention check posts). Participants saw each of the 15 facts at least once.
Whichever 3 facts/domains participants saw in the form of non-misinfo in the pre-survey, they saw in the form of misinfo in the post-survey, so these 3 facts/domains were repeated. Which facts/domains were repeated vary by participant.
For each posts, participants were asked 2 questions:
- Would you share this post? (yes/no)
- To the best of your knowledge, how accurate is the claim in the above post? (4-point Likert scale)
Data
In this script, we use one datasets.
intermediate_data_long_treatments.csv: all data at the
user–post level. Includes treatment conditions, covariates, and
individual-aggregated outcomes. Sample of 8,684 participants who
completed the pre- and post-surveys.
For more information about how these datasets are generated and a data dictionary, see memo.
Sample
This script primarily uses the sample of 8,684 participants who completed the post-survey.
For analyses involving the outcome referred to as the “Primary Misinfo Sharing Rate,” we drop participants who did not share any non-misinfo posts in the pre-survey, leaving a sample of 7,688 participants who completed the post-survey and intended to share at least 1 non-misinfo post in the pre-survey.
Outcomes and Covariates
All Outcomes
| Research Question: Do the treatment interventions… | Outcome Measures | Estimands | Notes |
|---|---|---|---|
| Sharing Outcomes | |||
| change misinfo sharing? | primary misinfo sharing rate | See equation below. | Defined only in post. |
| count of misinfo posts shared \(Y^M\) | \(\frac{E[Y^{M, post}(1)]}{6} - \frac{E[Y^{M, post}(0)]}{6}\), \((\frac{E[Y^{M, post}(1)]}{6} - \frac{E[Y^{M, pre}(1)]}{6}) - (\frac{E[Y^{M, post}(0)]}{6} - \frac{E[Y^{M, pre}(0)]}{6})\) | post and post-pre misinfo sharing rate | |
| change non-misinfo sharing? | count of non-misinfo posts shared \(Y^N\) | \(\frac{E[Y^{N, post}(1)]}{3} - \frac{E[Y^{N, post}(0)]}{3}\), \((\frac{E[Y^{N, post}(1)]}{6} - \frac{E[Y^{N, pre}(1)]}{6}) - (\frac{E[Y^{N, post}(0)]}{6} - \frac{E[Y^{N, pre}(0)]}{6})\) | post and post-pre non-misinfo sharing rate |
| change sharing discernment? | count of misinfo posts shared \(Y^M\), count of non-misinfo posts shared \(Y^N\) | \((\frac{E[Y^{N, post}(1)]}{3} - \frac{E[Y^{M, post}(1)]}{6} ) - (\frac{E[Y^{N, post}(0)]}{3} - \frac{E[Y^{M, post}(0)]}{6})\) | post sharing discernment |
| Accuracy Outcomes | |||
| change accuracy scoring of misinfo posts? | total score on misinfo posts \(A^{M_1}\) | \(\frac{E[A^{M_1, post}(1)]}{6} - \frac{E[A^{M_1, post}(0)]}{6}\), \((\frac{E[A^{M_1, post}(1)]}{6} - \frac{E[A^{M_1, pre}(1)]}{6}) - (\frac{E[A^{M_1, post}(0)]}{6} - \frac{E[A^{M_1, pre}(0)]}{6})\) | post and post-pre misinfo accuracy score |
| count of misinfo posts with accuracy score > 0 \(A^{M_2}\) | \(\frac{E[A^{M_2, post}(1)]}{6} - \frac{E[A^{M_2, post}(0)]}{6}\), \((\frac{E[A^{M_2, post}(1)]}{6} - \frac{E[A^{M_2, pre}(1)]}{6}) - (\frac{E[A^{M_2, post}(0)]}{6} - \frac{E[A^{M_2, pre}(0)]}{6})\) | post and post-pre misinfo binarized accuracy score | |
| change accuracy scoring of non-misinfo posts? | total score on non-misinfo posts \(A^{N_1}\) | \(\frac{E[A^{N_1, post}(1)]}{3} - \frac{E[A^{N_1, post}(0)]}{3}\), \((\frac{E[A^{N_1, post}(1)]}{6} - \frac{E[A^{N_1, pre}(1)]}{6}) - (\frac{E[A^{N_1, post}(0)]}{6} - \frac{E[A^{N_1, pre}(0)]}{6})\) | post and post-pre non-misinfo accuracy score |
| count of non-misinfo posts with accuracy score > 0 \(A^{N_2}\) | \(\frac{E[A^{N_2, post}(1)]}{3} - \frac{E[A^{N_2, post}(0)]}{3}\), \((\frac{E[A^{N_2, post}(1)]}{6} - \frac{E[A^{N_2, pre}(1)]}{6}) - (\frac{E[A^{N_2, post}(0)]}{6} - \frac{E[A^{N_2, pre}(0)]}{6})\) | post and post-pre non-misinfo binarized accuracy score | |
| change accuracy discernment? | total score on misinfo posts \(A^M\), total score on non-misinfo posts \(A^N\) | \((\frac{E[A^{N, post}(1)]}{3} - \frac{E[A^{M, post}(1)]}{6} ) - (\frac{E[A^{N, post}(0)]}{3} - \frac{E[A^{M, post}(0)]}{6})\) | post accuracy discernment |
\[\textit{Misinfo Sharing rate}_i = \frac{\sum^3_{j=1}[S^{M,post}_i(j)|S^{N,pre}_i(j) = 1]}{\sum^{3}_{j=1}S^{N,pre}_i(j)}\]
All Covariates
Demographics
covariates_binary_fancy = c("Man",
"High school or less",
"Some college",
"Bachelor\'s degree",
"Graduate degree",
"Married",
"Unemployed",
"Employed",
"Student",
"Mostly urban","Suburban","Mostly rural",
"Christian",
"Attends religious services",
"Uses social media",
"Social media share 0-20%","Social media share 20-40%",
"Social media share 40-60%","Social media share 60-80%",
"Social media share 80-100%","Passed att. check Pre")
covariates_continuous_fancy = c("Age","Hours per day spent in social media")
covariates_all_fancy = c("Age",covariates_binary_fancy[!endsWith(covariates_binary_fancy,"%")],"Hours per day spent in social media",covariates_binary_fancy[endsWith(covariates_binary_fancy,"%")])
#,"Any post sharing rate at pre-treatment","Misinformation sharing rate at pre-treatment","Non-Misinformation sharing rate at pre-treatment")
covariates_description = matrix(NA,23,2)
covariates_description[,1] = covariates_all_fancy
covariates_description[,1][19:23] = c("Social media share 0-20%","Social media share 20-40%","Social media share 40-60%","Social media share 60-80%","Social media share 80-100%")
covariates_description[,2] = c("Integer denoting age of participants",
"Complement to 'woman' and 'other'",
"including 'Less than a high school diploma' and 'High school degree or equivalent' (reference)","including 'Some college, no degree' and 'Associate degree'","Bachelor’s degree","including 'Master’s degree' and 'Doctorate or professional degree'",
"Complement to 'Single,' 'Widowed,' 'Divorced,' and 'Separated'",
"including 'Employed full time,' 'Employed part time,' and 'Self-employed'","including 'Unemployed and currently looking for work,' 'Unemployed not currently looking for work,' 'Retired,' 'Homemaker,' and 'Unable to work'","Student (reference)",
"Live in mostly urban area","Live in suburban area","Live in mostly rural area (reference)",
"Complement to 'None,' 'Hinduism,' 'Muslim,' 'Traditionalist,' and 'Other'","Frequency of attending religious services including 'Less than once a month,' 'One to three times per month,' 'Once a week,' 'More than once a week but less than daily,' and 'Daily'; Complement to 'Never' (reference)","1 if 'Yes' and 0 if 'No' (reference)",
"1 if the user passed the attention check and 0 otherwise",
"Integer denoting how many hours are spent on average each day on social media",
"Proportion of the content the participant sees on social media that they choose to share is between 0 and 20% (reference)","Proportion of the content the participant sees on social media that they choose to share is between 20 and 40%","Proportion of the content the participant sees on social media that they choose to share is between 40 and 60%","Proportion of the content the participant sees on social media that they choose to share is between 60 and 80%","Proportion of the content the participant sees on social media that they choose to share is between 80 and 100%")
#,"Proportion of all posts shared at the pre-treatment survey","Proportion of misinfo posts shared at the pre-treatment survey","Proportion of non-misinfo posts shared at the pre-treatment survey")
knitr::kable(covariates_description, col.names = c("Covariate", "Description"),
caption = "Demographics Definitions")| Covariate | Description |
|---|---|
| Age | Integer denoting age of participants |
| Man | Complement to ‘woman’ and ‘other’ |
| High school or less | including ‘Less than a high school diploma’ and ‘High school degree or equivalent’ (reference) |
| Some college | including ‘Some college, no degree’ and ‘Associate degree’ |
| Bachelor’s degree | Bachelor’s degree |
| Graduate degree | including ‘Master’s degree’ and ‘Doctorate or professional degree’ |
| Married | Complement to ‘Single,’ ‘Widowed,’ ‘Divorced,’ and ‘Separated’ |
| Unemployed | including ‘Employed full time,’ ‘Employed part time,’ and ‘Self-employed’ |
| Employed | including ‘Unemployed and currently looking for work,’ ‘Unemployed not currently looking for work,’ ‘Retired,’ ‘Homemaker,’ and ‘Unable to work’ |
| Student | Student (reference) |
| Mostly urban | Live in mostly urban area |
| Suburban | Live in suburban area |
| Mostly rural | Live in mostly rural area (reference) |
| Christian | Complement to ‘None,’ ‘Hinduism,’ ‘Muslim,’ ‘Traditionalist,’ and ‘Other’ |
| Attends religious services | Frequency of attending religious services including ‘Less than once a month,’ ‘One to three times per month,’ ‘Once a week,’ ‘More than once a week but less than daily,’ and ‘Daily’; Complement to ‘Never’ (reference) |
| Uses social media | 1 if ‘Yes’ and 0 if ‘No’ (reference) |
| Passed att. check Pre | 1 if the user passed the attention check and 0 otherwise |
| Hours per day spent in social media | Integer denoting how many hours are spent on average each day on social media |
| Social media share 0-20% | Proportion of the content the participant sees on social media that they choose to share is between 0 and 20% (reference) |
| Social media share 20-40% | Proportion of the content the participant sees on social media that they choose to share is between 20 and 40% |
| Social media share 40-60% | Proportion of the content the participant sees on social media that they choose to share is between 40 and 60% |
| Social media share 60-80% | Proportion of the content the participant sees on social media that they choose to share is between 60 and 80% |
| Social media share 80-100% | Proportion of the content the participant sees on social media that they choose to share is between 80 and 100% |
Pre-treatment variables
| Description | Variable Construction |
|---|---|
| Pre misinfo sharing rate | \(\frac{E[Y^{M, pre}]}{6}\) |
| Pre non-misinfo sharing rate | \(\frac{E[Y^{N, pre}]}{3}\) |
| Pre sharing discernment | \(\frac{E[Y^{N, pre}(1)]}{3} - \frac{E[Y^{M, pre}(1)]}{6}\) |
| Pre misinfo accuracy score | \(\frac{E[A^{M, pre}]}{6}\) |
| Pre non-misinfo accuracy score | \(\frac{E[A^{N, pre}]}{3}\) |
| Pre accuracy discernment | \(\frac{E[A^{N, pre}(1)]}{3} - \frac{E[A^{M, pre}(1)]}{6}\) |
| Pre attention check passing | \(\mathbb{I}[\text{passed attention check}]\) |
Setup
Load data
rm(list = ls())
if (file.exists("/home/users/jreg")) {
long <- read.csv("../panel/intermediate_data_long_treatments.csv") %>% filter(pre_post != "followup")
} else {
long <- read.csv("../../main_analysis/intermediate_outcomes/intermediate_data_long_treatments.csv") %>% filter(pre_post != "followup")}
# report data as panel
panel <- pdata.frame(long, index = c('user'))
# Define post
panel <- panel %>% mutate(post = ifelse(pre_post == "post",1, ifelse(pre_post == "pre",0,NA)))
panel$treatment <- factor(panel$treatment, levels = c('Facts Baseline','No-course Baseline','Emotions','Reasoning','Combo'))
# Add post-id
panel <- panel %>% mutate(post_id = as.integer(factor(interaction(fact, type))))
rm(long)Standard errors
robust_se <- function(model) {
sqrt(diag(vcovHC(model, type = "HC1")))
}
cluster_se <- function(model, cluster_var) {
cluster_var <- as.factor(cluster_var)
vcov_cluster <- vcovHC(model, type = "HC1", cluster = cluster_var)
se_cluster <- sqrt(diag(vcov_cluster))
return(se_cluster)
}
se_cont <- function(x, na.rm=FALSE) {
if (na.rm) x <- na.omit(x)
sqrt(var(x)/length(x))
}Remove intercept
remove_intercept <- function(model) {
coef <- coef(model)
se <- sqrt(diag(vcov(model)))
names(coef) <- names(se) <- setdiff(names(coef), "(Intercept)")
texreg::createTexreg(coef = coef, se = se, gof = numeric(), gof.names = NULL, gof.decimal = NULL)
}t-stat extraction function
extract_tstats <- function(model_list) {
tstats_list <- lapply(model_list, function(model) {
summary_model <- summary(model)
tstats <- summary_model$coefficients[, "t value"]
return(tstats)
})
return(tstats_list)
}Multicolumn header function
create_multicolumn_headers <- function(headers,ncols) {
modified_headers <- c("<td colspan='1' style='text-align:center;'></td>") # first column, which is not multicolumn
# Loop through each header and modify it for multicolumn
for (header in headers) {
modified_header <- sprintf(paste0("<td colspan='",ncols,"' style='text-align:center;'>%s</td>"), header)
modified_headers <- c(modified_headers, modified_header)
}
# Add an empty header for the last column if needed
if (length(headers) * 2 + 1 < length(modified_headers)) {
modified_headers <- c(modified_headers, "<td colspan='1' style='text-align:center;'></td>")
}
modified_headers <- paste0("<tr>",paste(modified_headers, collapse = ""),"</tr>")
return(modified_headers)
}
create_multicolumn_headers <- function(headers, ncols) {
modified_headers <- c() # Initialize an empty vector to store headers
# Loop through each header and modify it for multicolumn
for (header in headers) {
modified_header <- sprintf("<td colspan='%d' style='text-align:center;'>%s</td>", ncols, header)
modified_headers <- c(modified_headers, modified_header)
}
return(modified_headers)
}Define set of covariates
# Define covariates
# sharing at pre
panel$share_pre_cov <- panel$share_pre
panel$misinfo_pre_cov <- panel$misinfo_pre
panel$base_rate_pre_cov <- panel$base_rate_pre
# tactics
# Base posts already included (base_rate_pre)
combo_pre <- aggregate(share ~ user, data = panel %>% filter(pre_post == "pre" & type == "Combo posts"), mean)
panel <- panel %>% left_join(combo_pre %>% rename(combo_pre = share),
by = "user", suffix = c("",""))
emotions_pre <- aggregate(share ~ user, data = panel %>% filter(pre_post == "pre" & type == "Emotions posts"), mean)
panel <- panel %>% left_join(emotions_pre %>% rename(emotions_pre = share),
by = "user", suffix = c("",""))
reasoning_pre <- aggregate(share ~ user, data = panel %>% filter(pre_post == "pre" & type == "Reasoning posts"), mean)
panel <- panel %>% left_join(reasoning_pre %>% rename(reasoning_pre = share),
by = "user", suffix = c("",""))
# covariates
covariates <- c("age","gender_Man", "education_High_school_or_less",
"education_Some_college","education_Bachelor_degree","education_Graduate_degree",
"marital_Married_or_in_a_domestic_partnership",
"employment_Unemployed","employment_Employed","employment_Student",
"location_Mostly_urban","location_Suburban","location_Mostly_rural",
"religion_Christian","religiosity_Attends",
"social_media_bin_Yes","social_media_hours",
"social_media_share_80_100",
"social_media_share_60_80",
"social_media_share_40_60",
"social_media_share_20_40",
"social_media_share_0_20",
"att_check_pre")
covariates_sharing <- c("misinfo_pre","base_rate_pre")
covariates_tactic <- c('emotions_pre','reasoning_pre','combo_pre')
# demean
panel <- panel %>%
mutate(across(.cols = all_of(covariates),
.fns = ~ scale(., center = TRUE, scale = TRUE),
.names = "{.col}_cov"))
panel <- panel %>%
mutate(across(.cols = all_of(covariates_sharing),
.fns = ~ scale(., center = TRUE, scale = TRUE),
.names = "{.col}_cov"))
panel <- panel %>%
mutate(across(.cols = all_of(covariates_tactic),
.fns = ~ scale(., center = TRUE, scale = TRUE),
.names = "{.col}_cov"))
# list of covariates
covariates_demeaned <- c("age_cov","gender_Man_cov",
#"education_High_school_or_less",
"education_Some_college_cov",
"education_Bachelor_degree_cov","education_Graduate_degree_cov",
"marital_Married_or_in_a_domestic_partnership_cov",
"employment_Unemployed_cov","employment_Employed_cov",
#"employment_Student",
"location_Mostly_urban_cov","location_Suburban_cov",
#"location_Mostly_rural",
"religion_Christian_cov","religiosity_Attends_cov",
"social_media_bin_Yes_cov","social_media_hours_cov",
"social_media_share_80_100_cov","social_media_share_60_80_cov",
"social_media_share_40_60_cov","social_media_share_20_40_cov",
#"social_media_share_0_20",
"att_check_pre_cov")
sharing_demeaned <- c("misinfo_pre_cov","base_rate_pre_cov")
tactic_demeaned <- c('emotions_pre_cov','reasoning_pre_cov','combo_pre_cov')
# demographics
demographics_demeaned_treatment <- paste0(covariates_demeaned,'*treatment',collapse = "+")
# pre non-misinfo and pre-misinfo sharing
sharing_demeaned_treatment <- paste0(sharing_demeaned,'*treatment',collapse = "+")
# pre non-misinfo, pre-misinfo sharing and tactic sharing rates
sharing_tactic_demeaned_treatment <- paste0(c(sharing_demeaned,tactic_demeaned),'*treatment',collapse = "+")
# demographics and pre-non-misinfo and pre-misinfo sharing
demographics_sharing_demeaned_treatment <- paste0(c(covariates_demeaned,sharing_demeaned),'*treatment',collapse = "+")
# all
all_covariates_demeaned_treatment <- paste0(c(covariates_demeaned,sharing_demeaned,tactic_demeaned),'*treatment',collapse = "+")Variable labels
generate_variable_mapping <- function(treatments, post_domains, tactics, covariates) {
mapping <- list()
# Basic variables
basic_vars <- c("post_is_misinfo" = "1[Misinfo]",
"post" = "1[Post-treatment]",
"post:post_is_misinfo" = "1[Post-treatment] x 1[Misinfo]",
"(Intercept)" = "Facts Baseline") # intercept only used as Facts Baseline
# Covariates
cov_vars <- c("age" = "Age",
"gender_Man" = "Man",
"education_High_school_or_less" = "High school or less",
"education_Some_college" = "Some college",
"education_Bachelor_degree" = "Bachelor\'s degree",
"education_Graduate_degree" = "Graduate degree",
"marital_Married_or_in_a_domestic_partnership" = "Married",
"employment_Unemployed" = "Unemployed",
"employment_Employed" = "Employed",
"employment_Student" = "Student",
"location_Mostly_urban" = "Mostly urban",
"location_Suburban" = "Suburban",
"location_Mostly_rural" = "Mostly rural",
"religion_Christian" = "Christian",
"religiosity_Attends" = "Attends religious services",
"social_media_bin_Yes" = "Uses social media",
"social_media_share_0_20" = "Social Media Share: 0-20%",
"social_media_share_20_40" = "Social Media Share: 20-40%",
"social_media_share_40_60" = "Social Media Share: 40-60%",
"social_media_share_60_80" = "Social Media Share: 60-80%",
"social_media_share_80_100" = "Social Media Share: 80-100%",
"social_media_hours" = "Hours spent on Social Media",
"misinfo_pre_cov" = "Avg. Sharing of Misinfo posts at pre-treatment",
"base_rate_pre_cov" = "Avg. Sharing of Non-misinfo posts at pre-treatment",
"att_check_pre" = "Passed attention check at pre-treatment",
"emotions_pre" = "Emotions posts",
"reasoning_pre" = "Reasoning posts",
"combo_pre" = "Combo posts",
"age_cov" = "Age",
"gender_Man_cov" = "Man",
"education_High_school_or_less_cov" = "High school or less",
"education_Some_college_cov" = "Some college",
"education_Bachelor_degree_cov" = "Bachelor\'s degree",
"education_Graduate_degree_cov" = "Graduate degree",
"marital_Married_or_in_a_domestic_partnership_cov" = "Married",
"employment_Unemployed_cov" = "Unemployed",
"employment_Employed_cov" = "Employed",
"employment_Student_cov" = "Student",
"location_Mostly_urban_cov" = "Mostly urban",
"location_Suburban_cov" = "Suburban",
"location_Mostly_rural_cov" = "Mostly rural",
"religion_Christian_cov" = "Christian",
"religiosity_Attends_cov" = "Attends religious services",
"social_media_bin_Yes_cov" = "Uses social media",
"social_media_share_0_20_cov" = "Social Media Share: 0-20%",
"social_media_share_20_40_cov" = "Social Media Share: 20-40%",
"social_media_share_40_60_cov" = "Social Media Share: 40-60%",
"social_media_share_60_80_cov" = "Social Media Share: 60-80%",
"social_media_share_80_100_cov" = "Social Media Share: 80-100%",
"social_media_hours_cov" = "Hours spent on Social Media",
"misinfo_pre_cov" = "Avg. Sharing of Misinfo posts at pre-treatment",
"base_rate_pre_cov" = "Avg. Sharing of Non-misinfo posts at pre-treatment",
"att_check_pre_cov" = "Passed attention check at pre-treatment",
"emotions_pre_cov" = "Emotions posts",
"reasoning_pre_cov" = "Reasoning posts",
"combo_pre_cov" = "Combo posts")
# post characteristics
posts_vars <- c(
"picture" = "1[Post has a picture]",
"exclamation" = "1[Post includes exclamation(s) mark]",
"hashtag" = "1[Post includes hashtag(s)]",
"url" = "1[Post includes url(s)]",
"tags" = "1[Post includes tag(s)]",
"caps" = "1[Post includes word(s) in capital letters]",
"emojis" = "1[Post includes emoji(s)]",
"oneself" = "1[Post is written in first person]",
"insult" = "1[Post includes an insult]",
"references" = "1[Post makes a reference to an external source/data]",
"statistic" = "1[Post references a statistic]",
"gender_man" = "1[Post author is a man]",
"avatar_human" = "1[Post author's avatar is a human]",
"num_exclamation" = "Number of exclamation marks",
"num_hashtag" = "Number of hashtags",
"num_emojis" = "Number of emojis",
"length_words" = "Number of words in post",
"length_characters" = "Number of characters in post"
)
# Manipulation tactics:
manipulation_vars = c("factor(type)Combo posts" = "1[Combo manipulation tactic]",
"factor(type)Emotions posts" = "1[Emotions manipulation tactic]",
"factor(type)Reasoning posts" = "1[Reasoning manipulation tactic]",
"factor(type)Base posts" = "1[Non-misinfo tactic]")
# Treatment variables
for (treatment in treatments) {
mapping[paste0("treatment", treatment)] <- treatment
mapping[paste0("factor(treatment)", treatment)] <- treatment
mapping[paste0("treatment", treatment, ":post")] <- paste(treatment, "x 1[Post-treatment]")
mapping[paste0("factor(treatment)", treatment, ":post")] <- paste(treatment, "x 1[Post-treatment]")
mapping[paste0("treatment", treatment,":post_is_misinfo")] <- paste(treatment, "x 1[Misinfo]")
mapping[paste0("post_is_misinfo:treatment", treatment)] <- paste(treatment, "x 1[Misinfo]")
mapping[paste0("post_is_misinfo:treatment", treatment, ":post")] <- paste(treatment, "x 1[Misinfo] x 1[Post-treatment]")
mapping[paste0("treatment", treatment, ":post:post_is_misinfo")] <- paste(treatment, "x 1[Post-treatment] x 1[Misinfo]")
}
# Post domain variables
for (i in 1:length(post_domains)) {
domain <- post_domains[i]
mapping[paste0("factor(fact)", i)] <- paste("Post domain", i)
mapping[paste0("factor(fact)", i, ":post")] <- paste("Post domain", i, "x 1[Post-treatment]")
mapping[paste0("post:factor(fact)", i)] <- paste("Post domain", i, "x 1[Post-treatment]")
mapping[paste0("post_is_misinfo:factor(fact)", i)] <- paste("Post domain", i, "x 1[Misinfo]")
mapping[paste0("factor(fact)", i, ":factor(post)1")] <- paste("Post domain", i, "x 1[Post-treatment]")
mapping[paste0("post:post_is_misinfo:factor(fact)", i)] <- paste("Post domain", i, "x 1[Post-treatment] x 1[Misinfo]")
# Interaction with treatments
for (treatment in treatments) {
mapping[paste0("factor(fact)", i, ":treatment", treatment)] <- paste("Post domain", i, "x", treatment)
mapping[paste0("treatment", treatment, ":factor(fact)", i)] <- paste("Post domain", i, "x", treatment)
mapping[paste0("factor(fact)", i, ":factor(treatment)", treatment)] <- paste("Post domain", i, "x", treatment)
mapping[paste0("factor(fact)", i, ":treatment", treatment, ":post")] <- paste("Post domain", i, "x", treatment, "x 1[Post-treatment]")
mapping[paste0("factor(fact)", i, ":factor(treatment)", treatment, ":factor(post)1")] <- paste("Post domain", i, "x", treatment, "x 1[Post-treatment]")
mapping[paste0("treatment", treatment, ":post:factor(fact)", i)] <- paste("Post domain", i, "x", treatment, "x 1[Post-treatment]")
mapping[paste0("treatment", treatment, ":post_is_misinfo:factor(fact)", i)] <- paste("Post domain", i, "x", treatment, "x 1[Misinfo]")
mapping[paste0("treatment", treatment, ":post:post_is_misinfo:factor(fact)", i)] <- paste("Post domain", i, "x", treatment, "x 1[Post-treatment] x 1[Misinfo]")
}
for (tactic in tactics) {
mapping[paste0("factor(fact)", i, ":factor(type)", tactic)] <- paste("Post domain", i, "x", manipulation_vars[tactic])
mapping[paste0("factor(fact)", i, ":factor(type)", tactic)] <- paste("Post domain", i, "x", manipulation_vars[paste0("factor(type)",tactic)])
}
}
# Post ID
for (i in 1:length(post_ids)) {
mapping[paste0("factor\\\\(post\\\\_id\\\\)", i)] <- paste("Post", i,"indicator") # not working
}
post_id_vars <- c(
"factor(post_id)1" = "Post 1 indicator",
"factor(post_id)2" = "Post 2 indicator",
"factor(post_id)3" = "Post 3 indicator",
"factor(post_id)4" = "Post 4 indicator",
"factor(post_id)5" = "Post 5 indicator",
"factor(post_id)6" = "Post 6 indicator",
"factor(post_id)7" = "Post 7 indicator",
"factor(post_id)8" = "Post 8 indicator",
"factor(post_id)9" = "Post 9 indicator",
"factor(post_id)10" = "Post 10 indicator",
"factor(post_id)11" = "Post 11 indicator",
"factor(post_id)12" = "Post 12 indicator",
"factor(post_id)13" = "Post 13 indicator",
"factor(post_id)14" = "Post 14 indicator",
"factor(post_id)15" = "Post 15 indicator",
"factor(post_id)16" = "Post 16 indicator",
"factor(post_id)17" = "Post 17 indicator",
"factor(post_id)18" = "Post 18 indicator",
"factor(post_id)19" = "Post 19 indicator",
"factor(post_id)20" = "Post 20 indicator",
"factor(post_id)21" = "Post 21 indicator",
"factor(post_id)22" = "Post 22 indicator",
"factor(post_id)23" = "Post 23 indicator",
"factor(post_id)24" = "Post 24 indicator",
"factor(post_id)25" = "Post 25 indicator",
"factor(post_id)26" = "Post 26 indicator",
"factor(post_id)27" = "Post 27 indicator",
"factor(post_id)28" = "Post 28 indicator",
"factor(post_id)29" = "Post 29 indicator",
"factor(post_id)30" = "Post 30 indicator",
"factor(post_id)31" = "Post 31 indicator",
"factor(post_id)32" = "Post 32 indicator",
"factor(post_id)33" = "Post 33 indicator",
"factor(post_id)34" = "Post 34 indicator",
"factor(post_id)35" = "Post 35 indicator",
"factor(post_id)36" = "Post 36 indicator",
"factor(post_id)37" = "Post 37 indicator",
"factor(post_id)38" = "Post 38 indicator",
"factor(post_id)39" = "Post 39 indicator",
"factor(post_id)40" = "Post 40 indicator",
"factor(post_id)41" = "Post 41 indicator",
"factor(post_id)42" = "Post 42 indicator",
"factor(post_id)43" = "Post 43 indicator",
"factor(post_id)44" = "Post 44 indicator",
"factor(post_id)45" = "Post 45 indicator",
"factor(post_id)46" = "Post 46 indicator",
"factor(post_id)47" = "Post 47 indicator",
"factor(post_id)48" = "Post 48 indicator",
"factor(post_id)49" = "Post 49 indicator",
"factor(post_id)50" = "Post 50 indicator",
"factor(post_id)51" = "Post 51 indicator",
"factor(post_id)52" = "Post 52 indicator",
"factor(post_id)53" = "Post 53 indicator",
"factor(post_id)54" = "Post 54 indicator",
"factor(post_id)55" = "Post 55 indicator",
"factor(post_id)56" = "Post 56 indicator",
"factor(post_id)57" = "Post 57 indicator",
"factor(post_id)58" = "Post 58 indicator",
"factor(post_id)59" = "Post 59 indicator",
"factor(post_id)60" = "Post 60 indicator"
)
# Covariate variables
for (covariate in covariates) {
mapping[paste0("post:", covariate)] <- paste("Post-treatment x", cov_vars[covariate])
mapping[paste0("post_is_misinfo:", covariate)] <- paste("1[Misinfo] x", cov_vars[covariate])
mapping[paste0("post:post_is_misinfo:", covariate)] <- paste("Post-treatment x 1[Misinfo] x", cov_vars[covariate])
# Interaction with treatments
for (treatment in treatments) {
mapping[paste0("treatment", treatment, ":", covariate)] <- paste(treatment, "x", cov_vars[covariate])
mapping[paste0("factor(treatment)", treatment, ":", covariate)] <- paste(treatment, "x", cov_vars[covariate])
mapping[paste0("treatment", treatment, ":post:", covariate)] <- paste(treatment, "x 1[Post] x", cov_vars[covariate])
mapping[paste0("treatment", treatment, ":post_is_misinfo:", covariate)] <- paste(treatment, "x 1[Misinfo] x", cov_vars[covariate])
mapping[paste0("treatment", treatment, ":post:post_is_misinfo:", covariate)] <- paste(treatment, "x 1[Post] x 1[Misinfo] x", cov_vars[covariate])
}
# Interaction with domains
for (i in 1:length(post_domains)) {
mapping[paste0("factor(fact)",i,":",covariate)] <- paste("Post domain",i,"x", cov_vars[covariate])
mapping[paste0(covariate,":factor(fact)",i)] <- paste("Post domain",i,"x", cov_vars[covariate])
}
# Interaction with post
for (i in 1:length(post_domains)) {
mapping[paste0("factor(fact)",i,":",covariate)] <- paste("Post domain",i,"x", cov_vars[covariate])
mapping[paste0(covariate,":factor(fact)",i)] <- paste("Post domain",i,"x", cov_vars[covariate])
}
# Interaction with misinfo
}
# Post attributes
for (attribute in posts_vars) {
# Interaction with treatments
for (treatment in treatments) {
mapping[paste0("treatment", treatment, ":", attribute)] <- paste(treatment, "x", posts_vars[attribute])
mapping[paste0("treatment", treatment, ":post:", attribute)] <- paste(treatment, "x 1[Post-treatment] x ", posts_vars[attribute])
mapping[paste0("treatment", treatment, ":post_is_misinfo:", attribute)] <- paste(treatment, "x 1[Misinfo] x ", posts_vars[attribute])
mapping[paste0("treatment", treatment, ":post:post_is_misinfo:", attribute)] <- paste(treatment, "x 1[Post-treatment] x 1[Misinfo] x ", posts_vars[attribute])
}
for (i in 1:length(post_domains)) {
mapping[paste0(attribute,":factor(fact)", i)] <- paste("Post domain", i,posts_vars[attribute])
mapping[paste0("factor(fact)", i,attribute)] <- paste("Post domain", i,posts_vars[attribute])
mapping[paste0("post:",attribute,":","factor(fact)", i)] <- paste("Post domain", i,"x 1[Post-treatment] x",posts_vars[attribute])
mapping[paste0("post:factor(fact)",i,":",attribute)] <- paste("Post domain", i,"x 1[Post-treatment] x",posts_vars[attribute])
mapping[paste0("post_is_misinfo:factor(fact)",i,":",attribute)] <- paste("Post domain", i,"x 1[Misinfo] x",posts_vars[attribute])
for (treatment in treatments) {
mapping[paste0("treatment", treatment, ":", attribute,":factor(fact)",i)] <- paste("Post domain", i,"x",treatment,"x",posts_vars[attribute])
mapping[paste0("treatment", treatment, ":factor(fact)",i,":",attribute)] <- paste("Post domain", i,"x",treatment,"x",posts_vars[attribute])
mapping[paste0("treatment", treatment, ":post:",attribute,":factor(fact)",i)] <- paste("Post domain", i,"x",treatment,"x 1[Post-treatment]",posts_vars[attribute])
mapping[paste0("treatment", treatment, ":post:factor(fact)",i,":",attribute)] <- paste("Post domain", i,"x",treatment,"x 1[Post-treatment]",posts_vars[attribute])
mapping[paste0("treatment", treatment, ":post_is_misinfo:factor(fact)",i,":",attribute)] <- paste("Post domain", i,"x",treatment,"x 1[Misinfo]",posts_vars[attribute])
}
}
mapping[paste0("post:", attribute)] <- paste("Post-treatment x", posts_vars[attribute])
mapping[paste0("post_is_misinfo:", attribute)] <- paste("1[Misinfo] x", posts_vars[attribute])
}
# random cases not working
rndm_vars <- c(
"religiosity_Attends:factor(fact)3" = "Post domain 3 x Religiosity: Attends",
"religiosity_Attends:factor(fact)4" = "Post domain 4 x Religiosity: Attends",
"religiosity_Attends:factor(fact)5" = "Post domain 5 x Religiosity: Attends",
"religiosity_Attends:factor(fact)6" = "Post domain 6 x Religiosity: Attends",
"social_media_share_40_60:factor(fact)10" = "Post domain 10 x Social Media Share: 40-60%",
"social_media_share_40_60:factor(fact)11" = "Post domain 11 x Social Media Share: 40-60%",
"social_media_share_40_60:factor(fact)12" = "Post domain 12 x Social Media Share: 40-60%",
"social_media_share_40_60:factor(fact)13" = "Post domain 13 x Social Media Share: 40-60%",
"factor(post_id)19" = "Post 19 indicator",
"factor(post_id)20" = "Post 20 indicator",
"factor(post_id)21" = "Post 21 indicator"
)
return(c(mapping,basic_vars,cov_vars,manipulation_vars,post_id_vars,rndm_vars,posts_vars))
}
treatments <- c("No-course Baseline", "Reasoning", "Emotions", "Combo")
tactics <- c("Emotions posts","Combo posts","Reasoning posts","Base posts")
post_domains <- as.numeric(unique(panel$fact))
post_ids <- as.numeric(unique(panel$post_id))
covariates <- c("age_cov", "gender_Man_cov",
"education_Some_college_cov", "education_Bachelor_degree_cov",
"education_Graduate_degree_cov", "marital_Married_or_in_a_domestic_partnership_cov",
"employment_Unemployed_cov", "employment_Employed_cov", "location_Mostly_urban_cov",
"location_Suburban_cov", "religion_Christian_cov", "religiosity_Attends_cov",
"social_media_bin_Yes_cov", "social_media_hours_cov", "social_media_share_80_100_cov",
"social_media_share_60_80_cov", "social_media_share_40_60_cov",
"social_media_share_20_40_cov",
"misinfo_pre_cov", "base_rate_pre_cov","att_check_pre_cov",
"emotions_pre_cov","reasoning_pre_cov","combo_pre_cov")
variable_mapping <- generate_variable_mapping(treatments, post_domains, tactics, covariates)
variable_mapping_long <- as.character(variable_mapping)
names(variable_mapping_long) <- names(variable_mapping)# function to subset model for short tables
subset_model <- function(model, coeffs) {
new_model <- model
new_model$coefficients <- na.omit(model$coefficients[names(model$coefficients) %in% coeffs])
if (length(new_model$coefficients) == 0) {
new_model$coefficients <- na.omit(model$coefficients[names(model$coefficients[,1]) %in% coeffs])
}
# Std. errors
if (length(new_model$se) > 0) {
new_model$se <- model$se[names(model$coefficients) %in% coeffs]
if (length(new_model$se) == 0) {
new_model$se <- na.omit(model$se[names(model$coefficients[,1]) %in% coeffs])
}
}
return(new_model)
}
Balance tests
The following figures present balance in exposure to each post domain and post ID by treatment arm.
Clicking on the button below “User covariates” displays a section showing figures for balance by user characteristics.
# Plots standardized differences by treatment #
match_plot_cont <- function(data,title,legend_title){
num_grouping <- length(unique(data$grouping)) + 1
data$grouping <- factor(as.numeric(data$grouping), levels = (1:num_grouping))
colors <- c(RColorBrewer::brewer.pal(12, "Set3"),
RColorBrewer::brewer.pal(8, "Set2"),
RColorBrewer::brewer.pal(9, "Set1"))
colors <- rep(colors,10) # 290 colors
color_values <- setNames(colors, 1:length(colors))
shapes <- rep(21:25, length.out = num_grouping)
shape_values <- setNames(shapes, 1:length(shapes))
pic <- ggplot(data = data, aes(x = diff, y = factor(covariates, levels = unique(covariates)), group = grouping)) +
theme_bw()+
theme(axis.line.y = element_line(colour="black"),axis.line.x = element_line(colour="black"),
panel.border = element_blank(),
panel.grid.major = element_blank(), panel.grid.minor = element_blank(),
legend.title = element_text(face='bold', size=35, hjust=0.5, vjust=0.5),
legend.position = c(0.95,0.50),legend.justification = c("right", "bottom"),
legend.key = element_rect(colour = "transparent"),
legend.box.just = "right", legend.text = element_text(size=30), legend.margin = margin(0, 6, 6, 6),
legend.box.background = element_rect( fill="transparent", size=1),legend.background = element_blank()) +
geom_vline(aes(xintercept=-0.1),color="black", linetype="longdash", size=0.75) +
geom_vline(aes(xintercept=0.1),color="black", linetype="longdash", size=0.75) +
geom_vline(aes(xintercept=0.25),color="black", linetype="dashed", size=0.5) +
geom_vline(aes(xintercept=0), color="black", linetype="solid", size=0.5) +
geom_vline(aes(xintercept=-0.25),color="black", linetype="dashed", size=0.5) +
geom_vline(aes(xintercept=-0.5),color="black", linetype="dotted", size=0.5) +
geom_vline(aes(xintercept=0.5),color="black", linetype="dotted", size=0.5) +
geom_point(aes(color = grouping, fill = grouping, shape = grouping), size = 8) +
scale_x_continuous(breaks = c(-0.5, -0.25, -0.10, 0, 0.10, 0.25, 0.5),
labels = c('-0.50', '-0.25', '-0.10', '0', '0.10', '0.25', '0.50')) +
scale_color_manual(name = paste0("vs Post ",legend_title), values = color_values) +
scale_fill_manual(name = paste0("vs Post ",legend_title), values = color_values) +
scale_shape_manual(name = paste0("vs Post ",legend_title), values = shape_values) +
labs(y="Variable",x="\nStandardized Difference")+
theme(axis.text.x = element_text(color = "black", size = 30, angle = 0, hjust = .5, vjust = 0, face = "plain"),
axis.text.y = element_text(color = "black", size = 30, angle = 0, hjust = 1, vjust = .5, face = "plain",
margin=unit(rep(0.5,4),"cm")),
axis.title.x = element_text(color = "black", size = 35, angle = 0, hjust = .5, vjust = 0, face = "bold"),
axis.title.y = element_text(color = "black", size = 35, angle = 90, hjust = .5, vjust = .5, face = "bold"),
axis.ticks.length.y = unit(-0.25,"cm"), axis.ticks.x=element_blank()) +
ggtitle(title) +
theme(plot.title = element_text(face='bold', size=30, hjust=0.5, vjust=0.5)) +
facet_grid(rows = vars(cov_group),
scales = "free_y", # Let the x axis vary across facets.
space = "free_y", # Let the width of facets vary and force all bars to have the same width.
switch = "y")+
theme(strip.placement = "outside", # Place facet labels outside x axis labels.
strip.background = element_blank(), # Make facet label background white.
strip.text.y.left = element_text(size = 30,face = "bold",angle = 0, hjust=0.5),
axis.title = element_blank(),
panel.border = element_rect(color = "grey70", fill = NA, size = 2))
pic
}Post domain
Treatment arms
Balance in exposure to each post domain by treatment arm.
# define binary indicators for facts
panel$fact <- as.factor(factor(panel$fact, levels = (1:length(unique(panel$fact)))))
dummy_vars_domains <- model.matrix(~ fact - 1, data = panel)
data_domains <- as.data.frame(dummy_vars_domains)
data_covs <- cbind(data_covs, data_domains)
# define all_domains for test
all_domains <- c(paste0("fact",(1:length(unique(panel$fact)))))
all_domains_fancy <- c(paste0("Post domain ",(1:length(unique(panel$fact)))))
cov_group <- c("")knitr::include_graphics("balance_treat_domains.png")
Post ID
Treatment arms
Balance in exposure to each post by treatment arm.
# define binary indicators for facts
panel$post_id <- as.factor(factor(panel$post_id, levels = (1:length(unique(panel$post_id)))))
dummy_vars_post_id <- model.matrix(~ post_id - 1, data = panel)
data_post_id <- as.data.frame(dummy_vars_post_id)
data_covs <- cbind(data_covs, data_post_id)
# define all_post_id for test
all_post_id <- c(paste0("post_id",(1:length(unique(panel$post_id)))))
cov_group <- rep(unique(panel$type),length(unique(data_covs$fact)))
cov_group <- gsub("Base","Non-misinfo",cov_group)
all_post_id_fancy <- rep(unique(panel$type), each = length(unique(data_covs$fact)))knitr::include_graphics("balance_treat_post_id.png")
Variance decomposition
Within and between participant standard deviation and variance of variable \(\textit{share}_{ipt}\).
Within: standard deviation and variance of \(x_{ip} - \bar{x_i} + \bar{\bar{x}}\)
Between: standard deviation and variance of: \(\bar{x_i}\)
Where \(i\) is the participant, \(p\) is a post, \(\bar{x_i}\) is the between-participant mean, and \(\bar{\bar{x}}\) is the global mean.
Although we present all treatments, the focus should be on
No-course Baseline since estimates for users in such
treatment arm are not conflated with treatment effects.
Observations:
- While the within-user variance on misinfo is higher than on non-misinfo, the opposite is true for between-user: the variance on non-misinfo is slightly higher than on misinfo.
# data <- filtered_data
# treatment <- t
# period <- p
# share_type <- s
# Function to create a variance summary table
variance_summary_table <- function(data, treatment, period, share_type) {
# Filter data based on the period and share type
if (period != "pre and post") {
data <- data %>% filter(pre_post == period)
}
if (share_type == "misinfo share") {
data <- data %>% filter(post_is_misinfo == 1)
} else if (share_type == "non-misinfo share") {
data <- data %>% filter(post_is_misinfo == 0)
}
# Define data as panel
data <- pdata.frame(data, index = c('user'))
within_sd <- formatC(xtsum::within_sd(data, 'share', 'user'), format = "f", digits = 4)
within_variance <- formatC(xtsum::within_sd(data, 'share', 'user')^2, format = "f", digits = 4)
between_sd <- formatC(xtsum::between_sd(data, 'share', 'user'), format = "f", digits = 4)
between_variance <- formatC(xtsum::between_sd(data, 'share', 'user')^2, format = "f", digits = 4)
summary_table <- data.frame(
Outcome = share_type,
Treatment = treatment,
Period = period,
Within_user_Standard_Deviation = within_sd,
Within_user_Variance = within_variance,
Between_user_Standard_Deviation = between_sd,
Between_user_Variance = between_variance
)
return(summary_table)
}
# Initialize an empty list to store the variance summary tables
variance_summary_tables <- list()
treatments <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre", "post", "pre and post")
share_types <- c("share", "misinfo share", "non-misinfo share")
# Loop over each combination of treatment, period, and share type
for (t in treatments) {
# Empty dataframe to store all combinations for the current treatment
all_combinations <- NULL
for (p in periods) {
for (s in share_types) {
if (t != "All") {
# Filter the data for the current treatment
filtered_data <- panel %>% filter(treatment == t)
} else {
filtered_data <- panel
}
# Create a variance summary table for the current combination
variance_summary <- variance_summary_table(filtered_data, t, p, s)
# Store the variance summary table in all_combinations dataframe
all_combinations <- rbind(all_combinations, variance_summary)
}
}
# Store the all_combinations dataframe in the list with treatment name as key
variance_summary_tables[[t]] <- all_combinations
}
# Initialize an empty list to store tab panels
variance_tab_panels <- list()
# Loop through each key in variance_summary_tables
for (treatment in names(variance_summary_tables)) {
tab_name <- paste0(treatment)
tab_content <- DT::datatable(variance_summary_tables[[treatment]], options = list(pageLength = 9),
rownames = FALSE, colnames = c("Outcome","Treatment","Period",
"Within-user Std. Dev.", "Within-user Variance",
"Between-user Std. Dev.", "Between-user Variance"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
# Create the tabset with the variance tab panels
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
# Render the tabset
htmltools::tagList(variance_tabset)Click here to see variance decomposition by tactic
Observations:
- Within tactic, the within-user variance becomes lower than the between-user variance.
# Function to create a variance summary table
variance_summary_table <- function(data, treatment, period, share_type) {
# Filter data based on the period
if (period != "pre and post") {
data <- data %>% filter(pre_post == period)
}
# Filter data based on the type
if (share_type == "non-misinfo share") {
data <- data %>% filter(post_is_misinfo == 0)
} else if (share_type == "misinfo share") {
data <- data %>% filter(post_is_misinfo == 1)
} else if (share_type == "Emotions posts") {
data <- data %>% filter(type == "Emotions posts")
} else if (share_type == "Reasoning posts") {
data <- data %>% filter(type == "Reasoning posts")
} else if (share_type == "Combo posts") {
data <- data %>% filter(type == "Combo posts")
}
# Define data as panel
data <- pdata.frame(data, index = c('user'))
within_sd <- formatC(xtsum::within_sd(data, 'share', 'user'), format = "f", digits = 4)
within_variance <- formatC(xtsum::within_sd(data, 'share', 'user')^2, format = "f", digits = 4)
between_sd <- formatC(xtsum::between_sd(data, 'share', 'user'), format = "f", digits = 4)
between_variance <- formatC(xtsum::between_sd(data, 'share', 'user')^2, format = "f", digits = 4)
summary_table <- data.frame(
Outcome = share_type,
Treatment = treatment,
Period = period,
Within_user_Standard_Deviation = within_sd,
Within_user_Variance = within_variance,
Between_user_Standard_Deviation = between_sd,
Between_user_Variance = between_variance
)
return(summary_table)
}
# Initialize an empty list to store the variance summary tables
variance_summary_tables <- list()
treatments <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre", "post", "pre and post")
share_types <- c("share", "misinfo share", "non-misinfo share", "Emotions posts", "Reasoning posts", "Combo posts")
facts <- seq(1:15)
# Loop over each combination of treatment, period, and share type
for (t in treatments) {
# Empty dataframe to store all combinations for the current treatment
all_combinations <- NULL
for (p in periods) {
for (s in share_types) {
if (t != "All") {
# Filter the data for the current treatment
filtered_data <- panel %>% filter(treatment == t)
} else {
filtered_data <- panel
}
# Create a variance summary table for the current combination
variance_summary <- variance_summary_table(filtered_data, t, p, s)
# Store the variance summary table in all_combinations dataframe
all_combinations <- rbind(all_combinations, variance_summary)
}
}
# Store the all_combinations dataframe in the list with treatment name as key
variance_summary_tables[[t]] <- all_combinations
}
# Initialize an empty list to store tab panels
variance_tab_panels <- list()
# Loop through each key in variance_summary_tables
for (treatment in names(variance_summary_tables)) {
tab_name <- paste0(treatment)
tab_content <- DT::datatable(variance_summary_tables[[treatment]], options = list(pageLength = 90),
rownames = FALSE, colnames = c("Outcome","Treatment","Period",
"Within-user Std. Dev.", "Within-user Variance",
"Between-user Std. Dev.", "Between-user Variance"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
# Create the tabset with the variance tab panels
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
# Render the tabset
htmltools::tagList(variance_tabset)Click here to see variance decomposition by post domain
The within-user standard deviation and variance is not shown since:
- Participants are only expose to one post of each domain in the pre-treatment period.
- The tactic is different for those domains to which participants are exposed twice (once as a misinfo and once as a non-misinfo).
# Function to create a variance summary table
variance_summary_table <- function(data, treatment, period, share_type, facts) {
# Filter data based on the period
if (period != "pre and post") {
data <- data %>% filter(pre_post == period)
}
# Filter data based on the type
if (share_type == "non-misinfo share") {
data <- data %>% filter(post_is_misinfo == 0)
} else if (share_type == "misinfo share") {
data <- data %>% filter(post_is_misinfo == 1)
} else if (share_type == "Emotions posts") {
data <- data %>% filter(type == "Emotions posts")
} else if (share_type == "Reasoning posts") {
data <- data %>% filter(type == "Reasoning posts")
} else if (share_type == "Combo posts") {
data <- data %>% filter(type == "Combo posts")
}
# Filter database on the domain
data <- data %>% filter(fact == facts)
# Define data as panel
data <- pdata.frame(data, index = c('user'))
within_sd <- formatC(xtsum::within_sd(data, 'share', 'user'), format = "f", digits = 4)
within_variance <- formatC(xtsum::within_sd(data, 'share', 'user')^2, format = "f", digits = 4)
between_sd <- formatC(xtsum::between_sd(data, 'share', 'user'), format = "f", digits = 4)
between_variance <- formatC(xtsum::between_sd(data, 'share', 'user')^2, format = "f", digits = 4)
summary_table <- data.frame(
Outcome = share_type,
Treatment = treatment,
Period = period,
Fact = facts,
#Within_user_Standard_Deviation = within_sd,
#Within_user_Variance = within_variance,
Between_user_Standard_Deviation = between_sd,
Between_user_Variance = between_variance
)
return(summary_table)
}
# Initialize an empty list to store the variance summary tables
variance_summary_tables <- list()
treatments <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre and post")
share_types <- c("share")
facts <- seq(1:15)
# Loop over each combination of treatment, period, and share type
for (t in treatments) {
# Empty dataframe to store all combinations for the current treatment
all_combinations <- NULL
for (p in periods) {
for (f in facts) {
for (s in share_types) {
if (t != "All") {
# Filter the data for the current treatment
filtered_data <- panel %>% filter(treatment == t)
} else {
filtered_data <- panel
}
# Create a variance summary table for the current combination
variance_summary <- variance_summary_table(filtered_data, t, p, s, f)
# Store the variance summary table in all_combinations dataframe
all_combinations <- rbind(all_combinations, variance_summary)
}
}
}
# Store the all_combinations dataframe in the list with treatment name as key
variance_summary_tables[[t]] <- all_combinations
}
# Initialize an empty list to store tab panels
variance_tab_panels <- list()
# Loop through each key in variance_summary_tables
for (treatment in names(variance_summary_tables)) {
tab_name <- paste0(treatment)
tab_content <- DT::datatable(variance_summary_tables[[treatment]], options = list(pageLength = 15),
rownames = FALSE, colnames = c("Outcome","Treatment","Period","Post domain",
#"Within-user Std. Dev.", "Within-user Variance",
"Between-user Std. Dev.", "Between-user Variance"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
# Create the tabset with the variance tab panels
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
# Render the tabset
htmltools::tagList(variance_tabset)Click here to see variance decomposition by post domain and tactic
The within-user standard deviation and variance is not shown since:
- Participants are only expose to one post of each domain in the pre-treatment period.
- The tactic is different for those domains to which participants are exposed twice (once as a misinfo and once as a non-misinfo).
# Function to create a variance summary table
variance_summary_table <- function(data, treatment, period, share_type, facts) {
# Filter data based on the period
if (period != "pre and post") {
data <- data %>% filter(pre_post == period)
}
# Filter data based on the type
if (share_type == "non-misinfo share") {
data <- data %>% filter(post_is_misinfo == 0)
} else if (share_type == "misinfo share") {
data <- data %>% filter(post_is_misinfo == 1)
} else if (share_type == "Emotions posts") {
data <- data %>% filter(type == "Emotions posts")
} else if (share_type == "Reasoning posts") {
data <- data %>% filter(type == "Reasoning posts")
} else if (share_type == "Combo posts") {
data <- data %>% filter(type == "Combo posts")
}
# Filter database on the domain
data <- data %>% filter(fact == facts)
# Define data as panel
data <- pdata.frame(data, index = c('user'))
within_sd <- formatC(xtsum::within_sd(data, 'share', 'user'), format = "f", digits = 4)
within_variance <- formatC(xtsum::within_sd(data, 'share', 'user')^2, format = "f", digits = 4)
between_sd <- formatC(xtsum::between_sd(data, 'share', 'user'), format = "f", digits = 4)
between_variance <- formatC(xtsum::between_sd(data, 'share', 'user')^2, format = "f", digits = 4)
summary_table <- data.frame(
Outcome = share_type,
Treatment = treatment,
Period = period,
Fact = facts,
#Within_user_Standard_Deviation = within_sd,
#Within_user_Variance = within_variance,
Between_user_Standard_Deviation = between_sd,
Between_user_Variance = between_variance
)
return(summary_table)
}
# Initialize an empty list to store the variance summary tables
variance_summary_tables <- list()
treatments <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre", "post", "pre and post")
share_types <- c("share", "misinfo share", "non-misinfo share", "Emotions posts", "Reasoning posts", "Combo posts")
facts <- seq(1:15)
# Loop over each combination of treatment, period, and share type
for (t in treatments) {
# Empty dataframe to store all combinations for the current treatment
all_combinations <- NULL
for (f in facts) {
for (p in periods) {
for (s in share_types) {
if (t != "All") {
# Filter the data for the current treatment
filtered_data <- panel %>% filter(treatment == t)
} else {
filtered_data <- panel
}
# Create a variance summary table for the current combination
variance_summary <- variance_summary_table(filtered_data, t, p, s, f)
# Store the variance summary table in all_combinations dataframe
all_combinations <- rbind(all_combinations, variance_summary)
}
}
}
# Store the all_combinations dataframe in the list with treatment name as key
variance_summary_tables[[t]] <- all_combinations
}
# Initialize an empty list to store tab panels
variance_tab_panels <- list()
# Loop through each key in variance_summary_tables
for (treatment in names(variance_summary_tables)) {
tab_name <- paste0(treatment)
tab_content <- DT::datatable(variance_summary_tables[[treatment]], options = list(pageLength = 18),
rownames = FALSE, colnames = c("Outcome","Treatment","Period","Post domain",
#"Within-user Std. Dev.", "Within-user Variance",
"Between-user Std. Dev.", "Between-user Variance"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
# Create the tabset with the variance tab panels
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
# Render the tabset
htmltools::tagList(variance_tabset)
Variance-covariance
We present a variance-covariance matrix for user’s mean share by:
Pre vs. post
Tactics (Post type)
Domains (Post domain)
Pre vs. Post
# Add variance-covariance matrix
type_post <- c("post_is_misinfo == 1", "post_is_misinfo == 0", "type == 'Emotions posts'", "type == 'Reasoning posts'", "type == 'Combo posts'")
vcov_df_list <- list()
for (t in c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")) {
if (t == "All") {
treat <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
# User mean share by pre vs. post
vcov <- panel %>%
filter(treatment %in% treat) %>%
filter(pre_post %in% c("pre","post")) %>%
group_by(user, pre_post) %>%
select(-post) %>%
summarise(share = mean(share, na.rm = TRUE)) %>%
pivot_wider(names_from = pre_post, values_from = share) %>%
select(pre, post)
# generate dataframe and then print
vcov_df <- data.frame(type = "All", element_1 = "pre", element_2 = "post",
var1 = var(vcov$pre),
var2 = var(vcov$post),
cov = cov(vcov$pre, vcov$post)) %>%
mutate(across(c(var1, var2, cov), ~ round(., 5))) %>%
mutate(treatment = t)
for (tp in type_post) {
# User mean share by pre vs. post
vcov <- panel %>%
filter(treatment %in% treat) %>%
filter(pre_post %in% c("pre","post")) %>%
filter(!!rlang::parse_expr(tp)) %>%
group_by(user, pre_post) %>%
select(-post) %>%
summarise(share = mean(share, na.rm = TRUE)) %>%
pivot_wider(names_from = pre_post, values_from = share) %>%
select(pre, post)
# append to vcov_df
vcov_df <- rbind(vcov_df, data.frame(type = tp, element_1 = "pre", element_2 = "post",
var1 = var(vcov$pre),
var2 = var(vcov$post),
cov = cov(vcov$pre, vcov$post)) %>%
mutate(across(c(var1, var2, cov), ~ round(., 5))) %>%
mutate(treatment = t))
}
# replace vcov_df$type with actual types
vcov_df$type <- c("All", "Misinfo", "Non-misinfo", "Emotions", "Reasoning", "Combo")
# append to list
vcov_df_list[[t]] <- vcov_df
}
# Print
# knitr::kable(vcov_df, format = "markdown", col.names = c("Type of post", "Element 1", "Element 2",
# "Element 1 Variance", "Element 2 Variance",
# "Covariance (Element 1, Element 2)"))
# Generate a tabset to print results
variance_tab_panels <- list()
for (t in c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")) {
tab_name <- paste0(t)
tab_content <- DT::datatable(vcov_df_list[[t]] %>% select(-treatment), options = list(pageLength = 6),
rownames = FALSE, colnames = c("Type of post", "Element 1", "Element 2",
"Element 1 Variance", "Element 2 Variance",
"Covariance (Element 1, Element 2)"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
htmltools::tagList(variance_tabset)Post Tactics
vcov_df_list <- list()
for (t in c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")) {
if (t == "All") {
treat <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
# Post type
vcov <- panel %>%
filter(treatment %in% treat) %>%
filter(type %in% unique(panel$type)) %>%
group_by(user, type) %>%
summarise(share = mean(share, na.rm = TRUE)) %>%
pivot_wider(names_from = type, values_from = share) %>%
select(`Base posts`,`Combo posts`,`Emotions posts`,`Reasoning posts`)
# append to vcov_df
vcov_df <- data.frame(type = c("Emotions vs. Non-misinfo", "Emotions vs. Reasoning", "Emotions vs. Combo",
"Non-misinfo vs. Reasoning", "Non-misinfo vs. Combo", "Reasoning vs. Combo"),
element_1 = c("Emotions", "Emotions", "Emotions", "Non-misinfo", "Non-misinfo", "Reasoning"),
element_2 = c("Non-misinfo", "Reasoning", "Combo", "Reasoning", "Combo", "Combo"),
var_1 = c(var(vcov$`Emotions posts`),var(vcov$`Emotions posts`),var(vcov$`Emotions posts`),
var(vcov$`Base posts`), var(vcov$`Base posts`), var(vcov$`Reasoning posts`)),
var_2 = c(var(vcov$`Base posts`),var(vcov$`Reasoning posts`),var(vcov$`Combo posts`),
var(vcov$`Reasoning posts`),var(vcov$`Combo posts`),var(vcov$`Combo posts`)),
cov = c(cor(vcov$`Emotions posts`,vcov$`Base posts`),
cor(vcov$`Emotions posts`,vcov$`Reasoning posts`),
cor(vcov$`Emotions posts`,vcov$`Combo posts`),
cor(vcov$`Base posts`,vcov$`Reasoning posts`),
cor(vcov$`Base posts`,vcov$`Combo posts`),
cor(vcov$`Reasoning posts`,vcov$`Combo posts`))) %>%
mutate(across(c(var_1,var_2, cov), ~ round(., 5))) %>%
mutate(treatment = t)
# append to list
vcov_df_list[[t]] <- vcov_df
}
# # print variance-covariance matrix
# knitr::kable(vcov_df, format = "markdown", col.names = c("Elements", "Element 1", "Element 2",
# "Element 1 Variance", "Element 2 Variance",
# "Covariance (Element 1, Element 2)"))
# Generate a tabset to print results
variance_tab_panels <- list()
for (t in c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")) {
tab_name <- paste0(t)
tab_content <- DT::datatable(vcov_df_list[[t]] %>% select(-treatment), options = list(pageLength = 6),
rownames = FALSE, colnames = c("Elements", "Element 1", "Element 2",
"Element 1 Variance", "Element 2 Variance",
"Covariance (Element 1, Element 2)"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
htmltools::tagList(variance_tabset)Post Domain
vcov_df_list <- list()
for (t in c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")) {
if (t == "All") {
treat <- c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
# Post domain
vcov <- panel %>%
filter(treatment %in% treat) %>%
filter(fact %in% unique(panel$fact)) %>%
group_by(user, fact) %>%
summarise(share = mean(share, na.rm = TRUE)) %>%
pivot_wider(names_from = fact, values_from = share) %>%
select(1:16)
# append to vcov_df
vcov_df <- data.frame(type = c("Domain 1 vs. Domain 2", "Domain 1 vs. Domain 3", "Domain 1 vs. Domain 4",
"Domain 1 vs. Domain 5", "Domain 1 vs. Domain 6", "Domain 1 vs. Domain 7",
"Domain 1 vs. Domain 8", "Domain 1 vs. Domain 9", "Domain 1 vs. Domain 10",
"Domain 1 vs. Domain 11", "Domain 1 vs. Domain 12", "Domain 1 vs. Domain 13",
"Domain 1 vs. Domain 14", "Domain 1 vs. Domain 15",
"Domain 2 vs. Domain 3", "Domain 2 vs. Domain 4", "Domain 2 vs. Domain 5",
"Domain 2 vs. Domain 6", "Domain 2 vs. Domain 7", "Domain 2 vs. Domain 8",
"Domain 2 vs. Domain 9", "Domain 2 vs. Domain 10", "Domain 2 vs. Domain 11",
"Domain 2 vs. Domain 12", "Domain 2 vs. Domain 13", "Domain 2 vs. Domain 14",
"Domain 2 vs. Domain 15",
"Domain 3 vs. Domain 4", "Domain 3 vs. Domain 5", "Domain 3 vs. Domain 6",
"Domain 3 vs. Domain 7", "Domain 3 vs. Domain 8", "Domain 3 vs. Domain 9",
"Domain 3 vs. Domain 10", "Domain 3 vs. Domain 11", "Domain 3 vs. Domain 12",
"Domain 3 vs. Domain 13", "Domain 3 vs. Domain 14", "Domain 3 vs. Domain 15",
"Domain 4 vs. Domain 5", "Domain 4 vs. Domain 6", "Domain 4 vs. Domain 7",
"Domain 4 vs. Domain 8", "Domain 4 vs. Domain 9", "Domain 4 vs. Domain 10",
"Domain 4 vs. Domain 11", "Domain 4 vs. Domain 12", "Domain 4 vs. Domain 13",
"Domain 4 vs. Domain 14", "Domain 4 vs. Domain 15",
"Domain 5 vs. Domain 6", "Domain 5 vs. Domain 7", "Domain 5 vs. Domain 8",
"Domain 5 vs. Domain 9", "Domain 5 vs. Domain 10", "Domain 5 vs. Domain 11",
"Domain 5 vs. Domain 12", "Domain 5 vs. Domain 13", "Domain 5 vs. Domain 14",
"Domain 5 vs. Domain 15",
"Domain 6 vs. Domain 7", "Domain 6 vs. Domain 8", "Domain 6 vs. Domain 9",
"Domain 6 vs. Domain 10", "Domain 6 vs. Domain 11", "Domain 6 vs. Domain 12",
"Domain 6 vs. Domain 13", "Domain 6 vs. Domain 14", "Domain 6 vs. Domain 15",
"Domain 7 vs. Domain 8", "Domain 7 vs. Domain 9", "Domain 7 vs. Domain 10",
"Domain 7 vs. Domain 11", "Domain 7 vs. Domain 12", "Domain 7 vs. Domain 13",
"Domain 7 vs. Domain 14", "Domain 7 vs. Domain 15",
"Domain 8 vs. Domain 9", "Domain 8 vs. Domain 10", "Domain 8 vs. Domain 11",
"Domain 8 vs. Domain 12", "Domain 8 vs. Domain 13", "Domain 8 vs. Domain 14",
"Domain 8 vs. Domain 15",
"Domain 9 vs. Domain 10", "Domain 9 vs. Domain 11", "Domain 9 vs. Domain 12",
"Domain 9 vs. Domain 13", "Domain 9 vs. Domain 14", "Domain 9 vs. Domain 15",
"Domain 10 vs. Domain 11", "Domain 10 vs. Domain 12", "Domain 10 vs. Domain 13",
"Domain 10 vs. Domain 14", "Domain 10 vs. Domain 15",
"Domain 11 vs. Domain 12", "Domain 11 vs. Domain 13", "Domain 11 vs. Domain 14",
"Domain 11 vs. Domain 15",
"Domain 12 vs. Domain 13", "Domain 12 vs. Domain 14", "Domain 12 vs. Domain 15",
"Domain 13 vs. Domain 14", "Domain 13 vs. Domain 15",
"Domain 14 vs. Domain 15"),
element_1 = c(rep("Domain 1",14), rep("Domain 2",13), rep("Domain 3", 12),
rep("Domain 4",11), rep("Domain 5",10), rep("Domain 6",9),
rep("Domain 7",8), rep("Domain 8",7), rep("Domain 9",6),
rep("Domain 10",5), rep("Domain 11",4), rep("Domain 12",3),
rep("Domain 13",2), rep("Domain 14",1)),
element_2 = c("Domain 2","Domain 3","Domain 4", "Domain 5",
"Domain 6", "Domain 7", "Domain 8", "Domain 9", "Domain 10",
"Domain 11", "Domain 12", "Domain 13", "Domain 14", "Domain 15",
"Domain 3","Domain 4", "Domain 5", "Domain 6", "Domain 7",
"Domain 8", "Domain 9", "Domain 10", "Domain 11", "Domain 12",
"Domain 13", "Domain 14", "Domain 15",
"Domain 4", "Domain 5", "Domain 6", "Domain 7", "Domain 8",
"Domain 9", "Domain 10", "Domain 11", "Domain 12", "Domain 13",
"Domain 14", "Domain 15",
"Domain 5", "Domain 6", "Domain 7", "Domain 8", "Domain 9",
"Domain 10", "Domain 11", "Domain 12", "Domain 13", "Domain 14",
"Domain 15",
"Domain 6", "Domain 7", "Domain 8", "Domain 9", "Domain 10",
"Domain 11", "Domain 12", "Domain 13", "Domain 14", "Domain 15",
"Domain 7", "Domain 8", "Domain 9", "Domain 10", "Domain 11",
"Domain 12", "Domain 13", "Domain 14", "Domain 15",
"Domain 8", "Domain 9", "Domain 10", "Domain 11", "Domain 12",
"Domain 13", "Domain 14", "Domain 15",
"Domain 9", "Domain 10", "Domain 11", "Domain 12", "Domain 13",
"Domain 14", "Domain 15",
"Domain 10", "Domain 11", "Domain 12", "Domain 13", "Domain 14",
"Domain 15",
"Domain 11", "Domain 12", "Domain 13", "Domain 14", "Domain 15",
"Domain 12", "Domain 13", "Domain 14", "Domain 15",
"Domain 13", "Domain 14", "Domain 15",
"Domain 14", "Domain 15",
"Domain 15"),
var_1 = c(var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),
var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),
var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),var(vcov$`1`),
var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),
var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),
var(vcov$`2`),var(vcov$`2`),var(vcov$`2`),
var(vcov$`3`),var(vcov$`3`),var(vcov$`3`),var(vcov$`3`),var(vcov$`3`),
var(vcov$`3`),var(vcov$`3`),var(vcov$`3`),var(vcov$`3`),var(vcov$`3`),
var(vcov$`3`),var(vcov$`3`),
var(vcov$`4`),var(vcov$`4`),var(vcov$`4`),var(vcov$`4`),var(vcov$`4`),
var(vcov$`4`),var(vcov$`4`),var(vcov$`4`),var(vcov$`4`),var(vcov$`4`),
var(vcov$`4`),
var(vcov$`5`),var(vcov$`5`),var(vcov$`5`),var(vcov$`5`),var(vcov$`5`),
var(vcov$`5`),var(vcov$`5`),var(vcov$`5`),var(vcov$`5`),var(vcov$`5`),
var(vcov$`6`),var(vcov$`6`),var(vcov$`6`),var(vcov$`6`),var(vcov$`6`),
var(vcov$`6`),var(vcov$`6`),var(vcov$`6`),var(vcov$`6`),
var(vcov$`7`),var(vcov$`7`),var(vcov$`7`),var(vcov$`7`),var(vcov$`7`),
var(vcov$`7`),var(vcov$`7`),var(vcov$`7`),
var(vcov$`8`),var(vcov$`8`),var(vcov$`8`),var(vcov$`8`),var(vcov$`8`),
var(vcov$`8`),var(vcov$`8`),
var(vcov$`9`),var(vcov$`9`),var(vcov$`9`),var(vcov$`9`),var(vcov$`9`),
var(vcov$`9`),
var(vcov$`10`),var(vcov$`10`),var(vcov$`10`),var(vcov$`10`),var(vcov$`10`),
var(vcov$`11`),var(vcov$`11`),var(vcov$`11`),var(vcov$`11`),
var(vcov$`12`),var(vcov$`12`),var(vcov$`12`),
var(vcov$`13`),var(vcov$`13`),
var(vcov$`14`)),
var_2 = c(var(vcov$`2`),var(vcov$`3`),var(vcov$`4`),var(vcov$`5`),
var(vcov$`6`),var(vcov$`7`),var(vcov$`8`),var(vcov$`9`),var(vcov$`10`),
var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`3`),var(vcov$`4`),var(vcov$`5`),var(vcov$`6`),var(vcov$`7`),
var(vcov$`8`),var(vcov$`9`),var(vcov$`10`),var(vcov$`11`),var(vcov$`12`),
var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`4`),var(vcov$`5`),var(vcov$`6`),var(vcov$`7`),var(vcov$`8`),var(vcov$`9`),var(vcov$`10`),
var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`5`),var(vcov$`6`),var(vcov$`7`),var(vcov$`8`),var(vcov$`9`),
var(vcov$`10`),var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),
var(vcov$`15`),
var(vcov$`6`),var(vcov$`7`),var(vcov$`8`),var(vcov$`9`),
var(vcov$`10`),var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),
var(vcov$`15`),
var(vcov$`7`),var(vcov$`8`),var(vcov$`9`),var(vcov$`10`),
var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`8`),var(vcov$`9`),var(vcov$`10`),var(vcov$`11`),var(vcov$`12`),
var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`9`),var(vcov$`10`),
var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`10`),var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),
var(vcov$`15`),
var(vcov$`11`),var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),
var(vcov$`15`),
var(vcov$`12`),var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`13`),var(vcov$`14`),var(vcov$`15`),
var(vcov$`14`),var(vcov$`15`),
var(vcov$`15`)),
cov = c(cov(vcov$`1`,vcov$`2`),cov(vcov$`1`,vcov$`3`),cov(vcov$`1`,vcov$`4`),cov(vcov$`1`,vcov$`5`),cov(vcov$`1`,vcov$`6`),
cov(vcov$`1`,vcov$`7`),cov(vcov$`1`,vcov$`8`),cov(vcov$`1`,vcov$`9`),cov(vcov$`1`,vcov$`10`),cov(vcov$`1`,vcov$`11`),
cov(vcov$`1`,vcov$`12`),cov(vcov$`1`,vcov$`13`),cov(vcov$`1`,vcov$`14`),cov(vcov$`1`,vcov$`15`),
cov(vcov$`2`,vcov$`3`),cov(vcov$`2`,vcov$`4`),cov(vcov$`2`,vcov$`5`),cov(vcov$`2`,vcov$`6`),cov(vcov$`2`,vcov$`7`),
cov(vcov$`2`,vcov$`8`),cov(vcov$`2`,vcov$`9`),cov(vcov$`2`,vcov$`10`),cov(vcov$`2`,vcov$`11`),cov(vcov$`2`,vcov$`12`),
cov(vcov$`2`,vcov$`13`),cov(vcov$`2`,vcov$`14`),cov(vcov$`2`,vcov$`15`),
cov(vcov$`3`,vcov$`4`),cov(vcov$`3`,vcov$`5`),cov(vcov$`3`,vcov$`6`),cov(vcov$`3`,vcov$`7`),cov(vcov$`3`,vcov$`8`),
cov(vcov$`3`,vcov$`9`),cov(vcov$`3`,vcov$`10`),cov(vcov$`3`,vcov$`11`),cov(vcov$`3`,vcov$`12`),cov(vcov$`3`,vcov$`13`),
cov(vcov$`3`,vcov$`14`),cov(vcov$`3`,vcov$`15`),
cov(vcov$`4`,vcov$`5`),cov(vcov$`4`,vcov$`6`),cov(vcov$`4`,vcov$`7`),cov(vcov$`4`,vcov$`8`),cov(vcov$`4`,vcov$`9`),
cov(vcov$`4`,vcov$`10`),cov(vcov$`4`,vcov$`11`),cov(vcov$`4`,vcov$`12`),cov(vcov$`4`,vcov$`13`),cov(vcov$`4`,vcov$`14`),
cov(vcov$`4`,vcov$`15`),
cov(vcov$`5`,vcov$`6`),cov(vcov$`5`,vcov$`7`),cov(vcov$`5`,vcov$`8`),cov(vcov$`5`,vcov$`9`),cov(vcov$`5`,vcov$`10`),
cov(vcov$`5`,vcov$`11`),cov(vcov$`5`,vcov$`12`),cov(vcov$`5`,vcov$`13`),cov(vcov$`5`,vcov$`14`),cov(vcov$`5`,vcov$`15`),
cov(vcov$`6`,vcov$`7`),cov(vcov$`6`,vcov$`8`),cov(vcov$`6`,vcov$`9`),cov(vcov$`6`,vcov$`10`),cov(vcov$`6`,vcov$`11`),
cov(vcov$`6`,vcov$`12`),cov(vcov$`6`,vcov$`13`),cov(vcov$`6`,vcov$`14`),cov(vcov$`6`,vcov$`15`),
cov(vcov$`7`,vcov$`8`),cov(vcov$`7`,vcov$`9`),cov(vcov$`7`,vcov$`10`),cov(vcov$`7`,vcov$`11`),cov(vcov$`7`,vcov$`12`),
cov(vcov$`7`,vcov$`13`),cov(vcov$`7`,vcov$`14`),cov(vcov$`7`,vcov$`15`),
cov(vcov$`8`,vcov$`9`),cov(vcov$`8`,vcov$`10`),cov(vcov$`8`,vcov$`11`),cov(vcov$`8`,vcov$`12`),cov(vcov$`8`,vcov$`13`),
cov(vcov$`8`,vcov$`14`),cov(vcov$`8`,vcov$`15`),
cov(vcov$`9`,vcov$`10`),cov(vcov$`9`,vcov$`11`),cov(vcov$`9`,vcov$`12`),cov(vcov$`9`,vcov$`13`),cov(vcov$`9`,vcov$`14`),
cov(vcov$`9`,vcov$`15`),
cov(vcov$`10`,vcov$`11`),cov(vcov$`10`,vcov$`12`),cov(vcov$`10`,vcov$`13`),cov(vcov$`10`,vcov$`14`),cov(vcov$`10`,vcov$`15`),
cov(vcov$`11`,vcov$`12`),cov(vcov$`11`,vcov$`13`),cov(vcov$`11`,vcov$`14`),cov(vcov$`11`,vcov$`15`),
cov(vcov$`12`,vcov$`13`),cov(vcov$`12`,vcov$`14`),cov(vcov$`12`,vcov$`15`),
cov(vcov$`13`,vcov$`14`),cov(vcov$`13`,vcov$`15`),
cov(vcov$`14`,vcov$`15`))) %>%
mutate(across(c(var_1,var_2, cov), ~ round(., 5))) %>%
mutate(treatment = t)
# append to list
vcov_df_list[[t]] <- vcov_df
}
# # Print
# knitr::kable(vcov_df, format = "markdown", col.names = c("Elements", "Element 1", "Element 2",
# "Element 1 Variance", "Element 2 Variance",
# "Covariance (Element 1, Element 2)"))
# Generate a tabset to print results
variance_tab_panels <- list()
for (t in c("No-course Baseline", "Facts Baseline", "Emotions", "Reasoning", "Combo", "All")) {
tab_name <- paste0(t)
tab_content <- DT::datatable(vcov_df_list[[t]] %>% select(-treatment), options = list(pageLength = 5),
rownames = FALSE, colnames = c("Elements", "Element 1", "Element 2",
"Element 1 Variance", "Element 2 Variance",
"Covariance (Element 1, Element 2)"))
variance_tab_panels[[length(variance_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
variance_tabset <- tabsetPanel(!!!variance_tab_panels)
htmltools::tagList(variance_tabset)
Intra- and inter-personal correlations
While correlations are calculated for all treatment groups, the focus
should be on the No-course Baseline treatment group since
estimates for users in such treatment arm are not conflated with
treatment effects.
