Data Cleaning and Preparation- pre wide format 1. Recode Self harm variables to binary and then calculate a new combined binary variable of “any self harm yes/no” 2. Make Attendance variables and last session attended variables
### Code to create binary self-harm variables and presence variables without overwriting the originals ###
library(readxl)
library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionlibrary(tidyr)
library(readr)
NoDup_PurrbleAnon <- read_csv("Desktop/Purrbel/NoDup_PurrbleAnon.csv")## Warning: One or more parsing issues, call `problems()` on your data frame for details,
## e.g.:
## dat <- vroom(...)
## problems(dat)## Rows: 1941 Columns: 140## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (40): psid, randomization, so, gi, ethnicity, identity_group, startdate,...
## dbl (89): Week, age, source, status, progress, durationinseconds, finished, ...
## lgl (11): firstname, q45_3_text, q46_6_text, GAD7_Complete, PHQ9_Complete, S...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.#View(NoDup_PurrbleAnon)
NoDup_PurrbleAnon <- NoDup_PurrbleAnon %>%
mutate(
# If missing, then NA. If <= 1 then 0, else 1
SHQ1 = ifelse(is.na(shqscreener1), NA, ifelse(shqscreener1 <= 1, 0, 1)),
SHQ2 = ifelse(is.na(shqscreener2), NA, ifelse(shqscreener2 <= 1, 0, 1)),
SHQ3 = ifelse(is.na(shqscreener3), NA, ifelse(shqscreener3 <= 1, 0, 1))
) %>%
mutate(
# If any of SHQ1, SHQ2, or SHQ3 is missing, SHQ_Any is missing.
# If all three are 0, SHQ_Any is 0, else 1.
SHQ_Any = case_when(
is.na(SHQ1) | is.na(SHQ2) | is.na(SHQ3) ~ NA_real_,
SHQ1 == 0 & SHQ2 == 0 & SHQ3 == 0 ~ 0,
TRUE ~ 1
)
)Data Cleaning and preparation: 2. Transform from wide to long format 3. Calculate a new condition variable
Note: Do not get confused! In the datset for user ID, T means transgender and C means Cisgender- IT DOES NOT MEAN TREATMENT AND CONTROL. This is based on Seonaid’s original participant tracking database and aligns with all of that data!
Data Cleaning and Preparation: - Calculate pre-test (avg of weeks 1-3) and post-test (average of weeks 10-13) variables
library(readxl)
library(dplyr)
library(tidyr)
library(readr)
# Define the vector of outcomes for which we need pre/post lists
outcomes <- c("shqscreener1", "shqscreener2", "shqscreener3",
"DERS8_Sum", "GAD7_Sum",
"PHQ9_Sum", "SHS_Pathways", "SHS_Agency", "SHS_TotalHope",
"ucla_Sum", "pmerq_Focus_Avg", "pmerq_Distract_Avg", "pmerq_AD_Avg")
# Create the pretest list (weeks 1-3 for each outcome)
pretest_list <- unlist(lapply(outcomes, function(x) paste0(x, "_", 1:3)))
# Create the posttest list (weeks 11-13 for each outcome)
posttest_list <- unlist(lapply(outcomes, function(x) paste0(x, "_", 11:13)))
for (outcome in outcomes) {
# Identify the pre-test columns for this outcome
pre_cols <- pretest_list[grepl(paste0("^", outcome, "_"), pretest_list)]
# Identify the post-test columns for this outcome
post_cols <- posttest_list[grepl(paste0("^", outcome, "_"), posttest_list)]
# Create the names for the new Pre_ and Post_ columns
pre_colname <- paste0("Pre_", outcome)
post_colname <- paste0("Post_", outcome)
# Calculate row means for pre-test columns (weeks 1-3)
purrble_wide[[pre_colname]] <- rowMeans(purrble_wide[, pre_cols, drop = FALSE], na.rm = TRUE)
# Calculate row means for post-test columns (weeks 11-13)
purrble_wide[[post_colname]] <- rowMeans(purrble_wide[, post_cols, drop = FALSE], na.rm = TRUE)
}
#View(purrble_wide)Data Cleaning and Preparation: - Calculate Attendance Variables for Each Week (for later survival curve analyses!) - Calculate “Week of Dropout” Variable
library(dplyr)