All correlations are calculated using the Pearson correlation coefficien, ranging from -1 to 1 (perfect linear relationship).
Intra-personal correlations are calculated within users. We report the mean correlation across users.
Pre vs. Post
Focus should be on the No-course Baseline treatment
group since estimates for users in such treatment arm are not conflated
with treatment effects.
# Add intra- and inter-personal correlations
# Intra-personal correlations
# Correlation within users
# Mean across users
# Pre vs. post
correlations_list <- list()
treatments <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
for (t in treatments) {
if (t == "All") {
treat <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
cor_matrix_intra <- panel %>%
filter(treatment %in% treat) %>%
group_by(user) %>%
summarize(
all = cor(share[pre_post == "pre"], share[pre_post == "post"], use = "pairwise.complete.obs"),
cor_misinfo = cor(share[pre_post == "pre" & post_is_misinfo == 1], share[pre_post == "post" & post_is_misinfo == 1], use = "pairwise.complete.obs"),
cor_nonmisinfo = cor(share[pre_post == "pre" & post_is_misinfo == 0], share[pre_post == "post" & post_is_misinfo == 0], use = "pairwise.complete.obs"),
cor_emotions = cor(share[pre_post == "pre" & type == "Emotions posts"], share[pre_post == "post" & type == "Emotions posts"], use = "pairwise.complete.obs"),
cor_reasoning = cor(share[pre_post == "pre" & type == "Reasoning posts"], share[pre_post == "post" & type == "Reasoning posts"], use = "pairwise.complete.obs"),
cor_combo = cor(share[pre_post == "pre" & type == "Combo posts"], share[pre_post == "post" & type == "Combo posts"], use = "pairwise.complete.obs")
) %>%
summarize(
all = mean(all, na.rm = TRUE),
cor_misinfo = mean(cor_misinfo, na.rm = TRUE),
cor_nonmisinfo = mean(cor_nonmisinfo, na.rm = TRUE),
cor_emotions = mean(cor_emotions, na.rm = TRUE),
cor_reasoning = mean(cor_reasoning, na.rm = TRUE),
cor_combo = mean(cor_combo, na.rm = TRUE)
) %>%
mutate(type = "pre vs. post")
# pivot to long format
cor_matrix_intra <- cor_matrix_intra %>%
pivot_longer(cols = c("all","cor_misinfo","cor_nonmisinfo","cor_emotions","cor_reasoning","cor_combo"),
names_to = "cor_type", values_to = "cor") %>%
mutate(cor_type = str_replace(cor_type, "cor_", "")) %>%
rename('intra_cor' = 'cor')
# Inter-personal correlations
# Correlation between users
# Mean across users
# Pre vs. post
cor_matrix_inter <- panel %>%
filter(treatment %in% treat) %>%
summarize(
all = cor(share[pre_post == "pre"], share[pre_post == "post"], use = "pairwise.complete.obs"),
cor_misinfo = cor(share[pre_post == "pre" & post_is_misinfo == 1], share[pre_post == "post" & post_is_misinfo == 1], use = "pairwise.complete.obs"),
cor_nonmisinfo = cor(share[pre_post == "pre" & post_is_misinfo == 0], share[pre_post == "post" & post_is_misinfo == 0], use = "pairwise.complete.obs"),
cor_emotions = cor(share[pre_post == "pre" & type == "Emotions posts"], share[pre_post == "post" & type == "Emotions posts"], use = "pairwise.complete.obs"),
cor_reasoning = cor(share[pre_post == "pre" & type == "Reasoning posts"], share[pre_post == "post" & type == "Reasoning posts"], use = "pairwise.complete.obs"),
cor_combo = cor(share[pre_post == "pre" & type == "Combo posts"], share[pre_post == "post" & type == "Combo posts"], use = "pairwise.complete.obs")
) %>%
mutate(type = "pre vs. post")
# pivot to long format
cor_matrix_inter <- cor_matrix_inter %>%
pivot_longer(cols = c("all","cor_misinfo","cor_nonmisinfo","cor_emotions","cor_reasoning","cor_combo"),
names_to = "cor_type", values_to = "cor") %>%
mutate(cor_type = str_replace(cor_type, "cor_", "")) %>%
rename('inter_cor' = 'cor')
cor_matrix <- left_join(cor_matrix_intra, cor_matrix_inter) %>%
mutate(treatment = t) %>%
select(treatment, cor_type, type, intra_cor, inter_cor) %>%
# format to 5 digits
mutate(across(c(intra_cor, inter_cor), ~ round(., 5)))
correlations_list[[t]] <- cor_matrix
}
# Initialize an empty list to store tab panels
correlations_tab_panels <- list()
# Loop through each key in variance_summary_tables
for (treatment in names(correlations_list)) {
tab_name <- paste0(treatment)
tab_content <- DT::datatable(correlations_list[[treatment]], options = list(pageLength = 6),
rownames = FALSE, colnames = c("Treatment",
"Post type",
"Correlation",
"Intra-participant correlation",
"Inter-participant correlation"))
correlations_tab_panels[[length(correlations_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
# Create the tabset with the variance tab panels
correlations_tabset <- tabsetPanel(!!!correlations_tab_panels)
# Render the tabset
htmltools::tagList(correlations_tabset)Across tactics
Pearson correlations between tactics for different survey periods.
Focus should be on the No-course Baseline treatment
group since estimates for users in such treatment arm are not conflated
with treatment effects.
# Add intra- and inter-personal correlations for pairs of tactics
# Intra-personal correlations
# Correlation within users
# Mean across users
# Tactic x vs. Tactic y
correlations_list <- list()
treatments <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre","post","Pre and Post")
for (t in treatments) {
if (t == "All") {
treat <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
corr_list_treatment <- data.frame()
for (p in periods) {
if (p == "Pre and Post") {
period <- c("pre","post")
} else {
period <- p
}
cor_matrix_intra <- panel %>%
filter(treatment %in% treat) %>%
filter(pre_post %in% period) %>%
group_by(user) %>%
summarize(
cor_emotions_reasoning = cor(share[type == "Emotions posts"], share[type == "Reasoning posts"], use = "pairwise.complete.obs"),
cor_emotions_combo = cor(share[type == "Emotions posts"], share[type == "Combo posts"], use = "pairwise.complete.obs"),
cor_reasoning_combo = cor(share[type == "Reasoning posts"], share[type == "Combo posts"], use = "pairwise.complete.obs")
) %>%
summarize(
cor_emotions_reasoning = mean(cor_emotions_reasoning, na.rm = TRUE),
cor_emotions_combo = mean(cor_emotions_combo, na.rm = TRUE),
cor_reasoning_combo = mean(cor_reasoning_combo, na.rm = TRUE)
) %>%
mutate(type = "tactics")
# pivot to long format
cor_matrix_intra <- cor_matrix_intra %>%
pivot_longer(cols = c("cor_emotions_reasoning","cor_emotions_combo","cor_reasoning_combo"),
names_to = "cor_type", values_to = "cor") %>%
mutate(cor_type = str_replace(cor_type, "cor_", "")) %>%
rename('intra_cor' = 'cor') %>%
mutate(cor_type = ifelse(cor_type == "emotions_reasoning", "Emotions vs. Reasoning",
ifelse(cor_type == "emotions_combo", "Emotions vs. Combo", "Reasoning vs. Combo"))) %>%
mutate(across(c(intra_cor), ~ round(., 5))) %>%
mutate(Period = p)
# Inter-personal correlations
# Correlation between users
# Mean across users
# Tactic x vs. Tactic y
cor_matrix_inter <- panel %>%
filter(treatment %in% treat) %>%
filter(pre_post %in% period) %>%
summarize(
cor_emotions_reasoning = cor(share[type == "Emotions posts"], share[type == "Reasoning posts"], use = "pairwise.complete.obs"),
cor_emotions_combo = cor(share[type == "Emotions posts"], share[type == "Combo posts"], use = "pairwise.complete.obs"),
cor_reasoning_combo = cor(share[type == "Reasoning posts"], share[type == "Combo posts"], use = "pairwise.complete.obs")
) %>%
mutate(type = "tactics")
# pivot to long format
cor_matrix_inter <- cor_matrix_inter %>%
pivot_longer(cols = c("cor_emotions_reasoning","cor_emotions_combo","cor_reasoning_combo"),
names_to = "cor_type", values_to = "cor") %>%
mutate(cor_type = str_replace(cor_type, "cor_", "")) %>%
rename('inter_cor' = 'cor') %>%
mutate(cor_type = ifelse(cor_type == "emotions_reasoning", "Emotions vs. Reasoning",
ifelse(cor_type == "emotions_combo", "Emotions vs. Combo", "Reasoning vs. Combo")))
cor_matrix <- left_join(cor_matrix_intra, cor_matrix_inter) %>%
mutate(treatment = t, period = p) %>%
select(treatment, cor_type, type, intra_cor, inter_cor) %>%
# format to 5 digits
mutate(across(c(intra_cor, inter_cor), ~ round(., 5))) %>%
# add time period
mutate(Period = p)
corr_list_treatment <- rbind(corr_list_treatment, cor_matrix) %>%
select(treatment, Period, type, cor_type, intra_cor, inter_cor)
}
correlations_list[[t]] <- corr_list_treatment
}
# Initialize an empty list to store tab panels
correlations_tab_panels <- list()
# Loop through each key in variance_summary_tables
for (treatment in names(correlations_list)) {
tab_name <- paste0(treatment)
tab_content <- DT::datatable(correlations_list[[treatment]] %>% select(-type), options = list(pageLength = 9),
rownames = FALSE, colnames = c("Treatment",
"Period",
"Correlation",
"Intra-participant correlation",
"Inter-participant correlation"))
correlations_tab_panels[[length(correlations_tab_panels) + 1]] <- tabPanel(tab_name, tab_content)
}
# Create the tabset with the variance tab panels
correlations_tabset <- tabsetPanel(!!!correlations_tab_panels)
# Render the tabset
htmltools::tagList(correlations_tabset)cat("In 'No-course Baseline' tab, the correlation between Reasoning and Combo at the post-treatment period is -1.2569e-17.")## In 'No-course Baseline' tab, the correlation between Reasoning and Combo at the post-treatment period is -1.2569e-17.
Sharing distributions / histograms
Post type
# 15 pastel colors from package
colors <- c("#F98C8C", "#F9D5A7", "#F9F1A5", "#D0F2A5", "#A5F2A5", "#A5F2D0", "#A5F2F2", "#A5D0F2", "#A5A5F2", "#D0A5F2", "#F2A5F2", "#F2A5D0", "#F2A5A5", "#F2D0A5", "#F2F2A5")
# Add histograms of share by post type
# Loop by type of post
# Loop by pre or post
# Loop by treatment
treatments <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre","post","Pre and Post")
histograms_list <- list()
for (t in treatments) {
cat("<details>")
if (t == "All") {
cat("<summary> Click here to view sharing distribution for ",paste0("**",t,"**")," treatment groups</summary>")
} else {
cat("<summary> Click here to view sharing distribution for ",paste0("**",t,"**")," treatment group</summary>")
}
if (t == "All") {
treat <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
histograms_list_treatment <- data.frame()
for (p in periods) {
if (p == "Pre and Post") {
period <- c("pre","post")
} else {
period <- p
}
cat(paste0("\n<br><br><br><br>$$\\textit{Period: ",p,"}$$\n<br><br><br><br>"))
for (post_type in 1:length(unique(panel$type))) {
df <- panel %>%
filter(treatment %in% treat) %>%
filter(pre_post %in% period) %>%
filter(type %in% unique(panel$type)[post_type]) %>%
group_by(user) %>%
summarise(share = mean(share, na.rm = TRUE)) %>%
mutate(denominator = n()) %>%
group_by(share) %>%
mutate(numerator = n()) %>%
distinct(share, numerator, denominator) %>%
ungroup() %>%
mutate(ratio = (numerator/denominator)*100)
# generate histogram of variable share
# as a proportion of all posts seen by user of such type
g <- ggplot(df) +
geom_bar(aes(x = share, y = ratio), stat = "identity", fill = colors[post_type]) +
labs(title = paste(unique(panel$type)[post_type]),
x = paste0("\nProportion of posts shared at ", p),
y = paste0("Frequency (as a percentage of total ",unique(panel$type)[post_type]," shown)\n")) +
scale_x_continuous(labels = scales::percent_format(scale = 100),
breaks = (df$share)) +
theme_minimal() +
theme(legend.position = "none",
plot.title = element_text(hjust = 0.5))
print(g)
cat("<br><br>")
}
}
cat("</details>")
cat("<br><br>")
}Click here to view sharing distribution for No-course Baseline treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Facts Baseline treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Emotions treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Reasoning treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Combo treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for All treatment groups
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Post domain
# Add histograms of share by post domain
# Loop by domain
# Loop by pre or post
# Loop by treatment
treatments <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo", "All")
periods <- c("pre","post","Pre and Post")
histograms_list <- list()
for (t in treatments) {
cat("<details>")
if (t == "All") {
cat("<summary> Click here to view sharing distribution for ",paste0("**",t,"**")," treatment groups</summary>")
} else {
cat("<summary> Click here to view sharing distribution for ",paste0("**",t,"**")," treatment group</summary>")
}
if (t == "All") {
treat <- c("No-course Baseline","Facts Baseline", "Emotions", "Reasoning", "Combo")
} else {
treat <- t
}
histograms_list_treatment <- data.frame()
for (p in periods) {
if (p == "Pre and Post") {
period <- c("pre","post")
} else {
period <- p
}
cat(paste0("\n<br><br><br><br>$$\\textit{Period: ",p,"}$$\n<br><br><br><br>"))
for (domain in 1:length(unique(panel$fact))) {
df <- panel %>%
filter(treatment %in% treat) %>%
filter(pre_post %in% period) %>%
filter(fact %in% domain) %>%
group_by(user) %>%
summarise(share = mean(share, na.rm = TRUE)) %>%
mutate(denominator = n()) %>%
group_by(share) %>%
mutate(numerator = n()) %>%
distinct(share, numerator, denominator) %>%
ungroup() %>%
mutate(ratio = (numerator/denominator)*100)
# generate histogram of variable share
# as a proportion of all posts seen by user of such type
g <- ggplot(df) +
geom_bar(aes(x = share, y = ratio), stat = "identity", fill = colors[domain]) +
labs(title = paste("Post domain",domain),
x = paste0("\nProportion of posts shared at ", p),
y = paste0("Frequency (as a percentage of total posts domain ",domain," shown)\n")) +
scale_x_continuous(labels = scales::percent_format(scale = 100),
breaks = (df$share)) +
theme_minimal() +
theme(legend.position = "none",
plot.title = element_text(hjust = 0.5))
print(g)
cat("<br><br>")
}
}
cat("</details>")
cat("<br><br>")
}Click here to view sharing distribution for No-course Baseline treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Facts Baseline treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Emotions treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Reasoning treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for Combo treatment group
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Click here to view sharing distribution for All treatment groups
\[\textit{Period: pre}\]
\[\textit{Period: post}\]
\[\textit{Period: Pre and Post}\]
Analysis
Baseline estimation
Through the following specification, we estimate the treatment effects of each course on Misinfo and Non-misinfo posts sharing:
\(\textit{Share}_{ipt} = \alpha + \beta_1\mathbb{I}[\textit{Treatment}_{i}] \times \mathbb{I}[\textit{Post-treatment}_{t}] \times \mathbb{I}[\textit{Misinfo}_{p}]\)\(+\)
\(\beta_2\mathbb{I}[\textit{Treatment}_{i}] \times \mathbb{I}[\textit{Post-treatment}_{t}]\)\(+\)
\(\beta_3\mathbb{I}[\textit{Treatment}_{i}] \times \mathbb{I}[\textit{Misinfo}_{p}]\)\(+\)
\(\beta_4\mathbb{I}[\textit{Treatment}_{i}]\)\(+\)
\(\beta_5\mathbb{I}[\textit{Post-treatment}_{t}] \times \mathbb{I}[\textit{Misinfo}_{p}]\)\(+\)
\(\beta_6\mathbb{I}[\textit{Post-treatment}_{t}]\)\(+\)
\(\beta_7\mathbb{I}[\textit{Misinfo}_{p}]\)\(+\)
\(\epsilon_{ipt}\),
where:
\(\textit{Share}_{ipt}\): a binary indicator equal to 1 if the user \(i\) decided to share post \(p\) at time \(t\) (pre-treatment or post-treatment survey)
\(\alpha\): Average sharing of Non-misinfo posts for Facts Baseline, the baseline treatment group.
version 1: \(\textit{Treatment}_i = \{\textit{Facts Baseline}, \textit{Emotions course}\}\)
\(\alpha\): Pre Non-Misinfo for Facts Baseline
\(\alpha + \beta_4\): Pre Non-Misinfo for Emotions
\(\alpha + \beta_7\): Pre Misinfo for Facts Baseline
\(\alpha + \beta_3 + \beta_4 + \beta_7\): Pre Misinfo for Emotions
\(\alpha + \beta_6\): Post Non-Misinfo for Facts Baseline
\(\alpha + \beta_2 + \beta_4 + \beta_6\): Post Non-Misinfo for Emotions
\(\alpha + \beta_5 + \beta_6 + \beta_7\): Post Misinfo for Facts Baseline
\(\alpha + \beta_1 + \beta_2 + \beta_3 + \beta_4 + \beta_5 + \beta_6 + \beta_7\): Post Misinfo for the Emotions
version 2: \(\textit{Treatment}_i = \{\textit{No-course Baseline}, \textit{Facts Baseline}, \textit{Emotions course}, \textit{Reasoning course}, \textit{Combo course}\}\)
\(\epsilon_{ipt}\): standard errors clustered at the user level
We estimate two versions:
two treatment arms (Facts Baseline and Emotions)
five treatment arms (No-course Baseline, Facts Baseline, Emotions, Reasoning, and Combo).
For all estimations, we provide a short and a long table. The short table shows the treatment coefficients only. The long table also shows all covariates and its interactions. To access the long table, click on the button.
Fixed effects
We report estimates when adding the following sets of fixed effects to the model:
User
- User-specific sharing fixed effect.
Post-domain
- Post-domain-specific sharing fixed effect (for 15 different domains) and user-specific sharing fixed effect.
Post-tactic
- Post-tactic-specific sharing fixed effect (for 3 manipulation tactics used in misinformation posts: Emotions, Reasons, Combo, plus non-misinformation) and user-specific sharing fixed effect.
Post-domain x Post-tactic
- Post-domain \(\times\) Post-tactic-specific sharing fixed effect (for 60 different combinations: 15 domains \(\times\) 4 tactics) and user-specific sharing fixed effect.
Post-ID
- Post-specific sharing fixed effect (for 60 different posts) and user-specific sharing fixed effect.
Covariates
We report estimates when adding one of the following covariate sets in addition to the user-specific sharing fixed effect.
Each set of covariates is de-meaned and interacted with the treatment, post-treatment, and misinformation indicators:
\[\textit{De-meaned Cov}_i \times \mathbb{1}[Treatment_i] \times \mathbb{1}[\textit{Post-treatment}_t] \times \mathbb{1}[Misinformation_p]\]
Demographics:
- Include set of de-meaned demographics \({X}_i\).
Misinfo and Non-misinfo Sharing rates:
- \(\textit{De-meaned Misinfo Sharing rate}_{it-1} + \textit{De-meaned Non-Misinfo Sharing rate}_{it-1}\)
Misinfo and Non-Misinfo Sharing rates and Tactic Sharing rates:
- \(\textit{De-meaned Misinfo Sharing rate}_{it-1} + \textit{De-meaned Non-Misinfo Sharing rate}_{it-1}\)\(+\) \(\textit{De-meaned Sharing rate of Emotions posts}_{it-1} + \textit{De-meaned Sharing rate of Reasoning posts}_{it-1}\)\(+\) \(\textit{De-meaned Sharing rate of Combo posts}_{it-1}\)
Demographics, as well as Misinfo and Non-misinfo Sharing rates:
- \({X}_i + \textit{De-meaned Misinfo Sharing rate}_{it-1} + \textit{De-meaned Non-Misinfo Sharing rate}_{it-1}\).
Demographics, as well as Misinfo, Non-Misinfo Sharing rates, and Tactic Sharing rates:
- \({X}_i + \textit{De-meaned Misinfo Sharing rate}_{it-1} + \textit{De-meaned Non-Misinfo Sharing rate}_{it-1}\)\(+\) \(\textit{De-meaned Sharing rate of Emotions posts}_{it-1} + \textit{De-meaned Sharing rate of Reasoning posts}_{it-1}\)\(+\) \(\textit{De-meaned Sharing rate of Combo posts}_{it-1}\)
Linear
Baseline estimation
Below, we present the estimates based on 2 treatment arms: Facts Baseline and Emotions.
The sample includes all users who completed the post-survey and belong to the “Facts Baseline” (n = 1,790) or “Emotions course” (n = 1,836) treatment groups (n = 3,626). We consider all 9 posts in the pre-treatment period (“pre”) and all 9 posts in the post-treatment period (i.e., “post”).
# Define panel df with two treatment arms only
panel_2 <- panel %>% filter(treatment %in% c('Facts Baseline','Emotions'))
panel_2$treatment <- factor(panel_2$treatment, levels = c('Facts Baseline','Emotions'))
panel_2$emotions <- ifelse(panel_2$treatment == "Emotions",1,0)# version 1: 2 treatment arms
formula_baseline_lm <- "share ~ treatment*post*post_is_misinfo"
baseline_lm <- lm(as.formula(formula_baseline_lm),panel_2)
baseline_lm_robust <- lm_robust(as.formula(formula_baseline_lm),panel_2, cluster = user, se_type = 'CR2')# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
if (!file.exists("R2oos_baseline1.rds")) {
fitFunLM <- function(y, x){lm_robust.fit(y = y, x = cbind(1, x), cluster = user, se_type = "CR2")}
predFunLM <- function(mod, x) {cbind(1,x) %*% mod$coef}
# baseline
formula <- gsub("^share","",formula_baseline_lm)
X <- model.matrix(as.formula(formula), data = panel_2)
X <- X[, !(colnames(X) %in% "(Intercept)")]
R2oos_baseline1 <- oosse::R2oosse(panel_2$share,
X[,-1],
fitFun = fitFunLM,
predFun = predFunLM,
methodMSE = "bootstrap",
nBootstraps = 100,
nBootstrapsCor = 25)
saveRDS(R2oos_baseline1,"R2oos_baseline1.rds")
} else {
R2oos_baseline1 <- readRDS("R2oos_baseline1.rds")
}
R2out <- as.numeric(R2oos_baseline1$R2[1])# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_lm)
names <- unique(c(names(baseline_lm$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2out")) {
R2out <- as.numeric(R2out)
} else {
R2out <- rep("",length(model_list))
}
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# standard errors
se <- baseline_lm_robust$std.error
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
override.se = se,
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test. MSE for out-of-sample R$^2$ is calculated through a cross-validation process with 100 repetitions and 10 folds.",
caption = paste("Linear Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out)
),
reorder.gof = c(6,7,5,8,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 |
| (0.0136) | |
| Emotions x 1[Post-treatment] | -0.0938*** |
| (0.0132) | |
| Emotions x 1[Misinfo] | -0.0039 |
| (0.0100) | |
| R2 | 0.0422 |
| Adj. R2 | 0.0421 |
| Out-of-sample R2 | 0.0422 |
| Num. obs. | 65268 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. MSE for out-of-sample R\(^2\) is calculated through a cross-validation process with 100 repetitions and 10 folds. | |
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
override.se = se,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test. MSE for out-of-sample R$^2$ is calculated through a cross-validation process with 10 folds.",
caption = paste("Linear Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out)
),
reorder.gof = c(6,7,5,8,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| Facts Baseline | 0.6829*** |
| (0.0081) | |
| Emotions | 0.0059 |
| (0.0114) | |
| 1[Post-treatment] | -0.0758*** |
| (0.0089) | |
| 1[Misinfo] | -0.1442*** |
| (0.0069) | |
| Emotions x 1[Post-treatment] | -0.0938*** |
| (0.0132) | |
| Emotions x 1[Misinfo] | -0.0039 |
| (0.0100) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 |
| (0.0095) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 |
| (0.0136) | |
| R2 | 0.0422 |
| Adj. R2 | 0.0421 |
| Out-of-sample R2 | 0.0422 |
| Num. obs. | 65268 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. MSE for out-of-sample R\(^2\) is calculated through a cross-validation process with 10 folds. | |
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method.
Treatment effects in the last column refer to estimated marginal means (EMMs, or least-squares means) for specified factor combinations (Post-treatment and Misinfo). Point estimates are reported in the upper cell and standard errors, clustered at the user level, reported in parentheses in the bottom cell. t-ratios are reported in brackets.
# Table with prediction values
treatmentLevels <- unique(panel_2$treatment)
predictorVariables <- expand.grid(treatment = treatmentLevels,
post = c(0, 1),
post_is_misinfo = c(0, 1))
predictions <- predict(baseline_lm, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
# calculate difference between Facts Baseline and all other treatments
# prediction_df <- prediction_df %>%
# group_by(post,post_is_misinfo) %>%
# mutate(lag = ifelse(treatment == "Facts Baseline",
# prediction, NA)) %>%
# tidyr::fill(lag) %>%
# ungroup() %>%
# mutate(te = ifelse(treatment != "Facts Baseline",
# format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
# select(-lag) %>%
# mutate(te = ifelse(is.na(te),"",te))
emm_res <- emmeans(baseline_lm, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% select(post, post_is_misinfo, estimate, SE, t.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(t.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
k <- knitr::kable(prediction_df, col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Treatment effect<br>Emotions vs. Facts"))
print(k)| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Treatment effect Emotions vs. Facts |
|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006670) | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.006586) | (0.009374); [0.6346] |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004717) | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.004657) | (0.006628); [0.3059] |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006670) | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.006586) | (0.009374); [ -9.3698] |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004717) | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.004657) | (0.006628); [-16.8909] |
Click for All treatment arms
# version 2: all treatment arms
formula_baseline_all_lm <- "share ~ treatment*post*post_is_misinfo"
baseline_all_lm <- lm(as.formula(formula_baseline_all_lm),panel)
baseline_all_lm_robust <- lm_robust(as.formula(formula_baseline_all_lm),panel, cluster = user, se_type = 'CR2')# Calculate out-of-sample R2
if (!file.exists("R2oos_baseline1_all.rds")) {
# baseline all treatments
formula <- gsub("^share","",formula_baseline_all_lm)
X <- model.matrix(as.formula(formula), data = panel)
X <- X[, !(colnames(X) %in% "(Intercept)")]
R2oos_baseline1_all <- oosse::R2oosse(panel$share,
X[,-1],
fitFun = fitFunLM,
predFun = predFunLM,
methodMSE = "bootstrap",
nBootstraps = 100,
nBootstrapsCor = 25)
saveRDS(R2oos_baseline1_all,"R2oos_baseline1_all.rds")
} else {
R2oos_baseline1_all <- readRDS("R2oos_baseline1_all.rds")
}
R2out_all <- as.numeric(R2oos_baseline1_all$R2[1])Baseline estimation
The sample includes all users who completed the post-survey (n = 8,684). We consider all 9 posts in the pre-treatment period (“pre”) and all 9 posts in the post-treatment period (i.e., “post”).
# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_all_lm)
names <- unique(c(names(baseline_all_lm$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2out_all")) {
R2out_all <- as.numeric(R2out_all)
} else {
R2out_all <- rep("",length(model_list))
}
# keep variables
keep <- list("treatmentFacts Baseline:post:post_is_misinfo" = "Facts Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentReasoning:post:post_is_misinfo" = "Reasoning x 1[Post-treatment] x 1[Misinfo]",
"treatmentCombo:post:post_is_misinfo" = "Combo x 1[Post-treatment] x 1[Misinfo]",
"treatmentFacts Baseline:post" = "Facts Baseline x 1[Post-treatment]",
"treatmentNo-course Baseline" = "No-course Baseline x 1[Post-treatment]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentReasoning:post" = "Reasoning x 1[Post-treatment]",
"treatmentCombo:post" = "Combo x 1[Post-treatment]",
"treatmentFacts Baseline:post_is_misinfo" = "Facts Baseline x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Misinfo]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]",
"treatmentReasoning:post_is_misinfo" = "Reasoning x 1[Misinfo]",
"treatmentCombo:post_is_misinfo" = "Combo x 1[Misinfo]")
# standard errors
se <- baseline_all_lm_robust$std.error
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
override.se = se,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out_all)
),
reorder.gof = c(6,7,5,8,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0051 |
| (0.0101) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0051 |
| (0.0101) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 |
| (0.0136) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 |
| (0.0139) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 |
| (0.0138) | |
| No-course Baseline x 1[Post-treatment] | 0.0213** |
| (0.0116) | |
| Emotions x 1[Post-treatment] | -0.0938*** |
| (0.0132) | |
| Reasoning x 1[Post-treatment] | -0.0777*** |
| (0.0130) | |
| Combo x 1[Post-treatment] | -0.0884*** |
| (0.0131) | |
| Emotions x 1[Misinfo] | -0.0039 |
| (0.0100) | |
| Reasoning x 1[Misinfo] | -0.0196** |
| (0.0101) | |
| Combo x 1[Misinfo] | 0.0040 |
| (0.0099) | |
| R2 | 0.0429 |
| Adj. R2 | 0.0428 |
| Out-of-sample R2 | 0.0428 |
| Num. obs. | 156312 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | |
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test. MSE for out-of-sample R$^2$ is calculated through a cross-validation process with 100 repetitions and 10 folds.",
caption = paste("Linear Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out_all)
),
reorder.gof = c(6,7,5,8,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| Facts Baseline | 0.6829*** |
| (0.0067) | |
| No-course Baseline | 0.0213** |
| (0.0096) | |
| Emotions | 0.0059 |
| (0.0094) | |
| Reasoning | 0.0209** |
| (0.0096) | |
| Combo | 0.0035 |
| (0.0095) | |
| 1[Post-treatment] | -0.0758*** |
| (0.0094) | |
| 1[Misinfo] | -0.1442*** |
| (0.0082) | |
| No-course Baseline x 1[Post-treatment] | 0.0514*** |
| (0.0136) | |
| Emotions x 1[Post-treatment] | -0.0938*** |
| (0.0132) | |
| Reasoning x 1[Post-treatment] | -0.0777*** |
| (0.0135) | |
| Combo x 1[Post-treatment] | -0.0884*** |
| (0.0134) | |
| No-course Baseline x 1[Misinfo] | 0.0051 |
| (0.0118) | |
| Emotions x 1[Misinfo] | -0.0039 |
| (0.0115) | |
| Reasoning x 1[Misinfo] | -0.0196** |
| (0.0117) | |
| Combo x 1[Misinfo] | 0.0040 |
| (0.0116) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 |
| (0.0115) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 |
| (0.0167) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 |
| (0.0162) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 |
| (0.0166) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 |
| (0.0164) | |
| R2 | 0.0429 |
| Adj. R2 | 0.0428 |
| Out-of-sample R2 | 0.0428 |
| Num. obs. | 156312 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. MSE for out-of-sample R\(^2\) is calculated through a cross-validation process with 100 repetitions and 10 folds. | |
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method. Treatment effects standard errors are clustered at the user level and reported in parentheses. t-ratios are reported in brackets.
# Table with prediction values
treatmentLevels <- unique(panel$treatment)
predictorVariables <- expand.grid(treatment = treatmentLevels,
post = c(0, 1),
post_is_misinfo = c(0, 1))
predictions <- predict(baseline_all_lm, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
# calculate difference between Facts Baseline and all other treatments
# prediction_df <- prediction_df %>%
# group_by(post,post_is_misinfo) %>%
# mutate(lag = ifelse(treatment == "Facts Baseline",
# prediction, NA)) %>%
# tidyr::fill(lag) %>%
# ungroup() %>%
# mutate(te = ifelse(treatment != "Facts Baseline",
# format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
# select(-lag) %>%
# mutate(te = ifelse(is.na(te),"",te))
emm_res <- emmeans(baseline_all_lm, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res)
contrast_df <- contrast_df %>% filter(contrast == "Emotions - Facts Baseline") %>%
select(post, post_is_misinfo, estimate, SE, t.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(t.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
knitr::kable(prediction_df, col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts"))| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006659) | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.006575) | (0.009358); [0.6356] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006983) | |
| Reasoning | 0 | 0 | 0.7037 | (0.006853) | |
| Combo | 0 | 0 | 0.6864 | (0.006754) | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004709) | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.004649) | (0.006617); [0.3064] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.004937) | |
| Reasoning | 0 | 1 | 0.5399 | (0.004846) | |
| Combo | 0 | 1 | 0.5462 | (0.004776) | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006659) | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.006575) | (0.009358); [ -9.3854] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006983) | |
| Reasoning | 1 | 0 | 0.5503 | (0.006853) | |
| Combo | 1 | 0 | 0.5222 | (0.006754) | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004709) | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.004649) | (0.006617); [-16.9190] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.004937) | |
| Reasoning | 1 | 1 | 0.4040 | (0.004846) | |
| Combo | 1 | 1 | 0.3715 | (0.004776) |
Click for Comparison with the estimates in the IAM paper
knitr::include_graphics("figures/table5.png")
Fixed effects
# version 1: 2 treatment arms
formula_fe1_lm <- "share ~ treatment*post*post_is_misinfo | user"
fe1_lm <- feols(as.formula(formula_fe1_lm),panel_2)
formula_fe2_lm <- "share ~ treatment*post*post_is_misinfo + factor(fact) | user"
fe2_lm <- feols(as.formula(formula_fe2_lm),panel_2)
formula_fe3_lm <- "share ~ treatment*post*post_is_misinfo + factor(type) | user"
fe3_lm <- feols(as.formula(formula_fe3_lm),panel_2)
formula_fe4_lm <- "share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) | user"
fe4_lm <- feols(as.formula(formula_fe4_lm),panel_2)
formula_fe5_lm <- "share ~ treatment*post*post_is_misinfo + factor(post_id) | user"
fe5_lm <- feols(as.formula(formula_fe5_lm),panel_2)# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
# estimate out-of-sample R2 for all fixed effects models
# use R2oosse function from oosse package
# w user FE
models_fe <- c('fe1_lm','fe2_lm','fe3_lm','fe4_lm','fe5_lm')
# Estimate R2oosse for all models
R2oos_fe_2treat <- list()
for (m in seq_along(models_fe)) {
if (file.exists(paste0("R2oos_fe_2treat_", m, ".rds"))) {
R2oos_fe_2treat[[m]] <- readRDS(paste0("R2oos_fe_2treat_", m, ".rds"))
} else {
formula <- gsub("^share", "", gsub(" \\| user", "", get(paste0("formula_fe", m, "_lm"))))
X <- model.matrix(as.formula(formula),
data = panel_2 %>%
mutate(type = as.numeric(factor(panel_2$type, levels = unique(panel_2$type)))))
X <- X[, !(colnames(X) %in% "(Intercept)")]
X <- cbind(X, user = panel_2$user, treatment = panel_2$treatment)
# identify estimation dataset
estimation_data <- panel_2 %>% filter(row_number() %in% obs(get(models_fe[[m]])))
if ("treatment" %in% colnames(X)) {
X[,"treatment"] <- factor(estimation_data$treatment,
levels = levels(estimation_data$treatment))
}
# add fixed effects based on model
if (m == 2) {
X <- cbind(X, fact = estimation_data$fact)
}
if (m == 3) {
X <- cbind(X, type = as.numeric(factor(estimation_data$type, levels = unique(estimation_data$type))))
}
if (m == 4) {
X <- cbind(X, fact = estimation_data$fact,
type = as.numeric(factor(estimation_data$type), levels = unique(estimation_data$type)))
}
if (m == 5) {
X <- cbind(X, post_id = estimation_data$post_id)
}
# Define the fit function for models with fixed effects
fitFunLM_fe <- function(y, x) {
feols(as.formula(get(paste0("formula_fe", m, "_lm"))),
data = estimation_data %>%
mutate(type = as.numeric(factor(estimation_data$type, levels = unique(estimation_data$type)))))
}
# Define the prediction function for models with fixed effects
predFunLM_fe <- function(mod,x) {
predict(mod, newdata = x, type = "response")
}
# Calculate R2oosse
R2oos_fe_2treat[[m]] <- oosse::R2oosse(y = panel_2$share,
x = X,
fitFun = fitFunLM_fe,
predFun = predFunLM_fe,
methodMSE = "CV",
cvReps = 100,
nFold = 10,
nBootstrapsCor = 25)
# Save results
saveRDS(R2oos_fe_2treat[[m]], paste0("R2oos_fe_2treat_", m, ".rds"))
}
}
R2oos_fe_2treat_all <- as.numeric(unlist(lapply(R2oos_fe_2treat, function(x) x$R2[1])),4)# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_lm, fe1_lm,fe2_lm,fe3_lm,fe4_lm,fe5_lm)
names <- unique(c(names(baseline_lm$coefficients),
names(fe1_lm$coefficients),
names(fe2_lm$coefficients),
names(fe3_lm$coefficients),
names(fe4_lm$coefficients),
names(fe5_lm$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2oos_fe_2treat_all")) {
#R2oos_fe_2treat_all <- c(R2out, R2oos_fe_2treat_all)
R2oos_fe_2treat_all <- c(format(R2out, digits = 4),rep("",length(model_list)-1))
} else {
R2oos_fe_2treat_all <- rep("",length(model_list))
}
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
#na_coeff <- names(model_list[[m]]$coefficients[,1])[is.na(model_list[[m]]$coefficients[,1])] # felm
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_fe_2treat_all)),
reorder.gof = c(8,6,7,9,10,11,12,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0162) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0133) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0115) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 | 0.0422 | |||||
| Adj. R2 | 0.0421 | |||||
| R2 (full model) | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 | |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 | |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | 0.04216 | |||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_fe_2treat_all)),
reorder.gof = c(8,6,7,9,10,11,12,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6829*** | |||||
| (0.0067) | ||||||
| Emotions | 0.0059 | |||||
| (0.0094) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0743*** | -0.0743*** |
| (0.0094) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1460*** | -0.0394*** | -0.0286*** | -0.0747*** |
| (0.0082) | (0.0069) | (0.0067) | (0.0060) | (0.0066) | (0.0054) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0133) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0115) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0040 | 0.0040 |
| (0.0116) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0162) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Post domain 2 | -0.1989*** | -0.1730*** | ||||
| (0.0089) | (0.0150) | |||||
| Post domain 3 | -0.2873*** | -0.3172*** | ||||
| (0.0087) | (0.0148) | |||||
| Post domain 4 | -0.1939*** | -0.2624*** | ||||
| (0.0089) | (0.0151) | |||||
| Post domain 5 | 0.0573*** | -0.0069 | ||||
| (0.0080) | (0.0137) | |||||
| Post domain 6 | 0.0263*** | 0.0467*** | ||||
| (0.0085) | (0.0134) | |||||
| Post domain 7 | -0.2382*** | -0.2006*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 8 | -0.2172*** | -0.1220*** | ||||
| (0.0088) | (0.0147) | |||||
| Post domain 9 | 0.0456*** | -0.0596*** | ||||
| (0.0082) | (0.0143) | |||||
| Post domain 10 | -0.0577*** | -0.1001*** | ||||
| (0.0085) | (0.0143) | |||||
| Post domain 11 | -0.1629*** | -0.1861*** | ||||
| (0.0090) | (0.0149) | |||||
| Post domain 12 | -0.1110*** | -0.1377*** | ||||
| (0.0085) | (0.0144) | |||||
| Post domain 13 | -0.1564*** | -0.2181*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 14 | -0.2815*** | -0.3292*** | ||||
| (0.0090) | (0.0152) | |||||
| Post domain 15 | -0.1923*** | -0.2956*** | ||||
| (0.0087) | (0.0151) | |||||
| 1[Combo manipulation tactic] | -0.1294*** | -0.1532*** | ||||
| (0.0044) | (0.0192) | |||||
| 1[Emotions manipulation tactic] | -0.0960*** | -0.2038*** | ||||
| (0.0042) | (0.0198) | |||||
| 1[Reasoning manipulation tactic] | -0.0890*** | -0.1088*** | ||||
| (0.0041) | (0.0194) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0666*** | |||||
| (0.0245) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0541** | |||||
| (0.0233) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0984*** | |||||
| (0.0236) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.1031*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3263*** | |||||
| (0.0226) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0368* | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0191 | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1669*** | |||||
| (0.0234) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0193 | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0561** | |||||
| (0.0242) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0390* | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0474** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0208 | |||||
| (0.0235) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.1033*** | |||||
| (0.0237) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0815*** | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1354*** | |||||
| (0.0236) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0892*** | |||||
| (0.0238) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1411*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1656*** | |||||
| (0.0234) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.0071 | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1331*** | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1794*** | |||||
| (0.0241) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1221*** | |||||
| (0.0235) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0225 | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1420*** | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.2141*** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1270*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3714*** | |||||
| (0.0242) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0258 | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0549*** | |||||
| (0.0232) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.1223*** | |||||
| (0.0243) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0403** | |||||
| (0.0236) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0723*** | |||||
| (0.0227) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1413*** | |||||
| (0.0235) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2816*** | |||||
| (0.0237) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1259*** | |||||
| (0.0238) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0433** | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0733*** | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0160 | |||||
| (0.0237) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0170 | |||||
| (0.0237) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0646*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0141 | |||||
| (0.0237) | ||||||
| Post 2 indicator | -0.1730*** | |||||
| (0.0150) | ||||||
| Post 3 indicator | -0.3172*** | |||||
| (0.0148) | ||||||
| Post 4 indicator | -0.2624*** | |||||
| (0.0151) | ||||||
| Post 5 indicator | -0.0069 | |||||
| (0.0137) | ||||||
| Post 6 indicator | 0.0467*** | |||||
| (0.0134) | ||||||
| Post 7 indicator | -0.2006*** | |||||
| (0.0149) | ||||||
| Post 8 indicator | -0.1220*** | |||||
| (0.0147) | ||||||
| Post 9 indicator | -0.0596*** | |||||
| (0.0143) | ||||||
| Post 10 indicator | -0.1001*** | |||||
| (0.0143) | ||||||
| Post 11 indicator | -0.1861*** | |||||
| (0.0149) | ||||||
| Post 12 indicator | -0.1377*** | |||||
| (0.0144) | ||||||
| Post 13 indicator | -0.2181*** | |||||
| (0.0149) | ||||||
| Post 14 indicator | -0.3292*** | |||||
| (0.0152) | ||||||
| Post 15 indicator | -0.2956*** | |||||
| (0.0151) | ||||||
| Post 16 indicator | -0.1071*** | |||||
| (0.0157) | ||||||
| Post 17 indicator | -0.3466*** | |||||
| (0.0165) | ||||||
| Post 18 indicator | -0.3702*** | |||||
| (0.0157) | ||||||
| Post 19 indicator | -0.2710*** | |||||
| (0.0160) | ||||||
| Post 20 indicator | -0.0109 | |||||
| (0.0153) | ||||||
| Post 21 indicator | -0.3867*** | |||||
| (0.0148) | ||||||
| Post 22 indicator | -0.3444*** | |||||
| (0.0156) | ||||||
| Post 23 indicator | -0.2481*** | |||||
| (0.0157) | ||||||
| Post 24 indicator | 0.0002 | |||||
| (0.0146) | ||||||
| Post 25 indicator | -0.1878*** | |||||
| (0.0163) | ||||||
| Post 26 indicator | -0.2371*** | |||||
| (0.0167) | ||||||
| Post 27 indicator | -0.2838*** | |||||
| (0.0159) | ||||||
| Post 28 indicator | -0.2778*** | |||||
| (0.0160) | ||||||
| Post 29 indicator | -0.4155*** | |||||
| (0.0157) | ||||||
| Post 30 indicator | -0.2995*** | |||||
| (0.0153) | ||||||
| Post 31 indicator | -0.1578*** | |||||
| (0.0164) | ||||||
| Post 32 indicator | -0.4123*** | |||||
| (0.0157) | ||||||
| Post 33 indicator | -0.3396*** | |||||
| (0.0156) | ||||||
| Post 34 indicator | -0.3310*** | |||||
| (0.0155) | ||||||
| Post 35 indicator | -0.0236* | |||||
| (0.0157) | ||||||
| Post 36 indicator | 0.0545*** | |||||
| (0.0149) | ||||||
| Post 37 indicator | -0.3512*** | |||||
| (0.0157) | ||||||
| Post 38 indicator | -0.4129*** | |||||
| (0.0153) | ||||||
| Post 39 indicator | -0.0380*** | |||||
| (0.0156) | ||||||
| Post 40 indicator | -0.1357*** | |||||
| (0.0156) | ||||||
| Post 41 indicator | -0.3663*** | |||||
| (0.0152) | ||||||
| Post 42 indicator | -0.1535*** | |||||
| (0.0160) | ||||||
| Post 43 indicator | -0.1619*** | |||||
| (0.0160) | ||||||
| Post 44 indicator | -0.3599*** | |||||
| (0.0157) | ||||||
| Post 45 indicator | -0.0820*** | |||||
| (0.0154) | ||||||
| Post 46 indicator | -0.0627*** | |||||
| (0.0160) | ||||||
| Post 47 indicator | -0.2098*** | |||||
| (0.0158) | ||||||
| Post 48 indicator | -0.4348*** | |||||
| (0.0149) | ||||||
| Post 49 indicator | -0.2027*** | |||||
| (0.0164) | ||||||
| Post 50 indicator | -0.0293** | |||||
| (0.0154) | ||||||
| Post 51 indicator | 0.0563*** | |||||
| (0.0148) | ||||||
| Post 52 indicator | -0.4046*** | |||||
| (0.0156) | ||||||
| Post 53 indicator | -0.4662*** | |||||
| (0.0158) | ||||||
| Post 54 indicator | 0.0036 | |||||
| (0.0152) | ||||||
| Post 55 indicator | -0.1195*** | |||||
| (0.0157) | ||||||
| Post 56 indicator | -0.1755*** | |||||
| (0.0163) | ||||||
| Post 57 indicator | -0.1844*** | |||||
| (0.0154) | ||||||
| Post 58 indicator | -0.2638*** | |||||
| (0.0160) | ||||||
| Post 59 indicator | -0.3272*** | |||||
| (0.0160) | ||||||
| Post 60 indicator | -0.3724*** | |||||
| (0.0151) | ||||||
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 | 0.0422 | |||||
| Adj. R2 | 0.0421 | |||||
| R2 (full model) | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 | |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 | |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | 0.04216 | |||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method.
Treatment effects in the last column refer to estimated marginal means (EMMs, or least-squares means) for specified factor combinations (Post-treatment and Misinfo). Point estimates are reported in the upper cell and standard errors, clustered at the user level, reported in parentheses in the bottom cell. t-ratios are reported in brackets.
Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
# Models' predictions
models <- c("Model 1: Baseline",
"Model 2: User ID FE",
"Model 3: User ID FE + Post domain FE",
"Model 4: User ID FE + Post tactic",
"Model 5: User ID FE + Post domain x Post tactic",
"Model 6: User ID FE + Post ID FE")
list_models <- list(baseline_lm,fe1_lm,fe2_lm,fe3_lm,fe4_lm,fe5_lm)
treatmentLevels <- unique(panel_2$treatment)
second_fixed_effects <- list("",
"",
c(fact = unique(panel_2$fact)),
c(fact = unique(panel_2$type)),
list(type = unique(panel_2$type),fact = unique(panel_2$fact)),
c(user = unique(panel_2$post_id)))
second_fixed_effects_name <- list("","","fact","type",c("type","fact"),"post_id")
second_fixed_effects_feglm <- c("share ~ treatment",
"share ~ treatment",
"share ~ treatment + factor(fact)",
"share ~ treatment + factor(type)",
"share ~ treatment + factor(fact)*factor(type)",
"share ~ treatment + factor(post_id)")
for (m in 1:length(models)) {
# output df
prediction_df <- expand.grid(treatment = unique(panel_2$treatment),
post = c(0, 1),
post_is_misinfo = c(0, 1),
prediction = NA,
prediction_se = NA)
for (treatment in unique(panel_2$treatment)) {
for (post in c(0,1)) {
for (misinfo in c(0,1)) {
# Create new data
# 1 vs. 2 fixed effects
# type
if (m %in% c(2,3,4,6)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel_2$user[panel_2$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]])
}
if (m %in% c(5)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel_2$user[panel_2$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]][[1]],
second_fixed_effects_name2 = second_fixed_effects[[m]][[2]])
}
if (m %in% c(2)) {
predictorVariables <- predictorVariables %>% select(-second_fixed_effects_name)
}
if (m %in% c(3,4,6)) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[m]
}
if (m == 5) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[[m]][1]
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name2"] <- second_fixed_effects_name[[m]][2]
}
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
}
}
# predicted value
if (m == 1) {
predictorVariables <- expand.grid(treatment = unique(panel_2$treatment),
post = c(0,1),
post_is_misinfo = c(0,1))
predictions <- predict(baseline_lm, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_lm, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% select(post, post_is_misinfo, estimate, SE, t.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(t.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
# prediction_df <- prediction_df %>%
# group_by(post,post_is_misinfo) %>%
# mutate(lag = ifelse(treatment == "Facts Baseline",
# prediction, NA)) %>%
# tidyr::fill(lag) %>%
# ungroup() %>%
# mutate(te = ifelse(treatment != "Facts Baseline",
# format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
# select(-lag) %>%
# mutate(te = ifelse(is.na(te),"",te))
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
fixed_effect_model <- feglm(as.formula(second_fixed_effects_feglm[m]),panel_2 %>%
filter(post == p & post_is_misinfo == mis), cluster = "user")
# obtain data frame with estimate, std error, and z value
fixed_effect_model_df_add <- data.frame(
treatment = unique(panel_2$treatment),
estimate = format(coef(fixed_effect_model), digits = 4),
SE = sqrt(diag(vcov(fixed_effect_model))),
z.ratio = coef(fixed_effect_model) / sqrt(diag(vcov(fixed_effect_model))),
post = p,
post_is_misinfo = mis
)
# make rownames a column
fixed_effect_model_df_add <- rownames_to_column(fixed_effect_model_df_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>% select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
fixed_effect_model_df_add <- rbind(fixed_effect_model_df_add %>% mutate(treatment = "Facts Baseline"),fixed_effect_model_df_add %>% mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate)
contrast_df <- rbind(contrast_df, fixed_effect_model_df_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","Treatment effect<br>Emotions--Facts"))
print(k)
}| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006670) | -0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.006586) | (0.009374); [0.6346] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004717) | -0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.004657) | (0.006628); [0.3059] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006670) | 0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.006586) | (0.009374); [ -9.3698] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004717) | 0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.004657) | (0.006628); [-16.8909] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | -0.00595 | 0.005948 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01144); [0.5199] | |
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.00203 | 0.002028 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01097); [0.1849] | |
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | 0.08784 | -0.08783 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01252); [-7.016] | |
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | 0.11196 | -0.112 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01114); [-10.05] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68362 | (0.001872) | -0.00537 | 0.006384 |
| Emotions | 0 | 0 | 0.68899 | (0.001721) | (0.01131); [0.5643] | |
| Facts Baseline | 0 | 1 | 0.53765 | (0.001872) | -0.00391 | 0.004800 |
| Emotions | 0 | 1 | 0.54156 | (0.001721) | (0.01084); [0.4428] | |
| Facts Baseline | 1 | 0 | 0.6088 | (0.001872) | 0.08947 | -0.088912 |
| Emotions | 1 | 0 | 0.51933 | (0.001721) | (0.01227); [-7.244] | |
| Facts Baseline | 1 | 1 | 0.46512 | (0.001872) | 0.11302 | -0.11371 |
| Emotions | 1 | 1 | 0.3521 | (0.001721) | (0.01107); [-10.27] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.60427 | (0.003416) | -0.00594 | |
| Emotions | 0 | 0 | 0.61021 | (0.003107) | ||
| Facts Baseline | 0 | 1 | 0.56484 | (0.003416) | -0.00203 | |
| Emotions | 0 | 1 | 0.56687 | (0.003107) | ||
| Facts Baseline | 1 | 0 | 0.52847 | (0.003416) | 0.08783 | |
| Emotions | 1 | 0 | 0.44064 | (0.003107) | ||
| Facts Baseline | 1 | 1 | 0.49119 | (0.003416) | 0.11195 | |
| Emotions | 1 | 1 | 0.37924 | (0.003107) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.59559 | (0.000977) | -0.00655 | |
| Emotions | 0 | 0 | 0.60214 | (0.0009033) | ||
| Facts Baseline | 0 | 1 | 0.56695 | (0.000977) | -0.00416 | |
| Emotions | 0 | 1 | 0.57111 | (0.0009033) | ||
| Facts Baseline | 1 | 0 | 0.52128 | (0.000977) | 0.08907 | |
| Emotions | 1 | 0 | 0.43221 | (0.0009033) | ||
| Facts Baseline | 1 | 1 | 0.49666 | (0.000977) | 0.11610 | |
| Emotions | 1 | 1 | 0.38056 | (0.0009033) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.63011 | (0.0009693) | -0.00654 | 0.006384 |
| Emotions | 0 | 0 | 0.63665 | (0.0008952) | (0.01131); [0.5643] | |
| Facts Baseline | 0 | 1 | 0.55545 | (0.0009693) | -0.00416 | 0.004202 |
| Emotions | 0 | 1 | 0.55961 | (0.0008952) | (0.01074); [0.3911] | |
| Facts Baseline | 1 | 0 | 0.55579 | (0.0009693) | 0.08906 | -0.088912 |
| Emotions | 1 | 0 | 0.46673 | (0.0008952) | (0.01227); [-7.244] | |
| Facts Baseline | 1 | 1 | 0.48516 | (0.0009693) | 0.11610 | -0.11650 |
| Emotions | 1 | 1 | 0.36906 | (0.0008952) | (0.01099); [-10.6] |
Click for All treatment arms
# version 2: all treatment arms
formula_fe1_all_lm <- "share ~ treatment*post*post_is_misinfo | user"
fe1_all_lm <- feols(as.formula(formula_fe1_all_lm),panel)
formula_fe2_all_lm <- "share ~ treatment*post*post_is_misinfo + factor(fact) | user"
fe2_all_lm <- feols(as.formula(formula_fe2_all_lm),panel)
formula_fe3_all_lm <- "share ~ treatment*post*post_is_misinfo + factor(type) | user"
fe3_all_lm <- feols(as.formula(formula_fe3_all_lm),panel)
formula_fe4_all_lm <- "share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) | user"
fe4_all_lm <- feols(as.formula(formula_fe4_all_lm),panel)
formula_fe5_all_lm <- "share ~ treatment*post*post_is_misinfo + factor(post_id) | user"
fe5_all_lm <- feols(as.formula(formula_fe5_all_lm),panel)# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
# estimate out-of-sample R2 for all fixed effects models
# use R2oosse function from oosse package
# w user FE
models_fe <- c('fe1_all_lm','fe2_all_lm','fe3_all_lm','fe4_all_lm','fe5_all_lm')
# Estimate R2oosse for all models
R2oos_fe_alltreat <- list()
for (m in seq_along(models_fe)) {
if (file.exists(paste0("R2oos_fe_alltreat_", m, ".rds"))) {
R2oos_fe_alltreat[[m]] <- readRDS(paste0("R2oos_fe_alltreat_", m, ".rds"))
} else {
formula <- gsub("^share", "", gsub(" \\| user", "", get(paste0("formula_fe", m, "_all_lm"))))
X <- model.matrix(as.formula(formula),
data = panel %>%
mutate(type = as.numeric(factor(panel$type, levels = unique(panel$type)))))
X <- X[, !(colnames(X) %in% "(Intercept)")]
X <- cbind(X, user = panel$user, treatment = panel$treatment)
# add fixed effects based on model
if (m == 2) {
X <- cbind(X, fact = panel$fact)
}
if (m == 3) {
X <- cbind(X, type = panel$type)
}
if (m == 4) {
X <- cbind(X, fact = panel$fact, type = panel$type)
}
if (m == 5) {
X <- cbind(X, post_id = panel$post_id)
}
# Define the fit function for models with fixed effects
fitFunLM_fe <- function(y, x) {
feols(as.formula(get(paste0("formula_fe", m, "_all_lm"))),
data = panel %>% mutate(treatment = as.integer(treatment)))
}
# Define the prediction function for models with fixed effects
predFunLM_fe <- function(mod, x) {
predict(mod, newdata = x, type = "response")
}
# Calculate R2oosse
R2oos_fe_alltreat[[m]] <- oosse::R2oosse(y = panel$share,
x = X,
fitFun = fitFunLM_fe,
predFun = predFunLM_fe,
methodMSE = "CV",
cvReps = 100,
nFold = 10,
nBootstrapsCor = 25)
# Save results
saveRDS(R2oos_fe_alltreat[[m]], paste0("R2oos_fe_alltreat_", m, ".rds"))
}
}
R2oos_fe_alltreat_all <- as.numeric(unlist(lapply(R2oos_fe_alltreat, function(x) x$R2[1])),4)The sample includes all users who completed the post-survey (n = 8,684). We consider all 9 posts in the pre-treatment period (“pre”) and all 9 posts in the post-treatment period (i.e., “post”).
# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_all_lm,fe1_all_lm,fe2_all_lm,fe3_all_lm,fe4_all_lm,fe5_all_lm)
names <- unique(c(names(baseline_all_lm$coefficients),
names(fe1_all_lm$coefficients),
names(fe2_all_lm$coefficients),
names(fe3_all_lm$coefficients),
names(fe4_all_lm$coefficients),
names(fe5_all_lm$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2oos_fe_alltreat_all")) {
#R2oos_fe_alltreat_all <- c(R2out_all, R2oos_fe_alltreat_all)
R2oos_fe_alltreat_all <- c(R2out_all,rep("",length(model_list)-1))
} else {
R2oos_fe_alltreat_all <- rep("",length(model_list))
}
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
#na_coeff <- names(model_list[[m]]$coefficients[,1])[is.na(model_list[[m]]$coefficients[,1])] # felm
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# keep variables
keep <- list("treatmentFacts Baseline:post:post_is_misinfo" = "Facts Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentReasoning:post:post_is_misinfo" = "Reasoning x 1[Post-treatment] x 1[Misinfo]",
"treatmentCombo:post:post_is_misinfo" = "Combo x 1[Post-treatment] x 1[Misinfo]",
"treatmentFacts Baseline:post" = "Facts Baseline x 1[Post-treatment]",
"treatmentNo-course Baseline" = "No-course Baseline x 1[Post-treatment]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentReasoning:post" = "Reasoning x 1[Post-treatment]",
"treatmentCombo:post" = "Combo x 1[Post-treatment]",
"treatmentFacts Baseline:post_is_misinfo" = "Facts Baseline x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Misinfo]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]",
"treatmentReasoning:post_is_misinfo" = "Reasoning x 1[Misinfo]",
"treatmentCombo:post_is_misinfo" = "Combo x 1[Misinfo]")
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_fe_alltreat_all)),
reorder.gof = c(8,6,7,9,10,11,12,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0118) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0118) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0162) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0166) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0164) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| No-course Baseline x 1[Post-treatment] | 0.0213** | |||||
| (0.0096) | ||||||
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0132) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0135) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0134) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0115) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0117) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0116) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 | 0.0429 | |||||
| Adj. R2 | 0.0428 | |||||
| R2 (full model) | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 | |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 | |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | 0.0428187159122504 | |||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"\\textit{{Out-of-sample R}$^2$" = c(R2oos_fe_alltreat_all)),
reorder.gof = c(8,6,7,9,10,11,12,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6829*** | |||||
| (0.0067) | ||||||
| No-course Baseline | 0.0213** | |||||
| (0.0096) | ||||||
| Emotions | 0.0059 | |||||
| (0.0094) | ||||||
| Reasoning | 0.0209** | |||||
| (0.0096) | ||||||
| Combo | 0.0035 | |||||
| (0.0095) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0742*** | -0.0742*** |
| (0.0094) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1459*** | -0.1173*** | -0.1537*** | -0.4320*** |
| (0.0082) | (0.0069) | (0.0067) | (0.0095) | (0.0127) | (0.0123) | |
| No-course Baseline x 1[Post-treatment] | 0.0514*** | 0.0514*** | 0.0513*** | 0.0514*** | 0.0507*** | 0.0507*** |
| (0.0136) | (0.0126) | (0.0124) | (0.0126) | (0.0123) | (0.0123) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0132) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0135) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0134) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| No-course Baseline x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0118) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0115) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0117) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0116) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0039 | 0.0039 |
| (0.0115) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 | 0.0033 | 0.0046 | 0.0033 | 0.0042 | 0.0042 |
| (0.0167) | (0.0140) | (0.0133) | (0.0140) | (0.0131) | (0.0131) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0162) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0166) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0164) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| Post domain 2 | -0.2001*** | -0.1794*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 3 | -0.2813*** | -0.3043*** | ||||
| (0.0055) | (0.0095) | |||||
| Post domain 4 | -0.1804*** | -0.2332*** | ||||
| (0.0056) | (0.0097) | |||||
| Post domain 5 | 0.0545*** | -0.0010 | ||||
| (0.0051) | (0.0087) | |||||
| Post domain 6 | 0.0164*** | 0.0389*** | ||||
| (0.0054) | (0.0084) | |||||
| Post domain 7 | -0.2347*** | -0.1827*** | ||||
| (0.0056) | (0.0094) | |||||
| Post domain 8 | -0.2195*** | -0.1160*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 9 | 0.0369*** | -0.0608*** | ||||
| (0.0053) | (0.0091) | |||||
| Post domain 10 | -0.0565*** | -0.0963*** | ||||
| (0.0054) | (0.0091) | |||||
| Post domain 11 | -0.1586*** | -0.1739*** | ||||
| (0.0057) | (0.0094) | |||||
| Post domain 12 | -0.1047*** | -0.1343*** | ||||
| (0.0054) | (0.0092) | |||||
| Post domain 13 | -0.1541*** | -0.2155*** | ||||
| (0.0055) | (0.0093) | |||||
| Post domain 14 | -0.2759*** | -0.3151*** | ||||
| (0.0057) | (0.0097) | |||||
| Post domain 15 | -0.1868*** | -0.2755*** | ||||
| (0.0056) | (0.0095) | |||||
| 1[Combo manipulation tactic] | -0.0512*** | -0.0275*** | ||||
| (0.0041) | (0.0078) | |||||
| 1[Emotions manipulation tactic] | -0.0154*** | -0.0585*** | ||||
| (0.0041) | (0.0080) | |||||
| 1[Reasoning manipulation tactic] | -0.0143*** | 0.0208*** | ||||
| (0.0041) | (0.0079) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0536*** | |||||
| (0.0158) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0462*** | |||||
| (0.0152) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0789*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.0971*** | |||||
| (0.0149) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3110*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0471*** | |||||
| (0.0153) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0192 | |||||
| (0.0155) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1469*** | |||||
| (0.0150) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0318** | |||||
| (0.0154) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0595*** | |||||
| (0.0153) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0230* | |||||
| (0.0156) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0600*** | |||||
| (0.0154) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0296** | |||||
| (0.0153) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.0822*** | |||||
| (0.0153) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0845*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1118*** | |||||
| (0.0154) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0617*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1184*** | |||||
| (0.0148) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1477*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0324** | |||||
| (0.0151) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1539*** | |||||
| (0.0153) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1660*** | |||||
| (0.0153) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1090*** | |||||
| (0.0152) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0397*** | |||||
| (0.0154) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1449*** | |||||
| (0.0153) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.1875*** | |||||
| (0.0151) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1025*** | |||||
| (0.0154) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3384*** | |||||
| (0.0155) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0427*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0517*** | |||||
| (0.0150) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.0985*** | |||||
| (0.0155) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0326** | |||||
| (0.0150) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0629*** | |||||
| (0.0144) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1533*** | |||||
| (0.0149) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2920*** | |||||
| (0.0151) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1260*** | |||||
| (0.0151) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0377*** | |||||
| (0.0151) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0529*** | |||||
| (0.0155) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0132 | |||||
| (0.0151) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0267** | |||||
| (0.0153) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0441*** | |||||
| (0.0152) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0236* | |||||
| (0.0151) | ||||||
| Post 2 indicator | -0.1794*** | |||||
| (0.0095) | ||||||
| Post 3 indicator | -0.3043*** | |||||
| (0.0095) | ||||||
| Post 4 indicator | -0.2332*** | |||||
| (0.0097) | ||||||
| Post 5 indicator | -0.0010 | |||||
| (0.0087) | ||||||
| Post 6 indicator | 0.0389*** | |||||
| (0.0084) | ||||||
| Post 7 indicator | -0.1827*** | |||||
| (0.0094) | ||||||
| Post 8 indicator | -0.1160*** | |||||
| (0.0095) | ||||||
| Post 9 indicator | -0.0608*** | |||||
| (0.0091) | ||||||
| Post 10 indicator | -0.0963*** | |||||
| (0.0091) | ||||||
| Post 11 indicator | -0.1739*** | |||||
| (0.0094) | ||||||
| Post 12 indicator | -0.1343*** | |||||
| (0.0092) | ||||||
| Post 13 indicator | -0.2155*** | |||||
| (0.0093) | ||||||
| Post 14 indicator | -0.3151*** | |||||
| (0.0097) | ||||||
| Post 15 indicator | -0.2755*** | |||||
| (0.0095) | ||||||
| Post 16 indicator | 0.2508*** | |||||
| (0.0118) | ||||||
| Post 17 indicator | 0.0178* | |||||
| (0.0119) | ||||||
| Post 18 indicator | -0.0072 | |||||
| (0.0113) | ||||||
| Post 19 indicator | 0.0965*** | |||||
| (0.0114) | ||||||
| Post 20 indicator | 0.3469*** | |||||
| (0.0115) | ||||||
| Post 21 indicator | -0.0213** | |||||
| (0.0113) | ||||||
| Post 22 indicator | 0.0211** | |||||
| (0.0114) | ||||||
| Post 23 indicator | 0.1157*** | |||||
| (0.0118) | ||||||
| Post 24 indicator | 0.3370*** | |||||
| (0.0116) | ||||||
| Post 25 indicator | 0.1863*** | |||||
| (0.0118) | ||||||
| Post 26 indicator | 0.1364*** | |||||
| (0.0120) | ||||||
| Post 27 indicator | 0.0935*** | |||||
| (0.0116) | ||||||
| Post 28 indicator | 0.0953*** | |||||
| (0.0117) | ||||||
| Post 29 indicator | -0.0346*** | |||||
| (0.0113) | ||||||
| Post 30 indicator | 0.0576*** | |||||
| (0.0116) | ||||||
| Post 31 indicator | 0.2198*** | |||||
| (0.0117) | ||||||
| Post 32 indicator | -0.0441*** | |||||
| (0.0115) | ||||||
| Post 33 indicator | 0.0273*** | |||||
| (0.0113) | ||||||
| Post 34 indicator | 0.0482*** | |||||
| (0.0114) | ||||||
| Post 35 indicator | 0.3371*** | |||||
| (0.0115) | ||||||
| Post 36 indicator | 0.4064*** | |||||
| (0.0116) | ||||||
| Post 37 indicator | 0.0046 | |||||
| (0.0112) | ||||||
| Post 38 indicator | -0.0501*** | |||||
| (0.0112) | ||||||
| Post 39 indicator | 0.3250*** | |||||
| (0.0117) | ||||||
| Post 40 indicator | 0.2324*** | |||||
| (0.0118) | ||||||
| Post 41 indicator | 0.0062 | |||||
| (0.0116) | ||||||
| Post 42 indicator | 0.2304*** | |||||
| (0.0118) | ||||||
| Post 43 indicator | 0.1917*** | |||||
| (0.0115) | ||||||
| Post 44 indicator | 0.0072 | |||||
| (0.0114) | ||||||
| Post 45 indicator | 0.2826*** | |||||
| (0.0118) | ||||||
| Post 46 indicator | 0.2991*** | |||||
| (0.0118) | ||||||
| Post 47 indicator | 0.1624*** | |||||
| (0.0118) | ||||||
| Post 48 indicator | -0.0569*** | |||||
| (0.0110) | ||||||
| Post 49 indicator | 0.1643*** | |||||
| (0.0121) | ||||||
| Post 50 indicator | 0.3306*** | |||||
| (0.0118) | ||||||
| Post 51 indicator | 0.4009*** | |||||
| (0.0113) | ||||||
| Post 52 indicator | -0.0369*** | |||||
| (0.0113) | ||||||
| Post 53 indicator | -0.1089*** | |||||
| (0.0115) | ||||||
| Post 54 indicator | 0.3643*** | |||||
| (0.0116) | ||||||
| Post 55 indicator | 0.2405*** | |||||
| (0.0118) | ||||||
| Post 56 indicator | 0.1781*** | |||||
| (0.0122) | ||||||
| Post 57 indicator | 0.1780*** | |||||
| (0.0115) | ||||||
| Post 58 indicator | 0.1102*** | |||||
| (0.0117) | ||||||
| Post 59 indicator | 0.0281*** | |||||
| (0.0116) | ||||||
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 | 0.0429 | |||||
| Adj. R2 | 0.0428 | |||||
| R2 (full model) | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 | |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 | |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| \textit{{Out-of-sample R}\(^2\) | 0.0428187159122504 | |||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method. Treatment effects standard errors are clustered at the user level and reported in parentheses. t-ratios are reported in brackets. Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
# Models' predictions
models <- c("Model 1: Baseline",
"Model 2: User ID FE",
"Model 3: User ID FE + Post domain FE",
"Model 4: User ID FE + Post tactic",
"Model 5: User ID FE + Post domain x Post tactic",
"Model 6: User ID FE + Post ID FE")
list_models <- list(baseline_all_lm, fe1_all_lm,fe2_all_lm,fe3_all_lm,fe4_all_lm,fe5_all_lm)
treatmentLevels <- unique(panel$treatment)
second_fixed_effects <- list("",
"",
c(fact = unique(panel$fact)),
c(fact = unique(panel$type)),
list(type = unique(panel$type),fact = unique(panel$fact)),
c(user = unique(panel$post_id)))
second_fixed_effects_name <- list("","","fact","type",c("type","fact"),"post_id")
second_fixed_effects_feglm <- c("share ~ treatment",
"share ~ treatment",
"share ~ treatment + factor(fact)",
"share ~ treatment + factor(type)",
"share ~ treatment + factor(fact)*factor(type)",
"share ~ treatment + factor(post_id)")
for (m in 1:length(models)) {
# output df
prediction_df <- expand.grid(treatment = unique(panel$treatment),
post = c(0, 1),
post_is_misinfo = c(0, 1),
prediction = NA,
prediction_se = NA)
for (treatment in unique(panel$treatment)) {
for (post in c(0,1)) {
for (misinfo in c(0,1)) {
# Create new data
# 1 vs. 2 fixed effects
if (m %in% c(2,3,4,6)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel$user[panel$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]])
}
if (m %in% c(5)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel$user[panel$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]][[1]],
second_fixed_effects_name2 = second_fixed_effects[[m]][[2]])
}
# change name of second-fixed effect
if (m %in% c(2)) {
predictorVariables <- predictorVariables %>% select(-second_fixed_effects_name)
}
if (m %in% c(3,4,6)) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[m]
}
if (m %in% c(5)) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[[m]][1]
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name2"] <- second_fixed_effects_name[[m]][2]
}
# predict value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
}
}
# predicted value
if (m == 1) {
predictorVariables <- expand.grid(treatment = unique(panel$treatment),
post = c(0,1),
post_is_misinfo = c(0,1))
predictions <- predict(baseline_all_lm, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_all_lm, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res)
contrast_df <- contrast_df %>% filter(contrast == "Emotions - Facts Baseline") %>%
select(post, post_is_misinfo, estimate, SE, t.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(t.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
# prediction_df <- prediction_df %>%
# group_by(post,post_is_misinfo) %>%
# mutate(lag = ifelse(treatment == "Facts Baseline",
# prediction, NA)) %>%
# tidyr::fill(lag) %>%
# ungroup() %>%
# mutate(te = ifelse(treatment != "Facts Baseline",
# format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
# select(-lag) %>%
# mutate(te = ifelse(is.na(te),"",te))
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
fixed_effect_model <- feglm(as.formula(second_fixed_effects_feglm[m]),panel %>%
filter(post == p & post_is_misinfo == mis), cluster = "user")
# obtain data frame with estimate, std error, and z value
fixed_effect_model_df_add <- data.frame(
treatment = c("No-course Baseline","Emotions","Reasoning","Combo"),
estimate = format(coef(fixed_effect_model)[grepl("treatment",names(coef(fixed_effect_model)))], digits = 4),
SE = sqrt(diag(vcov(fixed_effect_model)))[grepl("treatment",names(diag(vcov(fixed_effect_model))))],
z.ratio = coef(fixed_effect_model)[grepl("treatment",names(coef(fixed_effect_model)))] / sqrt(diag(vcov(fixed_effect_model)))[grepl("treatment",names(diag(vcov(fixed_effect_model))))],
post = p,
post_is_misinfo = mis
)
# make rownames a column
fixed_effect_model_df_add <- rownames_to_column(fixed_effect_model_df_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>% select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
fixed_effect_model_df_add <- rbind(fixed_effect_model_df_add %>% mutate(treatment = "Facts Baseline"),fixed_effect_model_df_add %>% mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
contrast_df <- rbind(contrast_df, fixed_effect_model_df_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","Treatment effect<br>Emotions--Facts"))
print(k)
}| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006659) | -0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.006575) | (0.009358); [0.6356] | |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006983) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006853) | ||
| Combo | 0 | 0 | 0.6864 | (0.006754) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004709) | -0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.004649) | (0.006617); [0.3064] | |
| No-course Baseline | 0 | 1 | 0.5650 | (0.004937) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004846) | ||
| Combo | 0 | 1 | 0.5462 | (0.004776) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006659) | 0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.006575) | (0.009358); [ -9.3854] | |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006983) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006853) | ||
| Combo | 1 | 0 | 0.5222 | (0.006754) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004709) | 0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.004649) | (0.006617); [-16.9190] | |
| No-course Baseline | 1 | 1 | 0.5461 | (0.004937) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004846) | ||
| Combo | 1 | 1 | 0.3715 | (0.004776) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | -0.00595 | 0.005948 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01144); [0.5199] | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.00203 | 0.002028 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01097); [0.1849] | |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | 0.08784 | -0.08783 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01252); [-7.017] | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | 0.11196 | -0.11196 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01114); [-10.05] | |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68359 | (0.001863) | -0.00544 | 0.006456 |
| Emotions | 0 | 0 | 0.68903 | (0.001712) | (0.01131); [0.5707] | |
| No-course Baseline | 0 | 0 | 0.70468 | (0.001914) | ||
| Reasoning | 0 | 0 | 0.70177 | (0.001852) | ||
| Combo | 0 | 0 | 0.68565 | (0.001838) | ||
| Facts Baseline | 0 | 1 | 0.53767 | (0.001863) | -0.00388 | 0.004722 |
| Emotions | 0 | 1 | 0.54155 | (0.001712) | (0.01084); [0.4355] | |
| No-course Baseline | 0 | 1 | 0.56355 | (0.001914) | ||
| Reasoning | 0 | 1 | 0.54155 | (0.001852) | ||
| Combo | 0 | 1 | 0.54616 | (0.001838) | ||
| Facts Baseline | 1 | 0 | 0.60875 | (0.001863) | 0.08945 | -0.08884 |
| Emotions | 1 | 0 | 0.5193 | (0.001712) | (0.01227); [-7.239] | |
| No-course Baseline | 1 | 0 | 0.6811 | (0.001914) | ||
| Reasoning | 1 | 0 | 0.54844 | (0.001852) | ||
| Combo | 1 | 0 | 0.52264 | (0.001838) | ||
| Facts Baseline | 1 | 1 | 0.46512 | (0.001863) | 0.11300 | -0.11362 |
| Emotions | 1 | 1 | 0.35212 | (0.001712) | (0.01107); [-10.26] | |
| No-course Baseline | 1 | 1 | 0.54686 | (0.001914) | ||
| Reasoning | 1 | 1 | 0.40336 | (0.001852) | ||
| Combo | 1 | 1 | 0.37125 | (0.001838) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.66264 | (0.003376) | -0.00594 | |
| Emotions | 0 | 0 | 0.66858 | (0.003064) | ||
| No-course Baseline | 0 | 0 | 0.6839 | (0.003455) | ||
| Reasoning | 0 | 0 | 0.68351 | (0.003342) | ||
| Combo | 0 | 0 | 0.66617 | (0.003319) | ||
| Facts Baseline | 0 | 1 | 0.54538 | (0.003376) | -0.00203 | |
| Emotions | 0 | 1 | 0.54741 | (0.003064) | ||
| No-course Baseline | 0 | 1 | 0.57175 | (0.003455) | ||
| Reasoning | 0 | 1 | 0.54668 | (0.003342) | ||
| Combo | 0 | 1 | 0.55291 | (0.003319) | ||
| Facts Baseline | 1 | 0 | 0.58684 | (0.003376) | 0.08783 | |
| Emotions | 1 | 0 | 0.49901 | (0.003064) | ||
| No-course Baseline | 1 | 0 | 0.65954 | (0.003455) | ||
| Reasoning | 1 | 0 | 0.53006 | (0.003342) | ||
| Combo | 1 | 0 | 0.50199 | (0.003319) | ||
| Facts Baseline | 1 | 1 | 0.47173 | (0.003376) | 0.11195 | |
| Emotions | 1 | 1 | 0.35978 | (0.003064) | ||
| No-course Baseline | 1 | 1 | 0.55281 | (0.003455) | ||
| Reasoning | 1 | 1 | 0.41079 | (0.003342) | ||
| Combo | 1 | 1 | 0.3782 | (0.003319) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68932 | (0.0009589) | -0.00661 | |
| Emotions | 0 | 0 | 0.69593 | (0.0008842) | ||
| No-course Baseline | 0 | 0 | 0.71016 | (0.000986) | ||
| Reasoning | 0 | 0 | 0.70796 | (0.000956) | ||
| Combo | 0 | 0 | 0.6916 | (0.0009496) | ||
| Facts Baseline | 0 | 1 | 0.53563 | (0.0009589) | -0.00416 | |
| Emotions | 0 | 1 | 0.53979 | (0.0008842) | ||
| No-course Baseline | 0 | 1 | 0.56092 | (0.000986) | ||
| Reasoning | 0 | 1 | 0.53896 | (0.000956) | ||
| Combo | 0 | 1 | 0.54489 | (0.0009496) | ||
| Facts Baseline | 1 | 0 | 0.61511 | (0.0009589) | 0.08900 | |
| Emotions | 1 | 0 | 0.52611 | (0.0008842) | ||
| No-course Baseline | 1 | 0 | 0.68661 | (0.000986) | ||
| Reasoning | 1 | 0 | 0.55543 | (0.000956) | ||
| Combo | 1 | 0 | 0.5291 | (0.0009496) | ||
| Facts Baseline | 1 | 1 | 0.46532 | (0.0009589) | 0.11603 | |
| Emotions | 1 | 1 | 0.34929 | (0.0008842) | ||
| No-course Baseline | 1 | 1 | 0.54547 | (0.000986) | ||
| Reasoning | 1 | 1 | 0.40099 | (0.000956) | ||
| Combo | 1 | 1 | 0.36686 | (0.0009496) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.89807 | (0.001031) | -0.00662 | 0.006456 |
| Emotions | 0 | 0 | 0.90469 | (0.0009602) | (0.01131); [0.5707] | |
| No-course Baseline | 0 | 0 | 0.91892 | (0.001063) | ||
| Reasoning | 0 | 0 | 0.91672 | (0.001033) | ||
| Combo | 0 | 0 | 0.90035 | (0.001024) | ||
| Facts Baseline | 0 | 1 | 0.46604 | (0.001031) | -0.00416 | 0.004337 |
| Emotions | 0 | 1 | 0.4702 | (0.0009602) | (0.01075); [0.4033] | |
| No-course Baseline | 0 | 1 | 0.49133 | (0.001063) | ||
| Reasoning | 0 | 1 | 0.46937 | (0.001033) | ||
| Combo | 0 | 1 | 0.47529 | (0.001024) | ||
| Facts Baseline | 1 | 0 | 0.82387 | (0.001031) | 0.08901 | -0.08884 |
| Emotions | 1 | 0 | 0.73486 | (0.0009602) | (0.01227); [-7.239] | |
| No-course Baseline | 1 | 0 | 0.89536 | (0.001063) | ||
| Reasoning | 1 | 0 | 0.76419 | (0.001033) | ||
| Combo | 1 | 0 | 0.73785 | (0.001024) | ||
| Facts Baseline | 1 | 1 | 0.39573 | (0.001031) | 0.11603 | -0.11615 |
| Emotions | 1 | 1 | 0.2797 | (0.0009602) | (0.011); [-10.56] | |
| No-course Baseline | 1 | 1 | 0.47588 | (0.001063) | ||
| Reasoning | 1 | 1 | 0.3314 | (0.001033) | ||
| Combo | 1 | 1 | 0.29727 | (0.001024) |
Covariates
# version 1: 2 treatment arms
covariate_list <- c(demographics_demeaned_treatment,
sharing_demeaned_treatment,
sharing_tactic_demeaned_treatment,
demographics_sharing_demeaned_treatment,
all_covariates_demeaned_treatment)
cov_estimates_list <- list()
for (cov_list in covariate_list) {
# Demographics
formula_cov1_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list)
cov1_lm <- feols(as.formula(formula_cov1_lm),panel_2)
formula_cov2_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list," | user")
cov2_lm <- feols(as.formula(formula_cov2_lm),panel_2)
# Misinfo and Non-misinfo sharing rates
formula_cov3_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact) "," + ",cov_list," | user")
cov3_lm <- feols(as.formula(formula_cov3_lm),panel_2)
# Misinfo, Non-misinfo, and Tactic sharing rates
formula_cov4_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(type) "," + ",cov_list," | user")
cov4_lm <- feols(as.formula(formula_cov4_lm),panel_2)
# Demographics and Misinfo and Non-misinfo sharing rates
formula_cov5_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) "," + ",cov_list," | user")
cov5_lm <- feols(as.formula(formula_cov5_lm),panel_2)
# Demographics, Misinfo and Non-misinfo sharing rates, and Tactic sharing rates
formula_cov6_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(post_id) "," + ",cov_list," | user")
cov6_lm <- feols(as.formula(formula_cov6_lm),panel_2)
cov_estimates_list[[length(cov_estimates_list) + 1]] <- list(cov1_lm,cov2_lm,cov3_lm,cov4_lm,cov5_lm,cov6_lm)
}# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
# estimate out-of-sample R2 for all fixed effects models
# use R2oosse function from oosse package
# w user FE
if (file.exists("cov_roose_list.rds")) {
cov_roose_list <- readRDS("cov_roose_list.rds")
# extract each element from list
R2oos_cov_2treat_all <- (cov_roose_list)
} else {
cov_roose_list <- list()
models_cov <- list('cov1_lm','cov2_lm','cov3_lm','cov4_lm','cov5_lm','cov6_lm')
for (cov_list in covariate_list) {
# Demographics
formula_cov1_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list)
formula_cov2_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list," | user")
# Misinfo and Non-misinfo sharing rates
formula_cov3_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact) "," + ",cov_list," | user")
# Misinfo, Non-misinfo, and Tactic sharing rates
formula_cov4_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(type) "," + ",cov_list," | user")
# Demographics and Misinfo and Non-misinfo sharing rates
formula_cov5_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) "," + ",cov_list," | user")
# Demographics, Misinfo and Non-misinfo sharing rates, and Tactic sharing rates
formula_cov6_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(post_id) "," + ",cov_list," | user")
# Estimate R2oosse for all models
R2oos_cov_2treat <- list()
for (m in seq_along(models_cov)) {
#print(paste0('cov: ',length(cov_roose_list) + 1,'; model: ',m))
if (file.exists(paste0("R2oos_cov_2treat_", m,"_",length(cov_roose_list) + 1,".rds"))) {
R2oos_cov_2treat[[m]] <- readRDS(paste0("R2oos_cov_2treat_", m,"_",length(cov_roose_list) + 1,".rds"))
} else {
formula <- gsub("^share", "", gsub(" \\| user", "", get(paste0("formula_cov", m, "_lm"))))
X <- model.matrix(as.formula(formula),
data = panel_2 %>%
mutate(type = as.numeric(factor(panel_2$type, levels = unique(panel_2$type)))))
X <- X[, !(colnames(X) %in% "(Intercept)")]
X <- cbind(X, user = panel_2$user, treatment = panel_2$treatment)
# identify estimation dataset
estimation_data <- panel_2 %>% filter(row_number() %in% fixest::obs(get(models_cov[[m]])))
if ("treatment" %in% colnames(X)) {
X[,"treatment"] <- factor(estimation_data$treatment,
levels = levels(estimation_data$treatment))
}
# add fixed effects based on model
if (m == 3) {
X <- cbind(X, fact = estimation_data$fact)
}
if (m == 4) {
X <- cbind(X, type = as.numeric(factor(estimation_data$type, levels = unique(estimation_data$type))))
}
if (m == 5) {
X <- cbind(X, fact = estimation_data$fact,
type = as.numeric(factor(estimation_data$type), levels = unique(estimation_data$type)))
}
if (m == 6) {
X <- cbind(X, post_id = estimation_data$post_id)
}
# Define the fit function for models with fixed effects
fitFunLM_cov <- function(y, x) {
fixest::feols(as.formula(get(paste0("formula_cov", m, "_lm"))),
data = estimation_data %>%
mutate(type = as.numeric(factor(estimation_data$type, levels = unique(estimation_data$type)))))
}
# Define the prediction function for models with fixed effects
predFunLM_cov <- function(mod,x) {
predict(mod, newdata = x, type = "response")
}
# Calculate R2oosse
R2oos_cov_2treat[[m]] <- oosse::R2oosse(y = estimation_data$share,
x = X,
fitFun = fitFunLM_cov,
predFun = predFunLM_cov,
methodMSE = "CV",
cvReps = 100,
nFold = 10,
nBootstrapsCor = 25)
# Save results
saveRDS(R2oos_cov_2treat[[m]], paste0("R2oos_cov_2treat_", m,"_",length(cov_roose_list) + 1,".rds"))
}
}
R2oos_cov_2treat_all <- as.numeric(unlist(lapply(R2oos_cov_2treat, function(x) x$R2[1])),4)
cov_roose_list[[length(cov_roose_list) + 1]] <- R2oos_cov_2treat_all
}
saveRDS(cov_roose_list, "cov_roose_list.rds")
}covariate_list_names <- c("Demographics",
"Sharing rates",
"Tactic sharing rates",
"Demographics + Sharing rates",
"Demographics + Sharing rates + Tactic sharing rates")
for (cov in seq(length(covariate_list))) {
# Print model
headers <- c("Share (binary indicator)")
model_list <- cov_estimates_list[[cov]]
names <- unique(c(names(cov_estimates_list[[cov]][[1]]$coefficients),
names(cov_estimates_list[[cov]][[2]]$coefficients),
names(cov_estimates_list[[cov]][[3]]$coefficients),
names(cov_estimates_list[[cov]][[4]]$coefficients),
names(cov_estimates_list[[cov]][[5]]$coefficients),
names(cov_estimates_list[[cov]][[6]]$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# out-of-sample R2
if (any(ls() %in% "R2oos_cov_2treat_all")) {
# obtain first elements of list
#R2oos_cov_2treat_all <- cov_roose_list[[cov]]
R2oos_cov_2treat_all <- rep("",length(model_list))
} else {
R2oos_cov_2treat_all <- rep("",length(model_list))
}
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# Print model name
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_2treat_all)),
reorder.gof = c(6,7,8,9,10,5,1,2,3,4)
)
cat(html_table)
cat("<br>")
# Full table (all independent variables)
cat('<details>')
cat('<summary>Click to view the full table</summary>')
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_2treat_all)),
reorder.gof = c(6,7,8,9,10,5,1,2,3,4)
)
cat(html_table)
cat('</details>')
cat('<br>\n<br>\n<br>')
}| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0159) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0130) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0112) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.0865 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.0859 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6877*** | |||||
| (0.0065) | ||||||
| Emotions | 0.0004 | |||||
| (0.0092) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0743*** | -0.0743*** |
| (0.0092) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1460*** | -0.0394*** | -0.0286*** | -0.0747*** |
| (0.0080) | (0.0069) | (0.0067) | (0.0060) | (0.0066) | (0.0054) | |
| Age | -0.0036 | |||||
| (0.0030) | ||||||
| Man | 0.0136*** | |||||
| (0.0027) | ||||||
| Some college | -0.0063** | |||||
| (0.0033) | ||||||
| Bachelor’s degree | -0.0143*** | |||||
| (0.0034) | ||||||
| Graduate degree | -0.0108*** | |||||
| (0.0029) | ||||||
| Married | 0.0006 | |||||
| (0.0030) | ||||||
| Unemployed | -0.0155*** | |||||
| (0.0035) | ||||||
| Employed | -0.0090*** | |||||
| (0.0038) | ||||||
| Mostly urban | -0.0282*** | |||||
| (0.0032) | ||||||
| Suburban | -0.0174*** | |||||
| (0.0031) | ||||||
| Christian | 0.0089*** | |||||
| (0.0028) | ||||||
| Attends religious services | 0.0103*** | |||||
| (0.0027) | ||||||
| Uses social media | -0.0196*** | |||||
| (0.0025) | ||||||
| Hours spent on Social Media | 0.0076*** | |||||
| (0.0030) | ||||||
| Social Media Share: 80-100% | 0.0827*** | |||||
| (0.0033) | ||||||
| Social Media Share: 60-80% | 0.1135*** | |||||
| (0.0036) | ||||||
| Social Media Share: 40-60% | 0.0988*** | |||||
| (0.0037) | ||||||
| Social Media Share: 20-40% | 0.0524*** | |||||
| (0.0035) | ||||||
| Passed attention check at pre-treatment | -0.0217*** | |||||
| (0.0027) | ||||||
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0130) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0112) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0040 | 0.0040 |
| (0.0113) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| Emotions x Age | -0.0116*** | |||||
| (0.0044) | ||||||
| Emotions x Man | -0.0022 | |||||
| (0.0038) | ||||||
| Emotions x Some college | -0.0034 | |||||
| (0.0047) | ||||||
| Emotions x Bachelor’s degree | 0.0086** | |||||
| (0.0047) | ||||||
| Emotions x Graduate degree | 0.0109*** | |||||
| (0.0039) | ||||||
| Emotions x Married | 0.0051 | |||||
| (0.0043) | ||||||
| Emotions x Unemployed | -0.0071* | |||||
| (0.0049) | ||||||
| Emotions x Employed | -0.0189*** | |||||
| (0.0053) | ||||||
| Emotions x Mostly urban | 0.0286*** | |||||
| (0.0045) | ||||||
| Emotions x Suburban | 0.0149*** | |||||
| (0.0044) | ||||||
| Emotions x Christian | -0.0083** | |||||
| (0.0040) | ||||||
| Emotions x Attends religious services | -0.0214*** | |||||
| (0.0039) | ||||||
| Emotions x Uses social media | 0.0062* | |||||
| (0.0040) | ||||||
| Emotions x Hours spent on Social Media | -0.0128*** | |||||
| (0.0041) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0095** | |||||
| (0.0046) | ||||||
| Emotions x Social Media Share: 60-80% | -0.0077* | |||||
| (0.0051) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0166*** | |||||
| (0.0053) | ||||||
| Emotions x Social Media Share: 20-40% | -0.0092** | |||||
| (0.0049) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0039 | |||||
| (0.0038) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0159) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Post domain 2 | -0.1989*** | -0.1730*** | ||||
| (0.0089) | (0.0150) | |||||
| Post domain 3 | -0.2873*** | -0.3172*** | ||||
| (0.0087) | (0.0148) | |||||
| Post domain 4 | -0.1939*** | -0.2624*** | ||||
| (0.0089) | (0.0151) | |||||
| Post domain 5 | 0.0573*** | -0.0069 | ||||
| (0.0080) | (0.0137) | |||||
| Post domain 6 | 0.0263*** | 0.0467*** | ||||
| (0.0085) | (0.0134) | |||||
| Post domain 7 | -0.2382*** | -0.2006*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 8 | -0.2172*** | -0.1220*** | ||||
| (0.0088) | (0.0147) | |||||
| Post domain 9 | 0.0456*** | -0.0596*** | ||||
| (0.0082) | (0.0143) | |||||
| Post domain 10 | -0.0577*** | -0.1001*** | ||||
| (0.0085) | (0.0143) | |||||
| Post domain 11 | -0.1629*** | -0.1861*** | ||||
| (0.0090) | (0.0149) | |||||
| Post domain 12 | -0.1110*** | -0.1377*** | ||||
| (0.0085) | (0.0144) | |||||
| Post domain 13 | -0.1564*** | -0.2181*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 14 | -0.2815*** | -0.3292*** | ||||
| (0.0090) | (0.0152) | |||||
| Post domain 15 | -0.1923*** | -0.2956*** | ||||
| (0.0087) | (0.0151) | |||||
| 1[Combo manipulation tactic] | -0.1294*** | -0.1532*** | ||||
| (0.0044) | (0.0192) | |||||
| 1[Emotions manipulation tactic] | -0.0960*** | -0.2038*** | ||||
| (0.0042) | (0.0198) | |||||
| 1[Reasoning manipulation tactic] | -0.0890*** | -0.1088*** | ||||
| (0.0041) | (0.0194) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0666*** | |||||
| (0.0245) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0541** | |||||
| (0.0233) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0984*** | |||||
| (0.0236) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.1031*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3263*** | |||||
| (0.0226) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0368* | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0191 | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1669*** | |||||
| (0.0234) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0193 | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0561** | |||||
| (0.0242) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0390* | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0474** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0208 | |||||
| (0.0235) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.1033*** | |||||
| (0.0237) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0815*** | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1354*** | |||||
| (0.0236) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0892*** | |||||
| (0.0238) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1411*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1656*** | |||||
| (0.0234) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.0071 | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1331*** | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1794*** | |||||
| (0.0241) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1221*** | |||||
| (0.0235) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0225 | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1420*** | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.2141*** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1270*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3714*** | |||||
| (0.0242) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0258 | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0549*** | |||||
| (0.0232) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.1223*** | |||||
| (0.0243) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0403** | |||||
| (0.0236) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0723*** | |||||
| (0.0227) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1413*** | |||||
| (0.0235) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2816*** | |||||
| (0.0237) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1259*** | |||||
| (0.0238) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0433** | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0733*** | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0160 | |||||
| (0.0237) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0170 | |||||
| (0.0237) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0646*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0141 | |||||
| (0.0237) | ||||||
| Post 2 indicator | -0.1730*** | |||||
| (0.0150) | ||||||
| Post 3 indicator | -0.3172*** | |||||
| (0.0148) | ||||||
| Post 4 indicator | -0.2624*** | |||||
| (0.0151) | ||||||
| Post 5 indicator | -0.0069 | |||||
| (0.0137) | ||||||
| Post 6 indicator | 0.0467*** | |||||
| (0.0134) | ||||||
| Post 7 indicator | -0.2006*** | |||||
| (0.0149) | ||||||
| Post 8 indicator | -0.1220*** | |||||
| (0.0147) | ||||||
| Post 9 indicator | -0.0596*** | |||||
| (0.0143) | ||||||
| Post 10 indicator | -0.1001*** | |||||
| (0.0143) | ||||||
| Post 11 indicator | -0.1861*** | |||||
| (0.0149) | ||||||
| Post 12 indicator | -0.1377*** | |||||
| (0.0144) | ||||||
| Post 13 indicator | -0.2181*** | |||||
| (0.0149) | ||||||
| Post 14 indicator | -0.3292*** | |||||
| (0.0152) | ||||||
| Post 15 indicator | -0.2956*** | |||||
| (0.0151) | ||||||
| Post 16 indicator | -0.1071*** | |||||
| (0.0157) | ||||||
| Post 17 indicator | -0.3466*** | |||||
| (0.0165) | ||||||
| Post 18 indicator | -0.3702*** | |||||
| (0.0157) | ||||||
| Post 19 indicator | -0.2710*** | |||||
| (0.0160) | ||||||
| Post 20 indicator | -0.0109 | |||||
| (0.0153) | ||||||
| Post 21 indicator | -0.3867*** | |||||
| (0.0148) | ||||||
| Post 22 indicator | -0.3444*** | |||||
| (0.0156) | ||||||
| Post 23 indicator | -0.2481*** | |||||
| (0.0157) | ||||||
| Post 24 indicator | 0.0002 | |||||
| (0.0146) | ||||||
| Post 25 indicator | -0.1878*** | |||||
| (0.0163) | ||||||
| Post 26 indicator | -0.2371*** | |||||
| (0.0167) | ||||||
| Post 27 indicator | -0.2838*** | |||||
| (0.0159) | ||||||
| Post 28 indicator | -0.2778*** | |||||
| (0.0160) | ||||||
| Post 29 indicator | -0.4155*** | |||||
| (0.0157) | ||||||
| Post 30 indicator | -0.2995*** | |||||
| (0.0153) | ||||||
| Post 31 indicator | -0.1578*** | |||||
| (0.0164) | ||||||
| Post 32 indicator | -0.4123*** | |||||
| (0.0157) | ||||||
| Post 33 indicator | -0.3396*** | |||||
| (0.0156) | ||||||
| Post 34 indicator | -0.3310*** | |||||
| (0.0155) | ||||||
| Post 35 indicator | -0.0236* | |||||
| (0.0157) | ||||||
| Post 36 indicator | 0.0545*** | |||||
| (0.0149) | ||||||
| Post 37 indicator | -0.3512*** | |||||
| (0.0157) | ||||||
| Post 38 indicator | -0.4129*** | |||||
| (0.0153) | ||||||
| Post 39 indicator | -0.0380*** | |||||
| (0.0156) | ||||||
| Post 40 indicator | -0.1357*** | |||||
| (0.0156) | ||||||
| Post 41 indicator | -0.3663*** | |||||
| (0.0152) | ||||||
| Post 42 indicator | -0.1535*** | |||||
| (0.0160) | ||||||
| Post 43 indicator | -0.1619*** | |||||
| (0.0160) | ||||||
| Post 44 indicator | -0.3599*** | |||||
| (0.0157) | ||||||
| Post 45 indicator | -0.0820*** | |||||
| (0.0154) | ||||||
| Post 46 indicator | -0.0627*** | |||||
| (0.0160) | ||||||
| Post 47 indicator | -0.2098*** | |||||
| (0.0158) | ||||||
| Post 48 indicator | -0.4348*** | |||||
| (0.0149) | ||||||
| Post 49 indicator | -0.2027*** | |||||
| (0.0164) | ||||||
| Post 50 indicator | -0.0293** | |||||
| (0.0154) | ||||||
| Post 51 indicator | 0.0563*** | |||||
| (0.0148) | ||||||
| Post 52 indicator | -0.4046*** | |||||
| (0.0156) | ||||||
| Post 53 indicator | -0.4662*** | |||||
| (0.0158) | ||||||
| Post 54 indicator | 0.0036 | |||||
| (0.0152) | ||||||
| Post 55 indicator | -0.1195*** | |||||
| (0.0157) | ||||||
| Post 56 indicator | -0.1755*** | |||||
| (0.0163) | ||||||
| Post 57 indicator | -0.1844*** | |||||
| (0.0154) | ||||||
| Post 58 indicator | -0.2638*** | |||||
| (0.0160) | ||||||
| Post 59 indicator | -0.3272*** | |||||
| (0.0160) | ||||||
| Post 60 indicator | -0.3724*** | |||||
| (0.0151) | ||||||
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.0865 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.0859 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0142) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0116) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2712 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2711 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6897*** | |||||
| (0.0058) | ||||||
| Emotions | 0.0026 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0743*** | -0.0743*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1460*** | -0.0394*** | -0.0286*** | -0.0747*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0060) | (0.0066) | (0.0054) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.1802*** | |||||
| (0.0031) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.1019*** | |||||
| (0.0030) | ||||||
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0116) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0040 | 0.0040 |
| (0.0101) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0213*** | |||||
| (0.0042) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0144*** | |||||
| (0.0042) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0142) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Post domain 2 | -0.1989*** | -0.1730*** | ||||
| (0.0089) | (0.0150) | |||||
| Post domain 3 | -0.2873*** | -0.3172*** | ||||
| (0.0087) | (0.0148) | |||||
| Post domain 4 | -0.1939*** | -0.2624*** | ||||
| (0.0089) | (0.0151) | |||||
| Post domain 5 | 0.0573*** | -0.0069 | ||||
| (0.0080) | (0.0137) | |||||
| Post domain 6 | 0.0263*** | 0.0467*** | ||||
| (0.0085) | (0.0134) | |||||
| Post domain 7 | -0.2382*** | -0.2006*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 8 | -0.2172*** | -0.1220*** | ||||
| (0.0088) | (0.0147) | |||||
| Post domain 9 | 0.0456*** | -0.0596*** | ||||
| (0.0082) | (0.0143) | |||||
| Post domain 10 | -0.0577*** | -0.1001*** | ||||
| (0.0085) | (0.0143) | |||||
| Post domain 11 | -0.1629*** | -0.1861*** | ||||
| (0.0090) | (0.0149) | |||||
| Post domain 12 | -0.1110*** | -0.1377*** | ||||
| (0.0085) | (0.0144) | |||||
| Post domain 13 | -0.1564*** | -0.2181*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 14 | -0.2815*** | -0.3292*** | ||||
| (0.0090) | (0.0152) | |||||
| Post domain 15 | -0.1923*** | -0.2956*** | ||||
| (0.0087) | (0.0151) | |||||
| 1[Combo manipulation tactic] | -0.1294*** | -0.1532*** | ||||
| (0.0044) | (0.0192) | |||||
| 1[Emotions manipulation tactic] | -0.0960*** | -0.2038*** | ||||
| (0.0042) | (0.0198) | |||||
| 1[Reasoning manipulation tactic] | -0.0890*** | -0.1088*** | ||||
| (0.0041) | (0.0194) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0666*** | |||||
| (0.0245) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0541** | |||||
| (0.0233) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0984*** | |||||
| (0.0236) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.1031*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3263*** | |||||
| (0.0226) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0368* | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0191 | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1669*** | |||||
| (0.0234) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0193 | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0561** | |||||
| (0.0242) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0390* | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0474** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0208 | |||||
| (0.0235) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.1033*** | |||||
| (0.0237) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0815*** | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1354*** | |||||
| (0.0236) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0892*** | |||||
| (0.0238) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1411*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1656*** | |||||
| (0.0234) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.0071 | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1331*** | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1794*** | |||||
| (0.0241) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1221*** | |||||
| (0.0235) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0225 | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1420*** | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.2141*** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1270*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3714*** | |||||
| (0.0242) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0258 | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0549*** | |||||
| (0.0232) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.1223*** | |||||
| (0.0243) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0403** | |||||
| (0.0236) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0723*** | |||||
| (0.0227) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1413*** | |||||
| (0.0235) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2816*** | |||||
| (0.0237) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1259*** | |||||
| (0.0238) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0433** | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0733*** | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0160 | |||||
| (0.0237) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0170 | |||||
| (0.0237) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0646*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0141 | |||||
| (0.0237) | ||||||
| Post 2 indicator | -0.1730*** | |||||
| (0.0150) | ||||||
| Post 3 indicator | -0.3172*** | |||||
| (0.0148) | ||||||
| Post 4 indicator | -0.2624*** | |||||
| (0.0151) | ||||||
| Post 5 indicator | -0.0069 | |||||
| (0.0137) | ||||||
| Post 6 indicator | 0.0467*** | |||||
| (0.0134) | ||||||
| Post 7 indicator | -0.2006*** | |||||
| (0.0149) | ||||||
| Post 8 indicator | -0.1220*** | |||||
| (0.0147) | ||||||
| Post 9 indicator | -0.0596*** | |||||
| (0.0143) | ||||||
| Post 10 indicator | -0.1001*** | |||||
| (0.0143) | ||||||
| Post 11 indicator | -0.1861*** | |||||
| (0.0149) | ||||||
| Post 12 indicator | -0.1377*** | |||||
| (0.0144) | ||||||
| Post 13 indicator | -0.2181*** | |||||
| (0.0149) | ||||||
| Post 14 indicator | -0.3292*** | |||||
| (0.0152) | ||||||
| Post 15 indicator | -0.2956*** | |||||
| (0.0151) | ||||||
| Post 16 indicator | -0.1071*** | |||||
| (0.0157) | ||||||
| Post 17 indicator | -0.3466*** | |||||
| (0.0165) | ||||||
| Post 18 indicator | -0.3702*** | |||||
| (0.0157) | ||||||
| Post 19 indicator | -0.2710*** | |||||
| (0.0160) | ||||||
| Post 20 indicator | -0.0109 | |||||
| (0.0153) | ||||||
| Post 21 indicator | -0.3867*** | |||||
| (0.0148) | ||||||
| Post 22 indicator | -0.3444*** | |||||
| (0.0156) | ||||||
| Post 23 indicator | -0.2481*** | |||||
| (0.0157) | ||||||
| Post 24 indicator | 0.0002 | |||||
| (0.0146) | ||||||
| Post 25 indicator | -0.1878*** | |||||
| (0.0163) | ||||||
| Post 26 indicator | -0.2371*** | |||||
| (0.0167) | ||||||
| Post 27 indicator | -0.2838*** | |||||
| (0.0159) | ||||||
| Post 28 indicator | -0.2778*** | |||||
| (0.0160) | ||||||
| Post 29 indicator | -0.4155*** | |||||
| (0.0157) | ||||||
| Post 30 indicator | -0.2995*** | |||||
| (0.0153) | ||||||
| Post 31 indicator | -0.1578*** | |||||
| (0.0164) | ||||||
| Post 32 indicator | -0.4123*** | |||||
| (0.0157) | ||||||
| Post 33 indicator | -0.3396*** | |||||
| (0.0156) | ||||||
| Post 34 indicator | -0.3310*** | |||||
| (0.0155) | ||||||
| Post 35 indicator | -0.0236* | |||||
| (0.0157) | ||||||
| Post 36 indicator | 0.0545*** | |||||
| (0.0149) | ||||||
| Post 37 indicator | -0.3512*** | |||||
| (0.0157) | ||||||
| Post 38 indicator | -0.4129*** | |||||
| (0.0153) | ||||||
| Post 39 indicator | -0.0380*** | |||||
| (0.0156) | ||||||
| Post 40 indicator | -0.1357*** | |||||
| (0.0156) | ||||||
| Post 41 indicator | -0.3663*** | |||||
| (0.0152) | ||||||
| Post 42 indicator | -0.1535*** | |||||
| (0.0160) | ||||||
| Post 43 indicator | -0.1619*** | |||||
| (0.0160) | ||||||
| Post 44 indicator | -0.3599*** | |||||
| (0.0157) | ||||||
| Post 45 indicator | -0.0820*** | |||||
| (0.0154) | ||||||
| Post 46 indicator | -0.0627*** | |||||
| (0.0160) | ||||||
| Post 47 indicator | -0.2098*** | |||||
| (0.0158) | ||||||
| Post 48 indicator | -0.4348*** | |||||
| (0.0149) | ||||||
| Post 49 indicator | -0.2027*** | |||||
| (0.0164) | ||||||
| Post 50 indicator | -0.0293** | |||||
| (0.0154) | ||||||
| Post 51 indicator | 0.0563*** | |||||
| (0.0148) | ||||||
| Post 52 indicator | -0.4046*** | |||||
| (0.0156) | ||||||
| Post 53 indicator | -0.4662*** | |||||
| (0.0158) | ||||||
| Post 54 indicator | 0.0036 | |||||
| (0.0152) | ||||||
| Post 55 indicator | -0.1195*** | |||||
| (0.0157) | ||||||
| Post 56 indicator | -0.1755*** | |||||
| (0.0163) | ||||||
| Post 57 indicator | -0.1844*** | |||||
| (0.0154) | ||||||
| Post 58 indicator | -0.2638*** | |||||
| (0.0160) | ||||||
| Post 59 indicator | -0.3272*** | |||||
| (0.0160) | ||||||
| Post 60 indicator | -0.3724*** | |||||
| (0.0151) | ||||||
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2712 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2711 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0142) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0116) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2713 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2711 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6898*** | |||||
| (0.0058) | ||||||
| Emotions | 0.0023 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0743*** | -0.0743*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1460*** | -0.0394*** | -0.0286*** | -0.0747*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0060) | (0.0066) | (0.0054) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.0203 | |||||
| (1556.2355) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.1021*** | |||||
| (0.0030) | ||||||
| Emotions posts | 0.0682 | |||||
| (635.9552) | ||||||
| Reasoning posts | 0.0606 | |||||
| (629.9211) | ||||||
| Combo posts | 0.0672 | |||||
| (642.8073) | ||||||
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0116) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0040 | 0.0040 |
| (0.0101) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0371*** | |||||
| (0.0103) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0147*** | |||||
| (0.0042) | ||||||
| Emotions x Emotions posts | 0.0066 | |||||
| (0.0067) | ||||||
| Emotions x Reasoning posts | 0.0131** | |||||
| (0.0066) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0142) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Post domain 2 | -0.1989*** | -0.1730*** | ||||
| (0.0089) | (0.0150) | |||||
| Post domain 3 | -0.2873*** | -0.3172*** | ||||
| (0.0087) | (0.0148) | |||||
| Post domain 4 | -0.1939*** | -0.2624*** | ||||
| (0.0089) | (0.0151) | |||||
| Post domain 5 | 0.0573*** | -0.0069 | ||||
| (0.0080) | (0.0137) | |||||
| Post domain 6 | 0.0263*** | 0.0467*** | ||||
| (0.0085) | (0.0134) | |||||
| Post domain 7 | -0.2382*** | -0.2006*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 8 | -0.2172*** | -0.1220*** | ||||
| (0.0088) | (0.0147) | |||||
| Post domain 9 | 0.0456*** | -0.0596*** | ||||
| (0.0082) | (0.0143) | |||||
| Post domain 10 | -0.0577*** | -0.1001*** | ||||
| (0.0085) | (0.0143) | |||||
| Post domain 11 | -0.1629*** | -0.1861*** | ||||
| (0.0090) | (0.0149) | |||||
| Post domain 12 | -0.1110*** | -0.1377*** | ||||
| (0.0085) | (0.0144) | |||||
| Post domain 13 | -0.1564*** | -0.2181*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 14 | -0.2815*** | -0.3292*** | ||||
| (0.0090) | (0.0152) | |||||
| Post domain 15 | -0.1923*** | -0.2956*** | ||||
| (0.0087) | (0.0151) | |||||
| 1[Combo manipulation tactic] | -0.1294*** | -0.1532*** | ||||
| (0.0044) | (0.0192) | |||||
| 1[Emotions manipulation tactic] | -0.0960*** | -0.2038*** | ||||
| (0.0042) | (0.0198) | |||||
| 1[Reasoning manipulation tactic] | -0.0890*** | -0.1088*** | ||||
| (0.0041) | (0.0194) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0666*** | |||||
| (0.0245) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0541** | |||||
| (0.0233) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0984*** | |||||
| (0.0236) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.1031*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3263*** | |||||
| (0.0226) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0368* | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0191 | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1669*** | |||||
| (0.0234) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0193 | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0561** | |||||
| (0.0242) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0390* | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0474** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0208 | |||||
| (0.0235) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.1033*** | |||||
| (0.0237) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0815*** | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1354*** | |||||
| (0.0236) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0892*** | |||||
| (0.0238) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1411*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1656*** | |||||
| (0.0234) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.0071 | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1331*** | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1794*** | |||||
| (0.0241) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1221*** | |||||
| (0.0235) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0225 | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1420*** | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.2141*** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1270*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3714*** | |||||
| (0.0242) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0258 | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0549*** | |||||
| (0.0232) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.1223*** | |||||
| (0.0243) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0403** | |||||
| (0.0236) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0723*** | |||||
| (0.0227) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1413*** | |||||
| (0.0235) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2816*** | |||||
| (0.0237) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1259*** | |||||
| (0.0238) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0433** | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0733*** | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0160 | |||||
| (0.0237) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0170 | |||||
| (0.0237) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0646*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0141 | |||||
| (0.0237) | ||||||
| Post 2 indicator | -0.1730*** | |||||
| (0.0150) | ||||||
| Post 3 indicator | -0.3172*** | |||||
| (0.0148) | ||||||
| Post 4 indicator | -0.2624*** | |||||
| (0.0151) | ||||||
| Post 5 indicator | -0.0069 | |||||
| (0.0137) | ||||||
| Post 6 indicator | 0.0467*** | |||||
| (0.0134) | ||||||
| Post 7 indicator | -0.2006*** | |||||
| (0.0149) | ||||||
| Post 8 indicator | -0.1220*** | |||||
| (0.0147) | ||||||
| Post 9 indicator | -0.0596*** | |||||
| (0.0143) | ||||||
| Post 10 indicator | -0.1001*** | |||||
| (0.0143) | ||||||
| Post 11 indicator | -0.1861*** | |||||
| (0.0149) | ||||||
| Post 12 indicator | -0.1377*** | |||||
| (0.0144) | ||||||
| Post 13 indicator | -0.2181*** | |||||
| (0.0149) | ||||||
| Post 14 indicator | -0.3292*** | |||||
| (0.0152) | ||||||
| Post 15 indicator | -0.2956*** | |||||
| (0.0151) | ||||||
| Post 16 indicator | -0.1071*** | |||||
| (0.0157) | ||||||
| Post 17 indicator | -0.3466*** | |||||
| (0.0165) | ||||||
| Post 18 indicator | -0.3702*** | |||||
| (0.0157) | ||||||
| Post 19 indicator | -0.2710*** | |||||
| (0.0160) | ||||||
| Post 20 indicator | -0.0109 | |||||
| (0.0153) | ||||||
| Post 21 indicator | -0.3867*** | |||||
| (0.0148) | ||||||
| Post 22 indicator | -0.3444*** | |||||
| (0.0156) | ||||||
| Post 23 indicator | -0.2481*** | |||||
| (0.0157) | ||||||
| Post 24 indicator | 0.0002 | |||||
| (0.0146) | ||||||
| Post 25 indicator | -0.1878*** | |||||
| (0.0163) | ||||||
| Post 26 indicator | -0.2371*** | |||||
| (0.0167) | ||||||
| Post 27 indicator | -0.2838*** | |||||
| (0.0159) | ||||||
| Post 28 indicator | -0.2778*** | |||||
| (0.0160) | ||||||
| Post 29 indicator | -0.4155*** | |||||
| (0.0157) | ||||||
| Post 30 indicator | -0.2995*** | |||||
| (0.0153) | ||||||
| Post 31 indicator | -0.1578*** | |||||
| (0.0164) | ||||||
| Post 32 indicator | -0.4123*** | |||||
| (0.0157) | ||||||
| Post 33 indicator | -0.3396*** | |||||
| (0.0156) | ||||||
| Post 34 indicator | -0.3310*** | |||||
| (0.0155) | ||||||
| Post 35 indicator | -0.0236* | |||||
| (0.0157) | ||||||
| Post 36 indicator | 0.0545*** | |||||
| (0.0149) | ||||||
| Post 37 indicator | -0.3512*** | |||||
| (0.0157) | ||||||
| Post 38 indicator | -0.4129*** | |||||
| (0.0153) | ||||||
| Post 39 indicator | -0.0380*** | |||||
| (0.0156) | ||||||
| Post 40 indicator | -0.1357*** | |||||
| (0.0156) | ||||||
| Post 41 indicator | -0.3663*** | |||||
| (0.0152) | ||||||
| Post 42 indicator | -0.1535*** | |||||
| (0.0160) | ||||||
| Post 43 indicator | -0.1619*** | |||||
| (0.0160) | ||||||
| Post 44 indicator | -0.3599*** | |||||
| (0.0157) | ||||||
| Post 45 indicator | -0.0820*** | |||||
| (0.0154) | ||||||
| Post 46 indicator | -0.0627*** | |||||
| (0.0160) | ||||||
| Post 47 indicator | -0.2098*** | |||||
| (0.0158) | ||||||
| Post 48 indicator | -0.4348*** | |||||
| (0.0149) | ||||||
| Post 49 indicator | -0.2027*** | |||||
| (0.0164) | ||||||
| Post 50 indicator | -0.0293** | |||||
| (0.0154) | ||||||
| Post 51 indicator | 0.0563*** | |||||
| (0.0148) | ||||||
| Post 52 indicator | -0.4046*** | |||||
| (0.0156) | ||||||
| Post 53 indicator | -0.4662*** | |||||
| (0.0158) | ||||||
| Post 54 indicator | 0.0036 | |||||
| (0.0152) | ||||||
| Post 55 indicator | -0.1195*** | |||||
| (0.0157) | ||||||
| Post 56 indicator | -0.1755*** | |||||
| (0.0163) | ||||||
| Post 57 indicator | -0.1844*** | |||||
| (0.0154) | ||||||
| Post 58 indicator | -0.2638*** | |||||
| (0.0160) | ||||||
| Post 59 indicator | -0.3272*** | |||||
| (0.0160) | ||||||
| Post 60 indicator | -0.3724*** | |||||
| (0.0151) | ||||||
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2713 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2711 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2744 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2738 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6907*** | |||||
| (0.0058) | ||||||
| Emotions | 0.0016 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0743*** | -0.0743*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1460*** | -0.0394*** | -0.0286*** | -0.0747*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0060) | (0.0066) | (0.0054) | |
| Age | -0.0058** | |||||
| (0.0026) | ||||||
| Man | 0.0028 | |||||
| (0.0024) | ||||||
| Some college | -0.0018 | |||||
| (0.0030) | ||||||
| Bachelor’s degree | 0.0034 | |||||
| (0.0030) | ||||||
| Graduate degree | -0.0078*** | |||||
| (0.0026) | ||||||
| Married | -0.0028 | |||||
| (0.0027) | ||||||
| Unemployed | -0.0045* | |||||
| (0.0031) | ||||||
| Employed | -0.0084*** | |||||
| (0.0034) | ||||||
| Mostly urban | -0.0056** | |||||
| (0.0029) | ||||||
| Suburban | -0.0052** | |||||
| (0.0028) | ||||||
| Christian | 0.0031 | |||||
| (0.0025) | ||||||
| Attends religious services | 0.0007 | |||||
| (0.0024) | ||||||
| Uses social media | -0.0060*** | |||||
| (0.0023) | ||||||
| Hours spent on Social Media | 0.0052** | |||||
| (0.0026) | ||||||
| Social Media Share: 80-100% | 0.0130*** | |||||
| (0.0030) | ||||||
| Social Media Share: 60-80% | 0.0179*** | |||||
| (0.0034) | ||||||
| Social Media Share: 40-60% | 0.0149*** | |||||
| (0.0034) | ||||||
| Social Media Share: 20-40% | 0.0065** | |||||
| (0.0031) | ||||||
| Passed attention check at pre-treatment | -0.0046** | |||||
| (0.0024) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.1749*** | |||||
| (0.0032) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.0999*** | |||||
| (0.0030) | ||||||
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0040 | 0.0040 |
| (0.0101) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| Emotions x Age | -0.0040 | |||||
| (0.0039) | ||||||
| Emotions x Man | 0.0061** | |||||
| (0.0034) | ||||||
| Emotions x Some college | 0.0003 | |||||
| (0.0042) | ||||||
| Emotions x Bachelor’s degree | -0.0029 | |||||
| (0.0042) | ||||||
| Emotions x Graduate degree | 0.0094*** | |||||
| (0.0035) | ||||||
| Emotions x Married | -0.0050* | |||||
| (0.0038) | ||||||
| Emotions x Unemployed | -0.0017 | |||||
| (0.0044) | ||||||
| Emotions x Employed | 0.0020 | |||||
| (0.0048) | ||||||
| Emotions x Mostly urban | 0.0080** | |||||
| (0.0040) | ||||||
| Emotions x Suburban | -0.0013 | |||||
| (0.0039) | ||||||
| Emotions x Christian | -0.0036 | |||||
| (0.0035) | ||||||
| Emotions x Attends religious services | -0.0063** | |||||
| (0.0035) | ||||||
| Emotions x Uses social media | 0.0086*** | |||||
| (0.0036) | ||||||
| Emotions x Hours spent on Social Media | -0.0020 | |||||
| (0.0036) | ||||||
| Emotions x Social Media Share: 80-100% | 0.0011 | |||||
| (0.0043) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0051 | |||||
| (0.0048) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0016 | |||||
| (0.0048) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0044 | |||||
| (0.0044) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0009 | |||||
| (0.0034) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0230*** | |||||
| (0.0043) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0125*** | |||||
| (0.0042) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Post domain 2 | -0.1989*** | -0.1730*** | ||||
| (0.0089) | (0.0150) | |||||
| Post domain 3 | -0.2873*** | -0.3172*** | ||||
| (0.0087) | (0.0148) | |||||
| Post domain 4 | -0.1939*** | -0.2624*** | ||||
| (0.0089) | (0.0151) | |||||
| Post domain 5 | 0.0573*** | -0.0069 | ||||
| (0.0080) | (0.0137) | |||||
| Post domain 6 | 0.0263*** | 0.0467*** | ||||
| (0.0085) | (0.0134) | |||||
| Post domain 7 | -0.2382*** | -0.2006*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 8 | -0.2172*** | -0.1220*** | ||||
| (0.0088) | (0.0147) | |||||
| Post domain 9 | 0.0456*** | -0.0596*** | ||||
| (0.0082) | (0.0143) | |||||
| Post domain 10 | -0.0577*** | -0.1001*** | ||||
| (0.0085) | (0.0143) | |||||
| Post domain 11 | -0.1629*** | -0.1861*** | ||||
| (0.0090) | (0.0149) | |||||
| Post domain 12 | -0.1110*** | -0.1377*** | ||||
| (0.0085) | (0.0144) | |||||
| Post domain 13 | -0.1564*** | -0.2181*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 14 | -0.2815*** | -0.3292*** | ||||
| (0.0090) | (0.0152) | |||||
| Post domain 15 | -0.1923*** | -0.2956*** | ||||
| (0.0087) | (0.0151) | |||||
| 1[Combo manipulation tactic] | -0.1294*** | -0.1532*** | ||||
| (0.0044) | (0.0192) | |||||
| 1[Emotions manipulation tactic] | -0.0960*** | -0.2038*** | ||||
| (0.0042) | (0.0198) | |||||
| 1[Reasoning manipulation tactic] | -0.0890*** | -0.1088*** | ||||
| (0.0041) | (0.0194) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0666*** | |||||
| (0.0245) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0541** | |||||
| (0.0233) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0984*** | |||||
| (0.0236) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.1031*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3263*** | |||||
| (0.0226) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0368* | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0191 | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1669*** | |||||
| (0.0234) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0193 | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0561** | |||||
| (0.0242) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0390* | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0474** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0208 | |||||
| (0.0235) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.1033*** | |||||
| (0.0237) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0815*** | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1354*** | |||||
| (0.0236) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0892*** | |||||
| (0.0238) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1411*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1656*** | |||||
| (0.0234) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.0071 | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1331*** | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1794*** | |||||
| (0.0241) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1221*** | |||||
| (0.0235) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0225 | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1420*** | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.2141*** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1270*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3714*** | |||||
| (0.0242) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0258 | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0549*** | |||||
| (0.0232) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.1223*** | |||||
| (0.0243) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0403** | |||||
| (0.0236) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0723*** | |||||
| (0.0227) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1413*** | |||||
| (0.0235) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2816*** | |||||
| (0.0237) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1259*** | |||||
| (0.0238) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0433** | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0733*** | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0160 | |||||
| (0.0237) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0170 | |||||
| (0.0237) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0646*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0141 | |||||
| (0.0237) | ||||||
| Post 2 indicator | -0.1730*** | |||||
| (0.0150) | ||||||
| Post 3 indicator | -0.3172*** | |||||
| (0.0148) | ||||||
| Post 4 indicator | -0.2624*** | |||||
| (0.0151) | ||||||
| Post 5 indicator | -0.0069 | |||||
| (0.0137) | ||||||
| Post 6 indicator | 0.0467*** | |||||
| (0.0134) | ||||||
| Post 7 indicator | -0.2006*** | |||||
| (0.0149) | ||||||
| Post 8 indicator | -0.1220*** | |||||
| (0.0147) | ||||||
| Post 9 indicator | -0.0596*** | |||||
| (0.0143) | ||||||
| Post 10 indicator | -0.1001*** | |||||
| (0.0143) | ||||||
| Post 11 indicator | -0.1861*** | |||||
| (0.0149) | ||||||
| Post 12 indicator | -0.1377*** | |||||
| (0.0144) | ||||||
| Post 13 indicator | -0.2181*** | |||||
| (0.0149) | ||||||
| Post 14 indicator | -0.3292*** | |||||
| (0.0152) | ||||||
| Post 15 indicator | -0.2956*** | |||||
| (0.0151) | ||||||
| Post 16 indicator | -0.1071*** | |||||
| (0.0157) | ||||||
| Post 17 indicator | -0.3466*** | |||||
| (0.0165) | ||||||
| Post 18 indicator | -0.3702*** | |||||
| (0.0157) | ||||||
| Post 19 indicator | -0.2710*** | |||||
| (0.0160) | ||||||
| Post 20 indicator | -0.0109 | |||||
| (0.0153) | ||||||
| Post 21 indicator | -0.3867*** | |||||
| (0.0148) | ||||||
| Post 22 indicator | -0.3444*** | |||||
| (0.0156) | ||||||
| Post 23 indicator | -0.2481*** | |||||
| (0.0157) | ||||||
| Post 24 indicator | 0.0002 | |||||
| (0.0146) | ||||||
| Post 25 indicator | -0.1878*** | |||||
| (0.0163) | ||||||
| Post 26 indicator | -0.2371*** | |||||
| (0.0167) | ||||||
| Post 27 indicator | -0.2838*** | |||||
| (0.0159) | ||||||
| Post 28 indicator | -0.2778*** | |||||
| (0.0160) | ||||||
| Post 29 indicator | -0.4155*** | |||||
| (0.0157) | ||||||
| Post 30 indicator | -0.2995*** | |||||
| (0.0153) | ||||||
| Post 31 indicator | -0.1578*** | |||||
| (0.0164) | ||||||
| Post 32 indicator | -0.4123*** | |||||
| (0.0157) | ||||||
| Post 33 indicator | -0.3396*** | |||||
| (0.0156) | ||||||
| Post 34 indicator | -0.3310*** | |||||
| (0.0155) | ||||||
| Post 35 indicator | -0.0236* | |||||
| (0.0157) | ||||||
| Post 36 indicator | 0.0545*** | |||||
| (0.0149) | ||||||
| Post 37 indicator | -0.3512*** | |||||
| (0.0157) | ||||||
| Post 38 indicator | -0.4129*** | |||||
| (0.0153) | ||||||
| Post 39 indicator | -0.0380*** | |||||
| (0.0156) | ||||||
| Post 40 indicator | -0.1357*** | |||||
| (0.0156) | ||||||
| Post 41 indicator | -0.3663*** | |||||
| (0.0152) | ||||||
| Post 42 indicator | -0.1535*** | |||||
| (0.0160) | ||||||
| Post 43 indicator | -0.1619*** | |||||
| (0.0160) | ||||||
| Post 44 indicator | -0.3599*** | |||||
| (0.0157) | ||||||
| Post 45 indicator | -0.0820*** | |||||
| (0.0154) | ||||||
| Post 46 indicator | -0.0627*** | |||||
| (0.0160) | ||||||
| Post 47 indicator | -0.2098*** | |||||
| (0.0158) | ||||||
| Post 48 indicator | -0.4348*** | |||||
| (0.0149) | ||||||
| Post 49 indicator | -0.2027*** | |||||
| (0.0164) | ||||||
| Post 50 indicator | -0.0293** | |||||
| (0.0154) | ||||||
| Post 51 indicator | 0.0563*** | |||||
| (0.0148) | ||||||
| Post 52 indicator | -0.4046*** | |||||
| (0.0156) | ||||||
| Post 53 indicator | -0.4662*** | |||||
| (0.0158) | ||||||
| Post 54 indicator | 0.0036 | |||||
| (0.0152) | ||||||
| Post 55 indicator | -0.1195*** | |||||
| (0.0157) | ||||||
| Post 56 indicator | -0.1755*** | |||||
| (0.0163) | ||||||
| Post 57 indicator | -0.1844*** | |||||
| (0.0154) | ||||||
| Post 58 indicator | -0.2638*** | |||||
| (0.0160) | ||||||
| Post 59 indicator | -0.3272*** | |||||
| (0.0160) | ||||||
| Post 60 indicator | -0.3724*** | |||||
| (0.0151) | ||||||
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2744 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2738 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates + Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2744 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2738 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6908*** | |||||
| (0.0058) | ||||||
| Emotions | 0.0014 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0743*** | -0.0743*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1460*** | -0.0394*** | -0.0286*** | -0.0747*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0060) | (0.0066) | (0.0054) | |
| Age | -0.0058** | |||||
| (0.0026) | ||||||
| Man | 0.0028 | |||||
| (0.0024) | ||||||
| Some college | -0.0018 | |||||
| (0.0030) | ||||||
| Bachelor’s degree | 0.0034 | |||||
| (0.0030) | ||||||
| Graduate degree | -0.0079*** | |||||
| (0.0026) | ||||||
| Married | -0.0026 | |||||
| (0.0027) | ||||||
| Unemployed | -0.0045* | |||||
| (0.0031) | ||||||
| Employed | -0.0085*** | |||||
| (0.0034) | ||||||
| Mostly urban | -0.0054** | |||||
| (0.0029) | ||||||
| Suburban | -0.0051** | |||||
| (0.0028) | ||||||
| Christian | 0.0031 | |||||
| (0.0025) | ||||||
| Attends religious services | 0.0007 | |||||
| (0.0024) | ||||||
| Uses social media | -0.0060*** | |||||
| (0.0023) | ||||||
| Hours spent on Social Media | 0.0053** | |||||
| (0.0026) | ||||||
| Social Media Share: 80-100% | 0.0130*** | |||||
| (0.0030) | ||||||
| Social Media Share: 60-80% | 0.0180*** | |||||
| (0.0034) | ||||||
| Social Media Share: 40-60% | 0.0150*** | |||||
| (0.0034) | ||||||
| Social Media Share: 20-40% | 0.0066** | |||||
| (0.0031) | ||||||
| Passed attention check at pre-treatment | -0.0046** | |||||
| (0.0024) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.0176 | |||||
| (1553.4631) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.1001*** | |||||
| (0.0030) | ||||||
| Emotions posts | 0.0673 | |||||
| (634.8222) | ||||||
| Reasoning posts | 0.0599 | |||||
| (628.7989) | ||||||
| Combo posts | 0.0656 | |||||
| (641.6622) | ||||||
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0948*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0015 | -0.0039 | -0.0024 | -0.0024 |
| (0.0100) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0040 | 0.0040 |
| (0.0101) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| Emotions x Age | -0.0039 | |||||
| (0.0039) | ||||||
| Emotions x Man | 0.0059** | |||||
| (0.0034) | ||||||
| Emotions x Some college | 0.0002 | |||||
| (0.0042) | ||||||
| Emotions x Bachelor’s degree | -0.0030 | |||||
| (0.0042) | ||||||
| Emotions x Graduate degree | 0.0094*** | |||||
| (0.0035) | ||||||
| Emotions x Married | -0.0052* | |||||
| (0.0038) | ||||||
| Emotions x Unemployed | -0.0019 | |||||
| (0.0044) | ||||||
| Emotions x Employed | 0.0020 | |||||
| (0.0048) | ||||||
| Emotions x Mostly urban | 0.0077** | |||||
| (0.0040) | ||||||
| Emotions x Suburban | -0.0014 | |||||
| (0.0039) | ||||||
| Emotions x Christian | -0.0036 | |||||
| (0.0035) | ||||||
| Emotions x Attends religious services | -0.0064** | |||||
| (0.0035) | ||||||
| Emotions x Uses social media | 0.0086*** | |||||
| (0.0036) | ||||||
| Emotions x Hours spent on Social Media | -0.0022 | |||||
| (0.0036) | ||||||
| Emotions x Social Media Share: 80-100% | 0.0010 | |||||
| (0.0043) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0049 | |||||
| (0.0048) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0020 | |||||
| (0.0048) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0041 | |||||
| (0.0044) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0011 | |||||
| (0.0034) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0363*** | |||||
| (0.0103) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0128*** | |||||
| (0.0042) | ||||||
| Emotions x Emotions posts | 0.0052 | |||||
| (0.0067) | ||||||
| Emotions x Reasoning posts | 0.0114** | |||||
| (0.0066) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0221** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Post domain 2 | -0.1989*** | -0.1730*** | ||||
| (0.0089) | (0.0150) | |||||
| Post domain 3 | -0.2873*** | -0.3172*** | ||||
| (0.0087) | (0.0148) | |||||
| Post domain 4 | -0.1939*** | -0.2624*** | ||||
| (0.0089) | (0.0151) | |||||
| Post domain 5 | 0.0573*** | -0.0069 | ||||
| (0.0080) | (0.0137) | |||||
| Post domain 6 | 0.0263*** | 0.0467*** | ||||
| (0.0085) | (0.0134) | |||||
| Post domain 7 | -0.2382*** | -0.2006*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 8 | -0.2172*** | -0.1220*** | ||||
| (0.0088) | (0.0147) | |||||
| Post domain 9 | 0.0456*** | -0.0596*** | ||||
| (0.0082) | (0.0143) | |||||
| Post domain 10 | -0.0577*** | -0.1001*** | ||||
| (0.0085) | (0.0143) | |||||
| Post domain 11 | -0.1629*** | -0.1861*** | ||||
| (0.0090) | (0.0149) | |||||
| Post domain 12 | -0.1110*** | -0.1377*** | ||||
| (0.0085) | (0.0144) | |||||
| Post domain 13 | -0.1564*** | -0.2181*** | ||||
| (0.0088) | (0.0149) | |||||
| Post domain 14 | -0.2815*** | -0.3292*** | ||||
| (0.0090) | (0.0152) | |||||
| Post domain 15 | -0.1923*** | -0.2956*** | ||||
| (0.0087) | (0.0151) | |||||
| 1[Combo manipulation tactic] | -0.1294*** | -0.1532*** | ||||
| (0.0044) | (0.0192) | |||||
| 1[Emotions manipulation tactic] | -0.0960*** | -0.2038*** | ||||
| (0.0042) | (0.0198) | |||||
| 1[Reasoning manipulation tactic] | -0.0890*** | -0.1088*** | ||||
| (0.0041) | (0.0194) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0666*** | |||||
| (0.0245) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0541** | |||||
| (0.0233) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0984*** | |||||
| (0.0236) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.1031*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3263*** | |||||
| (0.0226) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0368* | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0191 | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1669*** | |||||
| (0.0234) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0193 | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0561** | |||||
| (0.0242) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0390* | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0474** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0208 | |||||
| (0.0235) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.1033*** | |||||
| (0.0237) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0815*** | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1354*** | |||||
| (0.0236) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0892*** | |||||
| (0.0238) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1411*** | |||||
| (0.0234) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1656*** | |||||
| (0.0234) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.0071 | |||||
| (0.0238) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1331*** | |||||
| (0.0241) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1794*** | |||||
| (0.0241) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1221*** | |||||
| (0.0235) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0225 | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1420*** | |||||
| (0.0240) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.2141*** | |||||
| (0.0239) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1270*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3714*** | |||||
| (0.0242) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0258 | |||||
| (0.0243) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0549*** | |||||
| (0.0232) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.1223*** | |||||
| (0.0243) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0403** | |||||
| (0.0236) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0723*** | |||||
| (0.0227) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1413*** | |||||
| (0.0235) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2816*** | |||||
| (0.0237) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1259*** | |||||
| (0.0238) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0433** | |||||
| (0.0239) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0733*** | |||||
| (0.0240) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0160 | |||||
| (0.0237) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0170 | |||||
| (0.0237) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0646*** | |||||
| (0.0239) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0141 | |||||
| (0.0237) | ||||||
| Post 2 indicator | -0.1730*** | |||||
| (0.0150) | ||||||
| Post 3 indicator | -0.3172*** | |||||
| (0.0148) | ||||||
| Post 4 indicator | -0.2624*** | |||||
| (0.0151) | ||||||
| Post 5 indicator | -0.0069 | |||||
| (0.0137) | ||||||
| Post 6 indicator | 0.0467*** | |||||
| (0.0134) | ||||||
| Post 7 indicator | -0.2006*** | |||||
| (0.0149) | ||||||
| Post 8 indicator | -0.1220*** | |||||
| (0.0147) | ||||||
| Post 9 indicator | -0.0596*** | |||||
| (0.0143) | ||||||
| Post 10 indicator | -0.1001*** | |||||
| (0.0143) | ||||||
| Post 11 indicator | -0.1861*** | |||||
| (0.0149) | ||||||
| Post 12 indicator | -0.1377*** | |||||
| (0.0144) | ||||||
| Post 13 indicator | -0.2181*** | |||||
| (0.0149) | ||||||
| Post 14 indicator | -0.3292*** | |||||
| (0.0152) | ||||||
| Post 15 indicator | -0.2956*** | |||||
| (0.0151) | ||||||
| Post 16 indicator | -0.1071*** | |||||
| (0.0157) | ||||||
| Post 17 indicator | -0.3466*** | |||||
| (0.0165) | ||||||
| Post 18 indicator | -0.3702*** | |||||
| (0.0157) | ||||||
| Post 19 indicator | -0.2710*** | |||||
| (0.0160) | ||||||
| Post 20 indicator | -0.0109 | |||||
| (0.0153) | ||||||
| Post 21 indicator | -0.3867*** | |||||
| (0.0148) | ||||||
| Post 22 indicator | -0.3444*** | |||||
| (0.0156) | ||||||
| Post 23 indicator | -0.2481*** | |||||
| (0.0157) | ||||||
| Post 24 indicator | 0.0002 | |||||
| (0.0146) | ||||||
| Post 25 indicator | -0.1878*** | |||||
| (0.0163) | ||||||
| Post 26 indicator | -0.2371*** | |||||
| (0.0167) | ||||||
| Post 27 indicator | -0.2838*** | |||||
| (0.0159) | ||||||
| Post 28 indicator | -0.2778*** | |||||
| (0.0160) | ||||||
| Post 29 indicator | -0.4155*** | |||||
| (0.0157) | ||||||
| Post 30 indicator | -0.2995*** | |||||
| (0.0153) | ||||||
| Post 31 indicator | -0.1578*** | |||||
| (0.0164) | ||||||
| Post 32 indicator | -0.4123*** | |||||
| (0.0157) | ||||||
| Post 33 indicator | -0.3396*** | |||||
| (0.0156) | ||||||
| Post 34 indicator | -0.3310*** | |||||
| (0.0155) | ||||||
| Post 35 indicator | -0.0236* | |||||
| (0.0157) | ||||||
| Post 36 indicator | 0.0545*** | |||||
| (0.0149) | ||||||
| Post 37 indicator | -0.3512*** | |||||
| (0.0157) | ||||||
| Post 38 indicator | -0.4129*** | |||||
| (0.0153) | ||||||
| Post 39 indicator | -0.0380*** | |||||
| (0.0156) | ||||||
| Post 40 indicator | -0.1357*** | |||||
| (0.0156) | ||||||
| Post 41 indicator | -0.3663*** | |||||
| (0.0152) | ||||||
| Post 42 indicator | -0.1535*** | |||||
| (0.0160) | ||||||
| Post 43 indicator | -0.1619*** | |||||
| (0.0160) | ||||||
| Post 44 indicator | -0.3599*** | |||||
| (0.0157) | ||||||
| Post 45 indicator | -0.0820*** | |||||
| (0.0154) | ||||||
| Post 46 indicator | -0.0627*** | |||||
| (0.0160) | ||||||
| Post 47 indicator | -0.2098*** | |||||
| (0.0158) | ||||||
| Post 48 indicator | -0.4348*** | |||||
| (0.0149) | ||||||
| Post 49 indicator | -0.2027*** | |||||
| (0.0164) | ||||||
| Post 50 indicator | -0.0293** | |||||
| (0.0154) | ||||||
| Post 51 indicator | 0.0563*** | |||||
| (0.0148) | ||||||
| Post 52 indicator | -0.4046*** | |||||
| (0.0156) | ||||||
| Post 53 indicator | -0.4662*** | |||||
| (0.0158) | ||||||
| Post 54 indicator | 0.0036 | |||||
| (0.0152) | ||||||
| Post 55 indicator | -0.1195*** | |||||
| (0.0157) | ||||||
| Post 56 indicator | -0.1755*** | |||||
| (0.0163) | ||||||
| Post 57 indicator | -0.1844*** | |||||
| (0.0154) | ||||||
| Post 58 indicator | -0.2638*** | |||||
| (0.0160) | ||||||
| Post 59 indicator | -0.3272*** | |||||
| (0.0160) | ||||||
| Post 60 indicator | -0.3724*** | |||||
| (0.0151) | ||||||
| Num. obs. | 65268 | 65268 | 65268 | 65268 | 65268 | 65268 |
| R2 (full model) | 0.2744 | 0.3423 | 0.3939 | 0.3431 | 0.4154 | 0.4154 |
| R2 (proj model) | 0.0569 | 0.1309 | 0.0581 | 0.1618 | 0.1618 | |
| Adj. R2 (full model) | 0.2738 | 0.3035 | 0.3580 | 0.3044 | 0.3804 | 0.3804 |
| Adj. R2 (proj model) | 0.0568 | 0.1306 | 0.0579 | 0.1609 | 0.1609 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Prediction
Standard errors of the share predicted value are estimated using the Delta method.
Treatment effects in the last column refer to estimated marginal means (EMMs, or least-squares means) for specified factor combinations (Post-treatment and Misinfo). Point estimates are reported in the upper cell and standard errors, clustered at the user level, reported in parentheses in the bottom cell. t-ratios are reported in brackets.
Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
cov_feglm <- c(paste0("share ~ treatment"),
paste0("share ~ treatment + ",demographics_demeaned_treatment),
paste0("share ~ treatment + factor(fact) + ",sharing_demeaned_treatment),
paste0("share ~ treatment + factor(type) + ",sharing_tactic_demeaned_treatment),
paste0("share ~ treatment + factor(fact)*factor(type) + ",demographics_sharing_demeaned_treatment),
paste0("share ~ treatment + factor(post_id) + ",all_covariates_demeaned_treatment))
if (file.exists("prediction_2treat_df.rds")) {
prediction_2treat_df <- readRDS("prediction_2treat_df.rds")
counter <- 1
for (cov in seq_along(covariate_list)) {
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
k <- knitr::kable(prediction_2treat_df[[counter]], caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","Treatment effect<br>Emotions--Facts"))
print(k)
cat('<br>')
counter <- counter + 1
}
cat('<br>\n<br>\n<br>')
}
} else {
prediction_2treat_df <- list()
for (cov in seq_along(covariate_list)) {
list_models <- cov_estimates_list[[cov]]
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
treatmentLevels <- unique(panel_2$treatment)
posts <- c(0, 1)
misinfos <- c(0, 1)
prediction_df <- expand.grid(treatment = treatmentLevels, post = posts, post_is_misinfo = misinfos) %>%
mutate(prediction = NA_real_, prediction_se = NA_real_)
for (row in 1:nrow(prediction_df)) {
treat <- prediction_df$treatment[row]
post <- prediction_df$post[row]
misinfo <- prediction_df$post_is_misinfo[row]
# Filter the relevant data
relevant_data <- panel_2 %>%
filter(row_number() %in% obs(list_models[[m]])) %>%
filter(treatment == treat & post == post & post_is_misinfo == misinfo)
# Prepare predictor variables based on model requirements
predictorVariables <- relevant_data %>%
distinct(user, .keep_all = TRUE) %>%
group_by(user, type, fact, post_id) %>%
summarise(age_cov = mean(age_cov, na.rm = TRUE),
gender_Man_cov = mean(gender_Man_cov, na.rm = TRUE),
education_Some_college_cov = mean(education_Some_college_cov, na.rm = TRUE),
education_Bachelor_degree_cov = mean(education_Bachelor_degree_cov, na.rm = TRUE),
education_Graduate_degree_cov = mean(education_Graduate_degree_cov, na.rm = TRUE),
marital_Married_or_in_a_domestic_partnership_cov = mean(marital_Married_or_in_a_domestic_partnership_cov, na.rm = TRUE),
employment_Unemployed_cov = mean(employment_Unemployed_cov, na.rm = TRUE),
employment_Employed_cov = mean(employment_Employed_cov, na.rm = TRUE),
location_Mostly_urban_cov = mean(location_Mostly_urban_cov, na.rm = TRUE),
location_Suburban_cov = mean(location_Suburban_cov, na.rm = TRUE),
religion_Christian_cov = mean(religion_Christian_cov, na.rm = TRUE),
religiosity_Attends_cov = mean(religiosity_Attends_cov, na.rm = TRUE),
social_media_bin_Yes_cov = mean(social_media_bin_Yes_cov, na.rm = TRUE),
social_media_hours_cov = mean(social_media_hours_cov, na.rm = TRUE),
att_check_pre_cov = mean(att_check_pre_cov, na.rm = TRUE),
social_media_share_80_100_cov = mean(social_media_share_80_100_cov, na.rm = TRUE),
social_media_share_60_80_cov = mean(social_media_share_60_80_cov, na.rm = TRUE),
social_media_share_40_60_cov = mean(social_media_share_40_60_cov, na.rm = TRUE),
social_media_share_20_40_cov = mean(social_media_share_20_40_cov, na.rm = TRUE),
misinfo_pre_cov = mean(misinfo_pre_cov, na.rm = TRUE),
base_rate_pre_cov = mean(base_rate_pre_cov, na.rm = TRUE),
emotions_pre_cov = mean(emotions_pre_cov, na.rm = TRUE),
reasoning_pre_cov = mean(reasoning_pre_cov, na.rm = TRUE),
combo_pre_cov = mean(combo_pre_cov, na.rm = TRUE)) %>% # now add the treatment and post variables and misinfo
mutate(treatment = treat,
post = post,
post_is_misinfo = misinfo)
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[row] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[row] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
# predicted value
if (m == 1) {
predictorVariables <- expand.grid(treatment = unique(panel_2$treatment),
post = c(0,1),
post_is_misinfo = c(0,1))
predictions <- predict(baseline_lm, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_lm, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% select(post, post_is_misinfo, estimate, SE, t.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(t.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
# prediction_df <- prediction_df %>%
# group_by(post,post_is_misinfo) %>%
# mutate(lag = ifelse(treatment == "Facts Baseline",
# prediction, NA)) %>%
# tidyr::fill(lag) %>%
# ungroup() %>%
# mutate(te = ifelse(treatment != "Facts Baseline",
# format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
# select(-lag) %>%
# mutate(te = ifelse(is.na(te),"",te))
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(lag) - as.numeric(prediction), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
cov_model <- feglm(as.formula(cov_feglm[m]),panel_2 %>%
filter(post == p & post_is_misinfo == mis), cluster = "user")
# obtain data frame with estimate, std error, and z value
cov_model_add <- data.frame(
treatment = unique(panel_2$treatment),
estimate = format(coef(cov_model)[grepl("^treatmentEmotions$",names(coef(cov_model)))], digits = 4),
SE = sqrt(diag(vcov(cov_model)))[grepl("^treatmentEmotions$",names(diag(vcov(cov_model))))],
z.ratio = coef(cov_model)[grepl("^treatmentEmotions$",names(coef(cov_model)))] / sqrt(diag(vcov(cov_model)))[grepl("^treatmentEmotions$",names(diag(vcov(cov_model))))],
post = p,
post_is_misinfo = mis
) %>% distinct(estimate, SE, z.ratio, post, post_is_misinfo, .keep_all = TRUE)
# make rownames a column
cov_model_add <- rownames_to_column(cov_model_add, var = "treatment2") %>%
filter((treatment == "Facts Baseline" & treatment2 == "treatmentEmotions")|(treatment == "Facts Baseline" & treatment2 == 1)) %>%
select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
cov_model_add <- rbind(cov_model_add %>%
mutate(treatment = "Facts Baseline"),cov_model_add %>%
mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate)
contrast_df <- rbind(contrast_df, cov_model_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","Treatment Effect<br>Emotions vs. Facts"))
print(k)
cat('<br>')
prediction_2treat_df[[length(prediction_2treat_df) + 1]] <- prediction_df
}
cat('<br>\n<br>\n<br>')
}
saveRDS(prediction_2treat_df, "prediction_2treat_df.rds")
}\[\text{Demographics}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | (0.01144); [0.5198] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | (0.01097); [0.1848] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | (0.01252); [ -7.0158] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | (0.01114); [-10.0482] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00595 | 0.0002093 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01107); [0.01891] | |
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.00203 | -0.003628 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01027); [-0.3531] | |
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.08784 | -0.09335 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01211); [-7.706] | |
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11196 | -0.1174 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01061); [-11.06] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68715 | (0.00733) | -0.00043 | 0.0007216 |
| Emotions | 0 | 0 | 0.68672 | (0.006659) | (0.001309); [0.5511] | |
| Facts Baseline | 0 | 1 | 0.5375 | (0.007286) | 0.00360 | 0.00231 |
| Emotions | 0 | 1 | 0.5411 | (0.006665) | (0.001163); [1.987] | |
| Facts Baseline | 1 | 0 | 0.61233 | (0.00733) | -0.09527 | -0.09243 |
| Emotions | 1 | 0 | 0.51706 | (0.006659) | (0.01082); [-8.545] | |
| Facts Baseline | 1 | 1 | 0.46497 | (0.007286) | -0.11333 | -0.117 |
| Emotions | 1 | 1 | 0.35164 | (0.006665) | (0.009287); [-12.6] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68285 | (0.006739) | 0.00595 | |
| Emotions | 0 | 0 | 0.6888 | (0.006114) | ||
| Facts Baseline | 0 | 1 | 0.53795 | (0.006743) | 0.00260 | |
| Emotions | 0 | 1 | 0.54055 | (0.00612) | ||
| Facts Baseline | 1 | 0 | 0.60706 | (0.006739) | -0.08783 | |
| Emotions | 1 | 0 | 0.51923 | (0.006114) | ||
| Facts Baseline | 1 | 1 | 0.4643 | (0.006743) | -0.11138 | |
| Emotions | 1 | 1 | 0.35292 | (0.00612) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68642 | (0.007268) | -0.00151 | |
| Emotions | 0 | 0 | 0.68491 | (0.006642) | ||
| Facts Baseline | 0 | 1 | 0.53772 | (0.007549) | 0.00260 | |
| Emotions | 0 | 1 | 0.54032 | (0.006976) | ||
| Facts Baseline | 1 | 0 | 0.6121 | (0.007268) | -0.09711 | |
| Emotions | 1 | 0 | 0.51499 | (0.006642) | ||
| Facts Baseline | 1 | 1 | 0.46742 | (0.007549) | -0.11765 | |
| Emotions | 1 | 1 | 0.34977 | (0.006976) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68638 | (0.007268) | -0.00151 | 0.0006994 |
| Emotions | 0 | 0 | 0.68487 | (0.006642) | (0.001323); [0.5287] | |
| Facts Baseline | 0 | 1 | 0.53774 | (0.007549) | 0.00260 | 0.001624 |
| Emotions | 0 | 1 | 0.54034 | (0.006976) | (0.001449); [1.121] | |
| Facts Baseline | 1 | 0 | 0.61207 | (0.007268) | -0.09712 | -0.09515 |
| Emotions | 1 | 0 | 0.51495 | (0.006642) | (0.01066); [-8.926] | |
| Facts Baseline | 1 | 1 | 0.46744 | (0.007549) | -0.11765 | -0.1217 |
| Emotions | 1 | 1 | 0.34979 | (0.006976) | (0.008987); [-13.54] |
\[\text{Sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | (0.01144); [0.5198] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | (0.01097); [0.1848] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | (0.01252); [ -7.0158] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | (0.01114); [-10.0482] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00595 | 0.0002093 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01107); [0.01891] | |
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.00203 | -0.003628 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01027); [-0.3531] | |
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.08784 | -0.09335 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01211); [-7.706] | |
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11196 | -0.1174 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01061); [-11.06] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68715 | (0.00733) | -0.00043 | 0.0007216 |
| Emotions | 0 | 0 | 0.68672 | (0.006659) | (0.001309); [0.5511] | |
| Facts Baseline | 0 | 1 | 0.5375 | (0.007286) | 0.00360 | 0.00231 |
| Emotions | 0 | 1 | 0.5411 | (0.006665) | (0.001163); [1.987] | |
| Facts Baseline | 1 | 0 | 0.61233 | (0.00733) | -0.09527 | -0.09243 |
| Emotions | 1 | 0 | 0.51706 | (0.006659) | (0.01082); [-8.545] | |
| Facts Baseline | 1 | 1 | 0.46497 | (0.007286) | -0.11333 | -0.117 |
| Emotions | 1 | 1 | 0.35164 | (0.006665) | (0.009287); [-12.6] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68285 | (0.006739) | 0.00595 | |
| Emotions | 0 | 0 | 0.6888 | (0.006114) | ||
| Facts Baseline | 0 | 1 | 0.53795 | (0.006743) | 0.00260 | |
| Emotions | 0 | 1 | 0.54055 | (0.00612) | ||
| Facts Baseline | 1 | 0 | 0.60706 | (0.006739) | -0.08783 | |
| Emotions | 1 | 0 | 0.51923 | (0.006114) | ||
| Facts Baseline | 1 | 1 | 0.4643 | (0.006743) | -0.11138 | |
| Emotions | 1 | 1 | 0.35292 | (0.00612) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68642 | (0.007268) | -0.00151 | |
| Emotions | 0 | 0 | 0.68491 | (0.006642) | ||
| Facts Baseline | 0 | 1 | 0.53772 | (0.007549) | 0.00260 | |
| Emotions | 0 | 1 | 0.54032 | (0.006976) | ||
| Facts Baseline | 1 | 0 | 0.6121 | (0.007268) | -0.09711 | |
| Emotions | 1 | 0 | 0.51499 | (0.006642) | ||
| Facts Baseline | 1 | 1 | 0.46742 | (0.007549) | -0.11765 | |
| Emotions | 1 | 1 | 0.34977 | (0.006976) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68638 | (0.007268) | -0.00151 | 0.0006994 |
| Emotions | 0 | 0 | 0.68487 | (0.006642) | (0.001323); [0.5287] | |
| Facts Baseline | 0 | 1 | 0.53774 | (0.007549) | 0.00260 | 0.001624 |
| Emotions | 0 | 1 | 0.54034 | (0.006976) | (0.001449); [1.121] | |
| Facts Baseline | 1 | 0 | 0.61207 | (0.007268) | -0.09712 | -0.09515 |
| Emotions | 1 | 0 | 0.51495 | (0.006642) | (0.01066); [-8.926] | |
| Facts Baseline | 1 | 1 | 0.46744 | (0.007549) | -0.11765 | -0.1217 |
| Emotions | 1 | 1 | 0.34979 | (0.006976) | (0.008987); [-13.54] |
\[\text{Tactic sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | (0.01144); [0.5198] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | (0.01097); [0.1848] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | (0.01252); [ -7.0158] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | (0.01114); [-10.0482] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00595 | 0.0002093 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01107); [0.01891] | |
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.00203 | -0.003628 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01027); [-0.3531] | |
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.08784 | -0.09335 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01211); [-7.706] | |
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11196 | -0.1174 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01061); [-11.06] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68715 | (0.00733) | -0.00043 | 0.0007216 |
| Emotions | 0 | 0 | 0.68672 | (0.006659) | (0.001309); [0.5511] | |
| Facts Baseline | 0 | 1 | 0.5375 | (0.007286) | 0.00360 | 0.00231 |
| Emotions | 0 | 1 | 0.5411 | (0.006665) | (0.001163); [1.987] | |
| Facts Baseline | 1 | 0 | 0.61233 | (0.00733) | -0.09527 | -0.09243 |
| Emotions | 1 | 0 | 0.51706 | (0.006659) | (0.01082); [-8.545] | |
| Facts Baseline | 1 | 1 | 0.46497 | (0.007286) | -0.11333 | -0.117 |
| Emotions | 1 | 1 | 0.35164 | (0.006665) | (0.009287); [-12.6] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68285 | (0.006739) | 0.00595 | |
| Emotions | 0 | 0 | 0.6888 | (0.006114) | ||
| Facts Baseline | 0 | 1 | 0.53795 | (0.006743) | 0.00260 | |
| Emotions | 0 | 1 | 0.54055 | (0.00612) | ||
| Facts Baseline | 1 | 0 | 0.60706 | (0.006739) | -0.08783 | |
| Emotions | 1 | 0 | 0.51923 | (0.006114) | ||
| Facts Baseline | 1 | 1 | 0.4643 | (0.006743) | -0.11138 | |
| Emotions | 1 | 1 | 0.35292 | (0.00612) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68642 | (0.007268) | -0.00151 | |
| Emotions | 0 | 0 | 0.68491 | (0.006642) | ||
| Facts Baseline | 0 | 1 | 0.53772 | (0.007549) | 0.00260 | |
| Emotions | 0 | 1 | 0.54032 | (0.006976) | ||
| Facts Baseline | 1 | 0 | 0.6121 | (0.007268) | -0.09711 | |
| Emotions | 1 | 0 | 0.51499 | (0.006642) | ||
| Facts Baseline | 1 | 1 | 0.46742 | (0.007549) | -0.11765 | |
| Emotions | 1 | 1 | 0.34977 | (0.006976) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68638 | (0.007268) | -0.00151 | 0.0006994 |
| Emotions | 0 | 0 | 0.68487 | (0.006642) | (0.001323); [0.5287] | |
| Facts Baseline | 0 | 1 | 0.53774 | (0.007549) | 0.00260 | 0.001624 |
| Emotions | 0 | 1 | 0.54034 | (0.006976) | (0.001449); [1.121] | |
| Facts Baseline | 1 | 0 | 0.61207 | (0.007268) | -0.09712 | -0.09515 |
| Emotions | 1 | 0 | 0.51495 | (0.006642) | (0.01066); [-8.926] | |
| Facts Baseline | 1 | 1 | 0.46744 | (0.007549) | -0.11765 | -0.1217 |
| Emotions | 1 | 1 | 0.34979 | (0.006976) | (0.008987); [-13.54] |
\[\text{Demographics +
Sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | (0.01144); [0.5198] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | (0.01097); [0.1848] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | (0.01252); [ -7.0158] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | (0.01114); [-10.0482] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00595 | 0.0002093 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01107); [0.01891] | |
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.00203 | -0.003628 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01027); [-0.3531] | |
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.08784 | -0.09335 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01211); [-7.706] | |
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11196 | -0.1174 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01061); [-11.06] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68715 | (0.00733) | -0.00043 | 0.0007216 |
| Emotions | 0 | 0 | 0.68672 | (0.006659) | (0.001309); [0.5511] | |
| Facts Baseline | 0 | 1 | 0.5375 | (0.007286) | 0.00360 | 0.00231 |
| Emotions | 0 | 1 | 0.5411 | (0.006665) | (0.001163); [1.987] | |
| Facts Baseline | 1 | 0 | 0.61233 | (0.00733) | -0.09527 | -0.09243 |
| Emotions | 1 | 0 | 0.51706 | (0.006659) | (0.01082); [-8.545] | |
| Facts Baseline | 1 | 1 | 0.46497 | (0.007286) | -0.11333 | -0.117 |
| Emotions | 1 | 1 | 0.35164 | (0.006665) | (0.009287); [-12.6] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68285 | (0.006739) | 0.00595 | |
| Emotions | 0 | 0 | 0.6888 | (0.006114) | ||
| Facts Baseline | 0 | 1 | 0.53795 | (0.006743) | 0.00260 | |
| Emotions | 0 | 1 | 0.54055 | (0.00612) | ||
| Facts Baseline | 1 | 0 | 0.60706 | (0.006739) | -0.08783 | |
| Emotions | 1 | 0 | 0.51923 | (0.006114) | ||
| Facts Baseline | 1 | 1 | 0.4643 | (0.006743) | -0.11138 | |
| Emotions | 1 | 1 | 0.35292 | (0.00612) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68642 | (0.007268) | -0.00151 | |
| Emotions | 0 | 0 | 0.68491 | (0.006642) | ||
| Facts Baseline | 0 | 1 | 0.53772 | (0.007549) | 0.00260 | |
| Emotions | 0 | 1 | 0.54032 | (0.006976) | ||
| Facts Baseline | 1 | 0 | 0.6121 | (0.007268) | -0.09711 | |
| Emotions | 1 | 0 | 0.51499 | (0.006642) | ||
| Facts Baseline | 1 | 1 | 0.46742 | (0.007549) | -0.11765 | |
| Emotions | 1 | 1 | 0.34977 | (0.006976) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68638 | (0.007268) | -0.00151 | 0.0006994 |
| Emotions | 0 | 0 | 0.68487 | (0.006642) | (0.001323); [0.5287] | |
| Facts Baseline | 0 | 1 | 0.53774 | (0.007549) | 0.00260 | 0.001624 |
| Emotions | 0 | 1 | 0.54034 | (0.006976) | (0.001449); [1.121] | |
| Facts Baseline | 1 | 0 | 0.61207 | (0.007268) | -0.09712 | -0.09515 |
| Emotions | 1 | 0 | 0.51495 | (0.006642) | (0.01066); [-8.926] | |
| Facts Baseline | 1 | 1 | 0.46744 | (0.007549) | -0.11765 | -0.1217 |
| Emotions | 1 | 1 | 0.34979 | (0.006976) | (0.008987); [-13.54] |
\[\text{Demographics +
Sharing rates + Tactic sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.0059 | 0.005948 |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | (0.01144); [0.5198] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.0021 | 0.002028 |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | (0.01097); [0.1848] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.0879 | -0.087832 |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | (0.01252); [ -7.0158] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.1120 | -0.111959 |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | (0.01114); [-10.0482] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00595 | 0.0002093 |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | (0.01107); [0.01891] | |
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.00203 | -0.003628 |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | (0.01027); [-0.3531] | |
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.08784 | -0.09335 |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | (0.01211); [-7.706] | |
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11196 | -0.1174 |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | (0.01061); [-11.06] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68715 | (0.00733) | -0.00043 | 0.0007216 |
| Emotions | 0 | 0 | 0.68672 | (0.006659) | (0.001309); [0.5511] | |
| Facts Baseline | 0 | 1 | 0.5375 | (0.007286) | 0.00360 | 0.00231 |
| Emotions | 0 | 1 | 0.5411 | (0.006665) | (0.001163); [1.987] | |
| Facts Baseline | 1 | 0 | 0.61233 | (0.00733) | -0.09527 | -0.09243 |
| Emotions | 1 | 0 | 0.51706 | (0.006659) | (0.01082); [-8.545] | |
| Facts Baseline | 1 | 1 | 0.46497 | (0.007286) | -0.11333 | -0.117 |
| Emotions | 1 | 1 | 0.35164 | (0.006665) | (0.009287); [-12.6] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68285 | (0.006739) | 0.00595 | |
| Emotions | 0 | 0 | 0.6888 | (0.006114) | ||
| Facts Baseline | 0 | 1 | 0.53795 | (0.006743) | 0.00260 | |
| Emotions | 0 | 1 | 0.54055 | (0.00612) | ||
| Facts Baseline | 1 | 0 | 0.60706 | (0.006739) | -0.08783 | |
| Emotions | 1 | 0 | 0.51923 | (0.006114) | ||
| Facts Baseline | 1 | 1 | 0.4643 | (0.006743) | -0.11138 | |
| Emotions | 1 | 1 | 0.35292 | (0.00612) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68642 | (0.007268) | -0.00151 | |
| Emotions | 0 | 0 | 0.68491 | (0.006642) | ||
| Facts Baseline | 0 | 1 | 0.53772 | (0.007549) | 0.00260 | |
| Emotions | 0 | 1 | 0.54032 | (0.006976) | ||
| Facts Baseline | 1 | 0 | 0.6121 | (0.007268) | -0.09711 | |
| Emotions | 1 | 0 | 0.51499 | (0.006642) | ||
| Facts Baseline | 1 | 1 | 0.46742 | (0.007549) | -0.11765 | |
| Emotions | 1 | 1 | 0.34977 | (0.006976) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment effect Emotions–Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68638 | (0.007268) | -0.00151 | 0.0006994 |
| Emotions | 0 | 0 | 0.68487 | (0.006642) | (0.001323); [0.5287] | |
| Facts Baseline | 0 | 1 | 0.53774 | (0.007549) | 0.00260 | 0.001624 |
| Emotions | 0 | 1 | 0.54034 | (0.006976) | (0.001449); [1.121] | |
| Facts Baseline | 1 | 0 | 0.61207 | (0.007268) | -0.09712 | -0.09515 |
| Emotions | 1 | 0 | 0.51495 | (0.006642) | (0.01066); [-8.926] | |
| Facts Baseline | 1 | 1 | 0.46744 | (0.007549) | -0.11765 | -0.1217 |
| Emotions | 1 | 1 | 0.34979 | (0.006976) | (0.008987); [-13.54] |
Click for All treatment arms
# version 2: all treatment arms
covariate_list <- c(demographics_demeaned_treatment,
sharing_demeaned_treatment,
sharing_tactic_demeaned_treatment,
demographics_sharing_demeaned_treatment,
all_covariates_demeaned_treatment)
cov_estimates_list <- list()
for (cov_list in covariate_list) {
# Demographics
formula_cov1_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list)
cov1_lm <- feols(as.formula(formula_cov1_lm),panel)
formula_cov2_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list," | user")
cov2_lm <- feols(as.formula(formula_cov2_lm),panel)
# Misinfo and Non-misinfo sharing rates
formula_cov3_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact) "," + ",cov_list," | user")
cov3_lm <- feols(as.formula(formula_cov3_lm),panel)
# Misinfo, Non-misinfo, and Tactic sharing rates
formula_cov4_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(type) "," + ",cov_list," | user")
cov4_lm <- feols(as.formula(formula_cov4_lm),panel)
# Demographics and Misinfo and Non-misinfo sharing rates
formula_cov5_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) "," + ",cov_list," | user")
cov5_lm <- feols(as.formula(formula_cov5_lm),panel)
# Demographics, Misinfo and Non-misinfo sharing rates, and Tactic sharing rates
formula_cov6_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(post_id) "," + ",cov_list," | user")
cov6_lm <- feols(as.formula(formula_cov6_lm),panel)
cov_estimates_list[[length(cov_estimates_list) + 1]] <- list(cov1_lm,cov2_lm,cov3_lm,cov4_lm,cov5_lm,cov6_lm)
}# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
# estimate out-of-sample R2 for all fixed effects models
# use R2oosse function from oosse package
# w user FE
if (file.exists("cov_roose_all_list.rds")) {
cov_roose_all_list <- readRDS("cov_roose_all_list.rds")
} else {
cov_roose_all_list <- list()
models_cov <- list('cov1_lm','cov2_lm','cov3_lm','cov4_lm','cov5_lm','cov6_lm')
for (cov_list in covariate_list) {
# Demographics
formula_cov1_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list)
formula_cov2_lm <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list," | user")
# Misinfo and Non-misinfo sharing rates
formula_cov3_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact) "," + ",cov_list," | user")
# Misinfo, Non-misinfo, and Tactic sharing rates
formula_cov4_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(type) "," + ",cov_list," | user")
# Demographics and Misinfo and Non-misinfo sharing rates
formula_cov5_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) "," + ",cov_list," | user")
# Demographics, Misinfo and Non-misinfo sharing rates, and Tactic sharing rates
formula_cov6_lm <- paste0("share ~ treatment*post*post_is_misinfo + factor(post_id) "," + ",cov_list," | user")
# Estimate R2oosse for all models
R2oos_cov_alltreat <- list()
for (m in seq_along(models_cov)) {
#print(paste0('cov: ',length(cov_roose_all_list) + 1,'; model: ',m))
if (file.exists(paste0("R2oos_cov_alltreat_", m,"_",length(cov_roose_all_list) + 1,".rds"))) {
R2oos_cov_alltreat[[m]] <- readRDS(paste0("R2oos_cov_alltreat_", m,"_",length(cov_roose_all_list) + 1,".rds"))
} else {
formula <- gsub("^share", "", gsub(" \\| user", "", get(paste0("formula_cov", m, "_lm"))))
X <- model.matrix(as.formula(formula),
data = panel %>%
mutate(type = as.numeric(factor(panel$type, levels = unique(panel$type)))))
X <- X[, !(colnames(X) %in% "(Intercept)")]
X <- cbind(X, user = panel$user, treatment = panel$treatment)
# identify estimation dataset
estimation_data <- panel %>% filter(row_number() %in% obs(get(models_cov[[m]])))
if ("treatment" %in% colnames(X)) {
X[,"treatment"] <- factor(estimation_data$treatment,
levels = levels(estimation_data$treatment))
}
# add fixed effects based on model
if (m == 3) {
X <- cbind(X, fact = estimation_data$fact)
}
if (m == 4) {
X <- cbind(X, type = as.numeric(factor(estimation_data$type, levels = unique(estimation_data$type))))
}
if (m == 5) {
X <- cbind(X, fact = estimation_data$fact,
type = as.numeric(factor(estimation_data$type), levels = unique(estimation_data$type)))
}
if (m == 6) {
X <- cbind(X, post_id = estimation_data$post_id)
}
# Define the fit function for models with fixed effects
fitFunLM_cov <- function(y, x) {
feols(as.formula(get(paste0("formula_cov", m, "_lm"))),
data = estimation_data %>%
mutate(type = as.numeric(factor(estimation_data$type, levels = unique(estimation_data$type)))))
}
# Define the prediction function for models with fixed effects
predFunLM_cov <- function(mod,x) {
predict(mod, newdata = x, type = "response")
}
# Calculate R2oosse
R2oos_cov_alltreat[[m]] <- oosse::R2oosse(y = estimation_data$share,
x = X,
fitFun = fitFunLM_cov,
predFun = predFunLM_cov,
methodMSE = "CV",
cvReps = 100,
nFold = 10,
nBootstrapsCor = 25)
# Save results
saveRDS(R2oos_cov_alltreat[[m]], paste0("R2oos_cov_alltreat_", m,"_",length(cov_roose_all_list) + 1,".rds"))
}
}
R2oos_cov_alltreat_all <- as.numeric(unlist(lapply(R2oos_cov_alltreat, function(x) x$R2[1])),4)
cov_roose_all_list[[length(cov_roose_all_list) + 1]] <- R2oos_cov_alltreat_all
}
saveRDS(cov_roose_all_list, "cov_roose_all_list.rds")
}covariate_list_names <- c("Demographics",
"Sharing rates",
"Tactic sharing rates",
"Demographics + Sharing rates",
"Demographics + Sharing rates + Tactic sharing rates")
for (cov in seq(length(covariate_list))) {
# Print model
headers <- c("Share (binary indicator)")
model_list <- cov_estimates_list[[cov]]
names <- unique(c(names(cov_estimates_list[[cov]][[1]]$coefficients),
names(cov_estimates_list[[cov]][[2]]$coefficients),
names(cov_estimates_list[[cov]][[3]]$coefficients),
names(cov_estimates_list[[cov]][[4]]$coefficients),
names(cov_estimates_list[[cov]][[5]]$coefficients),
names(cov_estimates_list[[cov]][[6]]$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# out-of-sample R2
if (any(ls() %in% "R2oos_cov_alltreat_all")) {
#R2oos_cov_alltreat_all <- c(R2oos_cov_alltreat_all)
R2oos_cov_alltreat_all <- rep("",length(model_list))
} else {
R2oos_cov_alltreat_all <- rep("",length(model_list))
}
# Print model name
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_alltreat_all)),
reorder.gof = c(6,7,8,9,10,5,1,2,3,4)
)
cat(html_table)
cat("<br>")
# Full table (all independent variables)
cat('<details>')
cat('<summary>Click to view the full table</summary>')
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_alltreat_all)),
reorder.gof = c(6,7,8,9,10,5,1,2,3,4)
)
cat(html_table)
cat('</details>')
cat('<br>\n<br>\n<br>')
}| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0159) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0129) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0112) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.0847 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.0840 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6877*** | |||||
| (0.0065) | ||||||
| No-course Baseline | 0.0147* | |||||
| (0.0094) | ||||||
| Emotions | 0.0004 | |||||
| (0.0092) | ||||||
| Reasoning | 0.0140* | |||||
| (0.0094) | ||||||
| Combo | -0.0012 | |||||
| (0.0093) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0742*** | -0.0742*** |
| (0.0092) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1459*** | -0.1173*** | -0.1537*** | -0.4320*** |
| (0.0080) | (0.0069) | (0.0067) | (0.0095) | (0.0127) | (0.0123) | |
| Age | -0.0036 | |||||
| (0.0030) | ||||||
| Man | 0.0136*** | |||||
| (0.0027) | ||||||
| Some college | -0.0063** | |||||
| (0.0033) | ||||||
| Bachelor’s degree | -0.0143*** | |||||
| (0.0034) | ||||||
| Graduate degree | -0.0108*** | |||||
| (0.0029) | ||||||
| Married | 0.0006 | |||||
| (0.0030) | ||||||
| Unemployed | -0.0155*** | |||||
| (0.0035) | ||||||
| Employed | -0.0090*** | |||||
| (0.0038) | ||||||
| Mostly urban | -0.0282*** | |||||
| (0.0032) | ||||||
| Suburban | -0.0174*** | |||||
| (0.0031) | ||||||
| Christian | 0.0089*** | |||||
| (0.0028) | ||||||
| Attends religious services | 0.0103*** | |||||
| (0.0027) | ||||||
| Uses social media | -0.0196*** | |||||
| (0.0025) | ||||||
| Hours spent on Social Media | 0.0076*** | |||||
| (0.0030) | ||||||
| Social Media Share: 80-100% | 0.0827*** | |||||
| (0.0033) | ||||||
| Social Media Share: 60-80% | 0.1135*** | |||||
| (0.0036) | ||||||
| Social Media Share: 40-60% | 0.0988*** | |||||
| (0.0037) | ||||||
| Social Media Share: 20-40% | 0.0524*** | |||||
| (0.0035) | ||||||
| Passed attention check at pre-treatment | -0.0217*** | |||||
| (0.0027) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.0514*** | 0.0514*** | 0.0513*** | 0.0514*** | 0.0507*** | 0.0507*** |
| (0.0133) | (0.0126) | (0.0124) | (0.0126) | (0.0123) | (0.0123) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0129) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0132) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0131) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| No-course Baseline x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0116) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0112) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0114) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0114) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0039 | 0.0039 |
| (0.0113) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| No-course Baseline x Age | -0.0049 | |||||
| (0.0042) | ||||||
| Emotions x Age | -0.0116*** | |||||
| (0.0044) | ||||||
| Reasoning x Age | -0.0070* | |||||
| (0.0044) | ||||||
| Combo x Age | 0.0006 | |||||
| (0.0043) | ||||||
| No-course Baseline x Man | -0.0232*** | |||||
| (0.0040) | ||||||
| Emotions x Man | -0.0022 | |||||
| (0.0038) | ||||||
| Reasoning x Man | -0.0033 | |||||
| (0.0039) | ||||||
| Combo x Man | -0.0060* | |||||
| (0.0039) | ||||||
| No-course Baseline x Some college | 0.0021 | |||||
| (0.0047) | ||||||
| Emotions x Some college | -0.0034 | |||||
| (0.0047) | ||||||
| Reasoning x Some college | -0.0098** | |||||
| (0.0047) | ||||||
| Combo x Some college | 0.0013 | |||||
| (0.0047) | ||||||
| No-course Baseline x Bachelor’s degree | 0.0173*** | |||||
| (0.0048) | ||||||
| Emotions x Bachelor’s degree | 0.0086** | |||||
| (0.0047) | ||||||
| Reasoning x Bachelor’s degree | -0.0026 | |||||
| (0.0048) | ||||||
| Combo x Bachelor’s degree | -0.0049 | |||||
| (0.0048) | ||||||
| No-course Baseline x Graduate degree | 0.0125*** | |||||
| (0.0043) | ||||||
| Emotions x Graduate degree | 0.0109*** | |||||
| (0.0039) | ||||||
| Reasoning x Graduate degree | 0.0150*** | |||||
| (0.0044) | ||||||
| Combo x Graduate degree | 0.0164*** | |||||
| (0.0038) | ||||||
| No-course Baseline x Married | 0.0082** | |||||
| (0.0044) | ||||||
| Emotions x Married | 0.0051 | |||||
| (0.0043) | ||||||
| Reasoning x Married | 0.0010 | |||||
| (0.0044) | ||||||
| Combo x Married | -0.0076** | |||||
| (0.0044) | ||||||
| No-course Baseline x Unemployed | 0.0052 | |||||
| (0.0050) | ||||||
| Emotions x Unemployed | -0.0071* | |||||
| (0.0049) | ||||||
| Reasoning x Unemployed | 0.0098** | |||||
| (0.0050) | ||||||
| Combo x Unemployed | 0.0013 | |||||
| (0.0050) | ||||||
| No-course Baseline x Employed | -0.0035 | |||||
| (0.0055) | ||||||
| Emotions x Employed | -0.0189*** | |||||
| (0.0053) | ||||||
| Reasoning x Employed | -0.0019 | |||||
| (0.0055) | ||||||
| Combo x Employed | -0.0011 | |||||
| (0.0054) | ||||||
| No-course Baseline x Mostly urban | 0.0294*** | |||||
| (0.0046) | ||||||
| Emotions x Mostly urban | 0.0286*** | |||||
| (0.0045) | ||||||
| Reasoning x Mostly urban | 0.0063* | |||||
| (0.0046) | ||||||
| Combo x Mostly urban | 0.0213*** | |||||
| (0.0045) | ||||||
| No-course Baseline x Suburban | 0.0043 | |||||
| (0.0045) | ||||||
| Emotions x Suburban | 0.0149*** | |||||
| (0.0044) | ||||||
| Reasoning x Suburban | 0.0040 | |||||
| (0.0045) | ||||||
| Combo x Suburban | 0.0111*** | |||||
| (0.0045) | ||||||
| No-course Baseline x Christian | -0.0116*** | |||||
| (0.0039) | ||||||
| Emotions x Christian | -0.0083** | |||||
| (0.0040) | ||||||
| Reasoning x Christian | -0.0078** | |||||
| (0.0041) | ||||||
| Combo x Christian | -0.0026 | |||||
| (0.0040) | ||||||
| No-course Baseline x Attends religious services | -0.0124*** | |||||
| (0.0038) | ||||||
| Emotions x Attends religious services | -0.0214*** | |||||
| (0.0039) | ||||||
| Reasoning x Attends religious services | -0.0076** | |||||
| (0.0040) | ||||||
| Combo x Attends religious services | -0.0147*** | |||||
| (0.0040) | ||||||
| No-course Baseline x Uses social media | 0.0154*** | |||||
| (0.0036) | ||||||
| Emotions x Uses social media | 0.0062* | |||||
| (0.0040) | ||||||
| Reasoning x Uses social media | 0.0119*** | |||||
| (0.0037) | ||||||
| Combo x Uses social media | 0.0042 | |||||
| (0.0042) | ||||||
| No-course Baseline x Hours spent on Social Media | -0.0140*** | |||||
| (0.0041) | ||||||
| Emotions x Hours spent on Social Media | -0.0128*** | |||||
| (0.0041) | ||||||
| Reasoning x Hours spent on Social Media | 0.0037 | |||||
| (0.0042) | ||||||
| Combo x Hours spent on Social Media | -0.0203*** | |||||
| (0.0040) | ||||||
| No-course Baseline x Social Media Share: 80-100% | -0.0080** | |||||
| (0.0046) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0095** | |||||
| (0.0046) | ||||||
| Reasoning x Social Media Share: 80-100% | 0.0007 | |||||
| (0.0046) | ||||||
| Combo x Social Media Share: 80-100% | -0.0124*** | |||||
| (0.0046) | ||||||
| No-course Baseline x Social Media Share: 60-80% | -0.0029 | |||||
| (0.0053) | ||||||
| Emotions x Social Media Share: 60-80% | -0.0077* | |||||
| (0.0051) | ||||||
| Reasoning x Social Media Share: 60-80% | -0.0105** | |||||
| (0.0053) | ||||||
| Combo x Social Media Share: 60-80% | -0.0115** | |||||
| (0.0052) | ||||||
| No-course Baseline x Social Media Share: 40-60% | -0.0054 | |||||
| (0.0054) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0166*** | |||||
| (0.0052) | ||||||
| Reasoning x Social Media Share: 40-60% | -0.0073* | |||||
| (0.0054) | ||||||
| Combo x Social Media Share: 40-60% | -0.0209*** | |||||
| (0.0053) | ||||||
| No-course Baseline x Social Media Share: 20-40% | -0.0124*** | |||||
| (0.0051) | ||||||
| Emotions x Social Media Share: 20-40% | -0.0092** | |||||
| (0.0049) | ||||||
| Reasoning x Social Media Share: 20-40% | -0.0049 | |||||
| (0.0051) | ||||||
| Combo x Social Media Share: 20-40% | -0.0112** | |||||
| (0.0050) | ||||||
| No-course Baseline x Passed attention check at pre-treatment | -0.0144*** | |||||
| (0.0039) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0039 | |||||
| (0.0038) | ||||||
| Reasoning x Passed attention check at pre-treatment | 0.0115*** | |||||
| (0.0039) | ||||||
| Combo x Passed attention check at pre-treatment | -0.0079** | |||||
| (0.0038) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 | 0.0033 | 0.0046 | 0.0033 | 0.0042 | 0.0042 |
| (0.0163) | (0.0140) | (0.0133) | (0.0140) | (0.0131) | (0.0131) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202 | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0159) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0162) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0161) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| Post domain 2 | -0.2001*** | -0.1794*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 3 | -0.2813*** | -0.3043*** | ||||
| (0.0055) | (0.0095) | |||||
| Post domain 4 | -0.1804*** | -0.2332*** | ||||
| (0.0056) | (0.0097) | |||||
| Post domain 5 | 0.0545*** | -0.0010 | ||||
| (0.0051) | (0.0087) | |||||
| Post domain 6 | 0.0164*** | 0.0389*** | ||||
| (0.0054) | (0.0084) | |||||
| Post domain 7 | -0.2347*** | -0.1827*** | ||||
| (0.0056) | (0.0094) | |||||
| Post domain 8 | -0.2195*** | -0.1160*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 9 | 0.0369*** | -0.0608*** | ||||
| (0.0053) | (0.0091) | |||||
| Post domain 10 | -0.0565*** | -0.0963*** | ||||
| (0.0054) | (0.0091) | |||||
| Post domain 11 | -0.1586*** | -0.1739*** | ||||
| (0.0057) | (0.0094) | |||||
| Post domain 12 | -0.1047*** | -0.1343*** | ||||
| (0.0054) | (0.0092) | |||||
| Post domain 13 | -0.1541*** | -0.2155*** | ||||
| (0.0055) | (0.0093) | |||||
| Post domain 14 | -0.2759*** | -0.3151*** | ||||
| (0.0057) | (0.0097) | |||||
| Post domain 15 | -0.1868*** | -0.2755*** | ||||
| (0.0056) | (0.0095) | |||||
| 1[Combo manipulation tactic] | -0.0512*** | -0.0275*** | ||||
| (0.0041) | (0.0078) | |||||
| 1[Emotions manipulation tactic] | -0.0154*** | -0.0585*** | ||||
| (0.0041) | (0.0080) | |||||
| 1[Reasoning manipulation tactic] | -0.0143*** | 0.0208*** | ||||
| (0.0041) | (0.0079) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0536*** | |||||
| (0.0158) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0462*** | |||||
| (0.0152) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0789*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.0971*** | |||||
| (0.0149) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3110*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0471*** | |||||
| (0.0153) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0192 | |||||
| (0.0155) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1469*** | |||||
| (0.0150) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0318** | |||||
| (0.0154) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0595*** | |||||
| (0.0153) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0230* | |||||
| (0.0156) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0600*** | |||||
| (0.0154) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0296** | |||||
| (0.0153) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.0822*** | |||||
| (0.0153) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0845*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1118*** | |||||
| (0.0154) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0617*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1184*** | |||||
| (0.0148) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1477*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0324** | |||||
| (0.0151) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1539*** | |||||
| (0.0153) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1660*** | |||||
| (0.0153) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1090*** | |||||
| (0.0152) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0397*** | |||||
| (0.0154) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1449*** | |||||
| (0.0153) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.1875*** | |||||
| (0.0151) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1025*** | |||||
| (0.0154) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3384*** | |||||
| (0.0155) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0427*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0517*** | |||||
| (0.0150) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.0985*** | |||||
| (0.0155) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0326** | |||||
| (0.0150) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0629*** | |||||
| (0.0144) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1533*** | |||||
| (0.0149) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2920*** | |||||
| (0.0151) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1260*** | |||||
| (0.0151) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0377*** | |||||
| (0.0151) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0529*** | |||||
| (0.0155) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0132 | |||||
| (0.0151) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0267** | |||||
| (0.0153) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0441*** | |||||
| (0.0152) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0236* | |||||
| (0.0151) | ||||||
| Post 2 indicator | -0.1794*** | |||||
| (0.0095) | ||||||
| Post 3 indicator | -0.3043*** | |||||
| (0.0095) | ||||||
| Post 4 indicator | -0.2332*** | |||||
| (0.0097) | ||||||
| Post 5 indicator | -0.0010 | |||||
| (0.0087) | ||||||
| Post 6 indicator | 0.0389*** | |||||
| (0.0084) | ||||||
| Post 7 indicator | -0.1827*** | |||||
| (0.0094) | ||||||
| Post 8 indicator | -0.1160*** | |||||
| (0.0095) | ||||||
| Post 9 indicator | -0.0608*** | |||||
| (0.0091) | ||||||
| Post 10 indicator | -0.0963*** | |||||
| (0.0091) | ||||||
| Post 11 indicator | -0.1739*** | |||||
| (0.0094) | ||||||
| Post 12 indicator | -0.1343*** | |||||
| (0.0092) | ||||||
| Post 13 indicator | -0.2155*** | |||||
| (0.0093) | ||||||
| Post 14 indicator | -0.3151*** | |||||
| (0.0097) | ||||||
| Post 15 indicator | -0.2755*** | |||||
| (0.0095) | ||||||
| Post 16 indicator | 0.2508*** | |||||
| (0.0118) | ||||||
| Post 17 indicator | 0.0178* | |||||
| (0.0119) | ||||||
| Post 18 indicator | -0.0072 | |||||
| (0.0113) | ||||||
| Post 19 indicator | 0.0965*** | |||||
| (0.0114) | ||||||
| Post 20 indicator | 0.3469*** | |||||
| (0.0115) | ||||||
| Post 21 indicator | -0.0213** | |||||
| (0.0113) | ||||||
| Post 22 indicator | 0.0211** | |||||
| (0.0114) | ||||||
| Post 23 indicator | 0.1157*** | |||||
| (0.0118) | ||||||
| Post 24 indicator | 0.3370*** | |||||
| (0.0116) | ||||||
| Post 25 indicator | 0.1863*** | |||||
| (0.0118) | ||||||
| Post 26 indicator | 0.1364*** | |||||
| (0.0120) | ||||||
| Post 27 indicator | 0.0935*** | |||||
| (0.0116) | ||||||
| Post 28 indicator | 0.0953*** | |||||
| (0.0117) | ||||||
| Post 29 indicator | -0.0346*** | |||||
| (0.0113) | ||||||
| Post 30 indicator | 0.0576*** | |||||
| (0.0116) | ||||||
| Post 31 indicator | 0.2198*** | |||||
| (0.0117) | ||||||
| Post 32 indicator | -0.0441*** | |||||
| (0.0115) | ||||||
| Post 33 indicator | 0.0273*** | |||||
| (0.0113) | ||||||
| Post 34 indicator | 0.0482*** | |||||
| (0.0114) | ||||||
| Post 35 indicator | 0.3371*** | |||||
| (0.0115) | ||||||
| Post 36 indicator | 0.4064*** | |||||
| (0.0116) | ||||||
| Post 37 indicator | 0.0046 | |||||
| (0.0112) | ||||||
| Post 38 indicator | -0.0501*** | |||||
| (0.0112) | ||||||
| Post 39 indicator | 0.3250*** | |||||
| (0.0117) | ||||||
| Post 40 indicator | 0.2324*** | |||||
| (0.0118) | ||||||
| Post 41 indicator | 0.0062 | |||||
| (0.0116) | ||||||
| Post 42 indicator | 0.2304*** | |||||
| (0.0118) | ||||||
| Post 43 indicator | 0.1917*** | |||||
| (0.0115) | ||||||
| Post 44 indicator | 0.0072 | |||||
| (0.0114) | ||||||
| Post 45 indicator | 0.2826*** | |||||
| (0.0118) | ||||||
| Post 46 indicator | 0.2991*** | |||||
| (0.0118) | ||||||
| Post 47 indicator | 0.1624*** | |||||
| (0.0118) | ||||||
| Post 48 indicator | -0.0569*** | |||||
| (0.0110) | ||||||
| Post 49 indicator | 0.1643*** | |||||
| (0.0121) | ||||||
| Post 50 indicator | 0.3306*** | |||||
| (0.0118) | ||||||
| Post 51 indicator | 0.4009*** | |||||
| (0.0113) | ||||||
| Post 52 indicator | -0.0369*** | |||||
| (0.0113) | ||||||
| Post 53 indicator | -0.1089*** | |||||
| (0.0115) | ||||||
| Post 54 indicator | 0.3643*** | |||||
| (0.0116) | ||||||
| Post 55 indicator | 0.2405*** | |||||
| (0.0118) | ||||||
| Post 56 indicator | 0.1781*** | |||||
| (0.0122) | ||||||
| Post 57 indicator | 0.1780*** | |||||
| (0.0115) | ||||||
| Post 58 indicator | 0.1102*** | |||||
| (0.0117) | ||||||
| Post 59 indicator | 0.0281*** | |||||
| (0.0116) | ||||||
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.0847 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.0840 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2790 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2788 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6897*** | |||||
| (0.0058) | ||||||
| No-course Baseline | 0.0006 | |||||
| (0.0084) | ||||||
| Emotions | 0.0026 | |||||
| (0.0081) | ||||||
| Reasoning | 0.0142** | |||||
| (0.0083) | ||||||
| Combo | -0.0018 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0742*** | -0.0742*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1459*** | -0.1173*** | -0.1537*** | -0.4320*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0095) | (0.0127) | (0.0123) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.1802*** | |||||
| (0.0030) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.1019*** | |||||
| (0.0030) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.0514*** | 0.0514*** | 0.0513*** | 0.0514*** | 0.0507*** | 0.0507*** |
| (0.0118) | (0.0126) | (0.0124) | (0.0126) | (0.0123) | (0.0123) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0117) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0116) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| No-course Baseline x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0103) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0102) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0101) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0039 | 0.0039 |
| (0.0100) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | 0.0028 | |||||
| (0.0043) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0213*** | |||||
| (0.0042) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | -0.0047 | |||||
| (0.0043) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.0097** | |||||
| (0.0043) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0039 | |||||
| (0.0044) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0144*** | |||||
| (0.0041) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0111*** | |||||
| (0.0043) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0089** | |||||
| (0.0043) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 | 0.0033 | 0.0046 | 0.0033 | 0.0042 | 0.0042 |
| (0.0145) | (0.0140) | (0.0133) | (0.0140) | (0.0131) | (0.0131) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0144) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0143) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| Post domain 2 | -0.2001*** | -0.1794*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 3 | -0.2813*** | -0.3043*** | ||||
| (0.0055) | (0.0095) | |||||
| Post domain 4 | -0.1804*** | -0.2332*** | ||||
| (0.0056) | (0.0097) | |||||
| Post domain 5 | 0.0545*** | -0.0010 | ||||
| (0.0051) | (0.0087) | |||||
| Post domain 6 | 0.0164*** | 0.0389*** | ||||
| (0.0054) | (0.0084) | |||||
| Post domain 7 | -0.2347*** | -0.1827*** | ||||
| (0.0056) | (0.0094) | |||||
| Post domain 8 | -0.2195*** | -0.1160*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 9 | 0.0369*** | -0.0608*** | ||||
| (0.0053) | (0.0091) | |||||
| Post domain 10 | -0.0565*** | -0.0963*** | ||||
| (0.0054) | (0.0091) | |||||
| Post domain 11 | -0.1586*** | -0.1739*** | ||||
| (0.0057) | (0.0094) | |||||
| Post domain 12 | -0.1047*** | -0.1343*** | ||||
| (0.0054) | (0.0092) | |||||
| Post domain 13 | -0.1541*** | -0.2155*** | ||||
| (0.0055) | (0.0093) | |||||
| Post domain 14 | -0.2759*** | -0.3151*** | ||||
| (0.0057) | (0.0097) | |||||
| Post domain 15 | -0.1868*** | -0.2755*** | ||||
| (0.0056) | (0.0095) | |||||
| 1[Combo manipulation tactic] | -0.0512*** | -0.0275*** | ||||
| (0.0041) | (0.0078) | |||||
| 1[Emotions manipulation tactic] | -0.0154*** | -0.0585*** | ||||
| (0.0041) | (0.0080) | |||||
| 1[Reasoning manipulation tactic] | -0.0143*** | 0.0208*** | ||||
| (0.0041) | (0.0079) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0536*** | |||||
| (0.0158) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0462*** | |||||
| (0.0152) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0789*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.0971*** | |||||
| (0.0149) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3110*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0471*** | |||||
| (0.0153) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0192 | |||||
| (0.0155) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1469*** | |||||
| (0.0150) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0318** | |||||
| (0.0154) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0595*** | |||||
| (0.0153) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0230* | |||||
| (0.0156) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0600*** | |||||
| (0.0154) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0296** | |||||
| (0.0153) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.0822*** | |||||
| (0.0153) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0845*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1118*** | |||||
| (0.0154) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0617*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1184*** | |||||
| (0.0148) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1477*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0324** | |||||
| (0.0151) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1539*** | |||||
| (0.0153) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1660*** | |||||
| (0.0153) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1090*** | |||||
| (0.0152) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0397*** | |||||
| (0.0154) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1449*** | |||||
| (0.0153) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.1875*** | |||||
| (0.0151) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1025*** | |||||
| (0.0154) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3384*** | |||||
| (0.0155) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0427*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0517*** | |||||
| (0.0150) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.0985*** | |||||
| (0.0155) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0326** | |||||
| (0.0150) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0629*** | |||||
| (0.0144) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1533*** | |||||
| (0.0149) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2920*** | |||||
| (0.0151) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1260*** | |||||
| (0.0151) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0377*** | |||||
| (0.0151) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0529*** | |||||
| (0.0155) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0132 | |||||
| (0.0151) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0267** | |||||
| (0.0153) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0441*** | |||||
| (0.0152) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0236* | |||||
| (0.0151) | ||||||
| Post 2 indicator | -0.1794*** | |||||
| (0.0095) | ||||||
| Post 3 indicator | -0.3043*** | |||||
| (0.0095) | ||||||
| Post 4 indicator | -0.2332*** | |||||
| (0.0097) | ||||||
| Post 5 indicator | -0.0010 | |||||
| (0.0087) | ||||||
| Post 6 indicator | 0.0389*** | |||||
| (0.0084) | ||||||
| Post 7 indicator | -0.1827*** | |||||
| (0.0094) | ||||||
| Post 8 indicator | -0.1160*** | |||||
| (0.0095) | ||||||
| Post 9 indicator | -0.0608*** | |||||
| (0.0091) | ||||||
| Post 10 indicator | -0.0963*** | |||||
| (0.0091) | ||||||
| Post 11 indicator | -0.1739*** | |||||
| (0.0094) | ||||||
| Post 12 indicator | -0.1343*** | |||||
| (0.0092) | ||||||
| Post 13 indicator | -0.2155*** | |||||
| (0.0093) | ||||||
| Post 14 indicator | -0.3151*** | |||||
| (0.0097) | ||||||
| Post 15 indicator | -0.2755*** | |||||
| (0.0095) | ||||||
| Post 16 indicator | 0.2508*** | |||||
| (0.0118) | ||||||
| Post 17 indicator | 0.0178* | |||||
| (0.0119) | ||||||
| Post 18 indicator | -0.0072 | |||||
| (0.0113) | ||||||
| Post 19 indicator | 0.0965*** | |||||
| (0.0114) | ||||||
| Post 20 indicator | 0.3469*** | |||||
| (0.0115) | ||||||
| Post 21 indicator | -0.0213** | |||||
| (0.0113) | ||||||
| Post 22 indicator | 0.0211** | |||||
| (0.0114) | ||||||
| Post 23 indicator | 0.1157*** | |||||
| (0.0118) | ||||||
| Post 24 indicator | 0.3370*** | |||||
| (0.0116) | ||||||
| Post 25 indicator | 0.1863*** | |||||
| (0.0118) | ||||||
| Post 26 indicator | 0.1364*** | |||||
| (0.0120) | ||||||
| Post 27 indicator | 0.0935*** | |||||
| (0.0116) | ||||||
| Post 28 indicator | 0.0953*** | |||||
| (0.0117) | ||||||
| Post 29 indicator | -0.0346*** | |||||
| (0.0113) | ||||||
| Post 30 indicator | 0.0576*** | |||||
| (0.0116) | ||||||
| Post 31 indicator | 0.2198*** | |||||
| (0.0117) | ||||||
| Post 32 indicator | -0.0441*** | |||||
| (0.0115) | ||||||
| Post 33 indicator | 0.0273*** | |||||
| (0.0113) | ||||||
| Post 34 indicator | 0.0482*** | |||||
| (0.0114) | ||||||
| Post 35 indicator | 0.3371*** | |||||
| (0.0115) | ||||||
| Post 36 indicator | 0.4064*** | |||||
| (0.0116) | ||||||
| Post 37 indicator | 0.0046 | |||||
| (0.0112) | ||||||
| Post 38 indicator | -0.0501*** | |||||
| (0.0112) | ||||||
| Post 39 indicator | 0.3250*** | |||||
| (0.0117) | ||||||
| Post 40 indicator | 0.2324*** | |||||
| (0.0118) | ||||||
| Post 41 indicator | 0.0062 | |||||
| (0.0116) | ||||||
| Post 42 indicator | 0.2304*** | |||||
| (0.0118) | ||||||
| Post 43 indicator | 0.1917*** | |||||
| (0.0115) | ||||||
| Post 44 indicator | 0.0072 | |||||
| (0.0114) | ||||||
| Post 45 indicator | 0.2826*** | |||||
| (0.0118) | ||||||
| Post 46 indicator | 0.2991*** | |||||
| (0.0118) | ||||||
| Post 47 indicator | 0.1624*** | |||||
| (0.0118) | ||||||
| Post 48 indicator | -0.0569*** | |||||
| (0.0110) | ||||||
| Post 49 indicator | 0.1643*** | |||||
| (0.0121) | ||||||
| Post 50 indicator | 0.3306*** | |||||
| (0.0118) | ||||||
| Post 51 indicator | 0.4009*** | |||||
| (0.0113) | ||||||
| Post 52 indicator | -0.0369*** | |||||
| (0.0113) | ||||||
| Post 53 indicator | -0.1089*** | |||||
| (0.0115) | ||||||
| Post 54 indicator | 0.3643*** | |||||
| (0.0116) | ||||||
| Post 55 indicator | 0.2405*** | |||||
| (0.0118) | ||||||
| Post 56 indicator | 0.1781*** | |||||
| (0.0122) | ||||||
| Post 57 indicator | 0.1780*** | |||||
| (0.0115) | ||||||
| Post 58 indicator | 0.1102*** | |||||
| (0.0117) | ||||||
| Post 59 indicator | 0.0281*** | |||||
| (0.0116) | ||||||
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2790 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2788 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2791 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2789 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6898*** | |||||
| (0.0058) | ||||||
| No-course Baseline | 0.0004 | |||||
| (0.0084) | ||||||
| Emotions | 0.0023 | |||||
| (0.0081) | ||||||
| Reasoning | 0.0139** | |||||
| (0.0083) | ||||||
| Combo | -0.0019 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0742*** | -0.0742*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1459*** | -0.1173*** | -0.1537*** | -0.4320*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0095) | (0.0127) | (0.0123) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.0901 | |||||
| (805.5739) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.1021*** | |||||
| (0.0030) | ||||||
| Emotions posts | 0.0397 | |||||
| (329.1976) | ||||||
| Reasoning posts | 0.0323 | |||||
| (326.0741) | ||||||
| Combo posts | 0.0384 | |||||
| (332.7445) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.0514*** | 0.0514*** | 0.0513*** | 0.0514*** | 0.0507*** | 0.0507*** |
| (0.0118) | (0.0126) | (0.0124) | (0.0126) | (0.0123) | (0.0123) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0117) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0116) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| No-course Baseline x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0103) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0102) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0101) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0039 | 0.0039 |
| (0.0100) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | -0.0017 | |||||
| (0.0104) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0371*** | |||||
| (0.0102) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | 0.0066 | |||||
| (0.0103) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.0082 | |||||
| (0.0103) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0042 | |||||
| (0.0044) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0147*** | |||||
| (0.0042) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0116*** | |||||
| (0.0043) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0090** | |||||
| (0.0043) | ||||||
| No-course Baseline x Emotions posts | 0.0025 | |||||
| (0.0067) | ||||||
| Emotions x Emotions posts | 0.0066 | |||||
| (0.0066) | ||||||
| Reasoning x Emotions posts | -0.0033 | |||||
| (0.0067) | ||||||
| Combo x Emotions posts | -0.0075 | |||||
| (0.0067) | ||||||
| No-course Baseline x Reasoning posts | 0.0032 | |||||
| (0.0067) | ||||||
| Emotions x Reasoning posts | 0.0131** | |||||
| (0.0066) | ||||||
| Reasoning x Reasoning posts | -0.0105* | |||||
| (0.0066) | ||||||
| Combo x Reasoning posts | 0.0058 | |||||
| (0.0067) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 | 0.0033 | 0.0046 | 0.0033 | 0.0042 | 0.0042 |
| (0.0145) | (0.0140) | (0.0133) | (0.0140) | (0.0131) | (0.0131) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0141) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0144) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0143) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| Post domain 2 | -0.2001*** | -0.1794*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 3 | -0.2813*** | -0.3043*** | ||||
| (0.0055) | (0.0095) | |||||
| Post domain 4 | -0.1804*** | -0.2332*** | ||||
| (0.0056) | (0.0097) | |||||
| Post domain 5 | 0.0545*** | -0.0010 | ||||
| (0.0051) | (0.0087) | |||||
| Post domain 6 | 0.0164*** | 0.0389*** | ||||
| (0.0054) | (0.0084) | |||||
| Post domain 7 | -0.2347*** | -0.1827*** | ||||
| (0.0056) | (0.0094) | |||||
| Post domain 8 | -0.2195*** | -0.1160*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 9 | 0.0369*** | -0.0608*** | ||||
| (0.0053) | (0.0091) | |||||
| Post domain 10 | -0.0565*** | -0.0963*** | ||||
| (0.0054) | (0.0091) | |||||
| Post domain 11 | -0.1586*** | -0.1739*** | ||||
| (0.0057) | (0.0094) | |||||
| Post domain 12 | -0.1047*** | -0.1343*** | ||||
| (0.0054) | (0.0092) | |||||
| Post domain 13 | -0.1541*** | -0.2155*** | ||||
| (0.0055) | (0.0093) | |||||
| Post domain 14 | -0.2759*** | -0.3151*** | ||||
| (0.0057) | (0.0097) | |||||
| Post domain 15 | -0.1868*** | -0.2755*** | ||||
| (0.0056) | (0.0095) | |||||
| 1[Combo manipulation tactic] | -0.0512*** | -0.0275*** | ||||
| (0.0041) | (0.0078) | |||||
| 1[Emotions manipulation tactic] | -0.0154*** | -0.0585*** | ||||
| (0.0041) | (0.0080) | |||||
| 1[Reasoning manipulation tactic] | -0.0143*** | 0.0208*** | ||||
| (0.0041) | (0.0079) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0536*** | |||||
| (0.0158) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0462*** | |||||
| (0.0152) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0789*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.0971*** | |||||
| (0.0149) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3110*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0471*** | |||||
| (0.0153) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0192 | |||||
| (0.0155) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1469*** | |||||
| (0.0150) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0318** | |||||
| (0.0154) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0595*** | |||||
| (0.0153) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0230* | |||||
| (0.0156) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0600*** | |||||
| (0.0154) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0296** | |||||
| (0.0153) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.0822*** | |||||
| (0.0153) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0845*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1118*** | |||||
| (0.0154) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0617*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1184*** | |||||
| (0.0148) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1477*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0324** | |||||
| (0.0151) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1539*** | |||||
| (0.0153) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1660*** | |||||
| (0.0153) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1090*** | |||||
| (0.0152) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0397*** | |||||
| (0.0154) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1449*** | |||||
| (0.0153) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.1875*** | |||||
| (0.0151) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1025*** | |||||
| (0.0154) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3384*** | |||||
| (0.0155) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0427*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0517*** | |||||
| (0.0150) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.0985*** | |||||
| (0.0155) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0326** | |||||
| (0.0150) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0629*** | |||||
| (0.0144) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1533*** | |||||
| (0.0149) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2920*** | |||||
| (0.0151) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1260*** | |||||
| (0.0151) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0377*** | |||||
| (0.0151) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0529*** | |||||
| (0.0155) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0132 | |||||
| (0.0151) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0267** | |||||
| (0.0153) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0441*** | |||||
| (0.0152) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0236* | |||||
| (0.0151) | ||||||
| Post 2 indicator | -0.1794*** | |||||
| (0.0095) | ||||||
| Post 3 indicator | -0.3043*** | |||||
| (0.0095) | ||||||
| Post 4 indicator | -0.2332*** | |||||
| (0.0097) | ||||||
| Post 5 indicator | -0.0010 | |||||
| (0.0087) | ||||||
| Post 6 indicator | 0.0389*** | |||||
| (0.0084) | ||||||
| Post 7 indicator | -0.1827*** | |||||
| (0.0094) | ||||||
| Post 8 indicator | -0.1160*** | |||||
| (0.0095) | ||||||
| Post 9 indicator | -0.0608*** | |||||
| (0.0091) | ||||||
| Post 10 indicator | -0.0963*** | |||||
| (0.0091) | ||||||
| Post 11 indicator | -0.1739*** | |||||
| (0.0094) | ||||||
| Post 12 indicator | -0.1343*** | |||||
| (0.0092) | ||||||
| Post 13 indicator | -0.2155*** | |||||
| (0.0093) | ||||||
| Post 14 indicator | -0.3151*** | |||||
| (0.0097) | ||||||
| Post 15 indicator | -0.2755*** | |||||
| (0.0095) | ||||||
| Post 16 indicator | 0.2508*** | |||||
| (0.0118) | ||||||
| Post 17 indicator | 0.0178* | |||||
| (0.0119) | ||||||
| Post 18 indicator | -0.0072 | |||||
| (0.0113) | ||||||
| Post 19 indicator | 0.0965*** | |||||
| (0.0114) | ||||||
| Post 20 indicator | 0.3469*** | |||||
| (0.0115) | ||||||
| Post 21 indicator | -0.0213** | |||||
| (0.0113) | ||||||
| Post 22 indicator | 0.0211** | |||||
| (0.0114) | ||||||
| Post 23 indicator | 0.1157*** | |||||
| (0.0118) | ||||||
| Post 24 indicator | 0.3370*** | |||||
| (0.0116) | ||||||
| Post 25 indicator | 0.1863*** | |||||
| (0.0118) | ||||||
| Post 26 indicator | 0.1364*** | |||||
| (0.0120) | ||||||
| Post 27 indicator | 0.0935*** | |||||
| (0.0116) | ||||||
| Post 28 indicator | 0.0953*** | |||||
| (0.0117) | ||||||
| Post 29 indicator | -0.0346*** | |||||
| (0.0113) | ||||||
| Post 30 indicator | 0.0576*** | |||||
| (0.0116) | ||||||
| Post 31 indicator | 0.2198*** | |||||
| (0.0117) | ||||||
| Post 32 indicator | -0.0441*** | |||||
| (0.0115) | ||||||
| Post 33 indicator | 0.0273*** | |||||
| (0.0113) | ||||||
| Post 34 indicator | 0.0482*** | |||||
| (0.0114) | ||||||
| Post 35 indicator | 0.3371*** | |||||
| (0.0115) | ||||||
| Post 36 indicator | 0.4064*** | |||||
| (0.0116) | ||||||
| Post 37 indicator | 0.0046 | |||||
| (0.0112) | ||||||
| Post 38 indicator | -0.0501*** | |||||
| (0.0112) | ||||||
| Post 39 indicator | 0.3250*** | |||||
| (0.0117) | ||||||
| Post 40 indicator | 0.2324*** | |||||
| (0.0118) | ||||||
| Post 41 indicator | 0.0062 | |||||
| (0.0116) | ||||||
| Post 42 indicator | 0.2304*** | |||||
| (0.0118) | ||||||
| Post 43 indicator | 0.1917*** | |||||
| (0.0115) | ||||||
| Post 44 indicator | 0.0072 | |||||
| (0.0114) | ||||||
| Post 45 indicator | 0.2826*** | |||||
| (0.0118) | ||||||
| Post 46 indicator | 0.2991*** | |||||
| (0.0118) | ||||||
| Post 47 indicator | 0.1624*** | |||||
| (0.0118) | ||||||
| Post 48 indicator | -0.0569*** | |||||
| (0.0110) | ||||||
| Post 49 indicator | 0.1643*** | |||||
| (0.0121) | ||||||
| Post 50 indicator | 0.3306*** | |||||
| (0.0118) | ||||||
| Post 51 indicator | 0.4009*** | |||||
| (0.0113) | ||||||
| Post 52 indicator | -0.0369*** | |||||
| (0.0113) | ||||||
| Post 53 indicator | -0.1089*** | |||||
| (0.0115) | ||||||
| Post 54 indicator | 0.3643*** | |||||
| (0.0116) | ||||||
| Post 55 indicator | 0.2405*** | |||||
| (0.0118) | ||||||
| Post 56 indicator | 0.1781*** | |||||
| (0.0122) | ||||||
| Post 57 indicator | 0.1780*** | |||||
| (0.0115) | ||||||
| Post 58 indicator | 0.1102*** | |||||
| (0.0117) | ||||||
| Post 59 indicator | 0.0281*** | |||||
| (0.0116) | ||||||
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2791 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2789 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0140) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2816 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2811 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6907*** | |||||
| (0.0058) | ||||||
| No-course Baseline | -0.0012 | |||||
| (0.0084) | ||||||
| Emotions | 0.0016 | |||||
| (0.0081) | ||||||
| Reasoning | 0.0122* | |||||
| (0.0083) | ||||||
| Combo | -0.0031 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0742*** | -0.0742*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1459*** | -0.1173*** | -0.1537*** | -0.4320*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0095) | (0.0127) | (0.0123) | |
| Age | -0.0058** | |||||
| (0.0026) | ||||||
| Man | 0.0028 | |||||
| (0.0024) | ||||||
| Some college | -0.0018 | |||||
| (0.0030) | ||||||
| Bachelor’s degree | 0.0034 | |||||
| (0.0030) | ||||||
| Graduate degree | -0.0078*** | |||||
| (0.0026) | ||||||
| Married | -0.0028 | |||||
| (0.0027) | ||||||
| Unemployed | -0.0045* | |||||
| (0.0031) | ||||||
| Employed | -0.0084*** | |||||
| (0.0034) | ||||||
| Mostly urban | -0.0056** | |||||
| (0.0028) | ||||||
| Suburban | -0.0052** | |||||
| (0.0028) | ||||||
| Christian | 0.0031 | |||||
| (0.0025) | ||||||
| Attends religious services | 0.0007 | |||||
| (0.0024) | ||||||
| Uses social media | -0.0060*** | |||||
| (0.0022) | ||||||
| Hours spent on Social Media | 0.0052** | |||||
| (0.0026) | ||||||
| Social Media Share: 80-100% | 0.0130*** | |||||
| (0.0030) | ||||||
| Social Media Share: 60-80% | 0.0179*** | |||||
| (0.0033) | ||||||
| Social Media Share: 40-60% | 0.0149*** | |||||
| (0.0034) | ||||||
| Social Media Share: 20-40% | 0.0065** | |||||
| (0.0031) | ||||||
| Passed attention check at pre-treatment | -0.0046** | |||||
| (0.0024) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.1749*** | |||||
| (0.0031) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.0999*** | |||||
| (0.0030) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.0514*** | 0.0514*** | 0.0513*** | 0.0514*** | 0.0507*** | 0.0507*** |
| (0.0118) | (0.0126) | (0.0124) | (0.0126) | (0.0123) | (0.0123) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0117) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0116) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| No-course Baseline x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0102) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0101) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0101) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0039 | 0.0039 |
| (0.0100) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| No-course Baseline x Age | -0.0006 | |||||
| (0.0037) | ||||||
| Emotions x Age | -0.0040 | |||||
| (0.0039) | ||||||
| Reasoning x Age | -0.0018 | |||||
| (0.0039) | ||||||
| Combo x Age | -0.0011 | |||||
| (0.0038) | ||||||
| No-course Baseline x Man | -0.0075** | |||||
| (0.0035) | ||||||
| Emotions x Man | 0.0061** | |||||
| (0.0034) | ||||||
| Reasoning x Man | -0.0067** | |||||
| (0.0034) | ||||||
| Combo x Man | 0.0034 | |||||
| (0.0034) | ||||||
| No-course Baseline x Some college | -0.0014 | |||||
| (0.0042) | ||||||
| Emotions x Some college | 0.0003 | |||||
| (0.0041) | ||||||
| Reasoning x Some college | -0.0013 | |||||
| (0.0042) | ||||||
| Combo x Some college | -0.0035 | |||||
| (0.0042) | ||||||
| No-course Baseline x Bachelor’s degree | -0.0047 | |||||
| (0.0043) | ||||||
| Emotions x Bachelor’s degree | -0.0029 | |||||
| (0.0042) | ||||||
| Reasoning x Bachelor’s degree | -0.0077** | |||||
| (0.0042) | ||||||
| Combo x Bachelor’s degree | -0.0104*** | |||||
| (0.0042) | ||||||
| No-course Baseline x Graduate degree | 0.0027 | |||||
| (0.0038) | ||||||
| Emotions x Graduate degree | 0.0094*** | |||||
| (0.0034) | ||||||
| Reasoning x Graduate degree | 0.0019 | |||||
| (0.0039) | ||||||
| Combo x Graduate degree | 0.0087*** | |||||
| (0.0034) | ||||||
| No-course Baseline x Married | 0.0030 | |||||
| (0.0039) | ||||||
| Emotions x Married | -0.0050* | |||||
| (0.0038) | ||||||
| Reasoning x Married | 0.0017 | |||||
| (0.0039) | ||||||
| Combo x Married | -0.0042 | |||||
| (0.0039) | ||||||
| No-course Baseline x Unemployed | 0.0024 | |||||
| (0.0044) | ||||||
| Emotions x Unemployed | -0.0017 | |||||
| (0.0043) | ||||||
| Reasoning x Unemployed | 0.0047 | |||||
| (0.0044) | ||||||
| Combo x Unemployed | 0.0010 | |||||
| (0.0044) | ||||||
| No-course Baseline x Employed | 0.0002 | |||||
| (0.0049) | ||||||
| Emotions x Employed | 0.0020 | |||||
| (0.0047) | ||||||
| Reasoning x Employed | 0.0028 | |||||
| (0.0049) | ||||||
| Combo x Employed | 0.0020 | |||||
| (0.0048) | ||||||
| No-course Baseline x Mostly urban | 0.0080** | |||||
| (0.0041) | ||||||
| Emotions x Mostly urban | 0.0080** | |||||
| (0.0040) | ||||||
| Reasoning x Mostly urban | -0.0010 | |||||
| (0.0041) | ||||||
| Combo x Mostly urban | 0.0098*** | |||||
| (0.0040) | ||||||
| No-course Baseline x Suburban | 0.0037 | |||||
| (0.0040) | ||||||
| Emotions x Suburban | -0.0013 | |||||
| (0.0039) | ||||||
| Reasoning x Suburban | -0.0031 | |||||
| (0.0040) | ||||||
| Combo x Suburban | 0.0090** | |||||
| (0.0040) | ||||||
| No-course Baseline x Christian | -0.0077** | |||||
| (0.0034) | ||||||
| Emotions x Christian | -0.0036 | |||||
| (0.0035) | ||||||
| Reasoning x Christian | 0.0026 | |||||
| (0.0036) | ||||||
| Combo x Christian | -0.0040 | |||||
| (0.0035) | ||||||
| No-course Baseline x Attends religious services | -0.0011 | |||||
| (0.0033) | ||||||
| Emotions x Attends religious services | -0.0063** | |||||
| (0.0035) | ||||||
| Reasoning x Attends religious services | -0.0030 | |||||
| (0.0036) | ||||||
| Combo x Attends religious services | -0.0033 | |||||
| (0.0035) | ||||||
| No-course Baseline x Uses social media | 0.0058** | |||||
| (0.0032) | ||||||
| Emotions x Uses social media | 0.0086*** | |||||
| (0.0036) | ||||||
| Reasoning x Uses social media | 0.0078*** | |||||
| (0.0032) | ||||||
| Combo x Uses social media | 0.0045 | |||||
| (0.0037) | ||||||
| No-course Baseline x Hours spent on Social Media | -0.0069** | |||||
| (0.0036) | ||||||
| Emotions x Hours spent on Social Media | -0.0020 | |||||
| (0.0036) | ||||||
| Reasoning x Hours spent on Social Media | -0.0001 | |||||
| (0.0037) | ||||||
| Combo x Hours spent on Social Media | -0.0101*** | |||||
| (0.0036) | ||||||
| No-course Baseline x Social Media Share: 80-100% | 0.0002 | |||||
| (0.0042) | ||||||
| Emotions x Social Media Share: 80-100% | 0.0011 | |||||
| (0.0042) | ||||||
| Reasoning x Social Media Share: 80-100% | 0.0017 | |||||
| (0.0042) | ||||||
| Combo x Social Media Share: 80-100% | 0.0007 | |||||
| (0.0042) | ||||||
| No-course Baseline x Social Media Share: 60-80% | -0.0036 | |||||
| (0.0049) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0051 | |||||
| (0.0047) | ||||||
| Reasoning x Social Media Share: 60-80% | -0.0003 | |||||
| (0.0049) | ||||||
| Combo x Social Media Share: 60-80% | 0.0004 | |||||
| (0.0048) | ||||||
| No-course Baseline x Social Media Share: 40-60% | -0.0036 | |||||
| (0.0050) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0016 | |||||
| (0.0048) | ||||||
| Reasoning x Social Media Share: 40-60% | -0.0048 | |||||
| (0.0050) | ||||||
| Combo x Social Media Share: 40-60% | -0.0035 | |||||
| (0.0049) | ||||||
| No-course Baseline x Social Media Share: 20-40% | -0.0008 | |||||
| (0.0045) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0044 | |||||
| (0.0044) | ||||||
| Reasoning x Social Media Share: 20-40% | -0.0033 | |||||
| (0.0045) | ||||||
| Combo x Social Media Share: 20-40% | -0.0042 | |||||
| (0.0045) | ||||||
| No-course Baseline x Passed attention check at pre-treatment | -0.0055* | |||||
| (0.0035) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0009 | |||||
| (0.0034) | ||||||
| Reasoning x Passed attention check at pre-treatment | 0.0009 | |||||
| (0.0034) | ||||||
| Combo x Passed attention check at pre-treatment | 0.0076** | |||||
| (0.0034) | ||||||
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | 0.0023 | |||||
| (0.0045) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0230*** | |||||
| (0.0043) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | -0.0042 | |||||
| (0.0044) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.0081** | |||||
| (0.0045) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0022 | |||||
| (0.0044) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0125*** | |||||
| (0.0042) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0106*** | |||||
| (0.0043) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0077** | |||||
| (0.0043) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 | 0.0033 | 0.0046 | 0.0033 | 0.0042 | 0.0042 |
| (0.0145) | (0.0140) | (0.0133) | (0.0140) | (0.0131) | (0.0131) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0140) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0143) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0142) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| Post domain 2 | -0.2001*** | -0.1794*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 3 | -0.2813*** | -0.3043*** | ||||
| (0.0055) | (0.0095) | |||||
| Post domain 4 | -0.1804*** | -0.2332*** | ||||
| (0.0056) | (0.0097) | |||||
| Post domain 5 | 0.0545*** | -0.0010 | ||||
| (0.0051) | (0.0087) | |||||
| Post domain 6 | 0.0164*** | 0.0389*** | ||||
| (0.0054) | (0.0084) | |||||
| Post domain 7 | -0.2347*** | -0.1827*** | ||||
| (0.0056) | (0.0094) | |||||
| Post domain 8 | -0.2195*** | -0.1160*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 9 | 0.0369*** | -0.0608*** | ||||
| (0.0053) | (0.0091) | |||||
| Post domain 10 | -0.0565*** | -0.0963*** | ||||
| (0.0054) | (0.0091) | |||||
| Post domain 11 | -0.1586*** | -0.1739*** | ||||
| (0.0057) | (0.0094) | |||||
| Post domain 12 | -0.1047*** | -0.1343*** | ||||
| (0.0054) | (0.0092) | |||||
| Post domain 13 | -0.1541*** | -0.2155*** | ||||
| (0.0055) | (0.0093) | |||||
| Post domain 14 | -0.2759*** | -0.3151*** | ||||
| (0.0057) | (0.0097) | |||||
| Post domain 15 | -0.1868*** | -0.2755*** | ||||
| (0.0056) | (0.0095) | |||||
| 1[Combo manipulation tactic] | -0.0512*** | -0.0275*** | ||||
| (0.0041) | (0.0078) | |||||
| 1[Emotions manipulation tactic] | -0.0154*** | -0.0585*** | ||||
| (0.0041) | (0.0080) | |||||
| 1[Reasoning manipulation tactic] | -0.0143*** | 0.0208*** | ||||
| (0.0041) | (0.0079) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0536*** | |||||
| (0.0158) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0462*** | |||||
| (0.0152) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0789*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.0971*** | |||||
| (0.0149) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3110*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0471*** | |||||
| (0.0153) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0192 | |||||
| (0.0155) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1469*** | |||||
| (0.0150) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0318** | |||||
| (0.0154) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0595*** | |||||
| (0.0153) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0230* | |||||
| (0.0156) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0600*** | |||||
| (0.0154) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0296** | |||||
| (0.0153) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.0822*** | |||||
| (0.0153) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0845*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1118*** | |||||
| (0.0154) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0617*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1184*** | |||||
| (0.0148) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1477*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0324** | |||||
| (0.0151) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1539*** | |||||
| (0.0153) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1660*** | |||||
| (0.0153) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1090*** | |||||
| (0.0152) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0397*** | |||||
| (0.0154) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1449*** | |||||
| (0.0153) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.1875*** | |||||
| (0.0151) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1025*** | |||||
| (0.0154) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3384*** | |||||
| (0.0155) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0427*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0517*** | |||||
| (0.0150) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.0985*** | |||||
| (0.0155) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0326** | |||||
| (0.0150) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0629*** | |||||
| (0.0144) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1533*** | |||||
| (0.0149) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2920*** | |||||
| (0.0151) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1260*** | |||||
| (0.0151) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0377*** | |||||
| (0.0151) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0529*** | |||||
| (0.0155) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0132 | |||||
| (0.0151) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0267** | |||||
| (0.0153) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0441*** | |||||
| (0.0152) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0236* | |||||
| (0.0151) | ||||||
| Post 2 indicator | -0.1794*** | |||||
| (0.0095) | ||||||
| Post 3 indicator | -0.3043*** | |||||
| (0.0095) | ||||||
| Post 4 indicator | -0.2332*** | |||||
| (0.0097) | ||||||
| Post 5 indicator | -0.0010 | |||||
| (0.0087) | ||||||
| Post 6 indicator | 0.0389*** | |||||
| (0.0084) | ||||||
| Post 7 indicator | -0.1827*** | |||||
| (0.0094) | ||||||
| Post 8 indicator | -0.1160*** | |||||
| (0.0095) | ||||||
| Post 9 indicator | -0.0608*** | |||||
| (0.0091) | ||||||
| Post 10 indicator | -0.0963*** | |||||
| (0.0091) | ||||||
| Post 11 indicator | -0.1739*** | |||||
| (0.0094) | ||||||
| Post 12 indicator | -0.1343*** | |||||
| (0.0092) | ||||||
| Post 13 indicator | -0.2155*** | |||||
| (0.0093) | ||||||
| Post 14 indicator | -0.3151*** | |||||
| (0.0097) | ||||||
| Post 15 indicator | -0.2755*** | |||||
| (0.0095) | ||||||
| Post 16 indicator | 0.2508*** | |||||
| (0.0118) | ||||||
| Post 17 indicator | 0.0178* | |||||
| (0.0119) | ||||||
| Post 18 indicator | -0.0072 | |||||
| (0.0113) | ||||||
| Post 19 indicator | 0.0965*** | |||||
| (0.0114) | ||||||
| Post 20 indicator | 0.3469*** | |||||
| (0.0115) | ||||||
| Post 21 indicator | -0.0213** | |||||
| (0.0113) | ||||||
| Post 22 indicator | 0.0211** | |||||
| (0.0114) | ||||||
| Post 23 indicator | 0.1157*** | |||||
| (0.0118) | ||||||
| Post 24 indicator | 0.3370*** | |||||
| (0.0116) | ||||||
| Post 25 indicator | 0.1863*** | |||||
| (0.0118) | ||||||
| Post 26 indicator | 0.1364*** | |||||
| (0.0120) | ||||||
| Post 27 indicator | 0.0935*** | |||||
| (0.0116) | ||||||
| Post 28 indicator | 0.0953*** | |||||
| (0.0117) | ||||||
| Post 29 indicator | -0.0346*** | |||||
| (0.0113) | ||||||
| Post 30 indicator | 0.0576*** | |||||
| (0.0116) | ||||||
| Post 31 indicator | 0.2198*** | |||||
| (0.0117) | ||||||
| Post 32 indicator | -0.0441*** | |||||
| (0.0115) | ||||||
| Post 33 indicator | 0.0273*** | |||||
| (0.0113) | ||||||
| Post 34 indicator | 0.0482*** | |||||
| (0.0114) | ||||||
| Post 35 indicator | 0.3371*** | |||||
| (0.0115) | ||||||
| Post 36 indicator | 0.4064*** | |||||
| (0.0116) | ||||||
| Post 37 indicator | 0.0046 | |||||
| (0.0112) | ||||||
| Post 38 indicator | -0.0501*** | |||||
| (0.0112) | ||||||
| Post 39 indicator | 0.3250*** | |||||
| (0.0117) | ||||||
| Post 40 indicator | 0.2324*** | |||||
| (0.0118) | ||||||
| Post 41 indicator | 0.0062 | |||||
| (0.0116) | ||||||
| Post 42 indicator | 0.2304*** | |||||
| (0.0118) | ||||||
| Post 43 indicator | 0.1917*** | |||||
| (0.0115) | ||||||
| Post 44 indicator | 0.0072 | |||||
| (0.0114) | ||||||
| Post 45 indicator | 0.2826*** | |||||
| (0.0118) | ||||||
| Post 46 indicator | 0.2991*** | |||||
| (0.0118) | ||||||
| Post 47 indicator | 0.1624*** | |||||
| (0.0118) | ||||||
| Post 48 indicator | -0.0569*** | |||||
| (0.0110) | ||||||
| Post 49 indicator | 0.1643*** | |||||
| (0.0121) | ||||||
| Post 50 indicator | 0.3306*** | |||||
| (0.0118) | ||||||
| Post 51 indicator | 0.4009*** | |||||
| (0.0113) | ||||||
| Post 52 indicator | -0.0369*** | |||||
| (0.0113) | ||||||
| Post 53 indicator | -0.1089*** | |||||
| (0.0115) | ||||||
| Post 54 indicator | 0.3643*** | |||||
| (0.0116) | ||||||
| Post 55 indicator | 0.2405*** | |||||
| (0.0118) | ||||||
| Post 56 indicator | 0.1781*** | |||||
| (0.0122) | ||||||
| Post 57 indicator | 0.1780*** | |||||
| (0.0115) | ||||||
| Post 58 indicator | 0.1102*** | |||||
| (0.0117) | ||||||
| Post 59 indicator | 0.0281*** | |||||
| (0.0116) | ||||||
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2816 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2811 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates + Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0140) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2817 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2811 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.6908*** | |||||
| (0.0058) | ||||||
| No-course Baseline | -0.0013 | |||||
| (0.0084) | ||||||
| Emotions | 0.0014 | |||||
| (0.0081) | ||||||
| Reasoning | 0.0119* | |||||
| (0.0083) | ||||||
| Combo | -0.0031 | |||||
| (0.0082) | ||||||
| 1[Post-treatment] | -0.0758*** | -0.0758*** | -0.0748*** | -0.0758*** | -0.0742*** | -0.0742*** |
| (0.0082) | (0.0089) | (0.0087) | (0.0089) | (0.0086) | (0.0086) | |
| 1[Misinfo] | -0.1442*** | -0.1442*** | -0.1459*** | -0.1173*** | -0.1537*** | -0.4320*** |
| (0.0071) | (0.0069) | (0.0067) | (0.0095) | (0.0127) | (0.0123) | |
| Age | -0.0058** | |||||
| (0.0026) | ||||||
| Man | 0.0028 | |||||
| (0.0024) | ||||||
| Some college | -0.0018 | |||||
| (0.0030) | ||||||
| Bachelor’s degree | 0.0034 | |||||
| (0.0030) | ||||||
| Graduate degree | -0.0079*** | |||||
| (0.0026) | ||||||
| Married | -0.0026 | |||||
| (0.0027) | ||||||
| Unemployed | -0.0045* | |||||
| (0.0031) | ||||||
| Employed | -0.0085*** | |||||
| (0.0034) | ||||||
| Mostly urban | -0.0054** | |||||
| (0.0028) | ||||||
| Suburban | -0.0051** | |||||
| (0.0028) | ||||||
| Christian | 0.0031 | |||||
| (0.0025) | ||||||
| Attends religious services | 0.0007 | |||||
| (0.0024) | ||||||
| Uses social media | -0.0060*** | |||||
| (0.0022) | ||||||
| Hours spent on Social Media | 0.0053** | |||||
| (0.0026) | ||||||
| Social Media Share: 80-100% | 0.0130*** | |||||
| (0.0030) | ||||||
| Social Media Share: 60-80% | 0.0180*** | |||||
| (0.0033) | ||||||
| Social Media Share: 40-60% | 0.0150*** | |||||
| (0.0034) | ||||||
| Social Media Share: 20-40% | 0.0066** | |||||
| (0.0031) | ||||||
| Passed attention check at pre-treatment | -0.0046** | |||||
| (0.0024) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.0849 | |||||
| (804.3693) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.1001*** | |||||
| (0.0030) | ||||||
| Emotions posts | 0.0398 | |||||
| (328.7053) | ||||||
| Reasoning posts | 0.0326 | |||||
| (325.5864) | ||||||
| Combo posts | 0.0378 | |||||
| (332.2469) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.0514*** | 0.0514*** | 0.0513*** | 0.0514*** | 0.0507*** | 0.0507*** |
| (0.0118) | (0.0126) | (0.0124) | (0.0126) | (0.0123) | (0.0123) | |
| Emotions x 1[Post-treatment] | -0.0938*** | -0.0938*** | -0.0949*** | -0.0938*** | -0.0956*** | -0.0956*** |
| (0.0115) | (0.0132) | (0.0129) | (0.0132) | (0.0128) | (0.0128) | |
| Reasoning x 1[Post-treatment] | -0.0777*** | -0.0777*** | -0.0785*** | -0.0777*** | -0.0783*** | -0.0783*** |
| (0.0117) | (0.0130) | (0.0128) | (0.0130) | (0.0127) | (0.0127) | |
| Combo x 1[Post-treatment] | -0.0884*** | -0.0884*** | -0.0882*** | -0.0884*** | -0.0883*** | -0.0883*** |
| (0.0116) | (0.0131) | (0.0128) | (0.0131) | (0.0127) | (0.0127) | |
| No-course Baseline x 1[Misinfo] | 0.0051 | 0.0051 | 0.0048 | 0.0051 | 0.0044 | 0.0044 |
| (0.0102) | (0.0101) | (0.0098) | (0.0101) | (0.0097) | (0.0097) | |
| Emotions x 1[Misinfo] | -0.0039 | -0.0039 | -0.0016 | -0.0039 | -0.0024 | -0.0024 |
| (0.0099) | (0.0100) | (0.0097) | (0.0100) | (0.0096) | (0.0096) | |
| Reasoning x 1[Misinfo] | -0.0196** | -0.0196** | -0.0143* | -0.0196** | -0.0153* | -0.0153* |
| (0.0101) | (0.0101) | (0.0097) | (0.0101) | (0.0096) | (0.0096) | |
| Combo x 1[Misinfo] | 0.0040 | 0.0040 | 0.0064 | 0.0040 | 0.0070 | 0.0070 |
| (0.0101) | (0.0099) | (0.0097) | (0.0099) | (0.0096) | (0.0096) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0021 | 0.0021 | 0.0023 | 0.0021 | 0.0039 | 0.0039 |
| (0.0100) | (0.0095) | (0.0090) | (0.0095) | (0.0089) | (0.0089) | |
| No-course Baseline x Age | -0.0006 | |||||
| (0.0037) | ||||||
| Emotions x Age | -0.0039 | |||||
| (0.0039) | ||||||
| Reasoning x Age | -0.0017 | |||||
| (0.0039) | ||||||
| Combo x Age | -0.0011 | |||||
| (0.0038) | ||||||
| No-course Baseline x Man | -0.0074** | |||||
| (0.0035) | ||||||
| Emotions x Man | 0.0059** | |||||
| (0.0034) | ||||||
| Reasoning x Man | -0.0066** | |||||
| (0.0034) | ||||||
| Combo x Man | 0.0034 | |||||
| (0.0034) | ||||||
| No-course Baseline x Some college | -0.0015 | |||||
| (0.0042) | ||||||
| Emotions x Some college | 0.0002 | |||||
| (0.0041) | ||||||
| Reasoning x Some college | -0.0013 | |||||
| (0.0042) | ||||||
| Combo x Some college | -0.0037 | |||||
| (0.0042) | ||||||
| No-course Baseline x Bachelor’s degree | -0.0048 | |||||
| (0.0043) | ||||||
| Emotions x Bachelor’s degree | -0.0030 | |||||
| (0.0042) | ||||||
| Reasoning x Bachelor’s degree | -0.0074** | |||||
| (0.0042) | ||||||
| Combo x Bachelor’s degree | -0.0102*** | |||||
| (0.0042) | ||||||
| No-course Baseline x Graduate degree | 0.0026 | |||||
| (0.0038) | ||||||
| Emotions x Graduate degree | 0.0094*** | |||||
| (0.0034) | ||||||
| Reasoning x Graduate degree | 0.0020 | |||||
| (0.0039) | ||||||
| Combo x Graduate degree | 0.0088*** | |||||
| (0.0034) | ||||||
| No-course Baseline x Married | 0.0028 | |||||
| (0.0039) | ||||||
| Emotions x Married | -0.0052* | |||||
| (0.0038) | ||||||
| Reasoning x Married | 0.0014 | |||||
| (0.0039) | ||||||
| Combo x Married | -0.0042 | |||||
| (0.0039) | ||||||
| No-course Baseline x Unemployed | 0.0024 | |||||
| (0.0044) | ||||||
| Emotions x Unemployed | -0.0019 | |||||
| (0.0043) | ||||||
| Reasoning x Unemployed | 0.0047 | |||||
| (0.0044) | ||||||
| Combo x Unemployed | 0.0010 | |||||
| (0.0044) | ||||||
| No-course Baseline x Employed | 0.0005 | |||||
| (0.0049) | ||||||
| Emotions x Employed | 0.0020 | |||||
| (0.0047) | ||||||
| Reasoning x Employed | 0.0031 | |||||
| (0.0049) | ||||||
| Combo x Employed | 0.0020 | |||||
| (0.0048) | ||||||
| No-course Baseline x Mostly urban | 0.0078** | |||||
| (0.0041) | ||||||
| Emotions x Mostly urban | 0.0077** | |||||
| (0.0040) | ||||||
| Reasoning x Mostly urban | -0.0012 | |||||
| (0.0041) | ||||||
| Combo x Mostly urban | 0.0096*** | |||||
| (0.0040) | ||||||
| No-course Baseline x Suburban | 0.0036 | |||||
| (0.0040) | ||||||
| Emotions x Suburban | -0.0014 | |||||
| (0.0039) | ||||||
| Reasoning x Suburban | -0.0030 | |||||
| (0.0040) | ||||||
| Combo x Suburban | 0.0089** | |||||
| (0.0040) | ||||||
| No-course Baseline x Christian | -0.0076** | |||||
| (0.0034) | ||||||
| Emotions x Christian | -0.0036 | |||||
| (0.0035) | ||||||
| Reasoning x Christian | 0.0028 | |||||
| (0.0036) | ||||||
| Combo x Christian | -0.0039 | |||||
| (0.0035) | ||||||
| No-course Baseline x Attends religious services | -0.0011 | |||||
| (0.0033) | ||||||
| Emotions x Attends religious services | -0.0064** | |||||
| (0.0035) | ||||||
| Reasoning x Attends religious services | -0.0032 | |||||
| (0.0036) | ||||||
| Combo x Attends religious services | -0.0034 | |||||
| (0.0035) | ||||||
| No-course Baseline x Uses social media | 0.0058** | |||||
| (0.0032) | ||||||
| Emotions x Uses social media | 0.0086*** | |||||
| (0.0036) | ||||||
| Reasoning x Uses social media | 0.0077*** | |||||
| (0.0032) | ||||||
| Combo x Uses social media | 0.0045 | |||||
| (0.0037) | ||||||
| No-course Baseline x Hours spent on Social Media | -0.0070** | |||||
| (0.0036) | ||||||
| Emotions x Hours spent on Social Media | -0.0022 | |||||
| (0.0036) | ||||||
| Reasoning x Hours spent on Social Media | -0.0002 | |||||
| (0.0037) | ||||||
| Combo x Hours spent on Social Media | -0.0102*** | |||||
| (0.0036) | ||||||
| No-course Baseline x Social Media Share: 80-100% | 0.0001 | |||||
| (0.0042) | ||||||
| Emotions x Social Media Share: 80-100% | 0.0010 | |||||
| (0.0042) | ||||||
| Reasoning x Social Media Share: 80-100% | 0.0017 | |||||
| (0.0042) | ||||||
| Combo x Social Media Share: 80-100% | 0.0007 | |||||
| (0.0042) | ||||||
| No-course Baseline x Social Media Share: 60-80% | -0.0036 | |||||
| (0.0049) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0049 | |||||
| (0.0047) | ||||||
| Reasoning x Social Media Share: 60-80% | -0.0007 | |||||
| (0.0049) | ||||||
| Combo x Social Media Share: 60-80% | 0.0003 | |||||
| (0.0048) | ||||||
| No-course Baseline x Social Media Share: 40-60% | -0.0038 | |||||
| (0.0050) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0020 | |||||
| (0.0048) | ||||||
| Reasoning x Social Media Share: 40-60% | -0.0051 | |||||
| (0.0050) | ||||||
| Combo x Social Media Share: 40-60% | -0.0036 | |||||
| (0.0049) | ||||||
| No-course Baseline x Social Media Share: 20-40% | -0.0008 | |||||
| (0.0045) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0041 | |||||
| (0.0044) | ||||||
| Reasoning x Social Media Share: 20-40% | -0.0035 | |||||
| (0.0045) | ||||||
| Combo x Social Media Share: 20-40% | -0.0042 | |||||
| (0.0045) | ||||||
| No-course Baseline x Passed attention check at pre-treatment | -0.0056* | |||||
| (0.0035) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0011 | |||||
| (0.0034) | ||||||
| Reasoning x Passed attention check at pre-treatment | 0.0012 | |||||
| (0.0034) | ||||||
| Combo x Passed attention check at pre-treatment | 0.0074** | |||||
| (0.0034) | ||||||
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | -0.0003 | |||||
| (0.0105) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.0363*** | |||||
| (0.0103) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | 0.0070 | |||||
| (0.0104) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.0059 | |||||
| (0.0104) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0024 | |||||
| (0.0044) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0128*** | |||||
| (0.0042) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0109*** | |||||
| (0.0043) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0078** | |||||
| (0.0043) | ||||||
| No-course Baseline x Emotions posts | 0.0011 | |||||
| (0.0068) | ||||||
| Emotions x Emotions posts | 0.0052 | |||||
| (0.0067) | ||||||
| Reasoning x Emotions posts | -0.0040 | |||||
| (0.0068) | ||||||
| Combo x Emotions posts | -0.0075 | |||||
| (0.0068) | ||||||
| No-course Baseline x Reasoning posts | 0.0023 | |||||
| (0.0067) | ||||||
| Emotions x Reasoning posts | 0.0114** | |||||
| (0.0066) | ||||||
| Reasoning x Reasoning posts | -0.0096* | |||||
| (0.0066) | ||||||
| Combo x Reasoning posts | 0.0049 | |||||
| (0.0067) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0033 | 0.0033 | 0.0046 | 0.0033 | 0.0042 | 0.0042 |
| (0.0145) | (0.0140) | (0.0133) | (0.0140) | (0.0131) | (0.0131) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0202* | -0.0202* | -0.0220** | -0.0202* | -0.0246** | -0.0246** |
| (0.0140) | (0.0136) | (0.0129) | (0.0136) | (0.0127) | (0.0127) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0154 | 0.0154 | 0.0129 | 0.0154 | 0.0107 | 0.0107 |
| (0.0143) | (0.0139) | (0.0132) | (0.0139) | (0.0130) | (0.0130) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0127 | -0.0127 | -0.0142 | -0.0127 | -0.0194* | -0.0194* |
| (0.0142) | (0.0138) | (0.0132) | (0.0138) | (0.0130) | (0.0130) | |
| Post domain 2 | -0.2001*** | -0.1794*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 3 | -0.2813*** | -0.3043*** | ||||
| (0.0055) | (0.0095) | |||||
| Post domain 4 | -0.1804*** | -0.2332*** | ||||
| (0.0056) | (0.0097) | |||||
| Post domain 5 | 0.0545*** | -0.0010 | ||||
| (0.0051) | (0.0087) | |||||
| Post domain 6 | 0.0164*** | 0.0389*** | ||||
| (0.0054) | (0.0084) | |||||
| Post domain 7 | -0.2347*** | -0.1827*** | ||||
| (0.0056) | (0.0094) | |||||
| Post domain 8 | -0.2195*** | -0.1160*** | ||||
| (0.0057) | (0.0095) | |||||
| Post domain 9 | 0.0369*** | -0.0608*** | ||||
| (0.0053) | (0.0091) | |||||
| Post domain 10 | -0.0565*** | -0.0963*** | ||||
| (0.0054) | (0.0091) | |||||
| Post domain 11 | -0.1586*** | -0.1739*** | ||||
| (0.0057) | (0.0094) | |||||
| Post domain 12 | -0.1047*** | -0.1343*** | ||||
| (0.0054) | (0.0092) | |||||
| Post domain 13 | -0.1541*** | -0.2155*** | ||||
| (0.0055) | (0.0093) | |||||
| Post domain 14 | -0.2759*** | -0.3151*** | ||||
| (0.0057) | (0.0097) | |||||
| Post domain 15 | -0.1868*** | -0.2755*** | ||||
| (0.0056) | (0.0095) | |||||
| 1[Combo manipulation tactic] | -0.0512*** | -0.0275*** | ||||
| (0.0041) | (0.0078) | |||||
| 1[Emotions manipulation tactic] | -0.0154*** | -0.0585*** | ||||
| (0.0041) | (0.0080) | |||||
| 1[Reasoning manipulation tactic] | -0.0143*** | 0.0208*** | ||||
| (0.0041) | (0.0079) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.0536*** | |||||
| (0.0158) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.0462*** | |||||
| (0.0152) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.0789*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.0971*** | |||||
| (0.0149) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -0.3110*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0471*** | |||||
| (0.0153) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | -0.0192 | |||||
| (0.0155) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 0.1469*** | |||||
| (0.0150) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.0318** | |||||
| (0.0154) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.0595*** | |||||
| (0.0153) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0230* | |||||
| (0.0156) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.0600*** | |||||
| (0.0154) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.0296** | |||||
| (0.0153) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.0822*** | |||||
| (0.0153) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.0845*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.1118*** | |||||
| (0.0154) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.0617*** | |||||
| (0.0153) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.1184*** | |||||
| (0.0148) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.1477*** | |||||
| (0.0147) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0324** | |||||
| (0.0151) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.1539*** | |||||
| (0.0153) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 0.1660*** | |||||
| (0.0153) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.1090*** | |||||
| (0.0152) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0397*** | |||||
| (0.0154) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 0.1449*** | |||||
| (0.0153) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 0.1875*** | |||||
| (0.0151) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.1025*** | |||||
| (0.0154) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 0.3384*** | |||||
| (0.0155) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.0427*** | |||||
| (0.0155) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.0517*** | |||||
| (0.0150) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.0985*** | |||||
| (0.0155) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.0326** | |||||
| (0.0150) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.0629*** | |||||
| (0.0144) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.1533*** | |||||
| (0.0149) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -0.2920*** | |||||
| (0.0151) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.1260*** | |||||
| (0.0151) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.0377*** | |||||
| (0.0151) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.0529*** | |||||
| (0.0155) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.0132 | |||||
| (0.0151) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.0267** | |||||
| (0.0153) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.0441*** | |||||
| (0.0152) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | -0.0236* | |||||
| (0.0151) | ||||||
| Post 2 indicator | -0.1794*** | |||||
| (0.0095) | ||||||
| Post 3 indicator | -0.3043*** | |||||
| (0.0095) | ||||||
| Post 4 indicator | -0.2332*** | |||||
| (0.0097) | ||||||
| Post 5 indicator | -0.0010 | |||||
| (0.0087) | ||||||
| Post 6 indicator | 0.0389*** | |||||
| (0.0084) | ||||||
| Post 7 indicator | -0.1827*** | |||||
| (0.0094) | ||||||
| Post 8 indicator | -0.1160*** | |||||
| (0.0095) | ||||||
| Post 9 indicator | -0.0608*** | |||||
| (0.0091) | ||||||
| Post 10 indicator | -0.0963*** | |||||
| (0.0091) | ||||||
| Post 11 indicator | -0.1739*** | |||||
| (0.0094) | ||||||
| Post 12 indicator | -0.1343*** | |||||
| (0.0092) | ||||||
| Post 13 indicator | -0.2155*** | |||||
| (0.0093) | ||||||
| Post 14 indicator | -0.3151*** | |||||
| (0.0097) | ||||||
| Post 15 indicator | -0.2755*** | |||||
| (0.0095) | ||||||
| Post 16 indicator | 0.2508*** | |||||
| (0.0118) | ||||||
| Post 17 indicator | 0.0178* | |||||
| (0.0119) | ||||||
| Post 18 indicator | -0.0072 | |||||
| (0.0113) | ||||||
| Post 19 indicator | 0.0965*** | |||||
| (0.0114) | ||||||
| Post 20 indicator | 0.3469*** | |||||
| (0.0115) | ||||||
| Post 21 indicator | -0.0213** | |||||
| (0.0113) | ||||||
| Post 22 indicator | 0.0211** | |||||
| (0.0114) | ||||||
| Post 23 indicator | 0.1157*** | |||||
| (0.0118) | ||||||
| Post 24 indicator | 0.3370*** | |||||
| (0.0116) | ||||||
| Post 25 indicator | 0.1863*** | |||||
| (0.0118) | ||||||
| Post 26 indicator | 0.1364*** | |||||
| (0.0120) | ||||||
| Post 27 indicator | 0.0935*** | |||||
| (0.0116) | ||||||
| Post 28 indicator | 0.0953*** | |||||
| (0.0117) | ||||||
| Post 29 indicator | -0.0346*** | |||||
| (0.0113) | ||||||
| Post 30 indicator | 0.0576*** | |||||
| (0.0116) | ||||||
| Post 31 indicator | 0.2198*** | |||||
| (0.0117) | ||||||
| Post 32 indicator | -0.0441*** | |||||
| (0.0115) | ||||||
| Post 33 indicator | 0.0273*** | |||||
| (0.0113) | ||||||
| Post 34 indicator | 0.0482*** | |||||
| (0.0114) | ||||||
| Post 35 indicator | 0.3371*** | |||||
| (0.0115) | ||||||
| Post 36 indicator | 0.4064*** | |||||
| (0.0116) | ||||||
| Post 37 indicator | 0.0046 | |||||
| (0.0112) | ||||||
| Post 38 indicator | -0.0501*** | |||||
| (0.0112) | ||||||
| Post 39 indicator | 0.3250*** | |||||
| (0.0117) | ||||||
| Post 40 indicator | 0.2324*** | |||||
| (0.0118) | ||||||
| Post 41 indicator | 0.0062 | |||||
| (0.0116) | ||||||
| Post 42 indicator | 0.2304*** | |||||
| (0.0118) | ||||||
| Post 43 indicator | 0.1917*** | |||||
| (0.0115) | ||||||
| Post 44 indicator | 0.0072 | |||||
| (0.0114) | ||||||
| Post 45 indicator | 0.2826*** | |||||
| (0.0118) | ||||||
| Post 46 indicator | 0.2991*** | |||||
| (0.0118) | ||||||
| Post 47 indicator | 0.1624*** | |||||
| (0.0118) | ||||||
| Post 48 indicator | -0.0569*** | |||||
| (0.0110) | ||||||
| Post 49 indicator | 0.1643*** | |||||
| (0.0121) | ||||||
| Post 50 indicator | 0.3306*** | |||||
| (0.0118) | ||||||
| Post 51 indicator | 0.4009*** | |||||
| (0.0113) | ||||||
| Post 52 indicator | -0.0369*** | |||||
| (0.0113) | ||||||
| Post 53 indicator | -0.1089*** | |||||
| (0.0115) | ||||||
| Post 54 indicator | 0.3643*** | |||||
| (0.0116) | ||||||
| Post 55 indicator | 0.2405*** | |||||
| (0.0118) | ||||||
| Post 56 indicator | 0.1781*** | |||||
| (0.0122) | ||||||
| Post 57 indicator | 0.1780*** | |||||
| (0.0115) | ||||||
| Post 58 indicator | 0.1102*** | |||||
| (0.0117) | ||||||
| Post 59 indicator | 0.0281*** | |||||
| (0.0116) | ||||||
| Num. obs. | 156312 | 156312 | 156312 | 156312 | 156312 | 156312 |
| R2 (full model) | 0.2817 | 0.3470 | 0.3953 | 0.3478 | 0.4153 | 0.4153 |
| R2 (proj model) | 0.0545 | 0.1244 | 0.0556 | 0.1534 | 0.1534 | |
| Adj. R2 (full model) | 0.2811 | 0.3085 | 0.3596 | 0.3094 | 0.3806 | 0.3806 |
| Adj. R2 (proj model) | 0.0544 | 0.1242 | 0.0555 | 0.1530 | 0.1530 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Prediction
Standard errors of the share predicted value are estimated using the Delta method. Treatment effects standard errors are clustered at the user level and reported in parentheses. t-ratios are reported in brackets. Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
cov_feglm_list <- list()
# covariate_list <- c(demographics_demeaned_treatment,
# sharing_demeaned_treatment,
# sharing_tactic_demeaned_treatment,
# demographics_sharing_demeaned_treatment,
# all_covariates_demeaned_treatment)
#
# cov_estimates_list <- list()
for (cov_list in covariate_list) {
cov_feglm_list[[length(cov_feglm_list) + 1]] <- c(paste0("share ~ treatment"),
paste0("share ~ treatment + ",cov_list),
paste0("share ~ treatment + factor(fact) + ",cov_list),
paste0("share ~ treatment + factor(type) + ",cov_list),
paste0("share ~ treatment + factor(fact)*factor(type) + ",cov_list),
paste0("share ~ treatment + factor(post_id) + ",cov_list))
}
if (file.exists("prediction_alltreat_df.rds")) {
prediction_alltreat_df <- readRDS("prediction_alltreat_df.rds")
counter <- 1
for (cov in seq_along(covariate_list)) {
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
k <- knitr::kable(prediction_alltreat_df[[counter]], caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","Treatment Effect<br>Emotions vs. Facts"))
print(k)
cat('<br>')
counter <- counter + 1
}
cat('<br>\n<br>\n<br>')
}
} else {
prediction_alltreat_df <- list()
for (cov in seq_along(covariate_list)) {
list_models <- cov_estimates_list[[cov]]
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
treatmentLevels <- unique(panel$treatment)
posts <- c(0, 1)
misinfos <- c(0, 1)
prediction_df <- expand.grid(treatment = treatmentLevels, post = posts, post_is_misinfo = misinfos) %>%
mutate(prediction = NA_real_, prediction_se = NA_real_)
for (row in 1:nrow(prediction_df)) {
treat <- as.character(prediction_df$treatment[row])
post <- prediction_df$post[row]
misinfo <- prediction_df$post_is_misinfo[row]
# Filter the relevant data
relevant_data <- panel %>%
filter(row_number() %in% obs(list_models[[m]])) %>%
filter(treatment == treat & post == post & post_is_misinfo == misinfo)
# Prepare predictor variables based on model requirements
predictorVariables <- relevant_data %>%
distinct(user, .keep_all = TRUE) %>%
group_by(user, type, fact, post_id) %>%
summarise(age_cov = mean(age_cov, na.rm = TRUE),
gender_Man_cov = mean(gender_Man_cov, na.rm = TRUE),
education_Some_college_cov = mean(education_Some_college_cov, na.rm = TRUE),
education_Bachelor_degree_cov = mean(education_Bachelor_degree_cov, na.rm = TRUE),
education_Graduate_degree_cov = mean(education_Graduate_degree_cov, na.rm = TRUE),
marital_Married_or_in_a_domestic_partnership_cov = mean(marital_Married_or_in_a_domestic_partnership_cov, na.rm = TRUE),
employment_Unemployed_cov = mean(employment_Unemployed_cov, na.rm = TRUE),
employment_Employed_cov = mean(employment_Employed_cov, na.rm = TRUE),
location_Mostly_urban_cov = mean(location_Mostly_urban_cov, na.rm = TRUE),
location_Suburban_cov = mean(location_Suburban_cov, na.rm = TRUE),
religion_Christian_cov = mean(religion_Christian_cov, na.rm = TRUE),
religiosity_Attends_cov = mean(religiosity_Attends_cov, na.rm = TRUE),
social_media_bin_Yes_cov = mean(social_media_bin_Yes_cov, na.rm = TRUE),
social_media_hours_cov = mean(social_media_hours_cov, na.rm = TRUE),
att_check_pre_cov = mean(att_check_pre_cov, na.rm = TRUE),
social_media_share_80_100_cov = mean(social_media_share_80_100_cov, na.rm = TRUE),
social_media_share_60_80_cov = mean(social_media_share_60_80_cov, na.rm = TRUE),
social_media_share_40_60_cov = mean(social_media_share_40_60_cov, na.rm = TRUE),
social_media_share_20_40_cov = mean(social_media_share_20_40_cov, na.rm = TRUE),
misinfo_pre_cov = mean(misinfo_pre_cov, na.rm = TRUE),
base_rate_pre_cov = mean(base_rate_pre_cov, na.rm = TRUE),
emotions_pre_cov = mean(emotions_pre_cov, na.rm = TRUE),
reasoning_pre_cov = mean(reasoning_pre_cov, na.rm = TRUE),
combo_pre_cov = mean(combo_pre_cov, na.rm = TRUE)) %>% # now add the treatment and post variables and misinfo
mutate(treatment = treat,
post = post,
post_is_misinfo = misinfo)
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[m], newdata = predictorVariables)
predictorVariables$prediction <- as.numeric(unlist(predicted_shares))
# mean of predicted value
prediction_df$prediction[row] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[row] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
# predicted value
if (m == 1) {
predictorVariables <- expand.grid(treatment = unique(panel$treatment),
post = c(0,1),
post_is_misinfo = c(0,1))
predictions <- predict(baseline_all_lm, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_all_lm, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res)
contrast_df <- contrast_df %>% filter(contrast == "Emotions - Facts Baseline")
contrast_df <- contrast_df %>% select(post, post_is_misinfo, estimate, SE, t.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(t.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
# prediction_df <- prediction_df %>%
# group_by(post,post_is_misinfo) %>%
# mutate(lag = ifelse(treatment == "Facts Baseline",
# prediction, NA)) %>%
# tidyr::fill(lag) %>%
# ungroup() %>%
# mutate(te = ifelse(treatment != "Facts Baseline",
# format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
# select(-lag) %>%
# mutate(te = ifelse(is.na(te),"",te))
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
cov_model <- feglm(as.formula(cov_feglm_list[[cov]][m]),panel %>%
filter(post == p & post_is_misinfo == mis), cluster = "user")
# obtain data frame with estimate, std error, and z value
cov_model_df_add <- data.frame(
treatment = c("No-course Baseline","Emotions","Reasoning","Combo"),
estimate = format(coef(cov_model)[grepl("^treatment",names(coef(cov_model))) & !grepl("\\:",names(coef(cov_model)))], digits = 4),
SE = sqrt(diag(vcov(cov_model)))[grepl("^treatment",names(diag(vcov(cov_model)))) & !grepl("\\:",names(diag(vcov(cov_model))))],
z.ratio = coef(cov_model)[grepl("^treatment",names(coef(cov_model))) & !grepl("\\:",names(coef(cov_model)))] / sqrt(diag(vcov(cov_model)))[grepl("^treatment",names(diag(vcov(cov_model)))) & !grepl("\\:",names(diag(vcov(cov_model))))],
post = p,
post_is_misinfo = mis
)
# make rownames a column
cov_model_df_add <- rownames_to_column(cov_model_df_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>% select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
cov_model_df_add <- rbind(cov_model_df_add %>% mutate(treatment = "Facts Baseline"),
cov_model_df_add %>% mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
contrast_df <- rbind(contrast_df, cov_model_df_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","Treatment effect<br>Emotions vs. Facts"))
print(k)
cat('<br>')
prediction_alltreat_df[[length(prediction_alltreat_df) + 1]] <- prediction_df
}
cat('<br>\n<br>\n<br>')
}
saveRDS(prediction_alltreat_df, "prediction_alltreat_df.rds")
}\[\text{Demographics}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.005948 | |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | 0.0059 | (0.01144); [0.5198] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.008371) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.008294) | ||
| Combo | 0 | 0 | 0.6864 | (0.008359) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.002028 | |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | 0.0021 | (0.01097); [0.1848] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.008148) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.008065) | ||
| Combo | 0 | 1 | 0.5462 | (0.007968) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.087832 | |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | -0.0879 | (0.01252); [ -7.0158] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.008775) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.009097) | ||
| Combo | 1 | 0 | 0.5222 | (0.009096) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.111959 | |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | -0.1120 | (0.01114); [-10.0482] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.008271) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.008156) | ||
| Combo | 1 | 1 | 0.3715 | (0.008052) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.0002093 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.01107); [0.01891] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.003628 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.01027); [-0.3531] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.09335 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.01211); [-7.706] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11737 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.01061); [-11.06] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0007917 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.01093); [0.07244] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | -0.001089 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.01014); [-0.1074] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.09431 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.01185); [-7.958] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.11893 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.01054); [-11.28] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0007917 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.01093); [0.07244] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | -0.001545 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.01004); [-0.1539] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.09431 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.01185); [-7.958] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.12131 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.01046); [-11.6] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.005948 | |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | 0.0059 | (0.01144); [0.5198] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.008371) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.008294) | ||
| Combo | 0 | 0 | 0.6864 | (0.008359) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.002028 | |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | 0.0021 | (0.01097); [0.1848] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.008148) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.008065) | ||
| Combo | 0 | 1 | 0.5462 | (0.007968) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.087832 | |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | -0.0879 | (0.01252); [ -7.0158] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.008775) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.009097) | ||
| Combo | 1 | 0 | 0.5222 | (0.009096) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.111959 | |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | -0.1120 | (0.01114); [-10.0482] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.008271) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.008156) | ||
| Combo | 1 | 1 | 0.3715 | (0.008052) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00000000000001223 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.000000000000007451); [1.641] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.000000000000008436 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.0000000000000137); [0.6158] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.09122 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.01116); [-8.172] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11537 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.00941); [-12.26] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0005929 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.001209); [0.4903] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0022597 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.001151); [1.964] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.09233 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.01082); [-8.531] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.11698 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.00929); [-12.59] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0005929 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.001209); [0.4903] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0018000 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.001422); [1.266] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.09233 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.01082); [-8.531] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.11951 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.009207); [-12.98] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Tactic sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.005948 | |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | 0.0059 | (0.01144); [0.5198] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.008371) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.008294) | ||
| Combo | 0 | 0 | 0.6864 | (0.008359) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.002028 | |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | 0.0021 | (0.01097); [0.1848] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.008148) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.008065) | ||
| Combo | 0 | 1 | 0.5462 | (0.007968) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.087832 | |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | -0.0879 | (0.01252); [ -7.0158] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.008775) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.009097) | ||
| Combo | 1 | 0 | 0.5222 | (0.009096) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.111959 | |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | -0.1120 | (0.01114); [-10.0482] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.008271) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.008156) | ||
| Combo | 1 | 1 | 0.3715 | (0.008052) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00000000000001265 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.000000000000008011); [1.579] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.000000000000007078 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.00005619); [0.000000000126] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.09208 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.01114); [-8.264] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11568 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.009407); [-12.3] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0005690 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.001213); [0.4692] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0023335 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.001152); [2.025] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.09317 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.01081); [-8.621] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.11731 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.009289); [-12.63] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0005690 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.001213); [0.4692] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0019440 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.001423); [1.366] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.09317 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.01081); [-8.621] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.11988 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.009208); [-13.02] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Demographics +
Sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.005948 | |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | 0.0059 | (0.01144); [0.5198] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.008371) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.008294) | ||
| Combo | 0 | 0 | 0.6864 | (0.008359) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.002028 | |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | 0.0021 | (0.01097); [0.1848] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.008148) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.008065) | ||
| Combo | 0 | 1 | 0.5462 | (0.007968) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.087832 | |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | -0.0879 | (0.01252); [ -7.0158] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.008775) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.009097) | ||
| Combo | 1 | 0 | 0.5222 | (0.009096) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.111959 | |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | -0.1120 | (0.01114); [-10.0482] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.008271) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.008156) | ||
| Combo | 1 | 1 | 0.3715 | (0.008052) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.00000000000001212 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.000000000000007777); [1.558] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.000000000000008801 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.00000000000001472); [0.598] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.09329 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.01102); [-8.466] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11724 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.009208); [-12.73] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0005882 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.001219); [0.4826] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0021004 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.001151); [1.824] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.09427 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.01068); [-8.83] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.11872 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.009087); [-13.07] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0005882 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.001219); [0.4826] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0015879 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.001417); [1.121] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.09427 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.01068); [-8.83] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.12107 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.008995); [-13.46] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Demographics +
Sharing rates + Tactic sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.008087) | 0.005948 | |
| Emotions | 0 | 0 | 0.6888 | (0.008097) | 0.0059 | (0.01144); [0.5198] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.008371) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.008294) | ||
| Combo | 0 | 0 | 0.6864 | (0.008359) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.007712) | 0.002028 | |
| Emotions | 0 | 1 | 0.5407 | (0.007800) | 0.0021 | (0.01097); [0.1848] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.008148) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.008065) | ||
| Combo | 0 | 1 | 0.5462 | (0.007968) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.008962) | -0.087832 | |
| Emotions | 1 | 0 | 0.5192 | (0.008741) | -0.0879 | (0.01252); [ -7.0158] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.008775) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.009097) | ||
| Combo | 1 | 0 | 0.5222 | (0.009096) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.008314) | -0.111959 | |
| Emotions | 1 | 1 | 0.3530 | (0.007418) | -0.1120 | (0.01114); [-10.0482] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.008271) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.008156) | ||
| Combo | 1 | 1 | 0.3715 | (0.008052) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.000000000000008093 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.000000000000008429); [0.9601] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | 0.00000000000001189 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.0001024); [0.0000000001161] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.09413 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.011); [-8.554] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.11754 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.009207); [-12.77] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0005702 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.001222); [0.4666] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0021652 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.001164); [1.86] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.09509 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.01067); [-8.915] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.11903 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.009087); [-13.1] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
Treatment Effect Emotions vs. Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0005702 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.001222); [0.4666] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0017194 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.001422); [1.209] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.09509 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.01067); [-8.915] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.12142 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.008998); [-13.49] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
Logistic
Baseline estimation
Below, we present the estimates based on 2 treatment arms: Facts Baseline and Emotions.
The sample includes all users who completed the post-survey and belong to the “Facts Baseline” (n = 1,790) or “Emotions course” (n = 1,836) treatment groups (n = 3,626). We consider all 9 posts in the pre-treatment period (“pre”) and all 9 posts in the post-treatment period (i.e., “post”).
# Define panel df with two treatment arms only
panel_2 <- panel %>% filter(treatment %in% c('Facts Baseline','Emotions'))
panel_2$treatment <- factor(panel_2$treatment, levels = c('Facts Baseline','Emotions'))
panel_2$emotions <- ifelse(panel_2$treatment == "Emotions",1,0)# version 1: 2 treatment arms
formula_baseline_logit <- "share ~ treatment*post*post_is_misinfo"
baseline_logit <- glm(as.formula(formula_baseline_logit),panel_2, family = binomial('logit'))
baseline_logit_se <- (feglm(as.formula(formula_baseline_logit),panel_2, family = binomial('logit'), cluster = "user"))$se# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
if (!file.exists("R2oos_baseline_logit.rds")) {
# baseline
formula <- gsub("^share","",formula_baseline_logit)
X <- model.matrix(as.formula(formula), data = panel_2)
X <- X[, !(colnames(X) %in% "(Intercept)")]
# fit and prediction functions
fitFunLogit <- function(y, x){feglm.fit(y = y, x = cbind(1, x), family = binomial(link = "logit"), cluster = "user")}
predFunLogit <- function(mod, x) {cbind(1,x) %*% mod$coef}
R2oos_baseline_logit <- oosse::R2oosse(panel_2$share,
X[,-1],
fitFun = fitFunLogit,
predFun = predFunLogit,
methodMSE = "bootstrap",
nBootstraps = 100,
nBootstrapsCor = 25)
saveRDS(R2oos_baseline_logit,"R2oos_baseline_logit.rds")
} else {
R2oos_baseline_logit <- readRDS("R2oos_baseline_logit.rds")
}
R2out <- as.numeric(R2oos_baseline_logit$R2[1])# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_logit)
names <- unique(c(names(baseline_logit$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# out-of-sample R2
if (any(ls() %in% "R2out")) {
R2out <- c(R2out)
} else {
R2out <- rep("",length(model_list))
}
# s.e.
standard_errors <- list(baseline_logit_se)
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test. MSE for out-of-sample R$^2$ is calculated through a cross-validation process with 100 repetitions and 10 folds.",
caption = paste("Logit Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out)
),
reorder.gof = c(6,7,5,8,9,10,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 |
| (0.0582) | |
| Emotions x 1[Post-treatment] | -0.3856*** |
| (0.0575) | |
| Emotions x 1[Misinfo] | -0.0194 |
| (0.0441) | |
| AIC | 87543.9346 |
| BIC | 87616.6247 |
| Out-of-sample R2 | -1.0971 |
| Log Likelihood | -43763.9673 |
| Deviance | 87527.9346 |
| Num. obs. | 65268 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. MSE for out-of-sample R\(^2\) is calculated through a cross-validation process with 100 repetitions and 10 folds. | |
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test. MSE for out-of-sample R$^2$ is calculated through a cross-validation process with 10 folds.",
caption = paste("Logit Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out)
),
reorder.gof = c(6,7,5,8,9,10,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| Facts Baseline | 0.7670*** |
| (0.0373) | |
| Emotions | 0.0276 |
| (0.0531) | |
| 1[Post-treatment] | -0.3319*** |
| (0.0391) | |
| 1[Misinfo] | -0.6121*** |
| (0.0305) | |
| Emotions x 1[Post-treatment] | -0.3856*** |
| (0.0575) | |
| Emotions x 1[Misinfo] | -0.0194 |
| (0.0441) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0368 |
| (0.0407) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 |
| (0.0582) | |
| AIC | 87543.9346 |
| BIC | 87616.6247 |
| Out-of-sample R2 | -1.0971 |
| Log Likelihood | -43763.9673 |
| Deviance | 87527.9346 |
| Num. obs. | 65268 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. MSE for out-of-sample R\(^2\) is calculated through a cross-validation process with 10 folds. | |
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method.
Treatment effects in the last column refer to estimated marginal means (EMMs, or least-squares means) for specified factor combinations (Post-treatment and Misinfo). Point estimates are reported in the upper cell and standard errors, clustered at the user level, reported in parentheses in the bottom cell. z-ratios are reported in brackets.
# Table with prediction values
treatmentLevels <- unique(panel_2$treatment)
predictorVariables <- expand.grid(treatment = treatmentLevels,
post = c(0, 1),
post_is_misinfo = c(0, 1))
predictions <- predict(baseline_logit, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(lag) - as.numeric(prediction), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
emm_res <- emmeans(baseline_logit, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% select(post, post_is_misinfo, estimate, SE, z.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
knitr::kable(prediction_df, col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
Click for All treatment arms
# version 2: all treatment arms
formula_baseline_all_logit <- "share ~ treatment*post*post_is_misinfo"
baseline_all_logit <- glm(as.formula(formula_baseline_all_logit),panel,family = binomial('logit'))
baseline_all_logit_se <- (feglm(as.formula(formula_baseline_all_logit),panel,family = binomial('logit'), cluster = "user"))$se# Calculate out-of-sample R2
BiocParallel::register(BiocParallel::MulticoreParam(workers = 8)) # define number of cores
if (!file.exists("R2oos_baseline_all_logit.rds")) {
# baseline all treatments
formula <- gsub("^share","",formula_baseline_all_logit)
X <- model.matrix(as.formula(formula), data = panel)
X <- X[, !(colnames(X) %in% "(Intercept)")]
# fit and prediction functions
fitFunLogit <- function(y, x){feglm.fit(y = y, x = cbind(1, x), family = binomial(link = "logit"), cluster = "user")}
predFunLogit <- function(mod, x) {cbind(1,x) %*% mod$coef}
R2oos_baseline_all_logit <- oosse::R2oosse(panel$share,
X[,-1],
fitFun = fitFunLogit,
predFun = predFunLogit,
methodMSE = "bootstrap",
nBootstraps = 100,
nBootstrapsCor = 25)
saveRDS(R2oos_baseline_all_logit,"R2oos_baseline_all_logit.rds")
} else {
R2oos_baseline_all_logit <- readRDS("R2oos_baseline_all_logit.rds")
}
R2out_all <- as.numeric(R2oos_baseline_all_logit$R2[1])Baseline estimation
The sample includes all users who completed the post-survey (n = 8,684). We consider all 9 posts in the pre-treatment period (“pre”) and all 9 posts in the post-treatment period (i.e., “post”).
# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_all_logit)
names <- unique(c(names(baseline_all_logit$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2out_all")) {
R2out_all <- c(R2out_all)
} else {
R2out_all <- rep("",length(model_list))
}
# standard errors
standard_errors <- list(baseline_all_logit_se)
# keep variables
keep <- list("treatmentFacts Baseline:post:post_is_misinfo" = "Facts Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentReasoning:post:post_is_misinfo" = "Reasoning x 1[Post-treatment] x 1[Misinfo]",
"treatmentCombo:post:post_is_misinfo" = "Combo x 1[Post-treatment] x 1[Misinfo]",
"treatmentFacts Baseline:post" = "Facts Baseline x 1[Post-treatment]",
"treatmentNo-course Baseline" = "No-course Baseline x 1[Post-treatment]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentReasoning:post" = "Reasoning x 1[Post-treatment]",
"treatmentCombo:post" = "Combo x 1[Post-treatment]",
"treatmentFacts Baseline:post_is_misinfo" = "Facts Baseline x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Misinfo]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]",
"treatmentReasoning:post_is_misinfo" = "Reasoning x 1[Misinfo]",
"treatmentCombo:post_is_misinfo" = "Combo x 1[Misinfo]")
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Logit Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out_all)
),
reorder.gof = c(6,7,5,8,9,10,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0066 |
| (0.0452) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0066 |
| (0.0452) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 |
| (0.0582) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0778 |
| (0.0596) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0536 |
| (0.0589) | |
| No-course Baseline x 1[Post-treatment] | 0.1001*** |
| (0.0548) | |
| Emotions x 1[Post-treatment] | -0.3856*** |
| (0.0575) | |
| Reasoning x 1[Post-treatment] | -0.3314*** |
| (0.0572) | |
| Combo x 1[Post-treatment] | -0.3624*** |
| (0.0567) | |
| Emotions x 1[Misinfo] | -0.0194 |
| (0.0441) | |
| Reasoning x 1[Misinfo] | -0.0930** |
| (0.0449) | |
| Combo x 1[Misinfo] | 0.0140 |
| (0.0438) | |
| AIC | 209149.6024 |
| BIC | 209348.7946 |
| Out-of-sample R2 | -0.9923 |
| Log Likelihood | -104554.8012 |
| Deviance | 209109.6024 |
| Num. obs. | 156312 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | |
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("(1)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test. MSE for out-of-sample R$^2$ is calculated through a cross-validation process with 100 repetitions and 10 folds.",
caption = paste("Logit Regression:", headers, "vs. Treatment"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.gof.rows = list("Fixed Effects" = rep("No",length(model_list)),
"Fixed Effects" = rep("No",length(model_list)),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2out_all)
),
reorder.gof = c(6,7,5,8,9,10,1,2,3,4)
)
html_table| Share (binary indicator) | |
|---|---|
| (1) | |
| Facts Baseline | 0.7670*** |
| (0.0373) | |
| No-course Baseline | 0.1001*** |
| (0.0548) | |
| Emotions | 0.0276 |
| (0.0531) | |
| Reasoning | 0.0982** |
| (0.0546) | |
| Combo | 0.0164 |
| (0.0539) | |
| 1[Post-treatment] | -0.3319*** |
| (0.0391) | |
| 1[Misinfo] | -0.6121*** |
| (0.0305) | |
| No-course Baseline x 1[Post-treatment] | 0.2176*** |
| (0.0572) | |
| Emotions x 1[Post-treatment] | -0.3856*** |
| (0.0575) | |
| Reasoning x 1[Post-treatment] | -0.3314*** |
| (0.0572) | |
| Combo x 1[Post-treatment] | -0.3624*** |
| (0.0567) | |
| No-course Baseline x 1[Misinfo] | 0.0066 |
| (0.0452) | |
| Emotions x 1[Misinfo] | -0.0194 |
| (0.0441) | |
| Reasoning x 1[Misinfo] | -0.0930** |
| (0.0449) | |
| Combo x 1[Misinfo] | 0.0140 |
| (0.0438) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0368 |
| (0.0407) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0008 |
| (0.0611) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 |
| (0.0582) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0778 |
| (0.0596) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0536 |
| (0.0589) | |
| AIC | 209149.6024 |
| BIC | 209348.7946 |
| Out-of-sample R2 | -0.9923 |
| Log Likelihood | -104554.8012 |
| Deviance | 209109.6024 |
| Num. obs. | 156312 |
| Fixed Effects | No |
| Fixed Effects | No |
| Covariates | No |
| Clustered Std. Errors | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. MSE for out-of-sample R\(^2\) is calculated through a cross-validation process with 100 repetitions and 10 folds. | |
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method. Treatment effects standard errors are clustered at the user level and reported in parentheses. z-ratios are reported in brackets.
# Table with prediction values
treatmentLevels <- unique(panel$treatment)
predictorVariables <- expand.grid(treatment = treatmentLevels,
post = c(0, 1),
post_is_misinfo = c(0, 1))
predictions <- predict(baseline_all_logit, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(lag) - as.numeric(prediction), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
emm_res <- emmeans(baseline_all_logit, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>%
filter(contrast == "Emotions - Facts Baseline") %>%
rename(treatment = contrast) %>%
select(post, post_is_misinfo, estimate, SE, z.ratio, treatment)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo")) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
knitr::kable(prediction_df, col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] | |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
Fixed effects
# version 1: 2 treatment arms
formula_fe1_logit <- "share ~ treatment*post*post_is_misinfo | user"
fe1_logit <- feglm(as.formula(formula_fe1_logit),panel_2, family = binomial(link = "logit"))
formula_fe2_logit <- "share ~ treatment*post*post_is_misinfo + factor(fact) | user"
fe2_logit <- feglm(as.formula(formula_fe2_logit),panel_2, family = binomial(link = "logit"))
formula_fe3_logit <- "share ~ treatment*post*post_is_misinfo + factor(type) | user"
fe3_logit <- feglm(as.formula(formula_fe3_logit),panel_2, family = binomial(link = "logit"))
formula_fe4_logit <- "share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) | user"
fe4_logit <- feglm(as.formula(formula_fe4_logit),panel_2, family = binomial(link = "logit"))
formula_fe5_logit <- "share ~ treatment*post*post_is_misinfo + factor(post_id) | user"
fe5_logit <- feglm(as.formula(formula_fe5_logit),panel_2, family = binomial(link = "logit"))# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_logit,fe1_logit,fe2_logit,fe3_logit,fe4_logit,fe5_logit)
names <- unique(c(names(baseline_logit$coefficients),
names(fe1_logit$coefficients),
names(fe2_logit$coefficients),
names(fe3_logit$coefficients),
names(fe4_logit$coefficients),
names(fe5_logit$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2oos_fe_2treat_logit_all")) {
if (length(R2oos_fe_2treat_logit_all)>0) {
R2oos_fe_2treat_logit_all <- c(R2out,R2oos_fe_2treat_logit_all, rep("",length(model_list)-length(R2oos_fe_2treat_logit_all)-1))
} else {
R2oos_fe_2treat_logit_all <- rep("",length(model_list))
}
} else {
R2oos_fe_2treat_logit_all <- rep("",length(model_list))
}
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
#na_coeff <- names(model_list[[m]]$coefficients[,1])[is.na(model_list[[m]]$coefficients[,1])] # felm
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# standard errors
standard_errors <- list(baseline_logit_se,fe1_logit$se,fe2_logit$se,fe3_logit$se,fe4_logit$se,fe5_logit$se)
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_fe_2treat_logit_all)),
reorder.gof = c(8,6,7,9,10,11,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0582) | (0.0836) | (0.0884) | (0.0836) | (0.0902) | (0.0902) | |
| Emotions x 1[Post-treatment] | -0.3856*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0575) | (0.0691) | (0.0730) | (0.0691) | (0.0740) | (0.0740) | |
| Emotions x 1[Misinfo] | -0.0194 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0441) | (0.0605) | (0.0638) | (0.0606) | (0.0650) | (0.0650) | |
| Log Likelihood | -43763.9673 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| AIC | 87543.9346 | |||||
| BIC | 87616.6247 | |||||
| Deviance | 87527.9346 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Pseudo R2 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_fe_2treat_logit_all)),
reorder.gof = c(8,6,7,9,10,11,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.7670*** | |||||
| (0.0373) | ||||||
| Emotions | 0.0276 | |||||
| (0.0531) | ||||||
| 1[Post-treatment] | -0.3319*** | -0.5002*** | -0.5601*** | -0.5005*** | -0.5646*** | -0.5646*** |
| (0.0391) | (0.0498) | (0.0527) | (0.0499) | (0.0534) | (0.0534) | |
| 1[Misinfo] | -0.6121*** | -0.9234*** | -1.0451*** | -0.6580*** | -1.1149*** | -3.1708*** |
| (0.0305) | (0.0435) | (0.0461) | (36168.5464) | (0.1288) | (0.1297) | |
| Emotions x 1[Post-treatment] | -0.3856*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0575) | (0.0691) | (0.0730) | (0.0691) | (0.0740) | (0.0740) | |
| Emotions x 1[Misinfo] | -0.0194 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0441) | (0.0605) | (0.0638) | (0.0606) | (0.0650) | (0.0650) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0368 | 0.0492 | 0.0615 | 0.0485 | 0.0524 | 0.0524 |
| (0.0407) | (0.0602) | (0.0637) | (0.0602) | (0.0650) | (0.0650) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0582) | (0.0836) | (0.0884) | (0.0836) | (0.0902) | (0.0902) | |
| Post domain 2 | -1.3255*** | -1.3401*** | ||||
| (0.0567) | (0.1096) | |||||
| Post domain 3 | -1.9197*** | -2.2672*** | ||||
| (0.0580) | (0.1079) | |||||
| Post domain 4 | -1.2883*** | -1.9129*** | ||||
| (0.0566) | (0.1084) | |||||
| Post domain 5 | 0.4122*** | -0.0798 | ||||
| (0.0589) | (0.1184) | |||||
| Post domain 6 | 0.1959*** | 0.5392*** | ||||
| (0.0581) | (0.1265) | |||||
| Post domain 7 | -1.5826*** | -1.5243*** | ||||
| (0.0571) | (0.1092) | |||||
| Post domain 8 | -1.4424*** | -0.9858*** | ||||
| (0.0567) | (0.1100) | |||||
| Post domain 9 | 0.3437*** | -0.5000*** | ||||
| (0.0585) | (0.1136) | |||||
| Post domain 10 | -0.4006*** | -0.8357*** | ||||
| (0.0567) | (0.1115) | |||||
| Post domain 11 | -1.0907*** | -1.4147*** | ||||
| (0.0563) | (0.1090) | |||||
| Post domain 12 | -0.7530*** | -1.1001*** | ||||
| (0.0563) | (0.1106) | |||||
| Post domain 13 | -1.0463*** | -1.6287*** | ||||
| (0.0564) | (0.1089) | |||||
| Post domain 14 | -1.8782*** | -2.3481*** | ||||
| (0.0578) | (0.1093) | |||||
| Post domain 15 | -1.2760*** | -2.1274*** | ||||
| (0.0565) | (0.1096) | |||||
| 1[Combo manipulation tactic] | -0.4163*** | -0.2959** | ||||
| (36168.5464) | (0.1228) | |||||
| 1[Emotions manipulation tactic] | -0.2120*** | -0.6319*** | ||||
| (36168.5464) | (0.1210) | |||||
| 1[Reasoning manipulation tactic] | -0.1699*** | |||||
| (36168.5464) | ||||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.2510* | |||||
| (0.1653) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4889*** | |||||
| (0.1655) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.8316*** | |||||
| (0.1642) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6964*** | |||||
| (0.1714) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.4606*** | |||||
| (0.1794) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0567 | |||||
| (0.1649) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0368 | |||||
| (0.1642) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.2625*** | |||||
| (0.1710) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.2924** | |||||
| (0.1650) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.5696*** | |||||
| (0.1634) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0738 | |||||
| (0.1638) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.4872*** | |||||
| (0.1637) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.2920** | |||||
| (0.1665) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.8359*** | |||||
| (0.1631) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4438*** | |||||
| (0.1679) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 1.0575*** | |||||
| (0.1640) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.7557*** | |||||
| (0.1616) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.9877*** | |||||
| (0.1714) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9932*** | |||||
| (0.1831) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.2112 | |||||
| (0.1643) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.7904*** | |||||
| (0.1681) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2790*** | |||||
| (0.1655) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.9799*** | |||||
| (0.1633) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0184 | |||||
| (0.1637) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1203*** | |||||
| (0.1629) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.5807*** | |||||
| (0.1624) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.9710*** | |||||
| (0.1645) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.6319*** | |||||
| (0.1632) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.3677** | |||||
| (0.1640) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.3345** | |||||
| (0.1711) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.9799*** | |||||
| (0.1638) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.3120** | |||||
| (0.1725) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3864** | |||||
| (0.1830) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.7992*** | |||||
| (0.1690) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.8581*** | |||||
| (0.1750) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.9999*** | |||||
| (0.1714) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4521*** | |||||
| (0.1660) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.6336*** | |||||
| (0.1643) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2927** | |||||
| (0.1653) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3094** | |||||
| (0.1634) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.5483*** | |||||
| (0.1664) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0715 | |||||
| (0.1666) | ||||||
| Post 2 indicator | -1.3401*** | |||||
| (0.1096) | ||||||
| Post 3 indicator | -2.2672*** | |||||
| (0.1079) | ||||||
| Post 4 indicator | -1.9129*** | |||||
| (0.1084) | ||||||
| Post 5 indicator | -0.0798 | |||||
| (0.1184) | ||||||
| Post 6 indicator | 0.5392*** | |||||
| (0.1265) | ||||||
| Post 7 indicator | -1.5243*** | |||||
| (0.1092) | ||||||
| Post 8 indicator | -0.9858*** | |||||
| (0.1100) | ||||||
| Post 9 indicator | -0.5000*** | |||||
| (0.1136) | ||||||
| Post 10 indicator | -0.8357*** | |||||
| (0.1115) | ||||||
| Post 11 indicator | -1.4147*** | |||||
| (0.1090) | ||||||
| Post 12 indicator | -1.1001*** | |||||
| (0.1106) | ||||||
| Post 13 indicator | -1.6287*** | |||||
| (0.1089) | ||||||
| Post 14 indicator | -2.3481*** | |||||
| (0.1093) | ||||||
| Post 15 indicator | -2.1274*** | |||||
| (0.1096) | ||||||
| Post 16 indicator | 1.7600*** | |||||
| (0.1231) | ||||||
| Post 17 indicator | 0.1690* | |||||
| (0.1254) | ||||||
| Post 18 indicator | -0.0183 | |||||
| (0.1271) | ||||||
| Post 19 indicator | 0.6788*** | |||||
| (0.1248) | ||||||
| Post 20 indicator | 2.3767*** | |||||
| (0.1257) | ||||||
| Post 21 indicator | -0.1614* | |||||
| (0.1283) | ||||||
| Post 22 indicator | 0.1790* | |||||
| (0.1252) | ||||||
| Post 23 indicator | 0.8110*** | |||||
| (0.1235) | ||||||
| Post 24 indicator | 2.5226*** | |||||
| (0.1296) | ||||||
| Post 25 indicator | 1.2168*** | |||||
| (0.1232) | ||||||
| Post 26 indicator | 0.9150*** | |||||
| (0.1235) | ||||||
| Post 27 indicator | 0.5861*** | |||||
| (0.1223) | ||||||
| Post 28 indicator | 0.6185*** | |||||
| (0.1237) | ||||||
| Post 29 indicator | -0.2961** | |||||
| (0.1271) | ||||||
| Post 30 indicator | 0.4685*** | |||||
| (0.1240) | ||||||
| Post 31 indicator | 1.4240*** | |||||
| (0.1212) | ||||||
| Post 32 indicator | -0.3599*** | |||||
| (0.1303) | ||||||
| Post 33 indicator | 0.2143** | |||||
| (0.1272) | ||||||
| Post 34 indicator | 0.2669** | |||||
| (0.1232) | ||||||
| Post 35 indicator | 2.3320*** | |||||
| (0.1275) | ||||||
| Post 36 indicator | 2.9565*** | |||||
| (0.1355) | ||||||
| Post 37 indicator | 0.1110 | |||||
| (0.1261) | ||||||
| Post 38 indicator | -0.3521*** | |||||
| (0.1303) | ||||||
| Post 39 indicator | 2.2031*** | |||||
| (0.1238) | ||||||
| Post 40 indicator | 1.5682*** | |||||
| (0.1228) | ||||||
| Post 41 indicator | -0.0090 | |||||
| (0.1254) | ||||||
| Post 42 indicator | 1.4442*** | |||||
| (0.1231) | ||||||
| Post 43 indicator | 1.3760*** | |||||
| (0.1241) | ||||||
| Post 44 indicator | 0.0469 | |||||
| (0.1262) | ||||||
| Post 45 indicator | 1.9285*** | |||||
| (0.1254) | ||||||
| Post 46 indicator | 2.0559*** | |||||
| (0.1252) | ||||||
| Post 47 indicator | 1.0835*** | |||||
| (0.1224) | ||||||
| Post 48 indicator | -0.5459*** | |||||
| (0.1330) | ||||||
| Post 49 indicator | 1.1230*** | |||||
| (0.1235) | ||||||
| Post 50 indicator | 2.2882*** | |||||
| (0.1257) | ||||||
| Post 51 indicator | 2.9815*** | |||||
| (0.1328) | ||||||
| Post 52 indicator | -0.2676** | |||||
| (0.1290) | ||||||
| Post 53 indicator | -0.7880*** | |||||
| (0.1360) | ||||||
| Post 54 indicator | 2.5558*** | |||||
| (0.1286) | ||||||
| Post 55 indicator | 1.6723*** | |||||
| (0.1233) | ||||||
| Post 56 indicator | 1.2748*** | |||||
| (0.1232) | ||||||
| Post 57 indicator | 1.2485*** | |||||
| (0.1233) | ||||||
| Post 58 indicator | 0.7366*** | |||||
| (0.1219) | ||||||
| Post 59 indicator | 0.2561** | |||||
| (0.1262) | ||||||
| Log Likelihood | -43763.9673 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| AIC | 87543.9346 | |||||
| BIC | 87616.6247 | |||||
| Deviance | 87527.9346 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Pseudo R2 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method.
Treatment effects in the last column refer to estimated marginal means (EMMs, or least-squares means) for specified factor combinations (Post-treatment and Misinfo). Point estimates are reported in the upper cell and standard errors, clustered at the user level, reported in parentheses in the bottom cell. z-ratios are reported in brackets.
Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
# Models' predictions
models <- c("Model 1: Baseline",
"Model 2: User ID FE",
"Model 3: User ID FE + Post domain FE",
"Model 4: User ID FE + Post tactic",
"Model 5: User ID FE + Post domain x Post tactic",
"Model 6: User ID FE + Post ID FE")
list_models <- list(baseline_logit,fe1_logit,fe2_logit,fe3_logit,fe4_logit,fe5_logit)
treatmentLevels <- unique(panel_2$treatment)
second_fixed_effects <- list("",
"",
c(fact = unique(panel_2$fact)),
c(fact = unique(panel_2$type)),
list(type = unique(panel_2$type),fact = unique(panel_2$fact)),
c(user = unique(panel_2$post_id)))
second_fixed_effects_name <- list("","","fact","type",c("type","fact"),"post_id")
second_fixed_effects_feglm <- c("share ~ treatment",
"share ~ treatment",
"share ~ treatment + factor(fact)",
"share ~ treatment + factor(type)",
"share ~ treatment + factor(fact)*factor(type)",
"share ~ treatment + factor(post_id)")
for (m in 1:length(models)) {
# output df
prediction_df <- expand.grid(treatment = unique(panel_2$treatment),
post = c(0, 1),
post_is_misinfo = c(0, 1),
prediction = NA,
prediction_se = NA)
for (treatment in unique(panel_2$treatment)) {
for (post in c(0,1)) {
for (misinfo in c(0,1)) {
# Create new data
# 1 vs. 2 fixed effects
# type
if (m %in% c(2,3,4,6)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel_2$user[panel_2$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]])
}
if (m %in% c(5)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel_2$user[panel_2$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]][[1]],
second_fixed_effects_name2 = second_fixed_effects[[m]][[2]])
}
if (m %in% c(2)) {
predictorVariables <- predictorVariables %>% select(-second_fixed_effects_name)
}
if (m %in% c(3,4,6)) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[m]
}
if (m == 5) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[[m]][1]
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name2"] <- second_fixed_effects_name[[m]][2]
}
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
}
}
# predicted value
if (m == 1) {
# Table with prediction values
treatmentLevels <- unique(panel_2$treatment)
predictorVariables <- expand.grid(treatment = treatmentLevels,
post = c(0, 1),
post_is_misinfo = c(0, 1))
predictions <- predict(baseline_logit, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_logit, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% select(post, post_is_misinfo, estimate, SE, z.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
fixed_effect_model <- feglm(as.formula(second_fixed_effects_feglm[m]),panel_2 %>% filter(post == p & post_is_misinfo == mis), family = binomial(link = "logit"), cluster = "user")
# obtain data frame with estimate, std error, and z value
fixed_effect_model_df_add <- data.frame(
treatment = unique(panel_2$treatment),
estimate = format(coef(fixed_effect_model), digits = 4),
SE = sqrt(diag(vcov(fixed_effect_model))),
z.ratio = coef(fixed_effect_model) / sqrt(diag(vcov(fixed_effect_model))),
post = p,
post_is_misinfo = mis
)
# make rownames a column
fixed_effect_model_df_add <- rownames_to_column(fixed_effect_model_df_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>% select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
fixed_effect_model_df_add <- rbind(fixed_effect_model_df_add %>% mutate(treatment = "Facts Baseline"),fixed_effect_model_df_add %>% mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate)
contrast_df <- rbind(contrast_df, fixed_effect_model_df_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))
print(k)
}| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | -0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | -0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | 0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | 0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | -0.01342 | 0.02761 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05311); [0.5198] | |
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | -0.01768 | 0.008161 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04415); [0.1849] | |
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | 0.08481 | -0.358 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.05135); [-6.972] | |
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | 0.10281 | -0.4655 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.0466); [-9.99] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69121 | (0.00166) | -0.00958 | 0.03113 |
| Emotions | 0 | 0 | 0.70079 | (0.001567) | (0.05517); [0.5643] | |
| Facts Baseline | 0 | 1 | 0.52177 | (0.001855) | -0.01927 | 0.02062 |
| Emotions | 0 | 1 | 0.54104 | (0.001752) | (0.04658); [0.4428] | |
| Facts Baseline | 1 | 0 | 0.60301 | (0.001798) | 0.08370 | -0.38807 |
| Emotions | 1 | 0 | 0.51931 | (0.001759) | (0.05393); [-7.196] | |
| Facts Baseline | 1 | 1 | 0.43752 | (0.001843) | 0.10386 | -0.5043 |
| Emotions | 1 | 1 | 0.33366 | (0.001639) | (0.04944); [-10.2] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.65503 | (0.002921) | -0.01387 | |
| Emotions | 0 | 0 | 0.6689 | (0.002743) | ||
| Facts Baseline | 0 | 1 | 0.53518 | (0.003128) | -0.01777 | |
| Emotions | 0 | 1 | 0.55295 | (0.00293) | ||
| Facts Baseline | 1 | 0 | 0.56467 | (0.0031) | 0.08605 | |
| Emotions | 1 | 0 | 0.47862 | (0.002942) | ||
| Facts Baseline | 1 | 1 | 0.4501 | (0.003131) | 0.10351 | |
| Emotions | 1 | 1 | 0.34659 | (0.002766) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69229 | (0.0008727) | -0.00994 | |
| Emotions | 0 | 0 | 0.70223 | (0.0008224) | ||
| Facts Baseline | 0 | 1 | 0.52038 | (0.0009758) | -0.01928 | |
| Emotions | 0 | 1 | 0.53966 | (0.0009235) | ||
| Facts Baseline | 1 | 0 | 0.60788 | (0.0009433) | 0.08208 | |
| Emotions | 1 | 0 | 0.5258 | (0.0009261) | ||
| Facts Baseline | 1 | 1 | 0.43821 | (0.0009701) | 0.10815 | |
| Emotions | 1 | 1 | 0.33006 | (0.0008621) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.85217 | (0.0006642) | -0.00605 | 0.03113 |
| Emotions | 0 | 0 | 0.85822 | (0.0006231) | (0.05517); [0.5643] | |
| Facts Baseline | 0 | 1 | 0.44451 | (0.001027) | -0.01704 | 0.0186 |
| Emotions | 0 | 1 | 0.46155 | (0.0009845) | (0.04755); [0.3911] | |
| Facts Baseline | 1 | 0 | 0.79877 | (0.0007798) | 0.05690 | -0.38807 |
| Emotions | 1 | 0 | 0.74187 | (0.0008368) | (0.05393); [-7.196] | |
| Facts Baseline | 1 | 1 | 0.3698 | (0.000989) | 0.09526 | -0.53448 |
| Emotions | 1 | 1 | 0.27454 | (0.0008516) | (0.05087); [-10.51] |
Click for All treatment arms
# version 2: all treatment arms
formula_fe1_all_logit <- "share ~ treatment*post*post_is_misinfo | user"
fe1_all_logit <- feglm(as.formula(formula_fe1_all_logit),panel, family = binomial(link = "logit"))
formula_fe2_all_logit <- "share ~ treatment*post*post_is_misinfo + factor(fact) | user"
fe2_all_logit <- feglm(as.formula(formula_fe2_all_logit),panel, family = binomial(link = "logit"))
formula_fe3_all_logit <- "share ~ treatment*post*post_is_misinfo + factor(type) | user"
fe3_all_logit <- feglm(as.formula(formula_fe3_all_logit),panel, family = binomial(link = "logit"))
formula_fe4_all_logit <- "share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) | user"
fe4_all_logit <- feglm(as.formula(formula_fe4_all_logit),panel, family = binomial(link = "logit"))
formula_fe5_all_logit <- "share ~ treatment*post*post_is_misinfo + factor(post_id) | user"
fe5_all_logit <- feglm(as.formula(formula_fe5_all_logit),panel, family = binomial(link = "logit"))The sample includes all users who completed the post-survey (n = 8,684). We consider all 9 posts in the pre-treatment period (“pre”) and all 9 posts in the post-treatment period (i.e., “post”).
# Print model
headers <- c("Share (binary indicator)")
model_list <- list(baseline_all_logit,fe1_all_logit,fe2_all_logit,fe3_all_logit,fe4_all_logit,fe5_all_logit)
names <- unique(c(names(baseline_all_logit$coefficients),
names(fe1_all_logit$coefficients),
names(fe2_all_logit$coefficients),
names(fe3_all_logit$coefficients),
names(fe4_all_logit$coefficients),
names(fe5_all_logit$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# out-of-sample R2
if (any(ls() %in% "R2oos_fe_alltreat_logit_all")) {
R2oos_fe_alltreat_logit_all <- c(R2out_all,R2oos_fe_alltreat_logit_all)
} else {
R2oos_fe_alltreat_logit_all <- rep("",length(model_list))
}
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
#na_coeff <- names(model_list[[m]]$coefficients[,1])[is.na(model_list[[m]]$coefficients[,1])] # felm
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# standard errors
standard_errors <- list(baseline_all_logit$se,fe1_all_logit$se,fe2_all_logit$se,fe3_all_logit$se,fe4_all_logit$se,fe5_all_logit$se)
# keep variables
keep <- list("treatmentFacts Baseline:post:post_is_misinfo" = "Facts Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentReasoning:post:post_is_misinfo" = "Reasoning x 1[Post-treatment] x 1[Misinfo]",
"treatmentCombo:post:post_is_misinfo" = "Combo x 1[Post-treatment] x 1[Misinfo]",
"treatmentFacts Baseline:post" = "Facts Baseline x 1[Post-treatment]",
"treatmentNo-course Baseline" = "No-course Baseline x 1[Post-treatment]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentReasoning:post" = "Reasoning x 1[Post-treatment]",
"treatmentCombo:post" = "Combo x 1[Post-treatment]",
"treatmentFacts Baseline:post_is_misinfo" = "Facts Baseline x 1[Misinfo]",
"treatmentNo-course Baseline:post_is_misinfo" = "No-course Baseline x 1[Misinfo]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]",
"treatmentReasoning:post_is_misinfo" = "Reasoning x 1[Misinfo]",
"treatmentCombo:post_is_misinfo" = "Combo x 1[Misinfo]")
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_fe_alltreat_logit_all)),
reorder.gof = c(8,6,7,9,10,11,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0066 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0513) | (0.0634) | (0.0667) | (0.0635) | (0.0679) | (0.0679) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0066 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0513) | (0.0634) | (0.0667) | (0.0635) | (0.0679) | (0.0679) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0688) | (0.0836) | (0.0881) | (0.0836) | (0.0898) | (0.0898) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0778 | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0703) | (0.0861) | (0.0908) | (0.0862) | (0.0925) | (0.0925) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0536 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0696) | (0.0855) | (0.0901) | (0.0855) | (0.0918) | (0.0918) | |
| No-course Baseline x 1[Post-treatment] | 0.1001*** | |||||
| (0.0429) | ||||||
| Emotions x 1[Post-treatment] | -0.3856*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0567) | (0.0691) | (0.0728) | (0.0691) | (0.0737) | (0.0737) | |
| Reasoning x 1[Post-treatment] | -0.3314*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0582) | (0.0714) | (0.0752) | (0.0714) | (0.0762) | (0.0762) | |
| Combo x 1[Post-treatment] | -0.3624*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0574) | (0.0707) | (0.0744) | (0.0707) | (0.0754) | (0.0754) | |
| Emotions x 1[Misinfo] | -0.0194 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0495) | (0.0605) | (0.0637) | (0.0606) | (0.0648) | (0.0648) | |
| Reasoning x 1[Misinfo] | -0.0930** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0508) | (0.0627) | (0.0660) | (0.0627) | (0.0671) | (0.0671) | |
| Combo x 1[Misinfo] | 0.0140 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0501) | (0.0619) | (0.0651) | (0.0619) | (0.0663) | (0.0663) | |
| Log Likelihood | -104554.8012 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| AIC | 209149.6024 | |||||
| BIC | 209348.7946 | |||||
| Deviance | 209109.6024 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Pseudo R2 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 | |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
# Full table (all independent variables)
cat('<details>')cat('<summary>Click to view the full table</summary>')Click to view the full table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
override.se = standard_errors,
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects"),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = rep("No",length(model_list)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"\\textit{{Out-of-sample R}$^2$" = c(R2oos_fe_alltreat_logit_all)),
reorder.gof = c(8,6,7,9,10,11,5,1,2,3,4)
)
html_table| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.7670*** | |||||
| (0.0293) | ||||||
| No-course Baseline | 0.1001*** | |||||
| (0.0429) | ||||||
| Emotions | 0.0276 | |||||
| (0.0413) | ||||||
| Reasoning | 0.0982** | |||||
| (0.0425) | ||||||
| Combo | 0.0164 | |||||
| (0.0418) | ||||||
| 1[Post-treatment] | -0.3319*** | -0.5002*** | -0.5575*** | -0.5005*** | -0.5596*** | -0.5596*** |
| (0.0405) | (0.0498) | (0.0525) | (0.0499) | (0.0532) | (0.0532) | |
| 1[Misinfo] | -0.6121*** | -0.9234*** | -1.0396*** | -0.8461*** | -1.0772*** | -3.0886*** |
| (0.0351) | (0.0435) | (0.0459) | (0.0449) | (0.0903) | (0.0913) | |
| No-course Baseline x 1[Post-treatment] | 0.2176*** | 0.3289*** | 0.3843*** | 0.3290*** | 0.3850*** | 0.3850*** |
| (0.0597) | (0.0733) | (0.0771) | (0.0733) | (0.0780) | (0.0780) | |
| Emotions x 1[Post-treatment] | -0.3856*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0567) | (0.0691) | (0.0728) | (0.0691) | (0.0737) | (0.0737) | |
| Reasoning x 1[Post-treatment] | -0.3314*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0582) | (0.0714) | (0.0752) | (0.0714) | (0.0762) | (0.0762) | |
| Combo x 1[Post-treatment] | -0.3624*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0574) | (0.0707) | (0.0744) | (0.0707) | (0.0754) | (0.0754) | |
| No-course Baseline x 1[Misinfo] | 0.0066 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0513) | (0.0634) | (0.0667) | (0.0635) | (0.0679) | (0.0679) | |
| Emotions x 1[Misinfo] | -0.0194 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0495) | (0.0605) | (0.0637) | (0.0606) | (0.0648) | (0.0648) | |
| Reasoning x 1[Misinfo] | -0.0930** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0508) | (0.0627) | (0.0660) | (0.0627) | (0.0671) | (0.0671) | |
| Combo x 1[Misinfo] | 0.0140 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0501) | (0.0619) | (0.0651) | (0.0619) | (0.0663) | (0.0663) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0368 | 0.0492 | 0.0613 | 0.0485 | 0.0507 | 0.0507 |
| (0.0489) | (0.0602) | (0.0635) | (0.0602) | (0.0647) | (0.0647) | |
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0008 | 0.0069 | -0.0022 | 0.0073 | 0.0198 | 0.0198 |
| (0.0717) | (0.0881) | (0.0928) | (0.0881) | (0.0945) | (0.0945) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0880 | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0688) | (0.0836) | (0.0881) | (0.0836) | (0.0898) | (0.0898) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0778 | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0703) | (0.0861) | (0.0908) | (0.0862) | (0.0925) | (0.0925) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0536 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0696) | (0.0855) | (0.0901) | (0.0855) | (0.0918) | (0.0918) | |
| Post domain 2 | -1.3458*** | -1.3896*** | ||||
| (0.0369) | (0.0708) | |||||
| Post domain 3 | -1.8926*** | -2.2086*** | ||||
| (0.0376) | (0.0705) | |||||
| Post domain 4 | -1.2128*** | -1.7401*** | ||||
| (0.0368) | (0.0704) | |||||
| Post domain 5 | 0.4049*** | 0.0051 | ||||
| (0.0385) | (0.0773) | |||||
| Post domain 6 | 0.1237*** | 0.4527*** | ||||
| (0.0378) | (0.0803) | |||||
| Post domain 7 | -1.5746*** | -1.4228*** | ||||
| (0.0371) | (0.0705) | |||||
| Post domain 8 | -1.4724*** | -0.9538*** | ||||
| (0.0369) | (0.0716) | |||||
| Post domain 9 | 0.2728*** | -0.5281*** | ||||
| (0.0382) | (0.0737) | |||||
| Post domain 10 | -0.3969*** | -0.8059*** | ||||
| (0.0370) | (0.0724) | |||||
| Post domain 11 | -1.0747*** | -1.3501*** | ||||
| (0.0367) | (0.0705) | |||||
| Post domain 12 | -0.7213*** | -1.0828*** | ||||
| (0.0367) | (0.0716) | |||||
| Post domain 13 | -1.0425*** | -1.6220*** | ||||
| (0.0367) | (0.0704) | |||||
| Post domain 14 | -1.8560*** | -2.2721*** | ||||
| (0.0376) | (0.0708) | |||||
| Post domain 15 | -1.2523*** | -2.0199*** | ||||
| (0.0368) | (0.0707) | |||||
| 1[Combo manipulation tactic] | -0.2264*** | -0.3419*** | ||||
| (0.0188) | (0.0802) | |||||
| 1[Emotions manipulation tactic] | -0.0068 | -0.5416*** | ||||
| (0.0187) | (0.0788) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.1530* | |||||
| (0.1077) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4818*** | |||||
| (0.1074) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.7230*** | |||||
| (0.1065) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6450*** | |||||
| (0.1126) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.2961*** | |||||
| (0.1154) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.1072 | |||||
| (0.1076) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0724 | |||||
| (0.1068) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.1394*** | |||||
| (0.1101) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.3736*** | |||||
| (0.1075) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.6002*** | |||||
| (0.1063) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | 0.0418 | |||||
| (0.1069) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.5858*** | |||||
| (0.1067) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.3450*** | |||||
| (0.1086) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.7389*** | |||||
| (0.1066) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4478*** | |||||
| (0.1086) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.9334*** | |||||
| (0.1066) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.5846*** | |||||
| (0.1055) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.8014*** | |||||
| (0.1117) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9194*** | |||||
| (0.1172) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0346 | |||||
| (0.1068) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.9456*** | |||||
| (0.1100) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2361*** | |||||
| (0.1085) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.8795*** | |||||
| (0.1065) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0799 | |||||
| (0.1062) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1423*** | |||||
| (0.1062) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.4266*** | |||||
| (0.1044) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.8221*** | |||||
| (0.1073) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.4403*** | |||||
| (0.1059) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.4651*** | |||||
| (0.1057) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.2843*** | |||||
| (0.1098) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.8423*** | |||||
| (0.1057) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.2026** | |||||
| (0.1121) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3527*** | |||||
| (0.1174) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.8764*** | |||||
| (0.1090) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.9427*** | |||||
| (0.1128) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 1.0028*** | |||||
| (0.1115) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4064*** | |||||
| (0.1076) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.5144*** | |||||
| (0.1063) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2580*** | |||||
| (0.1063) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3539*** | |||||
| (0.1060) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.4395*** | |||||
| (0.1069) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0085 | |||||
| (0.1077) | ||||||
| Post 2 indicator | -1.3896*** | |||||
| (0.0708) | ||||||
| Post 3 indicator | -2.2086*** | |||||
| (0.0705) | ||||||
| Post 4 indicator | -1.7401*** | |||||
| (0.0704) | ||||||
| Post 5 indicator | 0.0051 | |||||
| (0.0773) | ||||||
| Post 6 indicator | 0.4527*** | |||||
| (0.0803) | ||||||
| Post 7 indicator | -1.4228*** | |||||
| (0.0705) | ||||||
| Post 8 indicator | -0.9538*** | |||||
| (0.0716) | ||||||
| Post 9 indicator | -0.5281*** | |||||
| (0.0737) | ||||||
| Post 10 indicator | -0.8059*** | |||||
| (0.0724) | ||||||
| Post 11 indicator | -1.3501*** | |||||
| (0.0705) | ||||||
| Post 12 indicator | -1.0828*** | |||||
| (0.0716) | ||||||
| Post 13 indicator | -1.6220*** | |||||
| (0.0704) | ||||||
| Post 14 indicator | -2.2721*** | |||||
| (0.0708) | ||||||
| Post 15 indicator | -2.0199*** | |||||
| (0.0707) | ||||||
| Post 16 indicator | 1.6695*** | |||||
| (0.0809) | ||||||
| Post 17 indicator | 0.1269* | |||||
| (0.0815) | ||||||
| Post 18 indicator | -0.0573 | |||||
| (0.0815) | ||||||
| Post 19 indicator | 0.6525*** | |||||
| (0.0802) | ||||||
| Post 20 indicator | 2.3196*** | |||||
| (0.0822) | ||||||
| Post 21 indicator | -0.1740** | |||||
| (0.0828) | ||||||
| Post 22 indicator | 0.1395** | |||||
| (0.0816) | ||||||
| Post 23 indicator | 0.7881*** | |||||
| (0.0794) | ||||||
| Post 24 indicator | 2.2808*** | |||||
| (0.0823) | ||||||
| Post 25 indicator | 1.2372*** | |||||
| (0.0797) | ||||||
| Post 26 indicator | 0.9195*** | |||||
| (0.0800) | ||||||
| Post 27 indicator | 0.6285*** | |||||
| (0.0796) | ||||||
| Post 28 indicator | 0.6333*** | |||||
| (0.0805) | ||||||
| Post 29 indicator | -0.2576*** | |||||
| (0.0826) | ||||||
| Post 30 indicator | 0.3885*** | |||||
| (0.0810) | ||||||
| Post 31 indicator | 1.4698*** | |||||
| (0.0795) | ||||||
| Post 32 indicator | -0.3676*** | |||||
| (0.0839) | ||||||
| Post 33 indicator | 0.1946*** | |||||
| (0.0818) | ||||||
| Post 34 indicator | 0.3143*** | |||||
| (0.0803) | ||||||
| Post 35 indicator | 2.2763*** | |||||
| (0.0825) | ||||||
| Post 36 indicator | 2.8419*** | |||||
| (0.0871) | ||||||
| Post 37 indicator | 0.0124 | |||||
| (0.0819) | ||||||
| Post 38 indicator | -0.4296*** | |||||
| (0.0851) | ||||||
| Post 39 indicator | 2.1778*** | |||||
| (0.0814) | ||||||
| Post 40 indicator | 1.5434*** | |||||
| (0.0798) | ||||||
| Post 41 indicator | 0.0398 | |||||
| (0.0810) | ||||||
| Post 42 indicator | 1.5293*** | |||||
| (0.0802) | ||||||
| Post 43 indicator | 1.2744*** | |||||
| (0.0789) | ||||||
| Post 44 indicator | 0.0198 | |||||
| (0.0824) | ||||||
| Post 45 indicator | 1.8902*** | |||||
| (0.0813) | ||||||
| Post 46 indicator | 2.0114*** | |||||
| (0.0810) | ||||||
| Post 47 indicator | 1.0869*** | |||||
| (0.0793) | ||||||
| Post 48 indicator | -0.4814*** | |||||
| (0.0848) | ||||||
| Post 49 indicator | 1.1137*** | |||||
| (0.0797) | ||||||
| Post 50 indicator | 2.2191*** | |||||
| (0.0820) | ||||||
| Post 51 indicator | 2.8168*** | |||||
| (0.0864) | ||||||
| Post 52 indicator | -0.2878*** | |||||
| (0.0836) | ||||||
| Post 53 indicator | -0.8851*** | |||||
| (0.0876) | ||||||
| Post 54 indicator | 2.4861*** | |||||
| (0.0845) | ||||||
| Post 55 indicator | 1.6118*** | |||||
| (0.0804) | ||||||
| Post 56 indicator | 1.1756*** | |||||
| (0.0802) | ||||||
| Post 57 indicator | 1.1865*** | |||||
| (0.0794) | ||||||
| Post 58 indicator | 0.7433*** | |||||
| (0.0800) | ||||||
| Post 59 indicator | 0.1789** | |||||
| (0.0811) | ||||||
| Log Likelihood | -104554.8012 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| AIC | 209149.6024 | |||||
| BIC | 209348.7946 | |||||
| Deviance | 209109.6024 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Pseudo R2 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 | |
| \textit{{Out-of-sample R}\(^2\) | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | No | No | No | No | No | No |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
cat('</details>')Prediction
Standard errors of the share predicted value are estimated using the Delta method. Treatment effects standard errors are clustered at the user level and reported in parentheses. z-ratios are reported in brackets. Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
# Models' predictions
models <- c("Model 1: Baseline",
"Model 2: User ID FE",
"Model 3: User ID FE + Post domain FE",
"Model 4: User ID FE + Post tactic",
"Model 5: User ID FE + Post domain x Post tactic",
"Model 6: User ID FE + Post ID FE")
list_models <- list(baseline_all_logit,fe1_all_logit,fe2_all_logit,fe3_all_logit,fe4_all_logit,fe5_all_logit)
treatmentLevels <- unique(panel$treatment)
second_fixed_effects <- list("",
"",
c(fact = unique(panel$fact)),
c(fact = unique(panel$type)),
list(type = unique(panel$type),fact = unique(panel$fact)),
c(user = unique(panel$post_id)))
second_fixed_effects_name <- list("","","fact","type",c("type","fact"),"post_id")
second_fixed_effects_feglm <- c("share ~ treatment",
"share ~ treatment",
"share ~ treatment + factor(fact)",
"share ~ treatment + factor(type)",
"share ~ treatment + factor(fact)*factor(type)",
"share ~ treatment + factor(post_id)")
for (m in 1:length(models)) {
# output df
prediction_df <- expand.grid(treatment = unique(panel$treatment),
post = c(0, 1),
post_is_misinfo = c(0, 1),
prediction = NA,
prediction_se = NA)
for (treatment in unique(panel$treatment)) {
for (post in c(0,1)) {
for (misinfo in c(0,1)) {
# Create new data
# 1 vs. 2 fixed effects
# type
if (m %in% c(2,3,4,6)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel$user[panel$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]])
}
if (m %in% c(5)) {
predictorVariables <- expand.grid(treatment = treatment,
post = post,
post_is_misinfo = misinfo,
user = unique(panel$user[panel$treatment == treatment]),
second_fixed_effects_name = second_fixed_effects[[m]][[1]],
second_fixed_effects_name2 = second_fixed_effects[[m]][[2]])
}
if (m %in% c(2)) {
predictorVariables <- predictorVariables %>% select(-second_fixed_effects_name)
}
if (m %in% c(3,4,6)) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[m]
}
if (m == 5) {
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name"] <- second_fixed_effects_name[[m]][1]
names(predictorVariables)[names(predictorVariables) == "second_fixed_effects_name2"] <- second_fixed_effects_name[[m]][2]
}
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[prediction_df$treatment == treatment & prediction_df$post == post & prediction_df$post_is_misinfo == misinfo] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
}
}
# predicted value
if (m == 1) {
# Table with prediction values
treatmentLevels <- unique(panel$treatment)
predictorVariables <- expand.grid(treatment = treatmentLevels,
post = c(0, 1),
post_is_misinfo = c(0, 1))
predictions <- predict(baseline_all_logit, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_all_logit, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% filter(contrast == "Emotions - Facts Baseline") %>%
select(post, post_is_misinfo, estimate, SE, z.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
fixed_effect_model <- feglm(as.formula(second_fixed_effects_feglm[m]),panel %>% filter(post == p & post_is_misinfo == mis), family = binomial(link = "logit"), cluster = "user")
# obtain data frame with estimate, std error, and z value
fixed_effect_model_df_add <- data.frame(
treatment = c("No-course Baseline","Emotions","Reasoning","Combo"),
estimate = format(coef(fixed_effect_model)[grepl("treatment",names(coef(fixed_effect_model)))], digits = 4),
SE = sqrt(diag(vcov(fixed_effect_model)))[grepl("treatment",names(diag(vcov(fixed_effect_model))))],
z.ratio = coef(fixed_effect_model)[grepl("treatment",names(coef(fixed_effect_model)))] / sqrt(diag(vcov(fixed_effect_model)))[grepl("treatment",names(diag(vcov(fixed_effect_model))))],
post = p,
post_is_misinfo = mis
)
# make rownames a column
fixed_effect_model_df_add <- rownames_to_column(fixed_effect_model_df_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>% select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
fixed_effect_model_df_add <- rbind(fixed_effect_model_df_add %>% mutate(treatment = "Facts Baseline"),fixed_effect_model_df_add %>% mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
contrast_df <- rbind(contrast_df, fixed_effect_model_df_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))
print(k)
}| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | -0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | -0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | 0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | 0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] | |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | -0.01342 | 0.02761 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05311); [0.5198] | |
| No-course Baseline | 0 | 0 | 0.70429 | (0.005697) | ||
| Reasoning | 0 | 0 | 0.72166 | (0.005308) | ||
| Combo | 0 | 0 | 0.69983 | (0.005546) | ||
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | -0.01768 | 0.008161 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04415); [0.1849] | |
| No-course Baseline | 0 | 1 | 0.5404 | (0.006335) | ||
| Reasoning | 0 | 1 | 0.53448 | (0.006122) | ||
| Combo | 0 | 1 | 0.54118 | (0.006203) | ||
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | 0.08481 | -0.3580 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.05135); [-6.972] | |
| No-course Baseline | 1 | 0 | 0.67559 | (0.005877) | ||
| Reasoning | 1 | 0 | 0.54632 | (0.006106) | ||
| Combo | 1 | 0 | 0.51409 | (0.006231) | ||
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | 0.10281 | -0.4655 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.04659); [-9.99] | |
| No-course Baseline | 1 | 1 | 0.51809 | (0.006354) | ||
| Reasoning | 1 | 1 | 0.3792 | (0.005923) | ||
| Combo | 1 | 1 | 0.34352 | (0.005854) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69117 | (0.001652) | -0.00973 | 0.03138 |
| Emotions | 0 | 0 | 0.7009 | (0.001559) | (0.05497); [0.5709] | |
| No-course Baseline | 0 | 0 | 0.704 | (0.001697) | ||
| Reasoning | 0 | 0 | 0.72003 | (0.001587) | ||
| Combo | 0 | 0 | 0.7001 | (0.001636) | ||
| Facts Baseline | 0 | 1 | 0.5218 | (0.001844) | -0.01924 | 0.02028 |
| Emotions | 0 | 1 | 0.54104 | (0.001742) | (0.0465); [0.4361] | |
| No-course Baseline | 0 | 1 | 0.53852 | (0.001899) | ||
| Reasoning | 0 | 1 | 0.53724 | (0.001835) | ||
| Combo | 0 | 1 | 0.54042 | (0.001837) | ||
| Facts Baseline | 1 | 0 | 0.60298 | (0.001789) | 0.08375 | -0.3855 |
| Emotions | 1 | 0 | 0.51923 | (0.001749) | (0.05358); [-7.195] | |
| No-course Baseline | 1 | 0 | 0.67769 | (0.001748) | ||
| Reasoning | 1 | 0 | 0.54351 | (0.001832) | ||
| Combo | 1 | 0 | 0.51396 | (0.001846) | ||
| Facts Baseline | 1 | 1 | 0.43752 | (0.001833) | 0.10383 | -0.5018 |
| Emotions | 1 | 1 | 0.33369 | (0.001629) | (0.0492); [-10.2] | |
| No-course Baseline | 1 | 1 | 0.51867 | (0.001904) | ||
| Reasoning | 1 | 1 | 0.378 | (0.001778) | ||
| Combo | 1 | 1 | 0.34425 | (0.001736) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.67934 | (0.002842) | -0.01356 | |
| Emotions | 0 | 0 | 0.6929 | (0.002669) | ||
| No-course Baseline | 0 | 0 | 0.69447 | (0.002889) | ||
| Reasoning | 0 | 0 | 0.71206 | (0.002698) | ||
| Combo | 0 | 0 | 0.69006 | (0.002813) | ||
| Facts Baseline | 0 | 1 | 0.52636 | (0.003126) | -0.01773 | |
| Emotions | 0 | 1 | 0.54409 | (0.002927) | ||
| No-course Baseline | 0 | 1 | 0.54415 | (0.003175) | ||
| Reasoning | 0 | 1 | 0.53824 | (0.003068) | ||
| Combo | 0 | 1 | 0.54486 | (0.003107) | ||
| Facts Baseline | 1 | 0 | 0.59093 | (0.003054) | 0.08524 | |
| Emotions | 1 | 0 | 0.50569 | (0.002938) | ||
| No-course Baseline | 1 | 0 | 0.66533 | (0.002976) | ||
| Reasoning | 1 | 0 | 0.53503 | (0.00307) | ||
| Combo | 1 | 0 | 0.50301 | (0.003126) | ||
| Facts Baseline | 1 | 1 | 0.44124 | (0.003116) | 0.10308 | |
| Emotions | 1 | 1 | 0.33816 | (0.002738) | ||
| No-course Baseline | 1 | 1 | 0.52186 | (0.003185) | ||
| Reasoning | 1 | 1 | 0.38274 | (0.002977) | ||
| Combo | 1 | 1 | 0.34685 | (0.002943) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68894 | (0.0008706) | -0.01016 | |
| Emotions | 0 | 0 | 0.6991 | (0.0008202) | ||
| No-course Baseline | 0 | 0 | 0.70137 | (0.0008962) | ||
| Reasoning | 0 | 0 | 0.71754 | (0.0008404) | ||
| Combo | 0 | 0 | 0.69786 | (0.0008633) | ||
| Facts Baseline | 0 | 1 | 0.52146 | (0.0009691) | -0.01928 | |
| Emotions | 0 | 1 | 0.54074 | (0.0009167) | ||
| No-course Baseline | 0 | 1 | 0.53655 | (0.001002) | ||
| Reasoning | 0 | 1 | 0.53674 | (0.0009685) | ||
| Combo | 0 | 1 | 0.54125 | (0.0009662) | ||
| Facts Baseline | 1 | 0 | 0.60441 | (0.0009391) | 0.08204 | |
| Emotions | 1 | 0 | 0.52237 | (0.0009199) | ||
| No-course Baseline | 1 | 0 | 0.67592 | (0.0009226) | ||
| Reasoning | 1 | 0 | 0.54666 | (0.000966) | ||
| Combo | 1 | 0 | 0.51662 | (0.0009707) | ||
| Facts Baseline | 1 | 1 | 0.43918 | (0.0009635) | 0.10822 | |
| Emotions | 1 | 1 | 0.33096 | (0.0008554) | ||
| No-course Baseline | 1 | 1 | 0.51929 | (0.001005) | ||
| Reasoning | 1 | 1 | 0.37521 | (0.0009374) | ||
| Combo | 1 | 1 | 0.33953 | (0.0009105) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.8489 | (0.0006644) | -0.00631 | 0.03138 |
| Emotions | 0 | 0 | 0.85521 | (0.000623) | (0.05497); [0.5709] | |
| No-course Baseline | 0 | 0 | 0.85647 | (0.0006846) | ||
| Reasoning | 0 | 0 | 0.86734 | (0.000623) | ||
| Combo | 0 | 0 | 0.85431 | (0.0006533) | ||
| Facts Baseline | 0 | 1 | 0.44608 | (0.00102) | -0.01703 | 0.01916 |
| Emotions | 0 | 1 | 0.46311 | (0.0009768) | (0.04737); [0.4045] | |
| No-course Baseline | 0 | 1 | 0.45829 | (0.001066) | ||
| Reasoning | 0 | 1 | 0.45894 | (0.001033) | ||
| Combo | 0 | 1 | 0.46436 | (0.001027) | ||
| Facts Baseline | 1 | 0 | 0.79462 | (0.0007794) | 0.05764 | -0.3855 |
| Emotions | 1 | 0 | 0.73698 | (0.000835) | (0.05358); [-7.195] | |
| No-course Baseline | 1 | 0 | 0.84125 | (0.0007225) | ||
| Reasoning | 1 | 0 | 0.75676 | (0.0008513) | ||
| Combo | 1 | 0 | 0.73151 | (0.0008824) | ||
| Facts Baseline | 1 | 1 | 0.37102 | (0.0009821) | 0.09562 | -0.5277 |
| Emotions | 1 | 1 | 0.2754 | (0.0008449) | (0.05031); [-10.49] | |
| No-course Baseline | 1 | 1 | 0.44229 | (0.001061) | ||
| Reasoning | 1 | 1 | 0.31358 | (0.0009399) | ||
| Combo | 1 | 1 | 0.28317 | (0.0008995) |
Covariates
# version 1: 2 treatment arms
covariate_list <- c(demographics_demeaned_treatment,
sharing_demeaned_treatment,
sharing_tactic_demeaned_treatment,
demographics_sharing_demeaned_treatment,
all_covariates_demeaned_treatment)
cov_estimates_list <- list()
for (cov_list in covariate_list) {
# Demographics
formula_cov1_logit <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list)
cov1_logit <- feglm(as.formula(formula_cov1_logit),panel_2, family = binomial(link = "logit"))
formula_cov2_logit <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list," | user")
cov2_logit <- feglm(as.formula(formula_cov2_logit),panel_2, family = binomial(link = "logit"))
# Misinfo and Non-misinfo sharing rates
formula_cov3_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact) "," + ",cov_list," | user")
cov3_logit <- feglm(as.formula(formula_cov3_logit),panel_2, family = binomial(link = "logit"))
# Misinfo, Non-misinfo, and Tactic sharing rates
formula_cov4_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(type) "," + ",cov_list," | user")
cov4_logit <- feglm(as.formula(formula_cov4_logit),panel_2, family = binomial(link = "logit"))
# Demographics and Misinfo and Non-misinfo sharing rates
formula_cov5_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) "," + ",cov_list," | user")
cov5_logit <- feglm(as.formula(formula_cov5_logit),panel_2, family = binomial(link = "logit"))
# Demographics, Misinfo and Non-misinfo sharing rates, and Tactic sharing rates
formula_cov6_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(post_id) "," + ",cov_list," | user")
cov6_logit <- feglm(as.formula(formula_cov6_logit),panel_2, family = binomial(link = "logit"))
cov_estimates_list[[length(cov_estimates_list) + 1]] <- list(cov1_logit,cov2_logit,cov3_logit,cov4_logit,cov5_logit,cov6_logit)
}covariate_list_names <- c("Demographics",
"Sharing rates",
"Tactic sharing rates",
"Demographics + Sharing rates",
"Demographics + Sharing rates + Tactic sharing rates")
for (cov in seq(length(covariate_list))) {
# Print model
headers <- c("Share (binary indicator)")
model_list <- cov_estimates_list[[cov]]
names <- unique(c(names(cov_estimates_list[[cov]][[1]]$coefficients),
names(cov_estimates_list[[cov]][[2]]$coefficients),
names(cov_estimates_list[[cov]][[3]]$coefficients),
names(cov_estimates_list[[cov]][[4]]$coefficients),
names(cov_estimates_list[[cov]][[5]]$coefficients),
names(cov_estimates_list[[cov]][[6]]$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# out-of-sample R2
if (any(ls() %in% "R2oos_cov_2treat_logit_list")) {
R2oos_cov_2treat_logit_all <- c(R2oos_cov_2treat_logit_list[[cov]])
} else {
R2oos_cov_2treat_logit_all <- rep("",length(model_list))
}
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# Print model name
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_2treat_logit_all)),
reorder.gof = c(6,7,8,9,5,1,2,3,4)
)
cat(html_table)
cat("<br>")
# Full table (all independent variables)
cat('<details>')
cat('<summary>Click to view the full table</summary>')
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_2treat_logit_all)),
reorder.gof = c(6,7,8,9,5,1,2,3,4)
)
cat(html_table)
cat('</details>')
cat('<br>\n<br>\n<br>')
}| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0914* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0704) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Emotions x 1[Post-treatment] | -0.4006*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0581) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | -0.0151 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0507) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 84465.3730 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -42232.6865 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.0639 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.8245*** | |||||
| (0.0302) | ||||||
| Emotions | 0.0005 | |||||
| (0.0424) | ||||||
| 1[Post-treatment] | -0.3496*** | -0.5002*** | -0.5601*** | -0.5005*** | -0.5646*** | -0.5646*** |
| (0.0416) | (0.0589) | (0.0634) | (0.0589) | (0.0645) | (0.0645) | |
| 1[Misinfo] | -0.6451*** | -0.9234*** | -1.0451*** | -0.6580*** | -1.1149*** | -3.1708*** |
| (0.0361) | (0.0453) | (0.0487) | (0.0266) | (0.1374) | (0.1367) | |
| Age | -0.0152 | |||||
| (0.0130) | ||||||
| Man | 0.0593*** | |||||
| (0.0118) | ||||||
| Some college | -0.0274** | |||||
| (0.0147) | ||||||
| Bachelor’s degree | -0.0623*** | |||||
| (0.0148) | ||||||
| Graduate degree | -0.0465*** | |||||
| (0.0127) | ||||||
| Married | 0.0025 | |||||
| (0.0132) | ||||||
| Unemployed | -0.0678*** | |||||
| (0.0154) | ||||||
| Employed | -0.0396*** | |||||
| (0.0168) | ||||||
| Mostly urban | -0.1233*** | |||||
| (0.0140) | ||||||
| Suburban | -0.0765*** | |||||
| (0.0137) | ||||||
| Christian | 0.0387*** | |||||
| (0.0122) | ||||||
| Attends religious services | 0.0452*** | |||||
| (0.0119) | ||||||
| Uses social media | -0.0830*** | |||||
| (0.0111) | ||||||
| Hours spent on Social Media | 0.0327*** | |||||
| (0.0130) | ||||||
| Social Media Share: 80-100% | 0.3545*** | |||||
| (0.0148) | ||||||
| Social Media Share: 60-80% | 0.4825*** | |||||
| (0.0159) | ||||||
| Social Media Share: 40-60% | 0.4174*** | |||||
| (0.0164) | ||||||
| Social Media Share: 20-40% | 0.2228*** | |||||
| (0.0152) | ||||||
| Passed attention check at pre-treatment | -0.0945*** | |||||
| (0.0118) | ||||||
| Emotions x 1[Post-treatment] | -0.4006*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0581) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | -0.0151 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0507) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0386 | 0.0492 | 0.0615 | 0.0485 | 0.0524 | 0.0524 |
| (0.0502) | (0.0617) | (0.0642) | (0.0618) | (0.0653) | (0.0653) | |
| Emotions x Age | -0.0518*** | |||||
| (0.0196) | ||||||
| Emotions x Man | -0.0091 | |||||
| (0.0167) | ||||||
| Emotions x Some college | -0.0158 | |||||
| (0.0205) | ||||||
| Emotions x Bachelor’s degree | 0.0367** | |||||
| (0.0206) | ||||||
| Emotions x Graduate degree | 0.0470*** | |||||
| (0.0169) | ||||||
| Emotions x Married | 0.0228 | |||||
| (0.0188) | ||||||
| Emotions x Unemployed | -0.0320* | |||||
| (0.0216) | ||||||
| Emotions x Employed | -0.0835*** | |||||
| (0.0234) | ||||||
| Emotions x Mostly urban | 0.1246*** | |||||
| (0.0197) | ||||||
| Emotions x Suburban | 0.0657*** | |||||
| (0.0194) | ||||||
| Emotions x Christian | -0.0361** | |||||
| (0.0174) | ||||||
| Emotions x Attends religious services | -0.0942*** | |||||
| (0.0173) | ||||||
| Emotions x Uses social media | 0.0238* | |||||
| (0.0175) | ||||||
| Emotions x Hours spent on Social Media | -0.0555*** | |||||
| (0.0179) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0325* | |||||
| (0.0208) | ||||||
| Emotions x Social Media Share: 60-80% | -0.0176 | |||||
| (0.0228) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0546*** | |||||
| (0.0233) | ||||||
| Emotions x Social Media Share: 20-40% | -0.0290* | |||||
| (0.0217) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0182 | |||||
| (0.0167) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0914* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0704) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Post domain 2 | -1.3255*** | -1.3401*** | ||||
| (0.0603) | (0.1192) | |||||
| Post domain 3 | -1.9197*** | -2.2672*** | ||||
| (0.0609) | (0.1177) | |||||
| Post domain 4 | -1.2883*** | -1.9129*** | ||||
| (0.0601) | (0.1187) | |||||
| Post domain 5 | 0.4122*** | -0.0798 | ||||
| (0.0593) | (0.1262) | |||||
| Post domain 6 | 0.1959*** | 0.5392*** | ||||
| (0.0613) | (0.1363) | |||||
| Post domain 7 | -1.5826*** | -1.5243*** | ||||
| (0.0602) | (0.1183) | |||||
| Post domain 8 | -1.4424*** | -0.9858*** | ||||
| (0.0598) | (0.1199) | |||||
| Post domain 9 | 0.3437*** | -0.5000*** | ||||
| (0.0605) | (0.1233) | |||||
| Post domain 10 | -0.4006*** | -0.8357*** | ||||
| (0.0581) | (0.1185) | |||||
| Post domain 11 | -1.0907*** | -1.4147*** | ||||
| (0.0605) | (0.1181) | |||||
| Post domain 12 | -0.7530*** | -1.1001*** | ||||
| (0.0578) | (0.1181) | |||||
| Post domain 13 | -1.0463*** | -1.6287*** | ||||
| (0.0591) | (0.1174) | |||||
| Post domain 14 | -1.8782*** | -2.3481*** | ||||
| (0.0631) | (0.1200) | |||||
| Post domain 15 | -1.2760*** | -2.1274*** | ||||
| (0.0590) | (0.1188) | |||||
| 1[Combo manipulation tactic] | -0.4163*** | -0.2959** | ||||
| (0.0234) | (0.1277) | |||||
| 1[Emotions manipulation tactic] | -0.2120*** | -0.6319*** | ||||
| (0.0312) | (0.1279) | |||||
| 1[Reasoning manipulation tactic] | -0.1699*** | |||||
| (0.0317) | ||||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.2510* | |||||
| (0.1800) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4889*** | |||||
| (0.1724) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.8316*** | |||||
| (0.1703) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6964*** | |||||
| (0.1824) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.4606*** | |||||
| (0.1888) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0567 | |||||
| (0.1748) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0368 | |||||
| (0.1767) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.2625*** | |||||
| (0.1823) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.2924** | |||||
| (0.1725) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.5696*** | |||||
| (0.1735) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0738 | |||||
| (0.1737) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.4872*** | |||||
| (0.1721) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.2920** | |||||
| (0.1753) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.8359*** | |||||
| (0.1715) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4438*** | |||||
| (0.1815) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 1.0575*** | |||||
| (0.1695) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.7557*** | |||||
| (0.1693) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.9877*** | |||||
| (0.1800) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9932*** | |||||
| (0.1963) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.2112 | |||||
| (0.1731) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.7904*** | |||||
| (0.1827) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2790*** | |||||
| (0.1780) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.9799*** | |||||
| (0.1691) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0184 | |||||
| (0.1746) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1203*** | |||||
| (0.1709) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.5807*** | |||||
| (0.1671) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.9710*** | |||||
| (0.1742) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.6319*** | |||||
| (0.1723) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.3677** | |||||
| (0.1748) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.3345** | |||||
| (0.1785) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.9799*** | |||||
| (0.1729) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.3120** | |||||
| (0.1842) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3864** | |||||
| (0.1975) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.7992*** | |||||
| (0.1759) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.8581*** | |||||
| (0.1873) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.9999*** | |||||
| (0.1838) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4521*** | |||||
| (0.1747) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.6336*** | |||||
| (0.1722) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2927** | |||||
| (0.1730) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3094** | |||||
| (0.1694) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.5483*** | |||||
| (0.1744) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0715 | |||||
| (0.1739) | ||||||
| Post 2 indicator | -1.3401*** | |||||
| (0.1192) | ||||||
| Post 3 indicator | -2.2672*** | |||||
| (0.1177) | ||||||
| Post 4 indicator | -1.9129*** | |||||
| (0.1187) | ||||||
| Post 5 indicator | -0.0798 | |||||
| (0.1262) | ||||||
| Post 6 indicator | 0.5392*** | |||||
| (0.1363) | ||||||
| Post 7 indicator | -1.5243*** | |||||
| (0.1183) | ||||||
| Post 8 indicator | -0.9858*** | |||||
| (0.1199) | ||||||
| Post 9 indicator | -0.5000*** | |||||
| (0.1233) | ||||||
| Post 10 indicator | -0.8357*** | |||||
| (0.1185) | ||||||
| Post 11 indicator | -1.4147*** | |||||
| (0.1181) | ||||||
| Post 12 indicator | -1.1001*** | |||||
| (0.1181) | ||||||
| Post 13 indicator | -1.6287*** | |||||
| (0.1174) | ||||||
| Post 14 indicator | -2.3481*** | |||||
| (0.1200) | ||||||
| Post 15 indicator | -2.1274*** | |||||
| (0.1188) | ||||||
| Post 16 indicator | 1.7600*** | |||||
| (0.1235) | ||||||
| Post 17 indicator | 0.1690* | |||||
| (0.1305) | ||||||
| Post 18 indicator | -0.0183 | |||||
| (0.1260) | ||||||
| Post 19 indicator | 0.6788*** | |||||
| (0.1237) | ||||||
| Post 20 indicator | 2.3767*** | |||||
| (0.1258) | ||||||
| Post 21 indicator | -0.1614* | |||||
| (0.1231) | ||||||
| Post 22 indicator | 0.1790* | |||||
| (0.1245) | ||||||
| Post 23 indicator | 0.8110*** | |||||
| (0.1252) | ||||||
| Post 24 indicator | 2.5226*** | |||||
| (0.1322) | ||||||
| Post 25 indicator | 1.2168*** | |||||
| (0.1230) | ||||||
| Post 26 indicator | 0.9150*** | |||||
| (0.1271) | ||||||
| Post 27 indicator | 0.5861*** | |||||
| (0.1214) | ||||||
| Post 28 indicator | 0.6185*** | |||||
| (0.1237) | ||||||
| Post 29 indicator | -0.2961** | |||||
| (0.1281) | ||||||
| Post 30 indicator | 0.4685*** | |||||
| (0.1227) | ||||||
| Post 31 indicator | 1.4240*** | |||||
| (0.1231) | ||||||
| Post 32 indicator | -0.3599*** | |||||
| (0.1332) | ||||||
| Post 33 indicator | 0.2143** | |||||
| (0.1198) | ||||||
| Post 34 indicator | 0.2669** | |||||
| (0.1192) | ||||||
| Post 35 indicator | 2.3320*** | |||||
| (0.1294) | ||||||
| Post 36 indicator | 2.9565*** | |||||
| (0.1412) | ||||||
| Post 37 indicator | 0.1110 | |||||
| (0.1242) | ||||||
| Post 38 indicator | -0.3521*** | |||||
| (0.1324) | ||||||
| Post 39 indicator | 2.2031*** | |||||
| (0.1251) | ||||||
| Post 40 indicator | 1.5682*** | |||||
| (0.1241) | ||||||
| Post 41 indicator | -0.0090 | |||||
| (0.1260) | ||||||
| Post 42 indicator | 1.4442*** | |||||
| (0.1245) | ||||||
| Post 43 indicator | 1.3760*** | |||||
| (0.1199) | ||||||
| Post 44 indicator | 0.0469 | |||||
| (0.1266) | ||||||
| Post 45 indicator | 1.9285*** | |||||
| (0.1259) | ||||||
| Post 46 indicator | 2.0559*** | |||||
| (0.1269) | ||||||
| Post 47 indicator | 1.0835*** | |||||
| (0.1221) | ||||||
| Post 48 indicator | -0.5459*** | |||||
| (0.1312) | ||||||
| Post 49 indicator | 1.1230*** | |||||
| (0.1256) | ||||||
| Post 50 indicator | 2.2882*** | |||||
| (0.1289) | ||||||
| Post 51 indicator | 2.9815*** | |||||
| (0.1382) | ||||||
| Post 52 indicator | -0.2676** | |||||
| (0.1291) | ||||||
| Post 53 indicator | -0.7880*** | |||||
| (0.1427) | ||||||
| Post 54 indicator | 2.5558*** | |||||
| (0.1290) | ||||||
| Post 55 indicator | 1.6723*** | |||||
| (0.1267) | ||||||
| Post 56 indicator | 1.2748*** | |||||
| (0.1267) | ||||||
| Post 57 indicator | 1.2485*** | |||||
| (0.1206) | ||||||
| Post 58 indicator | 0.7366*** | |||||
| (0.1200) | ||||||
| Post 59 indicator | 0.2561** | |||||
| (0.1274) | ||||||
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 84465.3730 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -42232.6865 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.0639 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1042* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0793) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Emotions x 1[Post-treatment] | -0.4743*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0657) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0236 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0575) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70340.4988 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35170.2494 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2210 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0465*** | |||||
| (0.0347) | ||||||
| Emotions | -0.0305 | |||||
| (0.0482) | ||||||
| 1[Post-treatment] | -0.4565*** | -0.5002*** | -0.5601*** | -0.5005*** | -0.5646*** | -0.5646*** |
| (0.0476) | (0.0589) | (0.0634) | (0.0589) | (0.0645) | (0.0645) | |
| 1[Misinfo] | -0.8428*** | -0.9234*** | -1.0451*** | -0.6580*** | -1.1149*** | -3.1708*** |
| (0.0414) | (0.0453) | (0.0487) | (0.0266) | (0.1374) | (0.1367) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9274*** | |||||
| (0.0173) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5397*** | |||||
| (0.0166) | ||||||
| Emotions x 1[Post-treatment] | -0.4743*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0657) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0236 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0575) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0526 | 0.0492 | 0.0615 | 0.0485 | 0.0524 | 0.0524 |
| (0.0573) | (0.0617) | (0.0642) | (0.0618) | (0.0653) | (0.0653) | |
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.1364*** | |||||
| (0.0234) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0649*** | |||||
| (0.0229) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1042* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0793) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Post domain 2 | -1.3255*** | -1.3401*** | ||||
| (0.0603) | (0.1192) | |||||
| Post domain 3 | -1.9197*** | -2.2672*** | ||||
| (0.0609) | (0.1177) | |||||
| Post domain 4 | -1.2883*** | -1.9129*** | ||||
| (0.0601) | (0.1187) | |||||
| Post domain 5 | 0.4122*** | -0.0798 | ||||
| (0.0593) | (0.1262) | |||||
| Post domain 6 | 0.1959*** | 0.5392*** | ||||
| (0.0613) | (0.1363) | |||||
| Post domain 7 | -1.5826*** | -1.5243*** | ||||
| (0.0602) | (0.1183) | |||||
| Post domain 8 | -1.4424*** | -0.9858*** | ||||
| (0.0598) | (0.1199) | |||||
| Post domain 9 | 0.3437*** | -0.5000*** | ||||
| (0.0605) | (0.1233) | |||||
| Post domain 10 | -0.4006*** | -0.8357*** | ||||
| (0.0581) | (0.1185) | |||||
| Post domain 11 | -1.0907*** | -1.4147*** | ||||
| (0.0605) | (0.1181) | |||||
| Post domain 12 | -0.7530*** | -1.1001*** | ||||
| (0.0578) | (0.1181) | |||||
| Post domain 13 | -1.0463*** | -1.6287*** | ||||
| (0.0591) | (0.1174) | |||||
| Post domain 14 | -1.8782*** | -2.3481*** | ||||
| (0.0631) | (0.1200) | |||||
| Post domain 15 | -1.2760*** | -2.1274*** | ||||
| (0.0590) | (0.1188) | |||||
| 1[Combo manipulation tactic] | -0.4163*** | -0.2959** | ||||
| (0.0234) | (0.1277) | |||||
| 1[Emotions manipulation tactic] | -0.2120*** | -0.6319*** | ||||
| (0.0312) | (0.1279) | |||||
| 1[Reasoning manipulation tactic] | -0.1699*** | |||||
| (0.0317) | ||||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.2510* | |||||
| (0.1800) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4889*** | |||||
| (0.1724) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.8316*** | |||||
| (0.1703) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6964*** | |||||
| (0.1824) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.4606*** | |||||
| (0.1888) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0567 | |||||
| (0.1748) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0368 | |||||
| (0.1767) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.2625*** | |||||
| (0.1823) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.2924** | |||||
| (0.1725) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.5696*** | |||||
| (0.1735) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0738 | |||||
| (0.1737) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.4872*** | |||||
| (0.1721) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.2920** | |||||
| (0.1753) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.8359*** | |||||
| (0.1715) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4438*** | |||||
| (0.1815) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 1.0575*** | |||||
| (0.1695) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.7557*** | |||||
| (0.1693) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.9877*** | |||||
| (0.1800) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9932*** | |||||
| (0.1963) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.2112 | |||||
| (0.1731) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.7904*** | |||||
| (0.1827) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2790*** | |||||
| (0.1780) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.9799*** | |||||
| (0.1691) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0184 | |||||
| (0.1746) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1203*** | |||||
| (0.1709) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.5807*** | |||||
| (0.1671) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.9710*** | |||||
| (0.1742) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.6319*** | |||||
| (0.1723) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.3677** | |||||
| (0.1748) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.3345** | |||||
| (0.1785) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.9799*** | |||||
| (0.1729) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.3120** | |||||
| (0.1842) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3864** | |||||
| (0.1975) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.7992*** | |||||
| (0.1759) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.8581*** | |||||
| (0.1873) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.9999*** | |||||
| (0.1838) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4521*** | |||||
| (0.1747) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.6336*** | |||||
| (0.1722) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2927** | |||||
| (0.1730) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3094** | |||||
| (0.1694) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.5483*** | |||||
| (0.1744) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0715 | |||||
| (0.1739) | ||||||
| Post 2 indicator | -1.3401*** | |||||
| (0.1192) | ||||||
| Post 3 indicator | -2.2672*** | |||||
| (0.1177) | ||||||
| Post 4 indicator | -1.9129*** | |||||
| (0.1187) | ||||||
| Post 5 indicator | -0.0798 | |||||
| (0.1262) | ||||||
| Post 6 indicator | 0.5392*** | |||||
| (0.1363) | ||||||
| Post 7 indicator | -1.5243*** | |||||
| (0.1183) | ||||||
| Post 8 indicator | -0.9858*** | |||||
| (0.1199) | ||||||
| Post 9 indicator | -0.5000*** | |||||
| (0.1233) | ||||||
| Post 10 indicator | -0.8357*** | |||||
| (0.1185) | ||||||
| Post 11 indicator | -1.4147*** | |||||
| (0.1181) | ||||||
| Post 12 indicator | -1.1001*** | |||||
| (0.1181) | ||||||
| Post 13 indicator | -1.6287*** | |||||
| (0.1174) | ||||||
| Post 14 indicator | -2.3481*** | |||||
| (0.1200) | ||||||
| Post 15 indicator | -2.1274*** | |||||
| (0.1188) | ||||||
| Post 16 indicator | 1.7600*** | |||||
| (0.1235) | ||||||
| Post 17 indicator | 0.1690* | |||||
| (0.1305) | ||||||
| Post 18 indicator | -0.0183 | |||||
| (0.1260) | ||||||
| Post 19 indicator | 0.6788*** | |||||
| (0.1237) | ||||||
| Post 20 indicator | 2.3767*** | |||||
| (0.1258) | ||||||
| Post 21 indicator | -0.1614* | |||||
| (0.1231) | ||||||
| Post 22 indicator | 0.1790* | |||||
| (0.1245) | ||||||
| Post 23 indicator | 0.8110*** | |||||
| (0.1252) | ||||||
| Post 24 indicator | 2.5226*** | |||||
| (0.1322) | ||||||
| Post 25 indicator | 1.2168*** | |||||
| (0.1230) | ||||||
| Post 26 indicator | 0.9150*** | |||||
| (0.1271) | ||||||
| Post 27 indicator | 0.5861*** | |||||
| (0.1214) | ||||||
| Post 28 indicator | 0.6185*** | |||||
| (0.1237) | ||||||
| Post 29 indicator | -0.2961** | |||||
| (0.1281) | ||||||
| Post 30 indicator | 0.4685*** | |||||
| (0.1227) | ||||||
| Post 31 indicator | 1.4240*** | |||||
| (0.1231) | ||||||
| Post 32 indicator | -0.3599*** | |||||
| (0.1332) | ||||||
| Post 33 indicator | 0.2143** | |||||
| (0.1198) | ||||||
| Post 34 indicator | 0.2669** | |||||
| (0.1192) | ||||||
| Post 35 indicator | 2.3320*** | |||||
| (0.1294) | ||||||
| Post 36 indicator | 2.9565*** | |||||
| (0.1412) | ||||||
| Post 37 indicator | 0.1110 | |||||
| (0.1242) | ||||||
| Post 38 indicator | -0.3521*** | |||||
| (0.1324) | ||||||
| Post 39 indicator | 2.2031*** | |||||
| (0.1251) | ||||||
| Post 40 indicator | 1.5682*** | |||||
| (0.1241) | ||||||
| Post 41 indicator | -0.0090 | |||||
| (0.1260) | ||||||
| Post 42 indicator | 1.4442*** | |||||
| (0.1245) | ||||||
| Post 43 indicator | 1.3760*** | |||||
| (0.1199) | ||||||
| Post 44 indicator | 0.0469 | |||||
| (0.1266) | ||||||
| Post 45 indicator | 1.9285*** | |||||
| (0.1259) | ||||||
| Post 46 indicator | 2.0559*** | |||||
| (0.1269) | ||||||
| Post 47 indicator | 1.0835*** | |||||
| (0.1221) | ||||||
| Post 48 indicator | -0.5459*** | |||||
| (0.1312) | ||||||
| Post 49 indicator | 1.1230*** | |||||
| (0.1256) | ||||||
| Post 50 indicator | 2.2882*** | |||||
| (0.1289) | ||||||
| Post 51 indicator | 2.9815*** | |||||
| (0.1382) | ||||||
| Post 52 indicator | -0.2676** | |||||
| (0.1291) | ||||||
| Post 53 indicator | -0.7880*** | |||||
| (0.1427) | ||||||
| Post 54 indicator | 2.5558*** | |||||
| (0.1290) | ||||||
| Post 55 indicator | 1.6723*** | |||||
| (0.1267) | ||||||
| Post 56 indicator | 1.2748*** | |||||
| (0.1267) | ||||||
| Post 57 indicator | 1.2485*** | |||||
| (0.1206) | ||||||
| Post 58 indicator | 0.7366*** | |||||
| (0.1200) | ||||||
| Post 59 indicator | 0.2561** | |||||
| (0.1274) | ||||||
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70340.4988 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35170.2494 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2210 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1038* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0793) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Emotions x 1[Post-treatment] | -0.4747*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0657) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0233 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0575) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70333.2914 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35166.6457 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2210 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0473*** | |||||
| (0.0347) | ||||||
| Emotions | -0.0323 | |||||
| (0.0482) | ||||||
| 1[Post-treatment] | -0.4565*** | -0.5002*** | -0.5601*** | -0.5005*** | -0.5646*** | -0.5646*** |
| (0.0476) | (0.0589) | (0.0634) | (0.0589) | (0.0645) | (0.0645) | |
| 1[Misinfo] | -0.8428*** | -0.9234*** | -1.0451*** | -0.6580*** | -1.1149*** | -3.1708*** |
| (0.0414) | (0.0453) | (0.0487) | (0.0266) | (0.1374) | (0.1367) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9365*** | |||||
| (0.0379) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5408*** | |||||
| (0.0166) | ||||||
| Emotions posts | 0.0116 | |||||
| (0.0241) | ||||||
| Reasoning posts | -0.0236 | |||||
| (0.0237) | ||||||
| Emotions x 1[Post-treatment] | -0.4747*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0657) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0233 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0575) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0526 | 0.0492 | 0.0615 | 0.0485 | 0.0524 | 0.0524 |
| (0.0573) | (0.0617) | (0.0642) | (0.0618) | (0.0653) | (0.0653) | |
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.2211*** | |||||
| (0.0536) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0661*** | |||||
| (0.0229) | ||||||
| Emotions x Emotions posts | 0.0387 | |||||
| (0.0344) | ||||||
| Emotions x Reasoning posts | 0.0671** | |||||
| (0.0341) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1038* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0793) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Post domain 2 | -1.3255*** | -1.3401*** | ||||
| (0.0603) | (0.1192) | |||||
| Post domain 3 | -1.9197*** | -2.2672*** | ||||
| (0.0609) | (0.1177) | |||||
| Post domain 4 | -1.2883*** | -1.9129*** | ||||
| (0.0601) | (0.1187) | |||||
| Post domain 5 | 0.4122*** | -0.0798 | ||||
| (0.0593) | (0.1262) | |||||
| Post domain 6 | 0.1959*** | 0.5392*** | ||||
| (0.0613) | (0.1363) | |||||
| Post domain 7 | -1.5826*** | -1.5243*** | ||||
| (0.0602) | (0.1183) | |||||
| Post domain 8 | -1.4424*** | -0.9858*** | ||||
| (0.0598) | (0.1199) | |||||
| Post domain 9 | 0.3437*** | -0.5000*** | ||||
| (0.0605) | (0.1233) | |||||
| Post domain 10 | -0.4006*** | -0.8357*** | ||||
| (0.0581) | (0.1185) | |||||
| Post domain 11 | -1.0907*** | -1.4147*** | ||||
| (0.0605) | (0.1181) | |||||
| Post domain 12 | -0.7530*** | -1.1001*** | ||||
| (0.0578) | (0.1181) | |||||
| Post domain 13 | -1.0463*** | -1.6287*** | ||||
| (0.0591) | (0.1174) | |||||
| Post domain 14 | -1.8782*** | -2.3481*** | ||||
| (0.0631) | (0.1200) | |||||
| Post domain 15 | -1.2760*** | -2.1274*** | ||||
| (0.0590) | (0.1188) | |||||
| 1[Combo manipulation tactic] | -0.4163*** | -0.2959** | ||||
| (0.0234) | (0.1277) | |||||
| 1[Emotions manipulation tactic] | -0.2120*** | -0.6319*** | ||||
| (0.0312) | (0.1279) | |||||
| 1[Reasoning manipulation tactic] | -0.1699*** | |||||
| (0.0317) | ||||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.2510* | |||||
| (0.1800) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4889*** | |||||
| (0.1724) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.8316*** | |||||
| (0.1703) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6964*** | |||||
| (0.1824) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.4606*** | |||||
| (0.1888) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0567 | |||||
| (0.1748) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0368 | |||||
| (0.1767) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.2625*** | |||||
| (0.1823) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.2924** | |||||
| (0.1725) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.5696*** | |||||
| (0.1735) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0738 | |||||
| (0.1737) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.4872*** | |||||
| (0.1721) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.2920** | |||||
| (0.1753) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.8359*** | |||||
| (0.1715) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4438*** | |||||
| (0.1815) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 1.0575*** | |||||
| (0.1695) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.7557*** | |||||
| (0.1693) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.9877*** | |||||
| (0.1800) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9932*** | |||||
| (0.1963) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.2112 | |||||
| (0.1731) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.7904*** | |||||
| (0.1827) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2790*** | |||||
| (0.1780) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.9799*** | |||||
| (0.1691) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0184 | |||||
| (0.1746) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1203*** | |||||
| (0.1709) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.5807*** | |||||
| (0.1671) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.9710*** | |||||
| (0.1742) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.6319*** | |||||
| (0.1723) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.3677** | |||||
| (0.1748) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.3345** | |||||
| (0.1785) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.9799*** | |||||
| (0.1729) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.3120** | |||||
| (0.1842) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3864** | |||||
| (0.1975) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.7992*** | |||||
| (0.1759) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.8581*** | |||||
| (0.1873) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.9999*** | |||||
| (0.1838) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4521*** | |||||
| (0.1747) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.6336*** | |||||
| (0.1722) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2927** | |||||
| (0.1730) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3094** | |||||
| (0.1694) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.5483*** | |||||
| (0.1744) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0715 | |||||
| (0.1739) | ||||||
| Post 2 indicator | -1.3401*** | |||||
| (0.1192) | ||||||
| Post 3 indicator | -2.2672*** | |||||
| (0.1177) | ||||||
| Post 4 indicator | -1.9129*** | |||||
| (0.1187) | ||||||
| Post 5 indicator | -0.0798 | |||||
| (0.1262) | ||||||
| Post 6 indicator | 0.5392*** | |||||
| (0.1363) | ||||||
| Post 7 indicator | -1.5243*** | |||||
| (0.1183) | ||||||
| Post 8 indicator | -0.9858*** | |||||
| (0.1199) | ||||||
| Post 9 indicator | -0.5000*** | |||||
| (0.1233) | ||||||
| Post 10 indicator | -0.8357*** | |||||
| (0.1185) | ||||||
| Post 11 indicator | -1.4147*** | |||||
| (0.1181) | ||||||
| Post 12 indicator | -1.1001*** | |||||
| (0.1181) | ||||||
| Post 13 indicator | -1.6287*** | |||||
| (0.1174) | ||||||
| Post 14 indicator | -2.3481*** | |||||
| (0.1200) | ||||||
| Post 15 indicator | -2.1274*** | |||||
| (0.1188) | ||||||
| Post 16 indicator | 1.7600*** | |||||
| (0.1235) | ||||||
| Post 17 indicator | 0.1690* | |||||
| (0.1305) | ||||||
| Post 18 indicator | -0.0183 | |||||
| (0.1260) | ||||||
| Post 19 indicator | 0.6788*** | |||||
| (0.1237) | ||||||
| Post 20 indicator | 2.3767*** | |||||
| (0.1258) | ||||||
| Post 21 indicator | -0.1614* | |||||
| (0.1231) | ||||||
| Post 22 indicator | 0.1790* | |||||
| (0.1245) | ||||||
| Post 23 indicator | 0.8110*** | |||||
| (0.1252) | ||||||
| Post 24 indicator | 2.5226*** | |||||
| (0.1322) | ||||||
| Post 25 indicator | 1.2168*** | |||||
| (0.1230) | ||||||
| Post 26 indicator | 0.9150*** | |||||
| (0.1271) | ||||||
| Post 27 indicator | 0.5861*** | |||||
| (0.1214) | ||||||
| Post 28 indicator | 0.6185*** | |||||
| (0.1237) | ||||||
| Post 29 indicator | -0.2961** | |||||
| (0.1281) | ||||||
| Post 30 indicator | 0.4685*** | |||||
| (0.1227) | ||||||
| Post 31 indicator | 1.4240*** | |||||
| (0.1231) | ||||||
| Post 32 indicator | -0.3599*** | |||||
| (0.1332) | ||||||
| Post 33 indicator | 0.2143** | |||||
| (0.1198) | ||||||
| Post 34 indicator | 0.2669** | |||||
| (0.1192) | ||||||
| Post 35 indicator | 2.3320*** | |||||
| (0.1294) | ||||||
| Post 36 indicator | 2.9565*** | |||||
| (0.1412) | ||||||
| Post 37 indicator | 0.1110 | |||||
| (0.1242) | ||||||
| Post 38 indicator | -0.3521*** | |||||
| (0.1324) | ||||||
| Post 39 indicator | 2.2031*** | |||||
| (0.1251) | ||||||
| Post 40 indicator | 1.5682*** | |||||
| (0.1241) | ||||||
| Post 41 indicator | -0.0090 | |||||
| (0.1260) | ||||||
| Post 42 indicator | 1.4442*** | |||||
| (0.1245) | ||||||
| Post 43 indicator | 1.3760*** | |||||
| (0.1199) | ||||||
| Post 44 indicator | 0.0469 | |||||
| (0.1266) | ||||||
| Post 45 indicator | 1.9285*** | |||||
| (0.1259) | ||||||
| Post 46 indicator | 2.0559*** | |||||
| (0.1269) | ||||||
| Post 47 indicator | 1.0835*** | |||||
| (0.1221) | ||||||
| Post 48 indicator | -0.5459*** | |||||
| (0.1312) | ||||||
| Post 49 indicator | 1.1230*** | |||||
| (0.1256) | ||||||
| Post 50 indicator | 2.2882*** | |||||
| (0.1289) | ||||||
| Post 51 indicator | 2.9815*** | |||||
| (0.1382) | ||||||
| Post 52 indicator | -0.2676** | |||||
| (0.1291) | ||||||
| Post 53 indicator | -0.7880*** | |||||
| (0.1427) | ||||||
| Post 54 indicator | 2.5558*** | |||||
| (0.1290) | ||||||
| Post 55 indicator | 1.6723*** | |||||
| (0.1267) | ||||||
| Post 56 indicator | 1.2748*** | |||||
| (0.1267) | ||||||
| Post 57 indicator | 1.2485*** | |||||
| (0.1206) | ||||||
| Post 58 indicator | 0.7366*** | |||||
| (0.1200) | ||||||
| Post 59 indicator | 0.2561** | |||||
| (0.1274) | ||||||
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70333.2914 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35166.6457 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2210 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1044* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0795) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Emotions x 1[Post-treatment] | -0.4777*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0659) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0224 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0577) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70055.4124 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35027.7062 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2234 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0563*** | |||||
| (0.0349) | ||||||
| Emotions | -0.0355 | |||||
| (0.0484) | ||||||
| 1[Post-treatment] | -0.4587*** | -0.5002*** | -0.5601*** | -0.5005*** | -0.5646*** | -0.5646*** |
| (0.0477) | (0.0589) | (0.0634) | (0.0589) | (0.0645) | (0.0645) | |
| 1[Misinfo] | -0.8466*** | -0.9234*** | -1.0451*** | -0.6580*** | -1.1149*** | -3.1708*** |
| (0.0416) | (0.0453) | (0.0487) | (0.0266) | (0.1374) | (0.1367) | |
| Age | -0.0316** | |||||
| (0.0151) | ||||||
| Man | 0.0151 | |||||
| (0.0135) | ||||||
| Some college | -0.0110 | |||||
| (0.0168) | ||||||
| Bachelor’s degree | 0.0218* | |||||
| (0.0170) | ||||||
| Graduate degree | -0.0412*** | |||||
| (0.0146) | ||||||
| Married | -0.0154 | |||||
| (0.0151) | ||||||
| Unemployed | -0.0280* | |||||
| (0.0177) | ||||||
| Employed | -0.0483*** | |||||
| (0.0193) | ||||||
| Mostly urban | -0.0277** | |||||
| (0.0161) | ||||||
| Suburban | -0.0285** | |||||
| (0.0157) | ||||||
| Christian | 0.0142 | |||||
| (0.0138) | ||||||
| Attends religious services | 0.0051 | |||||
| (0.0139) | ||||||
| Uses social media | -0.0345*** | |||||
| (0.0123) | ||||||
| Hours spent on Social Media | 0.0297** | |||||
| (0.0149) | ||||||
| Social Media Share: 80-100% | 0.0846*** | |||||
| (0.0174) | ||||||
| Social Media Share: 60-80% | 0.1020*** | |||||
| (0.0189) | ||||||
| Social Media Share: 40-60% | 0.0893*** | |||||
| (0.0192) | ||||||
| Social Media Share: 20-40% | 0.0480*** | |||||
| (0.0176) | ||||||
| Passed attention check at pre-treatment | -0.0276** | |||||
| (0.0135) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9020*** | |||||
| (0.0178) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5294*** | |||||
| (0.0167) | ||||||
| Emotions x 1[Post-treatment] | -0.4777*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0659) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0224 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0577) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0533 | 0.0492 | 0.0615 | 0.0485 | 0.0524 | 0.0524 |
| (0.0574) | (0.0617) | (0.0642) | (0.0618) | (0.0653) | (0.0653) | |
| Emotions x Age | -0.0271 | |||||
| (0.0231) | ||||||
| Emotions x Man | 0.0319** | |||||
| (0.0188) | ||||||
| Emotions x Some college | 0.0018 | |||||
| (0.0231) | ||||||
| Emotions x Bachelor’s degree | -0.0197 | |||||
| (0.0232) | ||||||
| Emotions x Graduate degree | 0.0484*** | |||||
| (0.0193) | ||||||
| Emotions x Married | -0.0250 | |||||
| (0.0212) | ||||||
| Emotions x Unemployed | -0.0060 | |||||
| (0.0244) | ||||||
| Emotions x Employed | 0.0147 | |||||
| (0.0266) | ||||||
| Emotions x Mostly urban | 0.0429** | |||||
| (0.0222) | ||||||
| Emotions x Suburban | -0.0094 | |||||
| (0.0219) | ||||||
| Emotions x Christian | -0.0132 | |||||
| (0.0196) | ||||||
| Emotions x Attends religious services | -0.0395** | |||||
| (0.0196) | ||||||
| Emotions x Uses social media | 0.0530*** | |||||
| (0.0197) | ||||||
| Emotions x Hours spent on Social Media | -0.0167 | |||||
| (0.0203) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0054 | |||||
| (0.0241) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0296 | |||||
| (0.0265) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0100 | |||||
| (0.0269) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0225 | |||||
| (0.0248) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0025 | |||||
| (0.0188) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.1443*** | |||||
| (0.0242) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0526** | |||||
| (0.0231) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1044* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0795) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Post domain 2 | -1.3255*** | -1.3401*** | ||||
| (0.0603) | (0.1192) | |||||
| Post domain 3 | -1.9197*** | -2.2672*** | ||||
| (0.0609) | (0.1177) | |||||
| Post domain 4 | -1.2883*** | -1.9129*** | ||||
| (0.0601) | (0.1187) | |||||
| Post domain 5 | 0.4122*** | -0.0798 | ||||
| (0.0593) | (0.1262) | |||||
| Post domain 6 | 0.1959*** | 0.5392*** | ||||
| (0.0613) | (0.1363) | |||||
| Post domain 7 | -1.5826*** | -1.5243*** | ||||
| (0.0602) | (0.1183) | |||||
| Post domain 8 | -1.4424*** | -0.9858*** | ||||
| (0.0598) | (0.1199) | |||||
| Post domain 9 | 0.3437*** | -0.5000*** | ||||
| (0.0605) | (0.1233) | |||||
| Post domain 10 | -0.4006*** | -0.8357*** | ||||
| (0.0581) | (0.1185) | |||||
| Post domain 11 | -1.0907*** | -1.4147*** | ||||
| (0.0605) | (0.1181) | |||||
| Post domain 12 | -0.7530*** | -1.1001*** | ||||
| (0.0578) | (0.1181) | |||||
| Post domain 13 | -1.0463*** | -1.6287*** | ||||
| (0.0591) | (0.1174) | |||||
| Post domain 14 | -1.8782*** | -2.3481*** | ||||
| (0.0631) | (0.1200) | |||||
| Post domain 15 | -1.2760*** | -2.1274*** | ||||
| (0.0590) | (0.1188) | |||||
| 1[Combo manipulation tactic] | -0.4163*** | -0.2959** | ||||
| (0.0234) | (0.1277) | |||||
| 1[Emotions manipulation tactic] | -0.2120*** | -0.6319*** | ||||
| (0.0312) | (0.1279) | |||||
| 1[Reasoning manipulation tactic] | -0.1699*** | |||||
| (0.0317) | ||||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.2510* | |||||
| (0.1800) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4889*** | |||||
| (0.1724) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.8316*** | |||||
| (0.1703) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6964*** | |||||
| (0.1824) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.4606*** | |||||
| (0.1888) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0567 | |||||
| (0.1748) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0368 | |||||
| (0.1767) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.2625*** | |||||
| (0.1823) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.2924** | |||||
| (0.1725) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.5696*** | |||||
| (0.1735) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0738 | |||||
| (0.1737) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.4872*** | |||||
| (0.1721) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.2920** | |||||
| (0.1753) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.8359*** | |||||
| (0.1715) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4438*** | |||||
| (0.1815) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 1.0575*** | |||||
| (0.1695) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.7557*** | |||||
| (0.1693) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.9877*** | |||||
| (0.1800) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9932*** | |||||
| (0.1963) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.2112 | |||||
| (0.1731) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.7904*** | |||||
| (0.1827) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2790*** | |||||
| (0.1780) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.9799*** | |||||
| (0.1691) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0184 | |||||
| (0.1746) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1203*** | |||||
| (0.1709) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.5807*** | |||||
| (0.1671) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.9710*** | |||||
| (0.1742) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.6319*** | |||||
| (0.1723) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.3677** | |||||
| (0.1748) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.3345** | |||||
| (0.1785) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.9799*** | |||||
| (0.1729) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.3120** | |||||
| (0.1842) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3864** | |||||
| (0.1975) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.7992*** | |||||
| (0.1759) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.8581*** | |||||
| (0.1873) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.9999*** | |||||
| (0.1838) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4521*** | |||||
| (0.1747) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.6336*** | |||||
| (0.1722) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2927** | |||||
| (0.1730) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3094** | |||||
| (0.1694) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.5483*** | |||||
| (0.1744) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0715 | |||||
| (0.1739) | ||||||
| Post 2 indicator | -1.3401*** | |||||
| (0.1192) | ||||||
| Post 3 indicator | -2.2672*** | |||||
| (0.1177) | ||||||
| Post 4 indicator | -1.9129*** | |||||
| (0.1187) | ||||||
| Post 5 indicator | -0.0798 | |||||
| (0.1262) | ||||||
| Post 6 indicator | 0.5392*** | |||||
| (0.1363) | ||||||
| Post 7 indicator | -1.5243*** | |||||
| (0.1183) | ||||||
| Post 8 indicator | -0.9858*** | |||||
| (0.1199) | ||||||
| Post 9 indicator | -0.5000*** | |||||
| (0.1233) | ||||||
| Post 10 indicator | -0.8357*** | |||||
| (0.1185) | ||||||
| Post 11 indicator | -1.4147*** | |||||
| (0.1181) | ||||||
| Post 12 indicator | -1.1001*** | |||||
| (0.1181) | ||||||
| Post 13 indicator | -1.6287*** | |||||
| (0.1174) | ||||||
| Post 14 indicator | -2.3481*** | |||||
| (0.1200) | ||||||
| Post 15 indicator | -2.1274*** | |||||
| (0.1188) | ||||||
| Post 16 indicator | 1.7600*** | |||||
| (0.1235) | ||||||
| Post 17 indicator | 0.1690* | |||||
| (0.1305) | ||||||
| Post 18 indicator | -0.0183 | |||||
| (0.1260) | ||||||
| Post 19 indicator | 0.6788*** | |||||
| (0.1237) | ||||||
| Post 20 indicator | 2.3767*** | |||||
| (0.1258) | ||||||
| Post 21 indicator | -0.1614* | |||||
| (0.1231) | ||||||
| Post 22 indicator | 0.1790* | |||||
| (0.1245) | ||||||
| Post 23 indicator | 0.8110*** | |||||
| (0.1252) | ||||||
| Post 24 indicator | 2.5226*** | |||||
| (0.1322) | ||||||
| Post 25 indicator | 1.2168*** | |||||
| (0.1230) | ||||||
| Post 26 indicator | 0.9150*** | |||||
| (0.1271) | ||||||
| Post 27 indicator | 0.5861*** | |||||
| (0.1214) | ||||||
| Post 28 indicator | 0.6185*** | |||||
| (0.1237) | ||||||
| Post 29 indicator | -0.2961** | |||||
| (0.1281) | ||||||
| Post 30 indicator | 0.4685*** | |||||
| (0.1227) | ||||||
| Post 31 indicator | 1.4240*** | |||||
| (0.1231) | ||||||
| Post 32 indicator | -0.3599*** | |||||
| (0.1332) | ||||||
| Post 33 indicator | 0.2143** | |||||
| (0.1198) | ||||||
| Post 34 indicator | 0.2669** | |||||
| (0.1192) | ||||||
| Post 35 indicator | 2.3320*** | |||||
| (0.1294) | ||||||
| Post 36 indicator | 2.9565*** | |||||
| (0.1412) | ||||||
| Post 37 indicator | 0.1110 | |||||
| (0.1242) | ||||||
| Post 38 indicator | -0.3521*** | |||||
| (0.1324) | ||||||
| Post 39 indicator | 2.2031*** | |||||
| (0.1251) | ||||||
| Post 40 indicator | 1.5682*** | |||||
| (0.1241) | ||||||
| Post 41 indicator | -0.0090 | |||||
| (0.1260) | ||||||
| Post 42 indicator | 1.4442*** | |||||
| (0.1245) | ||||||
| Post 43 indicator | 1.3760*** | |||||
| (0.1199) | ||||||
| Post 44 indicator | 0.0469 | |||||
| (0.1266) | ||||||
| Post 45 indicator | 1.9285*** | |||||
| (0.1259) | ||||||
| Post 46 indicator | 2.0559*** | |||||
| (0.1269) | ||||||
| Post 47 indicator | 1.0835*** | |||||
| (0.1221) | ||||||
| Post 48 indicator | -0.5459*** | |||||
| (0.1312) | ||||||
| Post 49 indicator | 1.1230*** | |||||
| (0.1256) | ||||||
| Post 50 indicator | 2.2882*** | |||||
| (0.1289) | ||||||
| Post 51 indicator | 2.9815*** | |||||
| (0.1382) | ||||||
| Post 52 indicator | -0.2676** | |||||
| (0.1291) | ||||||
| Post 53 indicator | -0.7880*** | |||||
| (0.1427) | ||||||
| Post 54 indicator | 2.5558*** | |||||
| (0.1290) | ||||||
| Post 55 indicator | 1.6723*** | |||||
| (0.1267) | ||||||
| Post 56 indicator | 1.2748*** | |||||
| (0.1267) | ||||||
| Post 57 indicator | 1.2485*** | |||||
| (0.1206) | ||||||
| Post 58 indicator | 0.7366*** | |||||
| (0.1200) | ||||||
| Post 59 indicator | 0.2561** | |||||
| (0.1274) | ||||||
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70055.4124 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35027.7062 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2234 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates + Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1040* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0795) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Emotions x 1[Post-treatment] | -0.4780*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0659) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0222 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0577) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70049.3792 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35024.6896 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2233 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0571*** | |||||
| (0.0349) | ||||||
| Emotions | -0.0372 | |||||
| (0.0484) | ||||||
| 1[Post-treatment] | -0.4587*** | -0.5002*** | -0.5601*** | -0.5005*** | -0.5646*** | -0.5646*** |
| (0.0477) | (0.0589) | (0.0634) | (0.0589) | (0.0645) | (0.0645) | |
| 1[Misinfo] | -0.8466*** | -0.9234*** | -1.0451*** | -0.6580*** | -1.1149*** | -3.1708*** |
| (0.0416) | (0.0453) | (0.0487) | (0.0266) | (0.1374) | (0.1367) | |
| Age | -0.0315** | |||||
| (0.0152) | ||||||
| Man | 0.0152 | |||||
| (0.0135) | ||||||
| Some college | -0.0104 | |||||
| (0.0168) | ||||||
| Bachelor’s degree | 0.0221* | |||||
| (0.0170) | ||||||
| Graduate degree | -0.0415*** | |||||
| (0.0146) | ||||||
| Married | -0.0146 | |||||
| (0.0151) | ||||||
| Unemployed | -0.0281* | |||||
| (0.0177) | ||||||
| Employed | -0.0491*** | |||||
| (0.0193) | ||||||
| Mostly urban | -0.0265** | |||||
| (0.0161) | ||||||
| Suburban | -0.0280** | |||||
| (0.0157) | ||||||
| Christian | 0.0137 | |||||
| (0.0138) | ||||||
| Attends religious services | 0.0052 | |||||
| (0.0139) | ||||||
| Uses social media | -0.0347*** | |||||
| (0.0123) | ||||||
| Hours spent on Social Media | 0.0304** | |||||
| (0.0149) | ||||||
| Social Media Share: 80-100% | 0.0846*** | |||||
| (0.0174) | ||||||
| Social Media Share: 60-80% | 0.1021*** | |||||
| (0.0189) | ||||||
| Social Media Share: 40-60% | 0.0896*** | |||||
| (0.0192) | ||||||
| Social Media Share: 20-40% | 0.0480*** | |||||
| (0.0176) | ||||||
| Passed attention check at pre-treatment | -0.0272** | |||||
| (0.0135) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9030*** | |||||
| (0.0384) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5304*** | |||||
| (0.0167) | ||||||
| Emotions posts | 0.0165 | |||||
| (0.0243) | ||||||
| Reasoning posts | -0.0183 | |||||
| (0.0238) | ||||||
| Emotions x 1[Post-treatment] | -0.4780*** | -0.5257*** | -0.5530*** | -0.5260*** | -0.5747*** | -0.5747*** |
| (0.0659) | (0.0839) | (0.0904) | (0.0840) | (0.0921) | (0.0921) | |
| Emotions x 1[Misinfo] | 0.0222 | 0.0199 | 0.0586 | 0.0201 | 0.0607 | 0.0607 |
| (0.0577) | (0.0638) | (0.0680) | (0.0638) | (0.0697) | (0.0697) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0533 | 0.0492 | 0.0615 | 0.0485 | 0.0524 | 0.0524 |
| (0.0574) | (0.0617) | (0.0642) | (0.0618) | (0.0653) | (0.0653) | |
| Emotions x Age | -0.0262 | |||||
| (0.0231) | ||||||
| Emotions x Man | 0.0308* | |||||
| (0.0188) | ||||||
| Emotions x Some college | 0.0012 | |||||
| (0.0232) | ||||||
| Emotions x Bachelor’s degree | -0.0207 | |||||
| (0.0232) | ||||||
| Emotions x Graduate degree | 0.0486*** | |||||
| (0.0193) | ||||||
| Emotions x Married | -0.0266 | |||||
| (0.0212) | ||||||
| Emotions x Unemployed | -0.0072 | |||||
| (0.0244) | ||||||
| Emotions x Employed | 0.0150 | |||||
| (0.0266) | ||||||
| Emotions x Mostly urban | 0.0410** | |||||
| (0.0222) | ||||||
| Emotions x Suburban | -0.0099 | |||||
| (0.0219) | ||||||
| Emotions x Christian | -0.0129 | |||||
| (0.0196) | ||||||
| Emotions x Attends religious services | -0.0397** | |||||
| (0.0196) | ||||||
| Emotions x Uses social media | 0.0526*** | |||||
| (0.0197) | ||||||
| Emotions x Hours spent on Social Media | -0.0178 | |||||
| (0.0204) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0057 | |||||
| (0.0241) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0286 | |||||
| (0.0265) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0123 | |||||
| (0.0269) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0206 | |||||
| (0.0248) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0034 | |||||
| (0.0189) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.2131*** | |||||
| (0.0543) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0537** | |||||
| (0.0231) | ||||||
| Emotions x Emotions posts | 0.0285 | |||||
| (0.0346) | ||||||
| Emotions x Reasoning posts | 0.0574** | |||||
| (0.0344) | ||||||
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1040* | -0.1220* | -0.1895** | -0.1231* | -0.2363*** | -0.2363*** |
| (0.0795) | (0.0857) | (0.0897) | (0.0858) | (0.0915) | (0.0915) | |
| Post domain 2 | -1.3255*** | -1.3401*** | ||||
| (0.0603) | (0.1192) | |||||
| Post domain 3 | -1.9197*** | -2.2672*** | ||||
| (0.0609) | (0.1177) | |||||
| Post domain 4 | -1.2883*** | -1.9129*** | ||||
| (0.0601) | (0.1187) | |||||
| Post domain 5 | 0.4122*** | -0.0798 | ||||
| (0.0593) | (0.1262) | |||||
| Post domain 6 | 0.1959*** | 0.5392*** | ||||
| (0.0613) | (0.1363) | |||||
| Post domain 7 | -1.5826*** | -1.5243*** | ||||
| (0.0602) | (0.1183) | |||||
| Post domain 8 | -1.4424*** | -0.9858*** | ||||
| (0.0598) | (0.1199) | |||||
| Post domain 9 | 0.3437*** | -0.5000*** | ||||
| (0.0605) | (0.1233) | |||||
| Post domain 10 | -0.4006*** | -0.8357*** | ||||
| (0.0581) | (0.1185) | |||||
| Post domain 11 | -1.0907*** | -1.4147*** | ||||
| (0.0605) | (0.1181) | |||||
| Post domain 12 | -0.7530*** | -1.1001*** | ||||
| (0.0578) | (0.1181) | |||||
| Post domain 13 | -1.0463*** | -1.6287*** | ||||
| (0.0591) | (0.1174) | |||||
| Post domain 14 | -1.8782*** | -2.3481*** | ||||
| (0.0631) | (0.1200) | |||||
| Post domain 15 | -1.2760*** | -2.1274*** | ||||
| (0.0590) | (0.1188) | |||||
| 1[Combo manipulation tactic] | -0.4163*** | -0.2959** | ||||
| (0.0234) | (0.1277) | |||||
| 1[Emotions manipulation tactic] | -0.2120*** | -0.6319*** | ||||
| (0.0312) | (0.1279) | |||||
| 1[Reasoning manipulation tactic] | -0.1699*** | |||||
| (0.0317) | ||||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.2510* | |||||
| (0.1800) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4889*** | |||||
| (0.1724) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.8316*** | |||||
| (0.1703) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6964*** | |||||
| (0.1824) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.4606*** | |||||
| (0.1888) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.0567 | |||||
| (0.1748) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0368 | |||||
| (0.1767) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.2625*** | |||||
| (0.1823) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.2924** | |||||
| (0.1725) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.5696*** | |||||
| (0.1735) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | -0.0738 | |||||
| (0.1737) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.4872*** | |||||
| (0.1721) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.2920** | |||||
| (0.1753) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.8359*** | |||||
| (0.1715) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4438*** | |||||
| (0.1815) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 1.0575*** | |||||
| (0.1695) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.7557*** | |||||
| (0.1693) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.9877*** | |||||
| (0.1800) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9932*** | |||||
| (0.1963) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | 0.2112 | |||||
| (0.1731) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.7904*** | |||||
| (0.1827) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2790*** | |||||
| (0.1780) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.9799*** | |||||
| (0.1691) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0184 | |||||
| (0.1746) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1203*** | |||||
| (0.1709) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.5807*** | |||||
| (0.1671) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.9710*** | |||||
| (0.1742) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.6319*** | |||||
| (0.1723) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.3677** | |||||
| (0.1748) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.3345** | |||||
| (0.1785) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.9799*** | |||||
| (0.1729) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.3120** | |||||
| (0.1842) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3864** | |||||
| (0.1975) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.7992*** | |||||
| (0.1759) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.8581*** | |||||
| (0.1873) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 0.9999*** | |||||
| (0.1838) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4521*** | |||||
| (0.1747) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.6336*** | |||||
| (0.1722) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2927** | |||||
| (0.1730) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3094** | |||||
| (0.1694) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.5483*** | |||||
| (0.1744) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0715 | |||||
| (0.1739) | ||||||
| Post 2 indicator | -1.3401*** | |||||
| (0.1192) | ||||||
| Post 3 indicator | -2.2672*** | |||||
| (0.1177) | ||||||
| Post 4 indicator | -1.9129*** | |||||
| (0.1187) | ||||||
| Post 5 indicator | -0.0798 | |||||
| (0.1262) | ||||||
| Post 6 indicator | 0.5392*** | |||||
| (0.1363) | ||||||
| Post 7 indicator | -1.5243*** | |||||
| (0.1183) | ||||||
| Post 8 indicator | -0.9858*** | |||||
| (0.1199) | ||||||
| Post 9 indicator | -0.5000*** | |||||
| (0.1233) | ||||||
| Post 10 indicator | -0.8357*** | |||||
| (0.1185) | ||||||
| Post 11 indicator | -1.4147*** | |||||
| (0.1181) | ||||||
| Post 12 indicator | -1.1001*** | |||||
| (0.1181) | ||||||
| Post 13 indicator | -1.6287*** | |||||
| (0.1174) | ||||||
| Post 14 indicator | -2.3481*** | |||||
| (0.1200) | ||||||
| Post 15 indicator | -2.1274*** | |||||
| (0.1188) | ||||||
| Post 16 indicator | 1.7600*** | |||||
| (0.1235) | ||||||
| Post 17 indicator | 0.1690* | |||||
| (0.1305) | ||||||
| Post 18 indicator | -0.0183 | |||||
| (0.1260) | ||||||
| Post 19 indicator | 0.6788*** | |||||
| (0.1237) | ||||||
| Post 20 indicator | 2.3767*** | |||||
| (0.1258) | ||||||
| Post 21 indicator | -0.1614* | |||||
| (0.1231) | ||||||
| Post 22 indicator | 0.1790* | |||||
| (0.1245) | ||||||
| Post 23 indicator | 0.8110*** | |||||
| (0.1252) | ||||||
| Post 24 indicator | 2.5226*** | |||||
| (0.1322) | ||||||
| Post 25 indicator | 1.2168*** | |||||
| (0.1230) | ||||||
| Post 26 indicator | 0.9150*** | |||||
| (0.1271) | ||||||
| Post 27 indicator | 0.5861*** | |||||
| (0.1214) | ||||||
| Post 28 indicator | 0.6185*** | |||||
| (0.1237) | ||||||
| Post 29 indicator | -0.2961** | |||||
| (0.1281) | ||||||
| Post 30 indicator | 0.4685*** | |||||
| (0.1227) | ||||||
| Post 31 indicator | 1.4240*** | |||||
| (0.1231) | ||||||
| Post 32 indicator | -0.3599*** | |||||
| (0.1332) | ||||||
| Post 33 indicator | 0.2143** | |||||
| (0.1198) | ||||||
| Post 34 indicator | 0.2669** | |||||
| (0.1192) | ||||||
| Post 35 indicator | 2.3320*** | |||||
| (0.1294) | ||||||
| Post 36 indicator | 2.9565*** | |||||
| (0.1412) | ||||||
| Post 37 indicator | 0.1110 | |||||
| (0.1242) | ||||||
| Post 38 indicator | -0.3521*** | |||||
| (0.1324) | ||||||
| Post 39 indicator | 2.2031*** | |||||
| (0.1251) | ||||||
| Post 40 indicator | 1.5682*** | |||||
| (0.1241) | ||||||
| Post 41 indicator | -0.0090 | |||||
| (0.1260) | ||||||
| Post 42 indicator | 1.4442*** | |||||
| (0.1245) | ||||||
| Post 43 indicator | 1.3760*** | |||||
| (0.1199) | ||||||
| Post 44 indicator | 0.0469 | |||||
| (0.1266) | ||||||
| Post 45 indicator | 1.9285*** | |||||
| (0.1259) | ||||||
| Post 46 indicator | 2.0559*** | |||||
| (0.1269) | ||||||
| Post 47 indicator | 1.0835*** | |||||
| (0.1221) | ||||||
| Post 48 indicator | -0.5459*** | |||||
| (0.1312) | ||||||
| Post 49 indicator | 1.1230*** | |||||
| (0.1256) | ||||||
| Post 50 indicator | 2.2882*** | |||||
| (0.1289) | ||||||
| Post 51 indicator | 2.9815*** | |||||
| (0.1382) | ||||||
| Post 52 indicator | -0.2676** | |||||
| (0.1291) | ||||||
| Post 53 indicator | -0.7880*** | |||||
| (0.1427) | ||||||
| Post 54 indicator | 2.5558*** | |||||
| (0.1290) | ||||||
| Post 55 indicator | 1.6723*** | |||||
| (0.1267) | ||||||
| Post 56 indicator | 1.2748*** | |||||
| (0.1267) | ||||||
| Post 57 indicator | 1.2485*** | |||||
| (0.1206) | ||||||
| Post 58 indicator | 0.7366*** | |||||
| (0.1200) | ||||||
| Post 59 indicator | 0.2561** | |||||
| (0.1274) | ||||||
| Num. obs. | 65268 | 58068 | 58068 | 58068 | 58068 | 58068 |
| Deviance | 70049.3792 | 62633.2434 | 57225.1129 | 62550.9019 | 54895.4934 | 54895.4934 |
| Log Likelihood | -35024.6896 | -31316.6217 | -28612.5565 | -31275.4510 | -27447.7467 | -27447.7467 |
| Pseudo R2 | 0.2233 | 0.1412 | 0.2080 | 0.1421 | 0.2359 | 0.2359 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Prediction
Standard errors of the share predicted value are estimated using the Delta method.
Treatment effects in the last column refer to estimated marginal means (EMMs, or least-squares means) for specified factor combinations (Post-treatment and Misinfo). Point estimates are reported in the upper cell and standard errors, clustered at the user level, reported in parentheses in the bottom cell. z-ratios are reported in brackets.
Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
cov_feglm <- c(paste0("share ~ treatment"),
paste0("share ~ treatment + ",demographics_demeaned_treatment),
paste0("share ~ treatment + factor(fact) + ",sharing_demeaned_treatment),
paste0("share ~ treatment + factor(type) + ",sharing_tactic_demeaned_treatment),
paste0("share ~ treatment + factor(fact)*factor(type) + ",demographics_sharing_demeaned_treatment),
paste0("share ~ treatment + factor(post_id) + ",all_covariates_demeaned_treatment))
if (file.exists("prediction_2treat_logit_df.rds")) {
prediction_2treat_logit_df <- readRDS("prediction_2treat_logit_df.rds")
counter <- 1
for (cov in seq_along(covariate_list)) {
list_models <- cov_estimates_list[[cov]]
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
# obtain prediction_df
prediction_df <- prediction_2treat_logit_df[[counter]]
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))
print(k)
cat('<br>')
counter <- counter + 1
}
cat('<br>\n<br>\n<br>')
}
} else {
prediction_2treat_logit_df <- list()
for (cov in seq_along(covariate_list)) {
list_models <- cov_estimates_list[[cov]]
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
treatmentLevels <- unique(panel_2$treatment)
posts <- c(0, 1)
misinfos <- c(0, 1)
prediction_df <- expand.grid(treatment = treatmentLevels, post = posts, post_is_misinfo = misinfos) %>%
mutate(prediction = NA_real_, prediction_se = NA_real_)
for (row in 1:nrow(prediction_df)) {
treat <- prediction_df$treatment[row]
post <- prediction_df$post[row]
misinfo <- prediction_df$post_is_misinfo[row]
# Filter the relevant data
relevant_data <- panel_2 %>%
filter(row_number() %in% obs(list_models[[m]])) %>%
filter(treatment == treat & post == post & post_is_misinfo == misinfo)
# Prepare predictor variables based on model requirements
predictorVariables <- relevant_data %>%
distinct(user, .keep_all = TRUE) %>%
group_by(user, type, fact, post_id) %>%
summarise(age_cov = mean(age_cov, na.rm = TRUE),
gender_Man_cov = mean(gender_Man_cov, na.rm = TRUE),
education_Some_college_cov = mean(education_Some_college_cov, na.rm = TRUE),
education_Bachelor_degree_cov = mean(education_Bachelor_degree_cov, na.rm = TRUE),
education_Graduate_degree_cov = mean(education_Graduate_degree_cov, na.rm = TRUE),
marital_Married_or_in_a_domestic_partnership_cov = mean(marital_Married_or_in_a_domestic_partnership_cov, na.rm = TRUE),
employment_Unemployed_cov = mean(employment_Unemployed_cov, na.rm = TRUE),
employment_Employed_cov = mean(employment_Employed_cov, na.rm = TRUE),
location_Mostly_urban_cov = mean(location_Mostly_urban_cov, na.rm = TRUE),
location_Suburban_cov = mean(location_Suburban_cov, na.rm = TRUE),
religion_Christian_cov = mean(religion_Christian_cov, na.rm = TRUE),
religiosity_Attends_cov = mean(religiosity_Attends_cov, na.rm = TRUE),
social_media_bin_Yes_cov = mean(social_media_bin_Yes_cov, na.rm = TRUE),
social_media_hours_cov = mean(social_media_hours_cov, na.rm = TRUE),
att_check_pre_cov = mean(att_check_pre_cov, na.rm = TRUE),
social_media_share_80_100_cov = mean(social_media_share_80_100_cov, na.rm = TRUE),
social_media_share_60_80_cov = mean(social_media_share_60_80_cov, na.rm = TRUE),
social_media_share_40_60_cov = mean(social_media_share_40_60_cov, na.rm = TRUE),
social_media_share_20_40_cov = mean(social_media_share_20_40_cov, na.rm = TRUE),
misinfo_pre_cov = mean(misinfo_pre_cov, na.rm = TRUE),
base_rate_pre_cov = mean(base_rate_pre_cov, na.rm = TRUE),
emotions_pre_cov = mean(emotions_pre_cov, na.rm = TRUE),
reasoning_pre_cov = mean(reasoning_pre_cov, na.rm = TRUE),
combo_pre_cov = mean(combo_pre_cov, na.rm = TRUE)) %>% # now add the treatment and post variables and misinfo
mutate(treatment = treat,
post = post,
post_is_misinfo = misinfo)
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[row] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[row] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
# predicted value
if (m == 1) {
predictorVariables <- expand.grid(treatment = unique(panel_2$treatment),
post = c(0,1),
post_is_misinfo = c(0,1))
predictions <- predict(baseline_logit, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_logit, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% select(post, post_is_misinfo, estimate, SE, z.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate)
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(lag) - as.numeric(prediction), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
cov_model <- feglm(as.formula(cov_feglm[m]),panel_2 %>% filter(post == p & post_is_misinfo == mis), family = binomial(link = "logit"), cluster = "user")
# obtain data frame with estimate, std error, and z value
cov_model_add <- data.frame(
treatment = unique(panel_2$treatment),
estimate = format(coef(cov_model), digits = 4),
SE = sqrt(diag(vcov(cov_model))),
z.ratio = coef(cov_model) / sqrt(diag(vcov(cov_model))),
post = p,
post_is_misinfo = mis
)
# make rownames a column
cov_model_add <- rownames_to_column(cov_model_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>%
select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
cov_model_add <- rbind(cov_model_add %>%
mutate(treatment = "Facts Baseline"),cov_model_add %>%
mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate)
contrast_df <- rbind(contrast_df, cov_model_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo)
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))
print(k)
cat('<br>')
prediction_2treat_logit_df[[length(prediction_2treat_logit_df) + 1]] <- prediction_df
}
cat('<br>\n<br>\n<br>')
}
saveRDS(prediction_2treat_logit_df, "prediction_2treat_logit_df.rds")
}\[\text{Demographics}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | 0.01342 | 0.002590 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05418); [0.0478] | |
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | 0.01768 | -0.013730 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04402); [-0.3119] | |
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | -0.08481 | -0.401041 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.0522); [-7.683] | |
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | -0.10281 | -0.5110728 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.0467); [-10.94] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6932 | (0.006456) | 0.00759 | 0.005002 |
| Emotions | 0 | 0 | 0.70079 | (0.005974) | (0.03185); [0.157] | |
| Facts Baseline | 0 | 1 | 0.52286 | (0.007215) | 0.01568 | 0.02275 |
| Emotions | 0 | 1 | 0.53854 | (0.006702) | (0.01702); [1.337] | |
| Facts Baseline | 1 | 0 | 0.60565 | (0.006988) | -0.08874 | -0.47956 |
| Emotions | 1 | 0 | 0.51691 | (0.006733) | (0.05551); [-8.64] | |
| Facts Baseline | 1 | 1 | 0.439 | (0.007186) | -0.10921 | -0.5597 |
| Emotions | 1 | 1 | 0.32979 | (0.006297) | (0.04896); [-11.43] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68926 | (0.005609) | 0.01342 | |
| Emotions | 0 | 0 | 0.70268 | (0.005268) | ||
| Facts Baseline | 0 | 1 | 0.52206 | (0.006247) | 0.01865 | |
| Emotions | 0 | 1 | 0.54071 | (0.005857) | ||
| Facts Baseline | 1 | 0 | 0.60174 | (0.00606) | -0.08482 | |
| Emotions | 1 | 0 | 0.51692 | (0.005863) | ||
| Facts Baseline | 1 | 1 | 0.43688 | (0.006215) | -0.10195 | |
| Emotions | 1 | 1 | 0.33493 | (0.00545) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | ||
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | ||
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | ||
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | |
| Emotions | 1 | 1 | 0.32846 | (0.006672) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | 0.0051957 |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | (0.03257); [0.1595] | |
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | 0.0134475 |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | (0.01939); [0.6937] | |
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | -0.505778 |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | (0.05559); [-9.098] | |
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | -0.6307290 |
| Emotions | 1 | 1 | 0.32846 | (0.006672) | (0.05035); [-12.53] |
\[\text{Sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | 0.01342 | 0.002590 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05418); [0.0478] | |
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | 0.01768 | -0.013730 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04402); [-0.3119] | |
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | -0.08481 | -0.401041 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.0522); [-7.683] | |
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | -0.10281 | -0.5110728 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.0467); [-10.94] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6932 | (0.006456) | 0.00759 | 0.005002 |
| Emotions | 0 | 0 | 0.70079 | (0.005974) | (0.03185); [0.157] | |
| Facts Baseline | 0 | 1 | 0.52286 | (0.007215) | 0.01568 | 0.02275 |
| Emotions | 0 | 1 | 0.53854 | (0.006702) | (0.01702); [1.337] | |
| Facts Baseline | 1 | 0 | 0.60565 | (0.006988) | -0.08874 | -0.47956 |
| Emotions | 1 | 0 | 0.51691 | (0.006733) | (0.05551); [-8.64] | |
| Facts Baseline | 1 | 1 | 0.439 | (0.007186) | -0.10921 | -0.5597 |
| Emotions | 1 | 1 | 0.32979 | (0.006297) | (0.04896); [-11.43] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68926 | (0.005609) | 0.01342 | |
| Emotions | 0 | 0 | 0.70268 | (0.005268) | ||
| Facts Baseline | 0 | 1 | 0.52206 | (0.006247) | 0.01865 | |
| Emotions | 0 | 1 | 0.54071 | (0.005857) | ||
| Facts Baseline | 1 | 0 | 0.60174 | (0.00606) | -0.08482 | |
| Emotions | 1 | 0 | 0.51692 | (0.005863) | ||
| Facts Baseline | 1 | 1 | 0.43688 | (0.006215) | -0.10195 | |
| Emotions | 1 | 1 | 0.33493 | (0.00545) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | ||
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | ||
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | ||
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | |
| Emotions | 1 | 1 | 0.32846 | (0.006672) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | 0.0051957 |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | (0.03257); [0.1595] | |
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | 0.0134475 |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | (0.01939); [0.6937] | |
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | -0.505778 |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | (0.05559); [-9.098] | |
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | -0.6307290 |
| Emotions | 1 | 1 | 0.32846 | (0.006672) | (0.05035); [-12.53] |
\[\text{Tactic sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | 0.01342 | 0.002590 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05418); [0.0478] | |
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | 0.01768 | -0.013730 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04402); [-0.3119] | |
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | -0.08481 | -0.401041 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.0522); [-7.683] | |
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | -0.10281 | -0.5110728 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.0467); [-10.94] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6932 | (0.006456) | 0.00759 | 0.005002 |
| Emotions | 0 | 0 | 0.70079 | (0.005974) | (0.03185); [0.157] | |
| Facts Baseline | 0 | 1 | 0.52286 | (0.007215) | 0.01568 | 0.02275 |
| Emotions | 0 | 1 | 0.53854 | (0.006702) | (0.01702); [1.337] | |
| Facts Baseline | 1 | 0 | 0.60565 | (0.006988) | -0.08874 | -0.47956 |
| Emotions | 1 | 0 | 0.51691 | (0.006733) | (0.05551); [-8.64] | |
| Facts Baseline | 1 | 1 | 0.439 | (0.007186) | -0.10921 | -0.5597 |
| Emotions | 1 | 1 | 0.32979 | (0.006297) | (0.04896); [-11.43] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68926 | (0.005609) | 0.01342 | |
| Emotions | 0 | 0 | 0.70268 | (0.005268) | ||
| Facts Baseline | 0 | 1 | 0.52206 | (0.006247) | 0.01865 | |
| Emotions | 0 | 1 | 0.54071 | (0.005857) | ||
| Facts Baseline | 1 | 0 | 0.60174 | (0.00606) | -0.08482 | |
| Emotions | 1 | 0 | 0.51692 | (0.005863) | ||
| Facts Baseline | 1 | 1 | 0.43688 | (0.006215) | -0.10195 | |
| Emotions | 1 | 1 | 0.33493 | (0.00545) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | ||
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | ||
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | ||
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | |
| Emotions | 1 | 1 | 0.32846 | (0.006672) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | 0.0051957 |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | (0.03257); [0.1595] | |
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | 0.0134475 |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | (0.01939); [0.6937] | |
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | -0.505778 |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | (0.05559); [-9.098] | |
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | -0.6307290 |
| Emotions | 1 | 1 | 0.32846 | (0.006672) | (0.05035); [-12.53] |
\[\text{Demographics +
Sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | 0.01342 | 0.002590 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05418); [0.0478] | |
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | 0.01768 | -0.013730 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04402); [-0.3119] | |
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | -0.08481 | -0.401041 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.0522); [-7.683] | |
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | -0.10281 | -0.5110728 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.0467); [-10.94] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6932 | (0.006456) | 0.00759 | 0.005002 |
| Emotions | 0 | 0 | 0.70079 | (0.005974) | (0.03185); [0.157] | |
| Facts Baseline | 0 | 1 | 0.52286 | (0.007215) | 0.01568 | 0.02275 |
| Emotions | 0 | 1 | 0.53854 | (0.006702) | (0.01702); [1.337] | |
| Facts Baseline | 1 | 0 | 0.60565 | (0.006988) | -0.08874 | -0.47956 |
| Emotions | 1 | 0 | 0.51691 | (0.006733) | (0.05551); [-8.64] | |
| Facts Baseline | 1 | 1 | 0.439 | (0.007186) | -0.10921 | -0.5597 |
| Emotions | 1 | 1 | 0.32979 | (0.006297) | (0.04896); [-11.43] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68926 | (0.005609) | 0.01342 | |
| Emotions | 0 | 0 | 0.70268 | (0.005268) | ||
| Facts Baseline | 0 | 1 | 0.52206 | (0.006247) | 0.01865 | |
| Emotions | 0 | 1 | 0.54071 | (0.005857) | ||
| Facts Baseline | 1 | 0 | 0.60174 | (0.00606) | -0.08482 | |
| Emotions | 1 | 0 | 0.51692 | (0.005863) | ||
| Facts Baseline | 1 | 1 | 0.43688 | (0.006215) | -0.10195 | |
| Emotions | 1 | 1 | 0.33493 | (0.00545) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | ||
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | ||
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | ||
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | |
| Emotions | 1 | 1 | 0.32846 | (0.006672) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | 0.0051957 |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | (0.03257); [0.1595] | |
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | 0.0134475 |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | (0.01939); [0.6937] | |
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | -0.505778 |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | (0.05559); [-9.098] | |
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | -0.6307290 |
| Emotions | 1 | 1 | 0.32846 | (0.006672) | (0.05035); [-12.53] |
\[\text{Demographics +
Sharing rates + Tactic sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.0059 | 0.027609 |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | (0.04131); [0.6683] | |
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.0021 | 0.008161 |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | (0.02721); [0.3000] | |
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.0879 | -0.358022 |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | (0.03883); [ -9.2196] | |
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.1120 | -0.465472 |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | (0.02778); [-16.7557] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68925 | (0.005604) | 0.01342 | 0.002590 |
| Emotions | 0 | 0 | 0.70267 | (0.005263) | (0.05418); [0.0478] | |
| Facts Baseline | 0 | 1 | 0.52269 | (0.00624) | 0.01768 | -0.013730 |
| Emotions | 0 | 1 | 0.54037 | (0.005841) | (0.04402); [-0.3119] | |
| Facts Baseline | 1 | 0 | 0.60172 | (0.006054) | -0.08481 | -0.401041 |
| Emotions | 1 | 0 | 0.51691 | (0.005858) | (0.0522); [-7.683] | |
| Facts Baseline | 1 | 1 | 0.43763 | (0.006211) | -0.10281 | -0.5110728 |
| Emotions | 1 | 1 | 0.33482 | (0.005444) | (0.0467); [-10.94] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6932 | (0.006456) | 0.00759 | 0.005002 |
| Emotions | 0 | 0 | 0.70079 | (0.005974) | (0.03185); [0.157] | |
| Facts Baseline | 0 | 1 | 0.52286 | (0.007215) | 0.01568 | 0.02275 |
| Emotions | 0 | 1 | 0.53854 | (0.006702) | (0.01702); [1.337] | |
| Facts Baseline | 1 | 0 | 0.60565 | (0.006988) | -0.08874 | -0.47956 |
| Emotions | 1 | 0 | 0.51691 | (0.006733) | (0.05551); [-8.64] | |
| Facts Baseline | 1 | 1 | 0.439 | (0.007186) | -0.10921 | -0.5597 |
| Emotions | 1 | 1 | 0.32979 | (0.006297) | (0.04896); [-11.43] |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68926 | (0.005609) | 0.01342 | |
| Emotions | 0 | 0 | 0.70268 | (0.005268) | ||
| Facts Baseline | 0 | 1 | 0.52206 | (0.006247) | 0.01865 | |
| Emotions | 0 | 1 | 0.54071 | (0.005857) | ||
| Facts Baseline | 1 | 0 | 0.60174 | (0.00606) | -0.08482 | |
| Emotions | 1 | 0 | 0.51692 | (0.005863) | ||
| Facts Baseline | 1 | 1 | 0.43688 | (0.006215) | -0.10195 | |
| Emotions | 1 | 1 | 0.33493 | (0.00545) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | ||
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | ||
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | ||
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | |
| Emotions | 1 | 1 | 0.32846 | (0.006672) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.69277 | (0.006585) | 0.00548 | 0.0051957 |
| Emotions | 0 | 0 | 0.69825 | (0.006183) | (0.03257); [0.1595] | |
| Facts Baseline | 0 | 1 | 0.52407 | (0.007618) | 0.01378 | 0.0134475 |
| Emotions | 0 | 1 | 0.53785 | (0.007158) | (0.01939); [0.6937] | |
| Facts Baseline | 1 | 0 | 0.60601 | (0.007116) | -0.09149 | -0.505778 |
| Emotions | 1 | 0 | 0.51452 | (0.006926) | (0.05559); [-9.098] | |
| Facts Baseline | 1 | 1 | 0.44265 | (0.007596) | -0.11419 | -0.6307290 |
| Emotions | 1 | 1 | 0.32846 | (0.006672) | (0.05035); [-12.53] |
Click for All treatment arms
# version 2: all treatment arms
covariate_list <- c(demographics_demeaned_treatment,
sharing_demeaned_treatment,
sharing_tactic_demeaned_treatment,
demographics_sharing_demeaned_treatment,
all_covariates_demeaned_treatment)
cov_estimates_list <- list()
for (cov_list in covariate_list) {
# Demographics
formula_cov1_logit <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list)
cov1_logit <- feglm(as.formula(formula_cov1_logit),panel, family = binomial(link = "logit"))
formula_cov2_logit <- paste0("share ~ treatment*post*post_is_misinfo "," + ",cov_list," | user")
cov2_logit <- feglm(as.formula(formula_cov2_logit),panel, family = binomial(link = "logit"))
# Misinfo and Non-misinfo sharing rates
formula_cov3_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact) "," + ",cov_list," | user")
cov3_logit <- feglm(as.formula(formula_cov3_logit),panel, family = binomial(link = "logit"))
# Misinfo, Non-misinfo, and Tactic sharing rates
formula_cov4_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(type) "," + ",cov_list," | user")
cov4_logit <- feglm(as.formula(formula_cov4_logit),panel, family = binomial(link = "logit"))
# Demographics and Misinfo and Non-misinfo sharing rates
formula_cov5_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(fact)*factor(type) "," + ",cov_list," | user")
cov5_logit <- feglm(as.formula(formula_cov5_logit),panel, family = binomial(link = "logit"))
# Demographics, Misinfo and Non-misinfo sharing rates, and Tactic sharing rates
formula_cov6_logit <- paste0("share ~ treatment*post*post_is_misinfo + factor(post_id) "," + ",cov_list," | user")
cov6_logit <- feglm(as.formula(formula_cov6_logit),panel, family = binomial(link = "logit"))
cov_estimates_list[[length(cov_estimates_list) + 1]] <- list(cov1_logit,cov2_logit,cov3_logit,cov4_logit,cov5_logit,cov6_logit)
}covariate_list_names <- c("Demographics",
"Sharing rates",
"Tactic sharing rates",
"Demographics + Sharing rates",
"Demographics + Sharing rates + Tactic sharing rates")
for (cov in seq(length(covariate_list))) {
# Print model
headers <- c("Share (binary indicator)")
model_list <- cov_estimates_list[[cov]]
names <- unique(c(names(cov_estimates_list[[cov]][[1]]$coefficients),
names(cov_estimates_list[[cov]][[2]]$coefficients),
names(cov_estimates_list[[cov]][[3]]$coefficients),
names(cov_estimates_list[[cov]][[4]]$coefficients),
names(cov_estimates_list[[cov]][[5]]$coefficients),
names(cov_estimates_list[[cov]][[6]]$coefficients)))
variable_labels_wide <- variable_mapping_long[names]
# remove blankrows
coef_names <- names
matrix_coeff_na <- expand.grid(coef_name = coef_names, model = seq(length(model_list)))
matrix_coeff_na$var <- 0
matrix_coeff_na$var_na <- 0
for (m in 1:length(model_list)) {
na_coeff <- names(model_list[[m]]$coefficients)[is.na(model_list[[m]]$coefficients)]
for (c in na_coeff) {
matrix_coeff_na$var[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1 # 1 if variable is in the model (not all models have all variables)
matrix_coeff_na$var_na[matrix_coeff_na$coef_name == c & matrix_coeff_na$model == m] <- 1
}
}
matrix_coeff_na <- matrix_coeff_na %>% group_by(coef_name) %>% mutate(var = sum(var), var_na = sum(var_na)) %>% distinct(coef_name,var,var_na)
rows_to_omit <- as.vector(unique(matrix_coeff_na$coef_name[matrix_coeff_na$var_na != 0])) # omit if NA in all models
rows_to_omit <- gsub("\\^\\s","^",paste0("^",rows_to_omit,"$", collapse = "|"))
rows_to_omit <- gsub("\\(","\\\\(",gsub("\\)","\\\\)",rows_to_omit))
# keep variables
keep <- list("treatmentEmotions:post:post_is_misinfo" = "Emotions x 1[Post-treatment] x 1[Misinfo]",
"treatmentEmotions:post" = "Emotions x 1[Post-treatment]",
"treatmentEmotions:post_is_misinfo" = "Emotions x 1[Misinfo]")
# out-of-sample R2
if (any(ls() %in% "R2oos_cov_alltreat_logit_list")) {
R2oos_cov_alltreat_logit_all <- c(R2oos_cov_alltreat_logit_list[[cov]])
} else {
R2oos_cov_alltreat_logit_all <- rep("",length(model_list))
}
# Print model name
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
# Short table
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
custom.coef.map = keep,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_alltreat_logit_all)),
reorder.gof = c(6,7,8,9,5,1,2,3,4)
)
cat(html_table)
cat("<br>")
# Full table (all independent variables)
cat('<details>')
cat('<summary>Click to view the full table</summary>')
html_table <- texreg::htmlreg(model_list,
custom.header = list("Share (binary indicator)" = 1:length(model_list)),
custom.model.names = c("Baseline<br>(1)",
"User ID FE<br>(2)",
"User ID FE +<br>Post domain FE<br>(3)",
"User ID FE +<br>Post tactic FE<br>(4)",
"User ID FE +<br>Post domain x Post tactic<br>(5)",
"User ID FE +<br>Post ID FE<br>(6)"),
stars = c(0.20, 0.10, 0.02),
custom.note = "Significance levels: \\*p<0.10, \\*\\*p<0.05, \\*\\*\\*p<0.01 one-sided test.",
caption = paste("Linear Regression:", headers, "vs. Treatment for different set of fixed effects\n<br>Covariates:",covariate_list_names[cov]),
caption.above = TRUE,
doctype = FALSE,
html.tag = FALSE,
table.tag = FALSE,
single.row = FALSE,
include.groups = FALSE,
digits = 4,
custom.coef.names = variable_labels_wide,
omit.coef = rows_to_omit,
custom.gof.rows = list("Fixed Effects" = c("",rep("user",length(model_list)-1)),
"Fixed Effects" = c("","","Domain","Tactic","Domain x Tactic","Post ID"),
"Covariates" = c("",rep(covariate_list_names[cov],length(model_list)-1)),
"Clustered Std. Errors" = rep("user",length(model_list)),
"Out-of-sample R<sup>2</sup>" = c(R2oos_cov_alltreat_logit_all)),
reorder.gof = c(6,7,8,9,5,1,2,3,4)
)
cat(html_table)
cat('</details>')
cat('<br>\n<br>\n<br>')
}| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0914* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0704) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Emotions x 1[Post-treatment] | -0.4006*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0581) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Emotions x 1[Misinfo] | -0.0151 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0507) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 202191.4106 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -101095.7053 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.0625 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 0.8245*** | |||||
| (0.0302) | ||||||
| No-course Baseline | 0.0747** | |||||
| (0.0441) | ||||||
| Emotions | 0.0005 | |||||
| (0.0424) | ||||||
| Reasoning | 0.0663* | |||||
| (0.0436) | ||||||
| Combo | -0.0111 | |||||
| (0.0429) | ||||||
| 1[Post-treatment] | -0.3496*** | -0.5002*** | -0.5575*** | -0.5005*** | -0.5596*** | -0.5596*** |
| (0.0416) | (0.0589) | (0.0631) | (0.0589) | (0.0639) | (0.0639) | |
| 1[Misinfo] | -0.6451*** | -0.9234*** | -1.0396*** | -0.8461*** | -1.0772*** | -3.0886*** |
| (0.0361) | (0.0453) | (0.0483) | (0.0464) | (0.0958) | (0.0973) | |
| Age | -0.0152 | |||||
| (0.0130) | ||||||
| Man | 0.0593*** | |||||
| (0.0118) | ||||||
| Some college | -0.0274** | |||||
| (0.0147) | ||||||
| Bachelor’s degree | -0.0623*** | |||||
| (0.0148) | ||||||
| Graduate degree | -0.0465*** | |||||
| (0.0127) | ||||||
| Married | 0.0025 | |||||
| (0.0132) | ||||||
| Unemployed | -0.0678*** | |||||
| (0.0154) | ||||||
| Employed | -0.0396*** | |||||
| (0.0168) | ||||||
| Mostly urban | -0.1233*** | |||||
| (0.0140) | ||||||
| Suburban | -0.0765*** | |||||
| (0.0137) | ||||||
| Christian | 0.0387*** | |||||
| (0.0122) | ||||||
| Attends religious services | 0.0452*** | |||||
| (0.0119) | ||||||
| Uses social media | -0.0830*** | |||||
| (0.0111) | ||||||
| Hours spent on Social Media | 0.0327*** | |||||
| (0.0130) | ||||||
| Social Media Share: 80-100% | 0.3545*** | |||||
| (0.0148) | ||||||
| Social Media Share: 60-80% | 0.4825*** | |||||
| (0.0159) | ||||||
| Social Media Share: 40-60% | 0.4174*** | |||||
| (0.0164) | ||||||
| Social Media Share: 20-40% | 0.2228*** | |||||
| (0.0152) | ||||||
| Passed attention check at pre-treatment | -0.0945*** | |||||
| (0.0118) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.2300*** | 0.3289*** | 0.3843*** | 0.3290*** | 0.3850*** | 0.3850*** |
| (0.0612) | (0.0859) | (0.0921) | (0.0859) | (0.0937) | (0.0937) | |
| Emotions x 1[Post-treatment] | -0.4006*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0581) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Reasoning x 1[Post-treatment] | -0.3426*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0596) | (0.0848) | (0.0910) | (0.0848) | (0.0922) | (0.0922) | |
| Combo x 1[Post-treatment] | -0.3731*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0588) | (0.0845) | (0.0911) | (0.0846) | (0.0924) | (0.0924) | |
| No-course Baseline x 1[Misinfo] | 0.0101 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0526) | (0.0668) | (0.0708) | (0.0668) | (0.0722) | (0.0722) | |
| Emotions x 1[Misinfo] | -0.0151 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0507) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Reasoning x 1[Misinfo] | -0.0907** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0520) | (0.0657) | (0.0695) | (0.0657) | (0.0710) | (0.0710) | |
| Combo x 1[Misinfo] | 0.0227 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0513) | (0.0645) | (0.0687) | (0.0646) | (0.0703) | (0.0703) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0386 | 0.0492 | 0.0613 | 0.0485 | 0.0507 | 0.0507 |
| (0.0502) | (0.0617) | (0.0640) | (0.0618) | (0.0648) | (0.0648) | |
| No-course Baseline x Age | -0.0225 | |||||
| (0.0184) | ||||||
| Emotions x Age | -0.0518*** | |||||
| (0.0196) | ||||||
| Reasoning x Age | -0.0309* | |||||
| (0.0192) | ||||||
| Combo x Age | 0.0023 | |||||
| (0.0189) | ||||||
| No-course Baseline x Man | -0.1019*** | |||||
| (0.0174) | ||||||
| Emotions x Man | -0.0091 | |||||
| (0.0167) | ||||||
| Reasoning x Man | -0.0143 | |||||
| (0.0169) | ||||||
| Combo x Man | -0.0260* | |||||
| (0.0169) | ||||||
| No-course Baseline x Some college | 0.0082 | |||||
| (0.0209) | ||||||
| Emotions x Some college | -0.0158 | |||||
| (0.0205) | ||||||
| Reasoning x Some college | -0.0429** | |||||
| (0.0208) | ||||||
| Combo x Some college | 0.0058 | |||||
| (0.0208) | ||||||
| No-course Baseline x Bachelor’s degree | 0.0757*** | |||||
| (0.0211) | ||||||
| Emotions x Bachelor’s degree | 0.0367** | |||||
| (0.0206) | ||||||
| Reasoning x Bachelor’s degree | -0.0113 | |||||
| (0.0210) | ||||||
| Combo x Bachelor’s degree | -0.0213 | |||||
| (0.0209) | ||||||
| No-course Baseline x Graduate degree | 0.0544*** | |||||
| (0.0187) | ||||||
| Emotions x Graduate degree | 0.0470*** | |||||
| (0.0169) | ||||||
| Reasoning x Graduate degree | 0.0650*** | |||||
| (0.0191) | ||||||
| Combo x Graduate degree | 0.0714*** | |||||
| (0.0167) | ||||||
| No-course Baseline x Married | 0.0364** | |||||
| (0.0195) | ||||||
| Emotions x Married | 0.0228 | |||||
| (0.0188) | ||||||
| Reasoning x Married | 0.0045 | |||||
| (0.0194) | ||||||
| Combo x Married | -0.0335** | |||||
| (0.0192) | ||||||
| No-course Baseline x Unemployed | 0.0217 | |||||
| (0.0221) | ||||||
| Emotions x Unemployed | -0.0320* | |||||
| (0.0216) | ||||||
| Reasoning x Unemployed | 0.0428** | |||||
| (0.0218) | ||||||
| Combo x Unemployed | 0.0055 | |||||
| (0.0219) | ||||||
| No-course Baseline x Employed | -0.0170 | |||||
| (0.0242) | ||||||
| Emotions x Employed | -0.0835*** | |||||
| (0.0234) | ||||||
| Reasoning x Employed | -0.0076 | |||||
| (0.0242) | ||||||
| Combo x Employed | -0.0044 | |||||
| (0.0236) | ||||||
| No-course Baseline x Mostly urban | 0.1288*** | |||||
| (0.0202) | ||||||
| Emotions x Mostly urban | 0.1246*** | |||||
| (0.0197) | ||||||
| Reasoning x Mostly urban | 0.0276* | |||||
| (0.0202) | ||||||
| Combo x Mostly urban | 0.0930*** | |||||
| (0.0198) | ||||||
| No-course Baseline x Suburban | 0.0180 | |||||
| (0.0198) | ||||||
| Emotions x Suburban | 0.0657*** | |||||
| (0.0194) | ||||||
| Reasoning x Suburban | 0.0177 | |||||
| (0.0196) | ||||||
| Combo x Suburban | 0.0490*** | |||||
| (0.0195) | ||||||
| No-course Baseline x Christian | -0.0510*** | |||||
| (0.0170) | ||||||
| Emotions x Christian | -0.0361** | |||||
| (0.0174) | ||||||
| Reasoning x Christian | -0.0346** | |||||
| (0.0178) | ||||||
| Combo x Christian | -0.0112 | |||||
| (0.0173) | ||||||
| No-course Baseline x Attends religious services | -0.0547*** | |||||
| (0.0166) | ||||||
| Emotions x Attends religious services | -0.0942*** | |||||
| (0.0173) | ||||||
| Reasoning x Attends religious services | -0.0332** | |||||
| (0.0176) | ||||||
| Combo x Attends religious services | -0.0647*** | |||||
| (0.0174) | ||||||
| No-course Baseline x Uses social media | 0.0664*** | |||||
| (0.0155) | ||||||
| Emotions x Uses social media | 0.0238* | |||||
| (0.0175) | ||||||
| Reasoning x Uses social media | 0.0490*** | |||||
| (0.0159) | ||||||
| Combo x Uses social media | 0.0160 | |||||
| (0.0182) | ||||||
| No-course Baseline x Hours spent on Social Media | -0.0615*** | |||||
| (0.0181) | ||||||
| Emotions x Hours spent on Social Media | -0.0555*** | |||||
| (0.0179) | ||||||
| Reasoning x Hours spent on Social Media | 0.0167 | |||||
| (0.0184) | ||||||
| Combo x Hours spent on Social Media | -0.0883*** | |||||
| (0.0178) | ||||||
| No-course Baseline x Social Media Share: 80-100% | -0.0318* | |||||
| (0.0209) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0325* | |||||
| (0.0208) | ||||||
| Reasoning x Social Media Share: 80-100% | 0.0080 | |||||
| (0.0209) | ||||||
| Combo x Social Media Share: 80-100% | -0.0488*** | |||||
| (0.0206) | ||||||
| No-course Baseline x Social Media Share: 60-80% | -0.0074 | |||||
| (0.0237) | ||||||
| Emotions x Social Media Share: 60-80% | -0.0176 | |||||
| (0.0228) | ||||||
| Reasoning x Social Media Share: 60-80% | -0.0390** | |||||
| (0.0233) | ||||||
| Combo x Social Media Share: 60-80% | -0.0399** | |||||
| (0.0231) | ||||||
| No-course Baseline x Social Media Share: 40-60% | -0.0256 | |||||
| (0.0237) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0546*** | |||||
| (0.0233) | ||||||
| Reasoning x Social Media Share: 40-60% | -0.0227 | |||||
| (0.0240) | ||||||
| Combo x Social Media Share: 40-60% | -0.0781*** | |||||
| (0.0235) | ||||||
| No-course Baseline x Social Media Share: 20-40% | -0.0569*** | |||||
| (0.0220) | ||||||
| Emotions x Social Media Share: 20-40% | -0.0290* | |||||
| (0.0217) | ||||||
| Reasoning x Social Media Share: 20-40% | -0.0152 | |||||
| (0.0222) | ||||||
| Combo x Social Media Share: 20-40% | -0.0408** | |||||
| (0.0220) | ||||||
| No-course Baseline x Passed attention check at pre-treatment | -0.0655*** | |||||
| (0.0172) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0182 | |||||
| (0.0167) | ||||||
| Reasoning x Passed attention check at pre-treatment | 0.0498*** | |||||
| (0.0170) | ||||||
| Combo x Passed attention check at pre-treatment | -0.0344** | |||||
| (0.0168) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | 0.0004 | 0.0069 | -0.0022 | 0.0073 | 0.0198 | 0.0198 |
| (0.0735) | (0.0921) | (0.0958) | (0.0921) | (0.0971) | (0.0971) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.0914* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0704) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.0810 | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0720) | (0.0892) | (0.0922) | (0.0893) | (0.0935) | (0.0935) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0558 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0712) | (0.0887) | (0.0924) | (0.0887) | (0.0940) | (0.0940) | |
| Post domain 2 | -1.3458*** | -1.3896*** | ||||
| (0.0391) | (0.0766) | |||||
| Post domain 3 | -1.8926*** | -2.2086*** | ||||
| (0.0389) | (0.0755) | |||||
| Post domain 4 | -1.2128*** | -1.7401*** | ||||
| (0.0384) | (0.0763) | |||||
| Post domain 5 | 0.4049*** | 0.0051 | ||||
| (0.0388) | (0.0820) | |||||
| Post domain 6 | 0.1237*** | 0.4527*** | ||||
| (0.0395) | (0.0844) | |||||
| Post domain 7 | -1.5746*** | -1.4228*** | ||||
| (0.0388) | (0.0752) | |||||
| Post domain 8 | -1.4724*** | -0.9538*** | ||||
| (0.0391) | (0.0779) | |||||
| Post domain 9 | 0.2728*** | -0.5281*** | ||||
| (0.0392) | (0.0788) | |||||
| Post domain 10 | -0.3969*** | -0.8059*** | ||||
| (0.0377) | (0.0763) | |||||
| Post domain 11 | -1.0747*** | -1.3501*** | ||||
| (0.0388) | (0.0756) | |||||
| Post domain 12 | -0.7213*** | -1.0828*** | ||||
| (0.0372) | (0.0758) | |||||
| Post domain 13 | -1.0425*** | -1.6220*** | ||||
| (0.0378) | (0.0743) | |||||
| Post domain 14 | -1.8560*** | -2.2721*** | ||||
| (0.0403) | (0.0771) | |||||
| Post domain 15 | -1.2523*** | -2.0199*** | ||||
| (0.0381) | (0.0749) | |||||
| 1[Combo manipulation tactic] | -0.2264*** | -0.3419*** | ||||
| (0.0187) | (0.0832) | |||||
| 1[Emotions manipulation tactic] | -0.0068 | -0.5416*** | ||||
| (0.0184) | (0.0822) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.1530* | |||||
| (0.1163) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4818*** | |||||
| (0.1123) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.7230*** | |||||
| (0.1104) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6450*** | |||||
| (0.1186) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.2961*** | |||||
| (0.1206) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.1072 | |||||
| (0.1122) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0724 | |||||
| (0.1141) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.1394*** | |||||
| (0.1160) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.3736*** | |||||
| (0.1122) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.6002*** | |||||
| (0.1113) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | 0.0418 | |||||
| (0.1136) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.5858*** | |||||
| (0.1106) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.3450*** | |||||
| (0.1139) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.7389*** | |||||
| (0.1105) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4478*** | |||||
| (0.1167) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.9334*** | |||||
| (0.1107) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.5846*** | |||||
| (0.1097) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.8014*** | |||||
| (0.1170) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9194*** | |||||
| (0.1236) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0346 | |||||
| (0.1108) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.9456*** | |||||
| (0.1172) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2361*** | |||||
| (0.1149) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.8795*** | |||||
| (0.1109) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0799 | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1423*** | |||||
| (0.1105) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.4266*** | |||||
| (0.1071) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.8221*** | |||||
| (0.1124) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.4403*** | |||||
| (0.1113) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.4651*** | |||||
| (0.1126) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.2843*** | |||||
| (0.1157) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.8423*** | |||||
| (0.1111) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.2026** | |||||
| (0.1188) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3527*** | |||||
| (0.1241) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.8764*** | |||||
| (0.1125) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.9427*** | |||||
| (0.1203) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 1.0028*** | |||||
| (0.1186) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4064*** | |||||
| (0.1125) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.5144*** | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2580*** | |||||
| (0.1107) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3539*** | |||||
| (0.1101) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.4395*** | |||||
| (0.1117) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0085 | |||||
| (0.1116) | ||||||
| Post 2 indicator | -1.3896*** | |||||
| (0.0766) | ||||||
| Post 3 indicator | -2.2086*** | |||||
| (0.0755) | ||||||
| Post 4 indicator | -1.7401*** | |||||
| (0.0763) | ||||||
| Post 5 indicator | 0.0051 | |||||
| (0.0820) | ||||||
| Post 6 indicator | 0.4527*** | |||||
| (0.0844) | ||||||
| Post 7 indicator | -1.4228*** | |||||
| (0.0752) | ||||||
| Post 8 indicator | -0.9538*** | |||||
| (0.0779) | ||||||
| Post 9 indicator | -0.5281*** | |||||
| (0.0788) | ||||||
| Post 10 indicator | -0.8059*** | |||||
| (0.0763) | ||||||
| Post 11 indicator | -1.3501*** | |||||
| (0.0756) | ||||||
| Post 12 indicator | -1.0828*** | |||||
| (0.0758) | ||||||
| Post 13 indicator | -1.6220*** | |||||
| (0.0743) | ||||||
| Post 14 indicator | -2.2721*** | |||||
| (0.0771) | ||||||
| Post 15 indicator | -2.0199*** | |||||
| (0.0749) | ||||||
| Post 16 indicator | 1.6695*** | |||||
| (0.0824) | ||||||
| Post 17 indicator | 0.1269* | |||||
| (0.0846) | ||||||
| Post 18 indicator | -0.0573 | |||||
| (0.0815) | ||||||
| Post 19 indicator | 0.6525*** | |||||
| (0.0790) | ||||||
| Post 20 indicator | 2.3196*** | |||||
| (0.0838) | ||||||
| Post 21 indicator | -0.1740** | |||||
| (0.0827) | ||||||
| Post 22 indicator | 0.1395** | |||||
| (0.0819) | ||||||
| Post 23 indicator | 0.7881*** | |||||
| (0.0812) | ||||||
| Post 24 indicator | 2.2808*** | |||||
| (0.0839) | ||||||
| Post 25 indicator | 1.2372*** | |||||
| (0.0799) | ||||||
| Post 26 indicator | 0.9195*** | |||||
| (0.0829) | ||||||
| Post 27 indicator | 0.6285*** | |||||
| (0.0802) | ||||||
| Post 28 indicator | 0.6333*** | |||||
| (0.0813) | ||||||
| Post 29 indicator | -0.2576*** | |||||
| (0.0832) | ||||||
| Post 30 indicator | 0.3885*** | |||||
| (0.0809) | ||||||
| Post 31 indicator | 1.4698*** | |||||
| (0.0799) | ||||||
| Post 32 indicator | -0.3676*** | |||||
| (0.0867) | ||||||
| Post 33 indicator | 0.1946*** | |||||
| (0.0794) | ||||||
| Post 34 indicator | 0.3143*** | |||||
| (0.0795) | ||||||
| Post 35 indicator | 2.2763*** | |||||
| (0.0839) | ||||||
| Post 36 indicator | 2.8419*** | |||||
| (0.0910) | ||||||
| Post 37 indicator | 0.0124 | |||||
| (0.0810) | ||||||
| Post 38 indicator | -0.4296*** | |||||
| (0.0859) | ||||||
| Post 39 indicator | 2.1778*** | |||||
| (0.0835) | ||||||
| Post 40 indicator | 1.5434*** | |||||
| (0.0810) | ||||||
| Post 41 indicator | 0.0398 | |||||
| (0.0829) | ||||||
| Post 42 indicator | 1.5293*** | |||||
| (0.0807) | ||||||
| Post 43 indicator | 1.2744*** | |||||
| (0.0780) | ||||||
| Post 44 indicator | 0.0198 | |||||
| (0.0829) | ||||||
| Post 45 indicator | 1.8902*** | |||||
| (0.0825) | ||||||
| Post 46 indicator | 2.0114*** | |||||
| (0.0830) | ||||||
| Post 47 indicator | 1.0869*** | |||||
| (0.0811) | ||||||
| Post 48 indicator | -0.4814*** | |||||
| (0.0840) | ||||||
| Post 49 indicator | 1.1137*** | |||||
| (0.0820) | ||||||
| Post 50 indicator | 2.2191*** | |||||
| (0.0846) | ||||||
| Post 51 indicator | 2.8168*** | |||||
| (0.0890) | ||||||
| Post 52 indicator | -0.2878*** | |||||
| (0.0832) | ||||||
| Post 53 indicator | -0.8851*** | |||||
| (0.0926) | ||||||
| Post 54 indicator | 2.4861*** | |||||
| (0.0862) | ||||||
| Post 55 indicator | 1.6118*** | |||||
| (0.0820) | ||||||
| Post 56 indicator | 1.1756*** | |||||
| (0.0838) | ||||||
| Post 57 indicator | 1.1865*** | |||||
| (0.0791) | ||||||
| Post 58 indicator | 0.7433*** | |||||
| (0.0798) | ||||||
| Post 59 indicator | 0.1789** | |||||
| (0.0816) | ||||||
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 202191.4106 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -101095.7053 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.0625 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics | Demographics | Demographics | Demographics | Demographics | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1042* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0793) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Emotions x 1[Post-treatment] | -0.4743*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0657) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Emotions x 1[Misinfo] | 0.0236 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0575) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166604.6384 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83302.3192 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2281 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0465*** | |||||
| (0.0347) | ||||||
| No-course Baseline | 0.0168 | |||||
| (0.0508) | ||||||
| Emotions | -0.0305 | |||||
| (0.0482) | ||||||
| Reasoning | 0.0736* | |||||
| (0.0500) | ||||||
| Combo | -0.0193 | |||||
| (0.0493) | ||||||
| 1[Post-treatment] | -0.4565*** | -0.5002*** | -0.5575*** | -0.5005*** | -0.5596*** | -0.5596*** |
| (0.0476) | (0.0589) | (0.0631) | (0.0589) | (0.0639) | (0.0639) | |
| 1[Misinfo] | -0.8428*** | -0.9234*** | -1.0396*** | -0.8461*** | -1.0772*** | -3.0886*** |
| (0.0414) | (0.0453) | (0.0483) | (0.0464) | (0.0958) | (0.0973) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9274*** | |||||
| (0.0173) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5397*** | |||||
| (0.0166) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.2988*** | 0.3289*** | 0.3843*** | 0.3290*** | 0.3850*** | 0.3850*** |
| (0.0701) | (0.0859) | (0.0921) | (0.0859) | (0.0937) | (0.0937) | |
| Emotions x 1[Post-treatment] | -0.4743*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0657) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Reasoning x 1[Post-treatment] | -0.4360*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0680) | (0.0848) | (0.0910) | (0.0848) | (0.0922) | (0.0922) | |
| Combo x 1[Post-treatment] | -0.4866*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0673) | (0.0845) | (0.0911) | (0.0846) | (0.0924) | (0.0924) | |
| No-course Baseline x 1[Misinfo] | 0.0059 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0606) | (0.0668) | (0.0708) | (0.0668) | (0.0722) | (0.0722) | |
| Emotions x 1[Misinfo] | 0.0236 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0575) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Reasoning x 1[Misinfo] | -0.1060** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0596) | (0.0657) | (0.0695) | (0.0657) | (0.0710) | (0.0710) | |
| Combo x 1[Misinfo] | 0.0302 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0589) | (0.0645) | (0.0687) | (0.0646) | (0.0703) | (0.0703) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0526 | 0.0492 | 0.0613 | 0.0485 | 0.0507 | 0.0507 |
| (0.0573) | (0.0617) | (0.0640) | (0.0618) | (0.0648) | (0.0648) | |
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | 0.0346* | |||||
| (0.0249) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.1364*** | |||||
| (0.0234) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | -0.0396* | |||||
| (0.0243) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.0622*** | |||||
| (0.0243) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0299 | |||||
| (0.0239) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0649*** | |||||
| (0.0229) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0375* | |||||
| (0.0238) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0233 | |||||
| (0.0236) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | -0.0005 | 0.0069 | -0.0022 | 0.0073 | 0.0198 | 0.0198 |
| (0.0841) | (0.0921) | (0.0958) | (0.0921) | (0.0971) | (0.0971) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1042* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0793) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.1105* | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0818) | (0.0892) | (0.0922) | (0.0893) | (0.0935) | (0.0935) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0578 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0811) | (0.0887) | (0.0924) | (0.0887) | (0.0940) | (0.0940) | |
| Post domain 2 | -1.3458*** | -1.3896*** | ||||
| (0.0391) | (0.0766) | |||||
| Post domain 3 | -1.8926*** | -2.2086*** | ||||
| (0.0389) | (0.0755) | |||||
| Post domain 4 | -1.2128*** | -1.7401*** | ||||
| (0.0384) | (0.0763) | |||||
| Post domain 5 | 0.4049*** | 0.0051 | ||||
| (0.0388) | (0.0820) | |||||
| Post domain 6 | 0.1237*** | 0.4527*** | ||||
| (0.0395) | (0.0844) | |||||
| Post domain 7 | -1.5746*** | -1.4228*** | ||||
| (0.0388) | (0.0752) | |||||
| Post domain 8 | -1.4724*** | -0.9538*** | ||||
| (0.0391) | (0.0779) | |||||
| Post domain 9 | 0.2728*** | -0.5281*** | ||||
| (0.0392) | (0.0788) | |||||
| Post domain 10 | -0.3969*** | -0.8059*** | ||||
| (0.0377) | (0.0763) | |||||
| Post domain 11 | -1.0747*** | -1.3501*** | ||||
| (0.0388) | (0.0756) | |||||
| Post domain 12 | -0.7213*** | -1.0828*** | ||||
| (0.0372) | (0.0758) | |||||
| Post domain 13 | -1.0425*** | -1.6220*** | ||||
| (0.0378) | (0.0743) | |||||
| Post domain 14 | -1.8560*** | -2.2721*** | ||||
| (0.0403) | (0.0771) | |||||
| Post domain 15 | -1.2523*** | -2.0199*** | ||||
| (0.0381) | (0.0749) | |||||
| 1[Combo manipulation tactic] | -0.2264*** | -0.3419*** | ||||
| (0.0187) | (0.0832) | |||||
| 1[Emotions manipulation tactic] | -0.0068 | -0.5416*** | ||||
| (0.0184) | (0.0822) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.1530* | |||||
| (0.1163) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4818*** | |||||
| (0.1123) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.7230*** | |||||
| (0.1104) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6450*** | |||||
| (0.1186) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.2961*** | |||||
| (0.1206) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.1072 | |||||
| (0.1122) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0724 | |||||
| (0.1141) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.1394*** | |||||
| (0.1160) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.3736*** | |||||
| (0.1122) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.6002*** | |||||
| (0.1113) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | 0.0418 | |||||
| (0.1136) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.5858*** | |||||
| (0.1106) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.3450*** | |||||
| (0.1139) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.7389*** | |||||
| (0.1105) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4478*** | |||||
| (0.1167) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.9334*** | |||||
| (0.1107) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.5846*** | |||||
| (0.1097) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.8014*** | |||||
| (0.1170) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9194*** | |||||
| (0.1236) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0346 | |||||
| (0.1108) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.9456*** | |||||
| (0.1172) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2361*** | |||||
| (0.1149) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.8795*** | |||||
| (0.1109) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0799 | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1423*** | |||||
| (0.1105) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.4266*** | |||||
| (0.1071) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.8221*** | |||||
| (0.1124) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.4403*** | |||||
| (0.1113) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.4651*** | |||||
| (0.1126) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.2843*** | |||||
| (0.1157) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.8423*** | |||||
| (0.1111) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.2026** | |||||
| (0.1188) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3527*** | |||||
| (0.1241) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.8764*** | |||||
| (0.1125) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.9427*** | |||||
| (0.1203) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 1.0028*** | |||||
| (0.1186) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4064*** | |||||
| (0.1125) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.5144*** | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2580*** | |||||
| (0.1107) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3539*** | |||||
| (0.1101) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.4395*** | |||||
| (0.1117) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0085 | |||||
| (0.1116) | ||||||
| Post 2 indicator | -1.3896*** | |||||
| (0.0766) | ||||||
| Post 3 indicator | -2.2086*** | |||||
| (0.0755) | ||||||
| Post 4 indicator | -1.7401*** | |||||
| (0.0763) | ||||||
| Post 5 indicator | 0.0051 | |||||
| (0.0820) | ||||||
| Post 6 indicator | 0.4527*** | |||||
| (0.0844) | ||||||
| Post 7 indicator | -1.4228*** | |||||
| (0.0752) | ||||||
| Post 8 indicator | -0.9538*** | |||||
| (0.0779) | ||||||
| Post 9 indicator | -0.5281*** | |||||
| (0.0788) | ||||||
| Post 10 indicator | -0.8059*** | |||||
| (0.0763) | ||||||
| Post 11 indicator | -1.3501*** | |||||
| (0.0756) | ||||||
| Post 12 indicator | -1.0828*** | |||||
| (0.0758) | ||||||
| Post 13 indicator | -1.6220*** | |||||
| (0.0743) | ||||||
| Post 14 indicator | -2.2721*** | |||||
| (0.0771) | ||||||
| Post 15 indicator | -2.0199*** | |||||
| (0.0749) | ||||||
| Post 16 indicator | 1.6695*** | |||||
| (0.0824) | ||||||
| Post 17 indicator | 0.1269* | |||||
| (0.0846) | ||||||
| Post 18 indicator | -0.0573 | |||||
| (0.0815) | ||||||
| Post 19 indicator | 0.6525*** | |||||
| (0.0790) | ||||||
| Post 20 indicator | 2.3196*** | |||||
| (0.0838) | ||||||
| Post 21 indicator | -0.1740** | |||||
| (0.0827) | ||||||
| Post 22 indicator | 0.1395** | |||||
| (0.0819) | ||||||
| Post 23 indicator | 0.7881*** | |||||
| (0.0812) | ||||||
| Post 24 indicator | 2.2808*** | |||||
| (0.0839) | ||||||
| Post 25 indicator | 1.2372*** | |||||
| (0.0799) | ||||||
| Post 26 indicator | 0.9195*** | |||||
| (0.0829) | ||||||
| Post 27 indicator | 0.6285*** | |||||
| (0.0802) | ||||||
| Post 28 indicator | 0.6333*** | |||||
| (0.0813) | ||||||
| Post 29 indicator | -0.2576*** | |||||
| (0.0832) | ||||||
| Post 30 indicator | 0.3885*** | |||||
| (0.0809) | ||||||
| Post 31 indicator | 1.4698*** | |||||
| (0.0799) | ||||||
| Post 32 indicator | -0.3676*** | |||||
| (0.0867) | ||||||
| Post 33 indicator | 0.1946*** | |||||
| (0.0794) | ||||||
| Post 34 indicator | 0.3143*** | |||||
| (0.0795) | ||||||
| Post 35 indicator | 2.2763*** | |||||
| (0.0839) | ||||||
| Post 36 indicator | 2.8419*** | |||||
| (0.0910) | ||||||
| Post 37 indicator | 0.0124 | |||||
| (0.0810) | ||||||
| Post 38 indicator | -0.4296*** | |||||
| (0.0859) | ||||||
| Post 39 indicator | 2.1778*** | |||||
| (0.0835) | ||||||
| Post 40 indicator | 1.5434*** | |||||
| (0.0810) | ||||||
| Post 41 indicator | 0.0398 | |||||
| (0.0829) | ||||||
| Post 42 indicator | 1.5293*** | |||||
| (0.0807) | ||||||
| Post 43 indicator | 1.2744*** | |||||
| (0.0780) | ||||||
| Post 44 indicator | 0.0198 | |||||
| (0.0829) | ||||||
| Post 45 indicator | 1.8902*** | |||||
| (0.0825) | ||||||
| Post 46 indicator | 2.0114*** | |||||
| (0.0830) | ||||||
| Post 47 indicator | 1.0869*** | |||||
| (0.0811) | ||||||
| Post 48 indicator | -0.4814*** | |||||
| (0.0840) | ||||||
| Post 49 indicator | 1.1137*** | |||||
| (0.0820) | ||||||
| Post 50 indicator | 2.2191*** | |||||
| (0.0846) | ||||||
| Post 51 indicator | 2.8168*** | |||||
| (0.0890) | ||||||
| Post 52 indicator | -0.2878*** | |||||
| (0.0832) | ||||||
| Post 53 indicator | -0.8851*** | |||||
| (0.0926) | ||||||
| Post 54 indicator | 2.4861*** | |||||
| (0.0862) | ||||||
| Post 55 indicator | 1.6118*** | |||||
| (0.0820) | ||||||
| Post 56 indicator | 1.1756*** | |||||
| (0.0838) | ||||||
| Post 57 indicator | 1.1865*** | |||||
| (0.0791) | ||||||
| Post 58 indicator | 0.7433*** | |||||
| (0.0798) | ||||||
| Post 59 indicator | 0.1789** | |||||
| (0.0816) | ||||||
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166604.6384 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83302.3192 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2281 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1038* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0793) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Emotions x 1[Post-treatment] | -0.4747*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0657) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Emotions x 1[Misinfo] | 0.0233 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0575) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166585.4008 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83292.7004 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2281 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0473*** | |||||
| (0.0348) | ||||||
| No-course Baseline | 0.0159 | |||||
| (0.0508) | ||||||
| Emotions | -0.0323 | |||||
| (0.0482) | ||||||
| Reasoning | 0.0722* | |||||
| (0.0500) | ||||||
| Combo | -0.0200 | |||||
| (0.0494) | ||||||
| 1[Post-treatment] | -0.4565*** | -0.5002*** | -0.5575*** | -0.5005*** | -0.5596*** | -0.5596*** |
| (0.0476) | (0.0589) | (0.0631) | (0.0589) | (0.0639) | (0.0639) | |
| 1[Misinfo] | -0.8428*** | -0.9234*** | -1.0396*** | -0.8461*** | -1.0772*** | -3.0886*** |
| (0.0414) | (0.0453) | (0.0483) | (0.0464) | (0.0958) | (0.0973) | |
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9365*** | |||||
| (0.0379) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5408*** | |||||
| (0.0166) | ||||||
| Emotions posts | 0.0116 | |||||
| (0.0241) | ||||||
| Reasoning posts | -0.0236 | |||||
| (0.0237) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.2988*** | 0.3289*** | 0.3843*** | 0.3290*** | 0.3850*** | 0.3850*** |
| (0.0701) | (0.0859) | (0.0921) | (0.0859) | (0.0937) | (0.0937) | |
| Emotions x 1[Post-treatment] | -0.4747*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0657) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Reasoning x 1[Post-treatment] | -0.4364*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0680) | (0.0848) | (0.0910) | (0.0848) | (0.0922) | (0.0922) | |
| Combo x 1[Post-treatment] | -0.4866*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0674) | (0.0845) | (0.0911) | (0.0846) | (0.0924) | (0.0924) | |
| No-course Baseline x 1[Misinfo] | 0.0058 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0606) | (0.0668) | (0.0708) | (0.0668) | (0.0722) | (0.0722) | |
| Emotions x 1[Misinfo] | 0.0233 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0575) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Reasoning x 1[Misinfo] | -0.1063** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0596) | (0.0657) | (0.0695) | (0.0657) | (0.0710) | (0.0710) | |
| Combo x 1[Misinfo] | 0.0303 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0589) | (0.0645) | (0.0687) | (0.0646) | (0.0703) | (0.0703) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0526 | 0.0492 | 0.0613 | 0.0485 | 0.0507 | 0.0507 |
| (0.0573) | (0.0617) | (0.0640) | (0.0618) | (0.0648) | (0.0648) | |
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | 0.0243 | |||||
| (0.0550) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.2211*** | |||||
| (0.0536) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | 0.0061 | |||||
| (0.0542) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | 16.7667 | |||||
| (164617.3440) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0315* | |||||
| (0.0239) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0661*** | |||||
| (0.0229) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0397** | |||||
| (0.0239) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0243 | |||||
| (0.0237) | ||||||
| No-course Baseline x Emotions posts | 0.0015 | |||||
| (0.0349) | ||||||
| Emotions x Emotions posts | 0.0387 | |||||
| (0.0344) | ||||||
| Reasoning x Emotions posts | -0.0139 | |||||
| (0.0349) | ||||||
| Combo x Emotions posts | -6.9119 | |||||
| (67270.8308) | ||||||
| No-course Baseline x Reasoning posts | 0.0123 | |||||
| (0.0348) | ||||||
| Emotions x Reasoning posts | 0.0671** | |||||
| (0.0341) | ||||||
| Reasoning x Reasoning posts | -0.0414 | |||||
| (0.0339) | ||||||
| Combo x Reasoning posts | -6.7738 | |||||
| (66632.5497) | ||||||
| Combo x Combo posts | -6.9652 | |||||
| (67995.6434) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | -0.0004 | 0.0069 | -0.0022 | 0.0073 | 0.0198 | 0.0198 |
| (0.0841) | (0.0921) | (0.0958) | (0.0921) | (0.0971) | (0.0971) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1038* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0793) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.1106* | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0818) | (0.0892) | (0.0922) | (0.0893) | (0.0935) | (0.0935) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0579 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0811) | (0.0887) | (0.0924) | (0.0887) | (0.0940) | (0.0940) | |
| Post domain 2 | -1.3458*** | -1.3896*** | ||||
| (0.0391) | (0.0766) | |||||
| Post domain 3 | -1.8926*** | -2.2086*** | ||||
| (0.0389) | (0.0755) | |||||
| Post domain 4 | -1.2128*** | -1.7401*** | ||||
| (0.0384) | (0.0763) | |||||
| Post domain 5 | 0.4049*** | 0.0051 | ||||
| (0.0388) | (0.0820) | |||||
| Post domain 6 | 0.1237*** | 0.4527*** | ||||
| (0.0395) | (0.0844) | |||||
| Post domain 7 | -1.5746*** | -1.4228*** | ||||
| (0.0388) | (0.0752) | |||||
| Post domain 8 | -1.4724*** | -0.9538*** | ||||
| (0.0391) | (0.0779) | |||||
| Post domain 9 | 0.2728*** | -0.5281*** | ||||
| (0.0392) | (0.0788) | |||||
| Post domain 10 | -0.3969*** | -0.8059*** | ||||
| (0.0377) | (0.0763) | |||||
| Post domain 11 | -1.0747*** | -1.3501*** | ||||
| (0.0388) | (0.0756) | |||||
| Post domain 12 | -0.7213*** | -1.0828*** | ||||
| (0.0372) | (0.0758) | |||||
| Post domain 13 | -1.0425*** | -1.6220*** | ||||
| (0.0378) | (0.0743) | |||||
| Post domain 14 | -1.8560*** | -2.2721*** | ||||
| (0.0403) | (0.0771) | |||||
| Post domain 15 | -1.2523*** | -2.0199*** | ||||
| (0.0381) | (0.0749) | |||||
| 1[Combo manipulation tactic] | -0.2264*** | -0.3419*** | ||||
| (0.0187) | (0.0832) | |||||
| 1[Emotions manipulation tactic] | -0.0068 | -0.5416*** | ||||
| (0.0184) | (0.0822) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.1530* | |||||
| (0.1163) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4818*** | |||||
| (0.1123) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.7230*** | |||||
| (0.1104) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6450*** | |||||
| (0.1186) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.2961*** | |||||
| (0.1206) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.1072 | |||||
| (0.1122) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0724 | |||||
| (0.1141) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.1394*** | |||||
| (0.1160) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.3736*** | |||||
| (0.1122) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.6002*** | |||||
| (0.1113) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | 0.0418 | |||||
| (0.1136) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.5858*** | |||||
| (0.1106) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.3450*** | |||||
| (0.1139) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.7389*** | |||||
| (0.1105) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4478*** | |||||
| (0.1167) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.9334*** | |||||
| (0.1107) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.5846*** | |||||
| (0.1097) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.8014*** | |||||
| (0.1170) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9194*** | |||||
| (0.1236) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0346 | |||||
| (0.1108) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.9456*** | |||||
| (0.1172) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2361*** | |||||
| (0.1149) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.8795*** | |||||
| (0.1109) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0799 | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1423*** | |||||
| (0.1105) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.4266*** | |||||
| (0.1071) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.8221*** | |||||
| (0.1124) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.4403*** | |||||
| (0.1113) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.4651*** | |||||
| (0.1126) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.2843*** | |||||
| (0.1157) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.8423*** | |||||
| (0.1111) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.2026** | |||||
| (0.1188) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3527*** | |||||
| (0.1241) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.8764*** | |||||
| (0.1125) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.9427*** | |||||
| (0.1203) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 1.0028*** | |||||
| (0.1186) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4064*** | |||||
| (0.1125) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.5144*** | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2580*** | |||||
| (0.1107) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3539*** | |||||
| (0.1101) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.4395*** | |||||
| (0.1117) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0085 | |||||
| (0.1116) | ||||||
| Post 2 indicator | -1.3896*** | |||||
| (0.0766) | ||||||
| Post 3 indicator | -2.2086*** | |||||
| (0.0755) | ||||||
| Post 4 indicator | -1.7401*** | |||||
| (0.0763) | ||||||
| Post 5 indicator | 0.0051 | |||||
| (0.0820) | ||||||
| Post 6 indicator | 0.4527*** | |||||
| (0.0844) | ||||||
| Post 7 indicator | -1.4228*** | |||||
| (0.0752) | ||||||
| Post 8 indicator | -0.9538*** | |||||
| (0.0779) | ||||||
| Post 9 indicator | -0.5281*** | |||||
| (0.0788) | ||||||
| Post 10 indicator | -0.8059*** | |||||
| (0.0763) | ||||||
| Post 11 indicator | -1.3501*** | |||||
| (0.0756) | ||||||
| Post 12 indicator | -1.0828*** | |||||
| (0.0758) | ||||||
| Post 13 indicator | -1.6220*** | |||||
| (0.0743) | ||||||
| Post 14 indicator | -2.2721*** | |||||
| (0.0771) | ||||||
| Post 15 indicator | -2.0199*** | |||||
| (0.0749) | ||||||
| Post 16 indicator | 1.6695*** | |||||
| (0.0824) | ||||||
| Post 17 indicator | 0.1269* | |||||
| (0.0846) | ||||||
| Post 18 indicator | -0.0573 | |||||
| (0.0815) | ||||||
| Post 19 indicator | 0.6525*** | |||||
| (0.0790) | ||||||
| Post 20 indicator | 2.3196*** | |||||
| (0.0838) | ||||||
| Post 21 indicator | -0.1740** | |||||
| (0.0827) | ||||||
| Post 22 indicator | 0.1395** | |||||
| (0.0819) | ||||||
| Post 23 indicator | 0.7881*** | |||||
| (0.0812) | ||||||
| Post 24 indicator | 2.2808*** | |||||
| (0.0839) | ||||||
| Post 25 indicator | 1.2372*** | |||||
| (0.0799) | ||||||
| Post 26 indicator | 0.9195*** | |||||
| (0.0829) | ||||||
| Post 27 indicator | 0.6285*** | |||||
| (0.0802) | ||||||
| Post 28 indicator | 0.6333*** | |||||
| (0.0813) | ||||||
| Post 29 indicator | -0.2576*** | |||||
| (0.0832) | ||||||
| Post 30 indicator | 0.3885*** | |||||
| (0.0809) | ||||||
| Post 31 indicator | 1.4698*** | |||||
| (0.0799) | ||||||
| Post 32 indicator | -0.3676*** | |||||
| (0.0867) | ||||||
| Post 33 indicator | 0.1946*** | |||||
| (0.0794) | ||||||
| Post 34 indicator | 0.3143*** | |||||
| (0.0795) | ||||||
| Post 35 indicator | 2.2763*** | |||||
| (0.0839) | ||||||
| Post 36 indicator | 2.8419*** | |||||
| (0.0910) | ||||||
| Post 37 indicator | 0.0124 | |||||
| (0.0810) | ||||||
| Post 38 indicator | -0.4296*** | |||||
| (0.0859) | ||||||
| Post 39 indicator | 2.1778*** | |||||
| (0.0835) | ||||||
| Post 40 indicator | 1.5434*** | |||||
| (0.0810) | ||||||
| Post 41 indicator | 0.0398 | |||||
| (0.0829) | ||||||
| Post 42 indicator | 1.5293*** | |||||
| (0.0807) | ||||||
| Post 43 indicator | 1.2744*** | |||||
| (0.0780) | ||||||
| Post 44 indicator | 0.0198 | |||||
| (0.0829) | ||||||
| Post 45 indicator | 1.8902*** | |||||
| (0.0825) | ||||||
| Post 46 indicator | 2.0114*** | |||||
| (0.0830) | ||||||
| Post 47 indicator | 1.0869*** | |||||
| (0.0811) | ||||||
| Post 48 indicator | -0.4814*** | |||||
| (0.0840) | ||||||
| Post 49 indicator | 1.1137*** | |||||
| (0.0820) | ||||||
| Post 50 indicator | 2.2191*** | |||||
| (0.0846) | ||||||
| Post 51 indicator | 2.8168*** | |||||
| (0.0890) | ||||||
| Post 52 indicator | -0.2878*** | |||||
| (0.0832) | ||||||
| Post 53 indicator | -0.8851*** | |||||
| (0.0926) | ||||||
| Post 54 indicator | 2.4861*** | |||||
| (0.0862) | ||||||
| Post 55 indicator | 1.6118*** | |||||
| (0.0820) | ||||||
| Post 56 indicator | 1.1756*** | |||||
| (0.0838) | ||||||
| Post 57 indicator | 1.1865*** | |||||
| (0.0791) | ||||||
| Post 58 indicator | 0.7433*** | |||||
| (0.0798) | ||||||
| Post 59 indicator | 0.1789** | |||||
| (0.0816) | ||||||
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166585.4008 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83292.7004 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2281 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1044* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0795) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Emotions x 1[Post-treatment] | -0.4777*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0659) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Emotions x 1[Misinfo] | 0.0224 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0577) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166019.7932 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83009.8966 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2300 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0563*** | |||||
| (0.0349) | ||||||
| No-course Baseline | 0.0062 | |||||
| (0.0510) | ||||||
| Emotions | -0.0355 | |||||
| (0.0484) | ||||||
| Reasoning | 0.0638 | |||||
| (0.0502) | ||||||
| Combo | -0.0269 | |||||
| (0.0495) | ||||||
| 1[Post-treatment] | -0.4587*** | -0.5002*** | -0.5575*** | -0.5005*** | -0.5596*** | -0.5596*** |
| (0.0477) | (0.0589) | (0.0631) | (0.0589) | (0.0639) | (0.0639) | |
| 1[Misinfo] | -0.8466*** | -0.9234*** | -1.0396*** | -0.8461*** | -1.0772*** | -3.0886*** |
| (0.0416) | (0.0453) | (0.0483) | (0.0464) | (0.0958) | (0.0973) | |
| Age | -0.0316** | |||||
| (0.0151) | ||||||
| Man | 0.0151 | |||||
| (0.0135) | ||||||
| Some college | -0.0110 | |||||
| (0.0168) | ||||||
| Bachelor’s degree | 0.0218* | |||||
| (0.0170) | ||||||
| Graduate degree | -0.0412*** | |||||
| (0.0146) | ||||||
| Married | -0.0154 | |||||
| (0.0151) | ||||||
| Unemployed | -0.0280* | |||||
| (0.0177) | ||||||
| Employed | -0.0483*** | |||||
| (0.0193) | ||||||
| Mostly urban | -0.0277** | |||||
| (0.0161) | ||||||
| Suburban | -0.0285** | |||||
| (0.0157) | ||||||
| Christian | 0.0142 | |||||
| (0.0138) | ||||||
| Attends religious services | 0.0051 | |||||
| (0.0139) | ||||||
| Uses social media | -0.0345*** | |||||
| (0.0123) | ||||||
| Hours spent on Social Media | 0.0297** | |||||
| (0.0149) | ||||||
| Social Media Share: 80-100% | 0.0846*** | |||||
| (0.0174) | ||||||
| Social Media Share: 60-80% | 0.1020*** | |||||
| (0.0189) | ||||||
| Social Media Share: 40-60% | 0.0893*** | |||||
| (0.0192) | ||||||
| Social Media Share: 20-40% | 0.0480*** | |||||
| (0.0176) | ||||||
| Passed attention check at pre-treatment | -0.0276** | |||||
| (0.0135) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9020*** | |||||
| (0.0178) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5294*** | |||||
| (0.0167) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.3006*** | 0.3289*** | 0.3843*** | 0.3290*** | 0.3850*** | 0.3850*** |
| (0.0703) | (0.0859) | (0.0921) | (0.0859) | (0.0937) | (0.0937) | |
| Emotions x 1[Post-treatment] | -0.4777*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0659) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Reasoning x 1[Post-treatment] | -0.4372*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0682) | (0.0848) | (0.0910) | (0.0848) | (0.0922) | (0.0922) | |
| Combo x 1[Post-treatment] | -0.4875*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0675) | (0.0845) | (0.0911) | (0.0846) | (0.0924) | (0.0924) | |
| No-course Baseline x 1[Misinfo] | 0.0073 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0607) | (0.0668) | (0.0708) | (0.0668) | (0.0722) | (0.0722) | |
| Emotions x 1[Misinfo] | 0.0224 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0577) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Reasoning x 1[Misinfo] | -0.1057** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0597) | (0.0657) | (0.0695) | (0.0657) | (0.0710) | (0.0710) | |
| Combo x 1[Misinfo] | 0.0314 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0590) | (0.0645) | (0.0687) | (0.0646) | (0.0703) | (0.0703) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0533 | 0.0492 | 0.0613 | 0.0485 | 0.0507 | 0.0507 |
| (0.0574) | (0.0617) | (0.0640) | (0.0618) | (0.0648) | (0.0648) | |
| No-course Baseline x Age | 0.0023 | |||||
| (0.0209) | ||||||
| Emotions x Age | -0.0271 | |||||
| (0.0231) | ||||||
| Reasoning x Age | -0.0179 | |||||
| (0.0221) | ||||||
| Combo x Age | -0.0044 | |||||
| (0.0215) | ||||||
| No-course Baseline x Man | -0.0354** | |||||
| (0.0199) | ||||||
| Emotions x Man | 0.0319** | |||||
| (0.0188) | ||||||
| Reasoning x Man | -0.0325** | |||||
| (0.0193) | ||||||
| Combo x Man | 0.0220 | |||||
| (0.0192) | ||||||
| No-course Baseline x Some college | -0.0076 | |||||
| (0.0240) | ||||||
| Emotions x Some college | 0.0018 | |||||
| (0.0231) | ||||||
| Reasoning x Some college | -0.0029 | |||||
| (0.0237) | ||||||
| Combo x Some college | -0.0127 | |||||
| (0.0237) | ||||||
| No-course Baseline x Bachelor’s degree | -0.0243 | |||||
| (0.0243) | ||||||
| Emotions x Bachelor’s degree | -0.0197 | |||||
| (0.0232) | ||||||
| Reasoning x Bachelor’s degree | -0.0456** | |||||
| (0.0239) | ||||||
| Combo x Bachelor’s degree | -0.0497** | |||||
| (0.0238) | ||||||
| No-course Baseline x Graduate degree | 0.0144 | |||||
| (0.0214) | ||||||
| Emotions x Graduate degree | 0.0484*** | |||||
| (0.0193) | ||||||
| Reasoning x Graduate degree | 0.0047 | |||||
| (0.0212) | ||||||
| Combo x Graduate degree | 0.0493*** | |||||
| (0.0190) | ||||||
| No-course Baseline x Married | 0.0077 | |||||
| (0.0223) | ||||||
| Emotions x Married | -0.0250 | |||||
| (0.0212) | ||||||
| Reasoning x Married | 0.0117 | |||||
| (0.0220) | ||||||
| Combo x Married | -0.0264 | |||||
| (0.0218) | ||||||
| No-course Baseline x Unemployed | 0.0119 | |||||
| (0.0254) | ||||||
| Emotions x Unemployed | -0.0060 | |||||
| (0.0244) | ||||||
| Reasoning x Unemployed | 0.0328* | |||||
| (0.0250) | ||||||
| Combo x Unemployed | 0.0133 | |||||
| (0.0250) | ||||||
| No-course Baseline x Employed | 0.0059 | |||||
| (0.0278) | ||||||
| Emotions x Employed | 0.0147 | |||||
| (0.0266) | ||||||
| Reasoning x Employed | 0.0225 | |||||
| (0.0276) | ||||||
| Combo x Employed | 0.0136 | |||||
| (0.0269) | ||||||
| No-course Baseline x Mostly urban | 0.0390** | |||||
| (0.0231) | ||||||
| Emotions x Mostly urban | 0.0429** | |||||
| (0.0222) | ||||||
| Reasoning x Mostly urban | -0.0104 | |||||
| (0.0230) | ||||||
| Combo x Mostly urban | 0.0587*** | |||||
| (0.0225) | ||||||
| No-course Baseline x Suburban | 0.0165 | |||||
| (0.0228) | ||||||
| Emotions x Suburban | -0.0094 | |||||
| (0.0219) | ||||||
| Reasoning x Suburban | -0.0220 | |||||
| (0.0223) | ||||||
| Combo x Suburban | 0.0543*** | |||||
| (0.0223) | ||||||
| No-course Baseline x Christian | -0.0412** | |||||
| (0.0195) | ||||||
| Emotions x Christian | -0.0132 | |||||
| (0.0196) | ||||||
| Reasoning x Christian | 0.0204 | |||||
| (0.0202) | ||||||
| Combo x Christian | -0.0188 | |||||
| (0.0198) | ||||||
| No-course Baseline x Attends religious services | -0.0079 | |||||
| (0.0193) | ||||||
| Emotions x Attends religious services | -0.0395** | |||||
| (0.0196) | ||||||
| Reasoning x Attends religious services | -0.0192 | |||||
| (0.0202) | ||||||
| Combo x Attends religious services | -0.0142 | |||||
| (0.0201) | ||||||
| No-course Baseline x Uses social media | 0.0360** | |||||
| (0.0190) | ||||||
| Emotions x Uses social media | 0.0530*** | |||||
| (0.0197) | ||||||
| Reasoning x Uses social media | 0.0406** | |||||
| (0.0180) | ||||||
| Combo x Uses social media | 0.0305* | |||||
| (0.0203) | ||||||
| No-course Baseline x Hours spent on Social Media | -0.0381** | |||||
| (0.0207) | ||||||
| Emotions x Hours spent on Social Media | -0.0167 | |||||
| (0.0203) | ||||||
| Reasoning x Hours spent on Social Media | 0.0015 | |||||
| (0.0210) | ||||||
| Combo x Hours spent on Social Media | -0.0640*** | |||||
| (0.0203) | ||||||
| No-course Baseline x Social Media Share: 80-100% | -0.0005 | |||||
| (0.0247) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0054 | |||||
| (0.0241) | ||||||
| Reasoning x Social Media Share: 80-100% | -0.0018 | |||||
| (0.0244) | ||||||
| Combo x Social Media Share: 80-100% | -0.0085 | |||||
| (0.0241) | ||||||
| No-course Baseline x Social Media Share: 60-80% | -0.0288 | |||||
| (0.0279) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0296 | |||||
| (0.0265) | ||||||
| Reasoning x Social Media Share: 60-80% | -0.0010 | |||||
| (0.0274) | ||||||
| Combo x Social Media Share: 60-80% | 0.0020 | |||||
| (0.0270) | ||||||
| No-course Baseline x Social Media Share: 40-60% | -0.0405* | |||||
| (0.0278) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0100 | |||||
| (0.0269) | ||||||
| Reasoning x Social Media Share: 40-60% | -0.0329 | |||||
| (0.0280) | ||||||
| Combo x Social Media Share: 40-60% | -0.0237 | |||||
| (0.0274) | ||||||
| No-course Baseline x Social Media Share: 20-40% | -0.0228 | |||||
| (0.0255) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0225 | |||||
| (0.0248) | ||||||
| Reasoning x Social Media Share: 20-40% | -0.0249 | |||||
| (0.0256) | ||||||
| Combo x Social Media Share: 20-40% | -0.0272 | |||||
| (0.0254) | ||||||
| No-course Baseline x Passed attention check at pre-treatment | -0.0408** | |||||
| (0.0198) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0025 | |||||
| (0.0188) | ||||||
| Reasoning x Passed attention check at pre-treatment | 0.0044 | |||||
| (0.0193) | ||||||
| Combo x Passed attention check at pre-treatment | 0.0429** | |||||
| (0.0193) | ||||||
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | 0.0305 | |||||
| (0.0258) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -0.1443*** | |||||
| (0.0242) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | -0.0344* | |||||
| (0.0250) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.0531** | |||||
| (0.0251) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0188 | |||||
| (0.0241) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0526** | |||||
| (0.0231) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0347* | |||||
| (0.0241) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0167 | |||||
| (0.0238) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | -0.0011 | 0.0069 | -0.0022 | 0.0073 | 0.0198 | 0.0198 |
| (0.0843) | (0.0921) | (0.0958) | (0.0921) | (0.0971) | (0.0971) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1044* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0795) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.1100* | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0820) | (0.0892) | (0.0922) | (0.0893) | (0.0935) | (0.0935) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0589 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0813) | (0.0887) | (0.0924) | (0.0887) | (0.0940) | (0.0940) | |
| Post domain 2 | -1.3458*** | -1.3896*** | ||||
| (0.0391) | (0.0766) | |||||
| Post domain 3 | -1.8926*** | -2.2086*** | ||||
| (0.0389) | (0.0755) | |||||
| Post domain 4 | -1.2128*** | -1.7401*** | ||||
| (0.0384) | (0.0763) | |||||
| Post domain 5 | 0.4049*** | 0.0051 | ||||
| (0.0388) | (0.0820) | |||||
| Post domain 6 | 0.1237*** | 0.4527*** | ||||
| (0.0395) | (0.0844) | |||||
| Post domain 7 | -1.5746*** | -1.4228*** | ||||
| (0.0388) | (0.0752) | |||||
| Post domain 8 | -1.4724*** | -0.9538*** | ||||
| (0.0391) | (0.0779) | |||||
| Post domain 9 | 0.2728*** | -0.5281*** | ||||
| (0.0392) | (0.0788) | |||||
| Post domain 10 | -0.3969*** | -0.8059*** | ||||
| (0.0377) | (0.0763) | |||||
| Post domain 11 | -1.0747*** | -1.3501*** | ||||
| (0.0388) | (0.0756) | |||||
| Post domain 12 | -0.7213*** | -1.0828*** | ||||
| (0.0372) | (0.0758) | |||||
| Post domain 13 | -1.0425*** | -1.6220*** | ||||
| (0.0378) | (0.0743) | |||||
| Post domain 14 | -1.8560*** | -2.2721*** | ||||
| (0.0403) | (0.0771) | |||||
| Post domain 15 | -1.2523*** | -2.0199*** | ||||
| (0.0381) | (0.0749) | |||||
| 1[Combo manipulation tactic] | -0.2264*** | -0.3419*** | ||||
| (0.0187) | (0.0832) | |||||
| 1[Emotions manipulation tactic] | -0.0068 | -0.5416*** | ||||
| (0.0184) | (0.0822) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.1530* | |||||
| (0.1163) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4818*** | |||||
| (0.1123) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.7230*** | |||||
| (0.1104) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6450*** | |||||
| (0.1186) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.2961*** | |||||
| (0.1206) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.1072 | |||||
| (0.1122) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0724 | |||||
| (0.1141) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.1394*** | |||||
| (0.1160) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.3736*** | |||||
| (0.1122) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.6002*** | |||||
| (0.1113) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | 0.0418 | |||||
| (0.1136) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.5858*** | |||||
| (0.1106) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.3450*** | |||||
| (0.1139) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.7389*** | |||||
| (0.1105) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4478*** | |||||
| (0.1167) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.9334*** | |||||
| (0.1107) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.5846*** | |||||
| (0.1097) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.8014*** | |||||
| (0.1170) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9194*** | |||||
| (0.1236) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0346 | |||||
| (0.1108) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.9456*** | |||||
| (0.1172) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2361*** | |||||
| (0.1149) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.8795*** | |||||
| (0.1109) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0799 | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1423*** | |||||
| (0.1105) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.4266*** | |||||
| (0.1071) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.8221*** | |||||
| (0.1124) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.4403*** | |||||
| (0.1113) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.4651*** | |||||
| (0.1126) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.2843*** | |||||
| (0.1157) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.8423*** | |||||
| (0.1111) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.2026** | |||||
| (0.1188) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3527*** | |||||
| (0.1241) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.8764*** | |||||
| (0.1125) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.9427*** | |||||
| (0.1203) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 1.0028*** | |||||
| (0.1186) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4064*** | |||||
| (0.1125) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.5144*** | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2580*** | |||||
| (0.1107) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3539*** | |||||
| (0.1101) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.4395*** | |||||
| (0.1117) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0085 | |||||
| (0.1116) | ||||||
| Post 2 indicator | -1.3896*** | |||||
| (0.0766) | ||||||
| Post 3 indicator | -2.2086*** | |||||
| (0.0755) | ||||||
| Post 4 indicator | -1.7401*** | |||||
| (0.0763) | ||||||
| Post 5 indicator | 0.0051 | |||||
| (0.0820) | ||||||
| Post 6 indicator | 0.4527*** | |||||
| (0.0844) | ||||||
| Post 7 indicator | -1.4228*** | |||||
| (0.0752) | ||||||
| Post 8 indicator | -0.9538*** | |||||
| (0.0779) | ||||||
| Post 9 indicator | -0.5281*** | |||||
| (0.0788) | ||||||
| Post 10 indicator | -0.8059*** | |||||
| (0.0763) | ||||||
| Post 11 indicator | -1.3501*** | |||||
| (0.0756) | ||||||
| Post 12 indicator | -1.0828*** | |||||
| (0.0758) | ||||||
| Post 13 indicator | -1.6220*** | |||||
| (0.0743) | ||||||
| Post 14 indicator | -2.2721*** | |||||
| (0.0771) | ||||||
| Post 15 indicator | -2.0199*** | |||||
| (0.0749) | ||||||
| Post 16 indicator | 1.6695*** | |||||
| (0.0824) | ||||||
| Post 17 indicator | 0.1269* | |||||
| (0.0846) | ||||||
| Post 18 indicator | -0.0573 | |||||
| (0.0815) | ||||||
| Post 19 indicator | 0.6525*** | |||||
| (0.0790) | ||||||
| Post 20 indicator | 2.3196*** | |||||
| (0.0838) | ||||||
| Post 21 indicator | -0.1740** | |||||
| (0.0827) | ||||||
| Post 22 indicator | 0.1395** | |||||
| (0.0819) | ||||||
| Post 23 indicator | 0.7881*** | |||||
| (0.0812) | ||||||
| Post 24 indicator | 2.2808*** | |||||
| (0.0839) | ||||||
| Post 25 indicator | 1.2372*** | |||||
| (0.0799) | ||||||
| Post 26 indicator | 0.9195*** | |||||
| (0.0829) | ||||||
| Post 27 indicator | 0.6285*** | |||||
| (0.0802) | ||||||
| Post 28 indicator | 0.6333*** | |||||
| (0.0813) | ||||||
| Post 29 indicator | -0.2576*** | |||||
| (0.0832) | ||||||
| Post 30 indicator | 0.3885*** | |||||
| (0.0809) | ||||||
| Post 31 indicator | 1.4698*** | |||||
| (0.0799) | ||||||
| Post 32 indicator | -0.3676*** | |||||
| (0.0867) | ||||||
| Post 33 indicator | 0.1946*** | |||||
| (0.0794) | ||||||
| Post 34 indicator | 0.3143*** | |||||
| (0.0795) | ||||||
| Post 35 indicator | 2.2763*** | |||||
| (0.0839) | ||||||
| Post 36 indicator | 2.8419*** | |||||
| (0.0910) | ||||||
| Post 37 indicator | 0.0124 | |||||
| (0.0810) | ||||||
| Post 38 indicator | -0.4296*** | |||||
| (0.0859) | ||||||
| Post 39 indicator | 2.1778*** | |||||
| (0.0835) | ||||||
| Post 40 indicator | 1.5434*** | |||||
| (0.0810) | ||||||
| Post 41 indicator | 0.0398 | |||||
| (0.0829) | ||||||
| Post 42 indicator | 1.5293*** | |||||
| (0.0807) | ||||||
| Post 43 indicator | 1.2744*** | |||||
| (0.0780) | ||||||
| Post 44 indicator | 0.0198 | |||||
| (0.0829) | ||||||
| Post 45 indicator | 1.8902*** | |||||
| (0.0825) | ||||||
| Post 46 indicator | 2.0114*** | |||||
| (0.0830) | ||||||
| Post 47 indicator | 1.0869*** | |||||
| (0.0811) | ||||||
| Post 48 indicator | -0.4814*** | |||||
| (0.0840) | ||||||
| Post 49 indicator | 1.1137*** | |||||
| (0.0820) | ||||||
| Post 50 indicator | 2.2191*** | |||||
| (0.0846) | ||||||
| Post 51 indicator | 2.8168*** | |||||
| (0.0890) | ||||||
| Post 52 indicator | -0.2878*** | |||||
| (0.0832) | ||||||
| Post 53 indicator | -0.8851*** | |||||
| (0.0926) | ||||||
| Post 54 indicator | 2.4861*** | |||||
| (0.0862) | ||||||
| Post 55 indicator | 1.6118*** | |||||
| (0.0820) | ||||||
| Post 56 indicator | 1.1756*** | |||||
| (0.0838) | ||||||
| Post 57 indicator | 1.1865*** | |||||
| (0.0791) | ||||||
| Post 58 indicator | 0.7433*** | |||||
| (0.0798) | ||||||
| Post 59 indicator | 0.1789** | |||||
| (0.0816) | ||||||
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166019.7932 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83009.8966 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2300 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | Demographics + Sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
\[\text{Demographics + Sharing rates + Tactic sharing rates}\]
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1040* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0795) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Emotions x 1[Post-treatment] | -0.4780*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0659) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Emotions x 1[Misinfo] | 0.0222 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0577) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166005.0252 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83002.5126 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2299 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Click to view the full table
| Share (binary indicator) | ||||||
|---|---|---|---|---|---|---|
|
Baseline (1) |
User ID FE (2) |
User ID FE + Post domain FE (3) |
User ID FE + Post tactic FE (4) |
User ID FE + Post domain x Post tactic (5) |
User ID FE + Post ID FE (6) |
|
| Facts Baseline | 1.0571*** | |||||
| (0.0349) | ||||||
| No-course Baseline | 0.0053 | |||||
| (0.0510) | ||||||
| Emotions | -0.0372 | |||||
| (0.0484) | ||||||
| Reasoning | 0.0624 | |||||
| (0.0502) | ||||||
| Combo | -0.0276 | |||||
| (0.0495) | ||||||
| 1[Post-treatment] | -0.4587*** | -0.5002*** | -0.5575*** | -0.5005*** | -0.5596*** | -0.5596*** |
| (0.0477) | (0.0589) | (0.0631) | (0.0589) | (0.0639) | (0.0639) | |
| 1[Misinfo] | -0.8466*** | -0.9234*** | -1.0396*** | -0.8461*** | -1.0772*** | -3.0886*** |
| (0.0416) | (0.0453) | (0.0483) | (0.0464) | (0.0958) | (0.0973) | |
| Age | -0.0315** | |||||
| (0.0152) | ||||||
| Man | 0.0152 | |||||
| (0.0135) | ||||||
| Some college | -0.0104 | |||||
| (0.0168) | ||||||
| Bachelor’s degree | 0.0221* | |||||
| (0.0170) | ||||||
| Graduate degree | -0.0415*** | |||||
| (0.0146) | ||||||
| Married | -0.0146 | |||||
| (0.0151) | ||||||
| Unemployed | -0.0281* | |||||
| (0.0177) | ||||||
| Employed | -0.0491*** | |||||
| (0.0193) | ||||||
| Mostly urban | -0.0265** | |||||
| (0.0161) | ||||||
| Suburban | -0.0280** | |||||
| (0.0157) | ||||||
| Christian | 0.0137 | |||||
| (0.0138) | ||||||
| Attends religious services | 0.0052 | |||||
| (0.0139) | ||||||
| Uses social media | -0.0347*** | |||||
| (0.0123) | ||||||
| Hours spent on Social Media | 0.0304** | |||||
| (0.0149) | ||||||
| Social Media Share: 80-100% | 0.0846*** | |||||
| (0.0174) | ||||||
| Social Media Share: 60-80% | 0.1021*** | |||||
| (0.0189) | ||||||
| Social Media Share: 40-60% | 0.0896*** | |||||
| (0.0192) | ||||||
| Social Media Share: 20-40% | 0.0480*** | |||||
| (0.0176) | ||||||
| Passed attention check at pre-treatment | -0.0272** | |||||
| (0.0135) | ||||||
| Avg. Sharing of Misinfo posts at pre-treatment | 0.9030*** | |||||
| (0.0384) | ||||||
| Avg. Sharing of Non-misinfo posts at pre-treatment | 0.5304*** | |||||
| (0.0167) | ||||||
| Emotions posts | 0.0165 | |||||
| (0.0243) | ||||||
| Reasoning posts | -0.0183 | |||||
| (0.0238) | ||||||
| No-course Baseline x 1[Post-treatment] | 0.3006*** | 0.3289*** | 0.3843*** | 0.3290*** | 0.3850*** | 0.3850*** |
| (0.0703) | (0.0859) | (0.0921) | (0.0859) | (0.0937) | (0.0937) | |
| Emotions x 1[Post-treatment] | -0.4780*** | -0.5257*** | -0.5516*** | -0.5260*** | -0.5707*** | -0.5707*** |
| (0.0659) | (0.0839) | (0.0900) | (0.0840) | (0.0914) | (0.0914) | |
| Reasoning x 1[Post-treatment] | -0.4374*** | -0.4785*** | -0.5282*** | -0.4787*** | -0.5422*** | -0.5422*** |
| (0.0682) | (0.0848) | (0.0910) | (0.0848) | (0.0922) | (0.0922) | |
| Combo x 1[Post-treatment] | -0.4875*** | -0.5308*** | -0.5848*** | -0.5311*** | -0.6072*** | -0.6072*** |
| (0.0675) | (0.0845) | (0.0911) | (0.0846) | (0.0924) | (0.0924) | |
| No-course Baseline x 1[Misinfo] | 0.0072 | 0.0182 | 0.0295 | 0.0185 | 0.0243 | 0.0243 |
| (0.0607) | (0.0668) | (0.0708) | (0.0668) | (0.0722) | (0.0722) | |
| Emotions x 1[Misinfo] | 0.0222 | 0.0199 | 0.0571 | 0.0201 | 0.0588 | 0.0588 |
| (0.0577) | (0.0638) | (0.0677) | (0.0638) | (0.0692) | (0.0692) | |
| Reasoning x 1[Misinfo] | -0.1060** | -0.1165** | -0.0821 | -0.1164** | -0.0846 | -0.0846 |
| (0.0597) | (0.0657) | (0.0695) | (0.0657) | (0.0710) | (0.0710) | |
| Combo x 1[Misinfo] | 0.0315 | 0.0352 | 0.0522 | 0.0354 | 0.0620 | 0.0620 |
| (0.0590) | (0.0645) | (0.0687) | (0.0646) | (0.0703) | (0.0703) | |
| 1[Post-treatment] x 1[Misinfo] | 0.0533 | 0.0492 | 0.0613 | 0.0485 | 0.0507 | 0.0507 |
| (0.0574) | (0.0617) | (0.0640) | (0.0618) | (0.0648) | (0.0648) | |
| No-course Baseline x Age | 0.0022 | |||||
| (0.0209) | ||||||
| Emotions x Age | -0.0262 | |||||
| (0.0231) | ||||||
| Reasoning x Age | -0.0172 | |||||
| (0.0221) | ||||||
| Combo x Age | -0.0048 | |||||
| (0.0215) | ||||||
| No-course Baseline x Man | -0.0351** | |||||
| (0.0199) | ||||||
| Emotions x Man | 0.0308* | |||||
| (0.0188) | ||||||
| Reasoning x Man | -0.0324** | |||||
| (0.0193) | ||||||
| Combo x Man | 0.0217 | |||||
| (0.0192) | ||||||
| No-course Baseline x Some college | -0.0081 | |||||
| (0.0240) | ||||||
| Emotions x Some college | 0.0012 | |||||
| (0.0232) | ||||||
| Reasoning x Some college | -0.0030 | |||||
| (0.0237) | ||||||
| Combo x Some college | -0.0134 | |||||
| (0.0237) | ||||||
| No-course Baseline x Bachelor’s degree | -0.0248 | |||||
| (0.0243) | ||||||
| Emotions x Bachelor’s degree | -0.0207 | |||||
| (0.0232) | ||||||
| Reasoning x Bachelor’s degree | -0.0446** | |||||
| (0.0240) | ||||||
| Combo x Bachelor’s degree | -0.0491** | |||||
| (0.0239) | ||||||
| No-course Baseline x Graduate degree | 0.0137 | |||||
| (0.0214) | ||||||
| Emotions x Graduate degree | 0.0486*** | |||||
| (0.0193) | ||||||
| Reasoning x Graduate degree | 0.0054 | |||||
| (0.0212) | ||||||
| Combo x Graduate degree | 0.0496*** | |||||
| (0.0190) | ||||||
| No-course Baseline x Married | 0.0067 | |||||
| (0.0223) | ||||||
| Emotions x Married | -0.0266 | |||||
| (0.0212) | ||||||
| Reasoning x Married | 0.0103 | |||||
| (0.0220) | ||||||
| Combo x Married | -0.0268 | |||||
| (0.0219) | ||||||
| No-course Baseline x Unemployed | 0.0121 | |||||
| (0.0254) | ||||||
| Emotions x Unemployed | -0.0072 | |||||
| (0.0244) | ||||||
| Reasoning x Unemployed | 0.0330* | |||||
| (0.0250) | ||||||
| Combo x Unemployed | 0.0134 | |||||
| (0.0250) | ||||||
| No-course Baseline x Employed | 0.0071 | |||||
| (0.0278) | ||||||
| Emotions x Employed | 0.0150 | |||||
| (0.0266) | ||||||
| Reasoning x Employed | 0.0241 | |||||
| (0.0276) | ||||||
| Combo x Employed | 0.0140 | |||||
| (0.0269) | ||||||
| No-course Baseline x Mostly urban | 0.0380* | |||||
| (0.0231) | ||||||
| Emotions x Mostly urban | 0.0410** | |||||
| (0.0222) | ||||||
| Reasoning x Mostly urban | -0.0113 | |||||
| (0.0230) | ||||||
| Combo x Mostly urban | 0.0572*** | |||||
| (0.0226) | ||||||
| No-course Baseline x Suburban | 0.0160 | |||||
| (0.0228) | ||||||
| Emotions x Suburban | -0.0099 | |||||
| (0.0219) | ||||||
| Reasoning x Suburban | -0.0218 | |||||
| (0.0223) | ||||||
| Combo x Suburban | 0.0534*** | |||||
| (0.0223) | ||||||
| No-course Baseline x Christian | -0.0407** | |||||
| (0.0195) | ||||||
| Emotions x Christian | -0.0129 | |||||
| (0.0196) | ||||||
| Reasoning x Christian | 0.0216 | |||||
| (0.0202) | ||||||
| Combo x Christian | -0.0182 | |||||
| (0.0198) | ||||||
| No-course Baseline x Attends religious services | -0.0076 | |||||
| (0.0193) | ||||||
| Emotions x Attends religious services | -0.0397** | |||||
| (0.0196) | ||||||
| Reasoning x Attends religious services | -0.0201 | |||||
| (0.0203) | ||||||
| Combo x Attends religious services | -0.0145 | |||||
| (0.0201) | ||||||
| No-course Baseline x Uses social media | 0.0360** | |||||
| (0.0190) | ||||||
| Emotions x Uses social media | 0.0526*** | |||||
| (0.0197) | ||||||
| Reasoning x Uses social media | 0.0401** | |||||
| (0.0180) | ||||||
| Combo x Uses social media | 0.0305* | |||||
| (0.0203) | ||||||
| No-course Baseline x Hours spent on Social Media | -0.0388** | |||||
| (0.0207) | ||||||
| Emotions x Hours spent on Social Media | -0.0178 | |||||
| (0.0204) | ||||||
| Reasoning x Hours spent on Social Media | 0.0010 | |||||
| (0.0210) | ||||||
| Combo x Hours spent on Social Media | -0.0647*** | |||||
| (0.0203) | ||||||
| No-course Baseline x Social Media Share: 80-100% | -0.0008 | |||||
| (0.0247) | ||||||
| Emotions x Social Media Share: 80-100% | -0.0057 | |||||
| (0.0241) | ||||||
| Reasoning x Social Media Share: 80-100% | -0.0015 | |||||
| (0.0244) | ||||||
| Combo x Social Media Share: 80-100% | -0.0084 | |||||
| (0.0241) | ||||||
| No-course Baseline x Social Media Share: 60-80% | -0.0289 | |||||
| (0.0279) | ||||||
| Emotions x Social Media Share: 60-80% | 0.0286 | |||||
| (0.0265) | ||||||
| Reasoning x Social Media Share: 60-80% | -0.0028 | |||||
| (0.0274) | ||||||
| Combo x Social Media Share: 60-80% | 0.0016 | |||||
| (0.0270) | ||||||
| No-course Baseline x Social Media Share: 40-60% | -0.0416* | |||||
| (0.0278) | ||||||
| Emotions x Social Media Share: 40-60% | -0.0123 | |||||
| (0.0269) | ||||||
| Reasoning x Social Media Share: 40-60% | -0.0336 | |||||
| (0.0280) | ||||||
| Combo x Social Media Share: 40-60% | -0.0246 | |||||
| (0.0274) | ||||||
| No-course Baseline x Social Media Share: 20-40% | -0.0226 | |||||
| (0.0255) | ||||||
| Emotions x Social Media Share: 20-40% | 0.0206 | |||||
| (0.0248) | ||||||
| Reasoning x Social Media Share: 20-40% | -0.0257 | |||||
| (0.0256) | ||||||
| Combo x Social Media Share: 20-40% | -0.0274 | |||||
| (0.0254) | ||||||
| No-course Baseline x Passed attention check at pre-treatment | -0.0410** | |||||
| (0.0198) | ||||||
| Emotions x Passed attention check at pre-treatment | -0.0034 | |||||
| (0.0189) | ||||||
| Reasoning x Passed attention check at pre-treatment | 0.0056 | |||||
| (0.0193) | ||||||
| Combo x Passed attention check at pre-treatment | 0.0421** | |||||
| (0.0193) | ||||||
| No-course Baseline x Avg. Sharing of Misinfo posts at pre-treatment | -0.3381 | |||||
| (109778.9211) | ||||||
| Emotions x Avg. Sharing of Misinfo posts at pre-treatment | -1.4409 | |||||
| (122450.7479) | ||||||
| Reasoning x Avg. Sharing of Misinfo posts at pre-treatment | 0.0129 | |||||
| (0.0548) | ||||||
| Combo x Avg. Sharing of Misinfo posts at pre-treatment | -0.8856 | |||||
| (115784.0494) | ||||||
| No-course Baseline x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0201 | |||||
| (0.0241) | ||||||
| Emotions x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0537** | |||||
| (0.0231) | ||||||
| Reasoning x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0364* | |||||
| (0.0241) | ||||||
| Combo x Avg. Sharing of Non-misinfo posts at pre-treatment | -0.0175 | |||||
| (0.0238) | ||||||
| No-course Baseline x Emotions posts | 0.1433 | |||||
| (44861.1249) | ||||||
| Emotions x Emotions posts | 0.5209 | |||||
| (50039.4633) | ||||||
| Reasoning x Emotions posts | -0.0201 | |||||
| (0.0351) | ||||||
| Combo x Emotions posts | 0.3052 | |||||
| (47315.1189) | ||||||
| No-course Baseline x Reasoning posts | 0.1558 | |||||
| (44435.4722) | ||||||
| Emotions x Reasoning posts | 0.5377 | |||||
| (49564.6774) | ||||||
| Reasoning x Reasoning posts | -0.0373 | |||||
| (0.0342) | ||||||
| Combo x Reasoning posts | 0.3688 | |||||
| (46866.1821) | ||||||
| No-course Baseline x Combo posts | 0.1426 | |||||
| (45344.4831) | ||||||
| Emotions x Combo posts | 0.5060 | |||||
| (50578.6158) | ||||||
| Combo x Combo posts | 0.3537 | |||||
| (47824.9178) | ||||||
| No-course Baseline x 1[Post-treatment] x 1[Misinfo] | -0.0011 | 0.0069 | -0.0022 | 0.0073 | 0.0198 | 0.0198 |
| (0.0843) | (0.0921) | (0.0958) | (0.0921) | (0.0971) | (0.0971) | |
| Emotions x 1[Post-treatment] x 1[Misinfo] | -0.1040* | -0.1220* | -0.1864** | -0.1231* | -0.2332*** | -0.2332*** |
| (0.0795) | (0.0857) | (0.0893) | (0.0857) | (0.0907) | (0.0907) | |
| Reasoning x 1[Post-treatment] x 1[Misinfo] | 0.1100* | 0.1178* | 0.0982 | 0.1172* | 0.0620 | 0.0620 |
| (0.0820) | (0.0892) | (0.0922) | (0.0893) | (0.0935) | (0.0935) | |
| Combo x 1[Post-treatment] x 1[Misinfo] | -0.0590 | -0.0845 | -0.0937 | -0.0854 | -0.1563** | -0.1563** |
| (0.0813) | (0.0887) | (0.0924) | (0.0887) | (0.0940) | (0.0940) | |
| Post domain 2 | -1.3458*** | -1.3896*** | ||||
| (0.0391) | (0.0766) | |||||
| Post domain 3 | -1.8926*** | -2.2086*** | ||||
| (0.0389) | (0.0755) | |||||
| Post domain 4 | -1.2128*** | -1.7401*** | ||||
| (0.0384) | (0.0763) | |||||
| Post domain 5 | 0.4049*** | 0.0051 | ||||
| (0.0388) | (0.0820) | |||||
| Post domain 6 | 0.1237*** | 0.4527*** | ||||
| (0.0395) | (0.0844) | |||||
| Post domain 7 | -1.5746*** | -1.4228*** | ||||
| (0.0388) | (0.0752) | |||||
| Post domain 8 | -1.4724*** | -0.9538*** | ||||
| (0.0391) | (0.0779) | |||||
| Post domain 9 | 0.2728*** | -0.5281*** | ||||
| (0.0392) | (0.0788) | |||||
| Post domain 10 | -0.3969*** | -0.8059*** | ||||
| (0.0377) | (0.0763) | |||||
| Post domain 11 | -1.0747*** | -1.3501*** | ||||
| (0.0388) | (0.0756) | |||||
| Post domain 12 | -0.7213*** | -1.0828*** | ||||
| (0.0372) | (0.0758) | |||||
| Post domain 13 | -1.0425*** | -1.6220*** | ||||
| (0.0378) | (0.0743) | |||||
| Post domain 14 | -1.8560*** | -2.2721*** | ||||
| (0.0403) | (0.0771) | |||||
| Post domain 15 | -1.2523*** | -2.0199*** | ||||
| (0.0381) | (0.0749) | |||||
| 1[Combo manipulation tactic] | -0.2264*** | -0.3419*** | ||||
| (0.0187) | (0.0832) | |||||
| 1[Emotions manipulation tactic] | -0.0068 | -0.5416*** | ||||
| (0.0184) | (0.0822) | |||||
| Post domain 2 x 1[Combo manipulation tactic] | -0.1530* | |||||
| (0.1163) | ||||||
| Post domain 3 x 1[Combo manipulation tactic] | 0.4818*** | |||||
| (0.1123) | ||||||
| Post domain 4 x 1[Combo manipulation tactic] | 0.7230*** | |||||
| (0.1104) | ||||||
| Post domain 5 x 1[Combo manipulation tactic] | 0.6450*** | |||||
| (0.1186) | ||||||
| Post domain 6 x 1[Combo manipulation tactic] | -2.2961*** | |||||
| (0.1206) | ||||||
| Post domain 7 x 1[Combo manipulation tactic] | -0.1072 | |||||
| (0.1122) | ||||||
| Post domain 8 x 1[Combo manipulation tactic] | 0.0724 | |||||
| (0.1141) | ||||||
| Post domain 9 x 1[Combo manipulation tactic] | 1.1394*** | |||||
| (0.1160) | ||||||
| Post domain 10 x 1[Combo manipulation tactic] | 0.3736*** | |||||
| (0.1122) | ||||||
| Post domain 11 x 1[Combo manipulation tactic] | 0.6002*** | |||||
| (0.1113) | ||||||
| Post domain 12 x 1[Combo manipulation tactic] | 0.0418 | |||||
| (0.1136) | ||||||
| Post domain 13 x 1[Combo manipulation tactic] | 0.5858*** | |||||
| (0.1106) | ||||||
| Post domain 14 x 1[Combo manipulation tactic] | 0.3450*** | |||||
| (0.1139) | ||||||
| Post domain 15 x 1[Combo manipulation tactic] | 0.7389*** | |||||
| (0.1105) | ||||||
| Post domain 2 x 1[Emotions manipulation tactic] | -0.4478*** | |||||
| (0.1167) | ||||||
| Post domain 3 x 1[Emotions manipulation tactic] | 0.9334*** | |||||
| (0.1107) | ||||||
| Post domain 4 x 1[Emotions manipulation tactic] | 0.5846*** | |||||
| (0.1097) | ||||||
| Post domain 5 x 1[Emotions manipulation tactic] | 0.8014*** | |||||
| (0.1170) | ||||||
| Post domain 6 x 1[Emotions manipulation tactic] | 0.9194*** | |||||
| (0.1236) | ||||||
| Post domain 7 x 1[Emotions manipulation tactic] | -0.0346 | |||||
| (0.1108) | ||||||
| Post domain 8 x 1[Emotions manipulation tactic] | -0.9456*** | |||||
| (0.1172) | ||||||
| Post domain 9 x 1[Emotions manipulation tactic] | 1.2361*** | |||||
| (0.1149) | ||||||
| Post domain 10 x 1[Emotions manipulation tactic] | 0.8795*** | |||||
| (0.1109) | ||||||
| Post domain 11 x 1[Emotions manipulation tactic] | -0.0799 | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Emotions manipulation tactic] | 1.1423*** | |||||
| (0.1105) | ||||||
| Post domain 13 x 1[Emotions manipulation tactic] | 1.4266*** | |||||
| (0.1071) | ||||||
| Post domain 14 x 1[Emotions manipulation tactic] | 0.8221*** | |||||
| (0.1124) | ||||||
| Post domain 15 x 1[Emotions manipulation tactic] | 2.4403*** | |||||
| (0.1113) | ||||||
| Post domain 2 x 1[Reasoning manipulation tactic] | 0.4651*** | |||||
| (0.1126) | ||||||
| Post domain 3 x 1[Reasoning manipulation tactic] | -0.2843*** | |||||
| (0.1157) | ||||||
| Post domain 4 x 1[Reasoning manipulation tactic] | 0.8423*** | |||||
| (0.1111) | ||||||
| Post domain 5 x 1[Reasoning manipulation tactic] | 0.2026** | |||||
| (0.1188) | ||||||
| Post domain 6 x 1[Reasoning manipulation tactic] | 0.3527*** | |||||
| (0.1241) | ||||||
| Post domain 7 x 1[Reasoning manipulation tactic] | -0.8764*** | |||||
| (0.1125) | ||||||
| Post domain 8 x 1[Reasoning manipulation tactic] | -1.9427*** | |||||
| (0.1203) | ||||||
| Post domain 9 x 1[Reasoning manipulation tactic] | 1.0028*** | |||||
| (0.1186) | ||||||
| Post domain 10 x 1[Reasoning manipulation tactic] | 0.4064*** | |||||
| (0.1125) | ||||||
| Post domain 11 x 1[Reasoning manipulation tactic] | 0.5144*** | |||||
| (0.1124) | ||||||
| Post domain 12 x 1[Reasoning manipulation tactic] | 0.2580*** | |||||
| (0.1107) | ||||||
| Post domain 13 x 1[Reasoning manipulation tactic] | 0.3539*** | |||||
| (0.1101) | ||||||
| Post domain 14 x 1[Reasoning manipulation tactic] | 0.4395*** | |||||
| (0.1117) | ||||||
| Post domain 15 x 1[Reasoning manipulation tactic] | 0.0085 | |||||
| (0.1116) | ||||||
| Post 2 indicator | -1.3896*** | |||||
| (0.0766) | ||||||
| Post 3 indicator | -2.2086*** | |||||
| (0.0755) | ||||||
| Post 4 indicator | -1.7401*** | |||||
| (0.0763) | ||||||
| Post 5 indicator | 0.0051 | |||||
| (0.0820) | ||||||
| Post 6 indicator | 0.4527*** | |||||
| (0.0844) | ||||||
| Post 7 indicator | -1.4228*** | |||||
| (0.0752) | ||||||
| Post 8 indicator | -0.9538*** | |||||
| (0.0779) | ||||||
| Post 9 indicator | -0.5281*** | |||||
| (0.0788) | ||||||
| Post 10 indicator | -0.8059*** | |||||
| (0.0763) | ||||||
| Post 11 indicator | -1.3501*** | |||||
| (0.0756) | ||||||
| Post 12 indicator | -1.0828*** | |||||
| (0.0758) | ||||||
| Post 13 indicator | -1.6220*** | |||||
| (0.0743) | ||||||
| Post 14 indicator | -2.2721*** | |||||
| (0.0771) | ||||||
| Post 15 indicator | -2.0199*** | |||||
| (0.0749) | ||||||
| Post 16 indicator | 1.6695*** | |||||
| (0.0824) | ||||||
| Post 17 indicator | 0.1269* | |||||
| (0.0846) | ||||||
| Post 18 indicator | -0.0573 | |||||
| (0.0815) | ||||||
| Post 19 indicator | 0.6525*** | |||||
| (0.0790) | ||||||
| Post 20 indicator | 2.3196*** | |||||
| (0.0838) | ||||||
| Post 21 indicator | -0.1740** | |||||
| (0.0827) | ||||||
| Post 22 indicator | 0.1395** | |||||
| (0.0819) | ||||||
| Post 23 indicator | 0.7881*** | |||||
| (0.0812) | ||||||
| Post 24 indicator | 2.2808*** | |||||
| (0.0839) | ||||||
| Post 25 indicator | 1.2372*** | |||||
| (0.0799) | ||||||
| Post 26 indicator | 0.9195*** | |||||
| (0.0829) | ||||||
| Post 27 indicator | 0.6285*** | |||||
| (0.0802) | ||||||
| Post 28 indicator | 0.6333*** | |||||
| (0.0813) | ||||||
| Post 29 indicator | -0.2576*** | |||||
| (0.0832) | ||||||
| Post 30 indicator | 0.3885*** | |||||
| (0.0809) | ||||||
| Post 31 indicator | 1.4698*** | |||||
| (0.0799) | ||||||
| Post 32 indicator | -0.3676*** | |||||
| (0.0867) | ||||||
| Post 33 indicator | 0.1946*** | |||||
| (0.0794) | ||||||
| Post 34 indicator | 0.3143*** | |||||
| (0.0795) | ||||||
| Post 35 indicator | 2.2763*** | |||||
| (0.0839) | ||||||
| Post 36 indicator | 2.8419*** | |||||
| (0.0910) | ||||||
| Post 37 indicator | 0.0124 | |||||
| (0.0810) | ||||||
| Post 38 indicator | -0.4296*** | |||||
| (0.0859) | ||||||
| Post 39 indicator | 2.1778*** | |||||
| (0.0835) | ||||||
| Post 40 indicator | 1.5434*** | |||||
| (0.0810) | ||||||
| Post 41 indicator | 0.0398 | |||||
| (0.0829) | ||||||
| Post 42 indicator | 1.5293*** | |||||
| (0.0807) | ||||||
| Post 43 indicator | 1.2744*** | |||||
| (0.0780) | ||||||
| Post 44 indicator | 0.0198 | |||||
| (0.0829) | ||||||
| Post 45 indicator | 1.8902*** | |||||
| (0.0825) | ||||||
| Post 46 indicator | 2.0114*** | |||||
| (0.0830) | ||||||
| Post 47 indicator | 1.0869*** | |||||
| (0.0811) | ||||||
| Post 48 indicator | -0.4814*** | |||||
| (0.0840) | ||||||
| Post 49 indicator | 1.1137*** | |||||
| (0.0820) | ||||||
| Post 50 indicator | 2.2191*** | |||||
| (0.0846) | ||||||
| Post 51 indicator | 2.8168*** | |||||
| (0.0890) | ||||||
| Post 52 indicator | -0.2878*** | |||||
| (0.0832) | ||||||
| Post 53 indicator | -0.8851*** | |||||
| (0.0926) | ||||||
| Post 54 indicator | 2.4861*** | |||||
| (0.0862) | ||||||
| Post 55 indicator | 1.6118*** | |||||
| (0.0820) | ||||||
| Post 56 indicator | 1.1756*** | |||||
| (0.0838) | ||||||
| Post 57 indicator | 1.1865*** | |||||
| (0.0791) | ||||||
| Post 58 indicator | 0.7433*** | |||||
| (0.0798) | ||||||
| Post 59 indicator | 0.1789** | |||||
| (0.0816) | ||||||
| Num. obs. | 156312 | 137250 | 137250 | 137250 | 137250 | 137250 |
| Deviance | 166005.0252 | 148402.7617 | 136205.4546 | 148215.0340 | 130957.8466 | 130957.8466 |
| Log Likelihood | -83002.5126 | -74201.3809 | -68102.7273 | -74107.5170 | -65478.9233 | -65478.9233 |
| Pseudo R2 | 0.2299 | 0.1384 | 0.2024 | 0.1393 | 0.2296 | 0.2296 |
| Out-of-sample R2 | ||||||
| Fixed Effects | user | user | user | user | user | |
| Fixed Effects | Domain | Tactic | Domain x Tactic | Post ID | ||
| Covariates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | Demographics + Sharing rates + Tactic sharing rates | |
| Clustered Std. Errors | user | user | user | user | user | user |
| Significance levels: *p<0.10, **p<0.05, ***p<0.01 one-sided test. | ||||||
Prediction
Standard errors of the share predicted value are estimated using the Delta method. Treatment effects standard errors are clustered at the user level and reported in parentheses. z-ratios are reported in brackets. Treatment effects estimates for specifications that include Post-Tactic ID fixed effects are not reported since all tactics are misinformation.
cov_feglm <- c(paste0("share ~ treatment"),
paste0("share ~ treatment + ",demographics_demeaned_treatment),
paste0("share ~ treatment + factor(fact) + ",sharing_demeaned_treatment),
paste0("share ~ treatment + factor(type) + ",sharing_tactic_demeaned_treatment),
paste0("share ~ treatment + factor(fact)*factor(type) + ",demographics_sharing_demeaned_treatment),
paste0("share ~ treatment + factor(post_id) + ",all_covariates_demeaned_treatment))
if (file.exists("prediction_alltreat_logit_df.rds")) {
prediction_alltreat_logit_df <- readRDS("prediction_alltreat_logit_df.rds")
counter <- 1
for (cov in seq_along(covariate_list)) {
list_models <- cov_estimates_list[[cov]]
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
# obtain prediction_df
prediction_df <- prediction_alltreat_logit_df[[counter]]
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))
print(k)
cat('<br>')
counter <- counter + 1
}
cat('<br>\n<br>\n<br>')
}
} else {
prediction_alltreat_logit_df <- list()
for (cov in seq_along(covariate_list)) {
list_models <- cov_estimates_list[[cov]]
cat(paste0("$$\\text{",covariate_list_names[cov],"}$$"))
for (m in 1:length(models)) {
treatmentLevels <- unique(panel$treatment)
posts <- c(0, 1)
misinfos <- c(0, 1)
prediction_df <- expand.grid(treatment = treatmentLevels, post = posts, post_is_misinfo = misinfos) %>%
mutate(prediction = NA_real_, prediction_se = NA_real_)
for (row in 1:nrow(prediction_df)) {
treat <- as.character(prediction_df$treatment[row])
post <- prediction_df$post[row]
misinfo <- prediction_df$post_is_misinfo[row]
# Filter the relevant data
relevant_data <- panel %>%
filter(row_number() %in% obs(list_models[[m]])) %>%
filter(treatment == treat & post == post & post_is_misinfo == misinfo)
# Prepare predictor variables based on model requirements
predictorVariables <- relevant_data %>%
distinct(user, .keep_all = TRUE) %>%
group_by(user, type, fact, post_id) %>%
summarise(age_cov = mean(age_cov, na.rm = TRUE),
gender_Man_cov = mean(gender_Man_cov, na.rm = TRUE),
education_Some_college_cov = mean(education_Some_college_cov, na.rm = TRUE),
education_Bachelor_degree_cov = mean(education_Bachelor_degree_cov, na.rm = TRUE),
education_Graduate_degree_cov = mean(education_Graduate_degree_cov, na.rm = TRUE),
marital_Married_or_in_a_domestic_partnership_cov = mean(marital_Married_or_in_a_domestic_partnership_cov, na.rm = TRUE),
employment_Unemployed_cov = mean(employment_Unemployed_cov, na.rm = TRUE),
employment_Employed_cov = mean(employment_Employed_cov, na.rm = TRUE),
location_Mostly_urban_cov = mean(location_Mostly_urban_cov, na.rm = TRUE),
location_Suburban_cov = mean(location_Suburban_cov, na.rm = TRUE),
religion_Christian_cov = mean(religion_Christian_cov, na.rm = TRUE),
religiosity_Attends_cov = mean(religiosity_Attends_cov, na.rm = TRUE),
social_media_bin_Yes_cov = mean(social_media_bin_Yes_cov, na.rm = TRUE),
social_media_hours_cov = mean(social_media_hours_cov, na.rm = TRUE),
att_check_pre_cov = mean(att_check_pre_cov, na.rm = TRUE),
social_media_share_80_100_cov = mean(social_media_share_80_100_cov, na.rm = TRUE),
social_media_share_60_80_cov = mean(social_media_share_60_80_cov, na.rm = TRUE),
social_media_share_40_60_cov = mean(social_media_share_40_60_cov, na.rm = TRUE),
social_media_share_20_40_cov = mean(social_media_share_20_40_cov, na.rm = TRUE),
misinfo_pre_cov = mean(misinfo_pre_cov, na.rm = TRUE),
base_rate_pre_cov = mean(base_rate_pre_cov, na.rm = TRUE),
emotions_pre_cov = mean(emotions_pre_cov, na.rm = TRUE),
reasoning_pre_cov = mean(reasoning_pre_cov, na.rm = TRUE),
combo_pre_cov = mean(combo_pre_cov, na.rm = TRUE)) %>% # now add the treatment and post variables and misinfo
mutate(treatment = treat,
post = post,
post_is_misinfo = misinfo)
# predicted value
if (m != 1) {
predicted_shares <- predict(list_models[[m]], newdata = predictorVariables)
predictorVariables$prediction <- predicted_shares
# mean of predicted value
prediction_df$prediction[row] <- format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,mean)[,4], digits = 5)
# se of predicted value
prediction_df$prediction_se[row] <- paste0("(",format(aggregate(prediction~treatment*post*post_is_misinfo,predictorVariables,se_cont)[,4], digits = 4),")")
}
}
# predicted value
if (m == 1) {
predictorVariables <- expand.grid(treatment = unique(panel$treatment),
post = c(0,1),
post_is_misinfo = c(0,1))
predictions <- predict(baseline_all_logit, newdata = predictorVariables, type = "response", se.fit = TRUE)
prediction_df <- predictorVariables
prediction_df$prediction <- format(predictions$fit, digits = 4)
prediction_df$prediction_se <- paste0("(",format(predictions$se.fit, digits = 4),")")
emm_res <- emmeans(baseline_all_logit, specs = pairwise ~ treatment | post * post_is_misinfo)
contrast_res <- contrast(emm_res, method = "revpairwise")
contrast_df <- as.data.frame(contrast_res) %>% filter(contrast == "Emotions - Facts Baseline") %>%
select(post, post_is_misinfo, estimate, SE, z.ratio)
contrast_df <- rbind(contrast_df %>% mutate(treatment = "Facts Baseline"),contrast_df %>% mutate(treatment = "Emotions")) %>% mutate(se_tratio = gsub("\\[\\s ","\\[",paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))) %>%
mutate(estimate = format(estimate, digits = 4)) %>%
mutate(estimate = ifelse(treatment == "Emotions", se_tratio, estimate)) %>%
select(treatment, post, post_is_misinfo, estimate) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
# calculate difference between Facts Baseline and all other treatments
prediction_df <- prediction_df %>%
group_by(post,post_is_misinfo) %>%
mutate(lag = ifelse(treatment == "Emotions",
prediction, NA)) %>%
tidyr::fill(lag, .direction = "up") %>%
ungroup() %>%
mutate(te = ifelse(treatment == "Facts Baseline",
format(as.numeric(prediction) - as.numeric(lag), digits = 5), NA)) %>%
select(-lag) %>%
mutate(te = ifelse(is.na(te),"",te))
if (m %in% c(2,3,6)) {
# predicted treatment effects
contrast_df <- data.frame()
for (p in c(0,1)) {
for (mis in c(0,1)) {
cov_model <- feglm(as.formula(cov_feglm[m]),panel %>% filter(post == p & post_is_misinfo == mis), family = binomial(link = "logit"), cluster = "user")
# obtain data frame with estimate, std error, and z value
cov_model_df_add <- data.frame(
treatment = c("No-course Baseline","Emotions","Reasoning","Combo"),
estimate = format(coef(cov_model)[grepl("treatment",names(coef(cov_model)))], digits = 4),
SE = sqrt(diag(vcov(cov_model)))[grepl("treatment",names(diag(vcov(cov_model))))],
z.ratio = coef(cov_model)[grepl("treatment",names(coef(cov_model)))] / sqrt(diag(vcov(cov_model)))[grepl("treatment",names(diag(vcov(cov_model))))],
post = p,
post_is_misinfo = mis
)
# make rownames a column
cov_model_df_add <- rownames_to_column(cov_model_df_add, var = "treatment2") %>%
filter(treatment == "Emotions" & treatment2 == "treatmentEmotions") %>% select(-treatment2) %>%
mutate(zratio = paste0("(",format(SE, digits = 4),"); [",format(z.ratio, digits = 4),"]"))
cov_model_df_add <- rbind(cov_model_df_add %>% mutate(treatment = "Facts Baseline"),cov_model_df_add %>% mutate(treatment = "Emotions")) %>%
mutate(estimate = ifelse(treatment == "Emotions", zratio, estimate)) %>%
mutate(post = ifelse(post == 0,as.numeric(0),as.numeric(1)),
post_is_misinfo = ifelse(post_is_misinfo == 0,as.numeric(0),as.numeric(1))) %>%
select(treatment, post, post_is_misinfo, estimate) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
contrast_df <- rbind(contrast_df, cov_model_df_add)
}
}
prediction_df <- left_join(prediction_df, contrast_df,
by = c("treatment" = "treatment", "post" = "post", "post_is_misinfo" = "post_is_misinfo"))
}
if (m %in% c(4,5)) {
prediction_df$estimate <- ""
}
prediction_df <- prediction_df %>% select(treatment,post,post_is_misinfo,prediction,prediction_se,te,estimate)
prediction_df <- prediction_df %>% arrange(post,post_is_misinfo) %>%
mutate(estimate = ifelse(is.na(estimate),"",estimate))
k <- knitr::kable(prediction_df, caption = models[m],
col.names = c("Treatment","Post-treatment","1[Misinfo]","Predicted value<br>1[Share]","Prediction<br>Std. errors","Diff. in Predicted values<br>Emotions vs. Facts","log odds ratio<br>Emotions/Facts"))
print(k)
cat('<br>')
prediction_alltreat_logit_df[[length(prediction_alltreat_logit_df) + 1]] <- prediction_df
}
cat('<br>\n<br>')
}
saveRDS(prediction_alltreat_logit_df, "prediction_alltreat_logit_df.rds")
}\[\text{Demographics}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.027609 | |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | 0.0059 | (0.04131); [0.6683] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.008161 | |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | 0.0021 | (0.02721); [0.3000] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.358022 | |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | -0.0879 | (0.03883); [ -9.2196] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.465472 | |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | -0.1120 | (0.02778); [-16.7557] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.002590 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.05418); [0.0478] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.013730 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.04402); [-0.3119] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.401041 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.0522); [-7.683] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.5110728 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.0467); [-10.94] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0055615 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.03105); [0.1791] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0222133 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.01701); [1.306] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.47662 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.05522); [-8.631] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.557139 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.04873); [-11.43] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0058603 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.03174); [0.1846] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0141366 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.01931); [0.732] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.502190 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.05525); [-9.089] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.6220643 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.04981); [-12.49] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.027609 | |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | 0.0059 | (0.04131); [0.6683] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.008161 | |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | 0.0021 | (0.02721); [0.3000] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.358022 | |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | -0.0879 | (0.03883); [ -9.2196] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.465472 | |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | -0.1120 | (0.02778); [-16.7557] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.002590 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.05418); [0.0478] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.013730 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.04402); [-0.3119] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.401041 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.0522); [-7.683] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.5110728 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.0467); [-10.94] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0055615 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.03105); [0.1791] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0222133 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.01701); [1.306] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.47662 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.05522); [-8.631] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.557139 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.04873); [-11.43] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0058603 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.03174); [0.1846] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0141366 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.01931); [0.732] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.502190 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.05525); [-9.089] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.6220643 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.04981); [-12.49] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Tactic sharing
rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.027609 | |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | 0.0059 | (0.04131); [0.6683] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.008161 | |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | 0.0021 | (0.02721); [0.3000] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.358022 | |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | -0.0879 | (0.03883); [ -9.2196] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.465472 | |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | -0.1120 | (0.02778); [-16.7557] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.002590 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.05418); [0.0478] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.013730 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.04402); [-0.3119] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.401041 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.0522); [-7.683] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.5110728 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.0467); [-10.94] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0055615 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.03105); [0.1791] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0222133 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.01701); [1.306] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.47662 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.05522); [-8.631] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.557139 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.04873); [-11.43] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0058603 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.03174); [0.1846] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0141366 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.01931); [0.732] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.502190 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.05525); [-9.089] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.6220643 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.04981); [-12.49] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Demographics +
Sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.027609 | |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | 0.0059 | (0.04131); [0.6683] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.008161 | |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | 0.0021 | (0.02721); [0.3000] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.358022 | |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | -0.0879 | (0.03883); [ -9.2196] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.465472 | |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | -0.1120 | (0.02778); [-16.7557] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.002590 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.05418); [0.0478] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.013730 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.04402); [-0.3119] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.401041 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.0522); [-7.683] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.5110728 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.0467); [-10.94] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0055615 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.03105); [0.1791] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0222133 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.01701); [1.306] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.47662 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.05522); [-8.631] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.557139 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.04873); [-11.43] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0058603 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.03174); [0.1846] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0141366 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.01931); [0.732] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.502190 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.05525); [-9.089] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.6220643 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.04981); [-12.49] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |
\[\text{Demographics +
Sharing rates + Tactic sharing rates}\]
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.6829 | (0.006350) | 0.027609 | |
| Emotions | 0 | 0 | 0.6888 | (0.006238) | 0.0059 | (0.04131); [0.6683] |
| No-course Baseline | 0 | 0 | 0.7041 | (0.006531) | ||
| Reasoning | 0 | 0 | 0.7037 | (0.006413) | ||
| Combo | 0 | 0 | 0.6864 | (0.006422) | ||
| Facts Baseline | 0 | 1 | 0.5386 | (0.004810) | 0.008161 | |
| Emotions | 0 | 1 | 0.5407 | (0.004748) | 0.0021 | (0.02721); [0.3000] |
| No-course Baseline | 0 | 1 | 0.5650 | (0.005016) | ||
| Reasoning | 0 | 1 | 0.5399 | (0.004949) | ||
| Combo | 0 | 1 | 0.5462 | (0.004873) | ||
| Facts Baseline | 1 | 0 | 0.6071 | (0.006665) | -0.358022 | |
| Emotions | 1 | 0 | 0.5192 | (0.006732) | -0.0879 | (0.03883); [ -9.2196] |
| No-course Baseline | 1 | 0 | 0.6798 | (0.006676) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.006986) | ||
| Combo | 1 | 0 | 0.5222 | (0.006914) | ||
| Facts Baseline | 1 | 1 | 0.4650 | (0.004813) | -0.465472 | |
| Emotions | 1 | 1 | 0.3530 | (0.004553) | -0.1120 | (0.02778); [-16.7557] |
| No-course Baseline | 1 | 1 | 0.5461 | (0.005038) | ||
| Reasoning | 1 | 1 | 0.4040 | (0.004873) | ||
| Combo | 1 | 1 | 0.3715 | (0.004729) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | 0.002590 | |
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | (0.05418); [0.0478] |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53864 | (0.006739) | -0.013730 | |
| Emotions | 0 | 1 | 0.54067 | (0.006114) | 0.00203 | (0.04402); [-0.3119] |
| No-course Baseline | 0 | 1 | 0.56501 | (0.006896) | ||
| Reasoning | 0 | 1 | 0.53994 | (0.00667) | ||
| Combo | 0 | 1 | 0.54617 | (0.006623) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | -0.401041 | |
| Emotions | 1 | 0 | 0.51924 | (0.006114) | -0.08784 | (0.0522); [-7.683] |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46499 | (0.006739) | -0.5110728 | |
| Emotions | 1 | 1 | 0.35303 | (0.006114) | -0.11196 | (0.0467); [-10.94] |
| No-course Baseline | 1 | 1 | 0.54607 | (0.006896) | ||
| Reasoning | 1 | 1 | 0.40404 | (0.00667) | ||
| Combo | 1 | 1 | 0.37146 | (0.006623) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68708 | (0.007296) | 0.0055615 | |
| Emotions | 0 | 0 | 0.68691 | (0.006623) | -0.00017 | (0.03105); [0.1791] |
| No-course Baseline | 0 | 0 | 0.70681 | (0.007465) | ||
| Reasoning | 0 | 0 | 0.70467 | (0.007186) | ||
| Combo | 0 | 0 | 0.68609 | (0.007109) | ||
| Facts Baseline | 0 | 1 | 0.53759 | (0.007252) | 0.0222133 | |
| Emotions | 0 | 1 | 0.54124 | (0.00663) | 0.00365 | (0.01701); [1.306] |
| No-course Baseline | 0 | 1 | 0.56465 | (0.007482) | ||
| Reasoning | 0 | 1 | 0.53678 | (0.007188) | ||
| Combo | 0 | 1 | 0.5456 | (0.007157) | ||
| Facts Baseline | 1 | 0 | 0.61224 | (0.007296) | -0.47662 | |
| Emotions | 1 | 0 | 0.51718 | (0.006623) | -0.09506 | (0.05522); [-8.631] |
| No-course Baseline | 1 | 0 | 0.68323 | (0.007465) | ||
| Reasoning | 1 | 0 | 0.55133 | (0.007186) | ||
| Combo | 1 | 0 | 0.52308 | (0.007109) | ||
| Facts Baseline | 1 | 1 | 0.46504 | (0.007252) | -0.557139 | |
| Emotions | 1 | 1 | 0.35181 | (0.00663) | -0.11323 | (0.04873); [-11.43] |
| No-course Baseline | 1 | 1 | 0.54795 | (0.007482) | ||
| Reasoning | 1 | 1 | 0.39859 | (0.007188) | ||
| Combo | 1 | 1 | 0.37069 | (0.007157) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68287 | (0.006739) | ||
| Emotions | 0 | 0 | 0.68882 | (0.006114) | 0.00595 | |
| No-course Baseline | 0 | 0 | 0.70414 | (0.006896) | ||
| Reasoning | 0 | 0 | 0.70375 | (0.00667) | ||
| Combo | 0 | 0 | 0.6864 | (0.006623) | ||
| Facts Baseline | 0 | 1 | 0.53798 | (0.006743) | ||
| Emotions | 0 | 1 | 0.54051 | (0.006119) | 0.00253 | |
| No-course Baseline | 0 | 1 | 0.56442 | (0.006913) | ||
| Reasoning | 0 | 1 | 0.53984 | (0.006692) | ||
| Combo | 0 | 1 | 0.54628 | (0.00664) | ||
| Facts Baseline | 1 | 0 | 0.60708 | (0.006739) | ||
| Emotions | 1 | 0 | 0.51925 | (0.006114) | -0.08783 | |
| No-course Baseline | 1 | 0 | 0.67977 | (0.006896) | ||
| Reasoning | 1 | 0 | 0.5503 | (0.00667) | ||
| Combo | 1 | 0 | 0.52222 | (0.006623) | ||
| Facts Baseline | 1 | 1 | 0.46433 | (0.006743) | ||
| Emotions | 1 | 1 | 0.35288 | (0.006119) | -0.11145 | |
| No-course Baseline | 1 | 1 | 0.54548 | (0.006913) | ||
| Reasoning | 1 | 1 | 0.40394 | (0.006692) | ||
| Combo | 1 | 1 | 0.37156 | (0.00664) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68625 | (0.007209) | ||
| Emotions | 0 | 0 | 0.68537 | (0.006579) | -0.00088 | |
| No-course Baseline | 0 | 0 | 0.70608 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70338 | (0.007231) | ||
| Combo | 0 | 0 | 0.68525 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53805 | (0.007503) | ||
| Emotions | 0 | 1 | 0.54073 | (0.006932) | 0.00268 | |
| No-course Baseline | 0 | 1 | 0.56257 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53858 | (0.007523) | ||
| Combo | 0 | 1 | 0.54627 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61204 | (0.007209) | ||
| Emotions | 1 | 0 | 0.51555 | (0.006579) | -0.09649 | |
| No-course Baseline | 1 | 0 | 0.68253 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55085 | (0.007231) | ||
| Combo | 1 | 0 | 0.52275 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46774 | (0.007503) | ||
| Emotions | 1 | 1 | 0.35023 | (0.006932) | -0.11751 | |
| No-course Baseline | 1 | 1 | 0.54712 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.40062 | (0.007523) | ||
| Combo | 1 | 1 | 0.36824 | (0.007512) |
| Treatment | Post-treatment | 1[Misinfo] | Predicted value 1[Share] |
Prediction Std. errors |
Diff. in Predicted values Emotions vs. Facts |
log odds ratio Emotions/Facts |
|---|---|---|---|---|---|---|
| Facts Baseline | 0 | 0 | 0.68628 | (0.007209) | 0.0058603 | |
| Emotions | 0 | 0 | 0.6854 | (0.006579) | -0.00088 | (0.03174); [0.1846] |
| No-course Baseline | 0 | 0 | 0.70612 | (0.007419) | ||
| Reasoning | 0 | 0 | 0.70341 | (0.007231) | ||
| Combo | 0 | 0 | 0.68528 | (0.007031) | ||
| Facts Baseline | 0 | 1 | 0.53803 | (0.007503) | 0.0141366 | |
| Emotions | 0 | 1 | 0.54071 | (0.006932) | 0.00268 | (0.01931); [0.732] |
| No-course Baseline | 0 | 1 | 0.56255 | (0.007866) | ||
| Reasoning | 0 | 1 | 0.53857 | (0.007523) | ||
| Combo | 0 | 1 | 0.54625 | (0.007512) | ||
| Facts Baseline | 1 | 0 | 0.61207 | (0.007209) | -0.502190 | |
| Emotions | 1 | 0 | 0.51558 | (0.006579) | -0.09649 | (0.05525); [-9.089] |
| No-course Baseline | 1 | 0 | 0.68256 | (0.007419) | ||
| Reasoning | 1 | 0 | 0.55088 | (0.007231) | ||
| Combo | 1 | 0 | 0.52278 | (0.007031) | ||
| Facts Baseline | 1 | 1 | 0.46772 | (0.007503) | -0.6220643 | |
| Emotions | 1 | 1 | 0.35021 | (0.006932) | -0.11751 | (0.04981); [-12.49] |
| No-course Baseline | 1 | 1 | 0.54711 | (0.007866) | ||
| Reasoning | 1 | 1 | 0.4006 | (0.007523) | ||
| Combo | 1 | 1 | 0.36822 | (0.007512) |