# Create a vector of weeks
weeks <- 1:13
# For each week, check if any of that week's columns are non-NA.
# We use `rowSums(!is.na(across(...))) > 0` to check for the presence of any data.
for (w in weeks) {
purrble_wide <- purrble_wide %>%
mutate(!!paste0("ATT", w) := if_else(
rowSums(!is.na(across(ends_with(paste0("_", w))))) > 0,
1,
0
))
}
purrble_wide <- purrble_wide %>%
rowwise() %>%
mutate(Last_Session = {
# Extract the attendance values for all weeks
att_values <- c_across(starts_with("ATT"))
# Check if there is any attendance
if (all(att_values == 0 | is.na(att_values))) {
# If no attendance, set Last_Session to NA
NA_integer_
} else {
# If there is attendance, find the last (max) week attended
max(which(att_values == 1))
}
}) %>%
ungroup()
#View(purrble_wide)
#colnames(purrble_wide)Reporting baseline descriptive statistics of the sample (characteristics as reported at screening amongst those who were randomized) by condition: identity_group, Age
Problem note for Jess: I need Jess to code the ethnicity data and sexuality data into something usable- need this for reporting before any sort of publication and for all descriptive tables. People were able to write anything they wanted under those variables, so I do not have the ability to engage with the dataset for those accurately.
library(dplyr)
library(psych)
# 1. Continuous variable (e.g., age) by condition
age_desc <- describeBy(purrble_wide$age, purrble_wide$condition, mat = TRUE)
age_desc <- age_desc %>%
select(group1, n, mean, sd, min, max) %>%
rename(
Condition = group1,
N = n,
Mean = mean,
SD = sd,
Min = min,
Max = max
)
cat("**Age by Condition (Baseline)**\n")## **Age by Condition (Baseline)**print(age_desc)## Condition N Mean SD Min Max
## X11 Waitlist Control 78 20.06410 2.456583 16 25
## X12 Purrble Treatment 76 20.40789 2.281389 16 25cat("\n")# T-test for age by condition
t_result <- t.test(age ~ condition, data = purrble_wide)
cat("T-test for Age by Condition:\n")## T-test for Age by Condition:print(t_result)##
## Welch Two Sample t-test
##
## data: age by condition
## t = -0.9002, df = 151.65, p-value = 0.3694
## alternative hypothesis: true difference in means between group Waitlist Control and group Purrble Treatment is not equal to 0
## 95 percent confidence interval:
## -1.098336 0.410752
## sample estimates:
## mean in group Waitlist Control mean in group Purrble Treatment
## 20.06410 20.40789cat("\n")# Effect size (Cohen's d) using psych
cohen <- cohen.d(purrble_wide$age, purrble_wide$condition)
cat("Cohen's d for Age by Condition:\n")## Cohen's d for Age by Condition:print(cohen)## Call: cohen.d(x = purrble_wide$age, group = purrble_wide$condition)
## Cohen d statistic of difference between two means
## lower effect upper
## [1,] -0.17 0.15 0.46
##
## Multivariate (Mahalanobis) distance between groups
## [1] 0.15
## r equivalent of difference between two means
## data
## 0.07cat("\n")# Categorical Variable (identity_group) by condition
cat("Frequency of 'identity_group' by Condition (Baseline):\n")## Frequency of 'identity_group' by Condition (Baseline):print(table(purrble_wide$identity_group, purrble_wide$condition))##
## Waitlist Control Purrble Treatment
## c 0 1
## C 38 38
## TGD 40 37cat("\n")Attrition Analysis by Condition Attrition is defined as dropping out of study prior to Week 11. (In other words, someone could have participated in any data collection during weeks 11-13 and count as enrolled).
Result: No difference in attrition by group
# Create an attrition variable: 1 if last session <= 10, else 0
purrble_wide$attrition <- ifelse(purrble_wide$Last_Session <= 10, 1, 0)
# Identify condition levels
cond_levels <- unique(purrble_wide$condition)
cond_levels <- cond_levels[!is.na(cond_levels)]
if (length(cond_levels) != 2) {
stop("This code expects exactly two conditions for the comparison.")
}
# Extract data by condition
cond1_data <- purrble_wide[purrble_wide$condition == cond_levels[1], "attrition", drop = TRUE]
cond2_data <- purrble_wide[purrble_wide$condition == cond_levels[2], "attrition", drop = TRUE]
# Calculate descriptive statistics
cond1_n <- sum(!is.na(cond1_data))
cond1_mean <- mean(cond1_data, na.rm = TRUE)
cond1_sd <- sd(cond1_data, na.rm = TRUE)
cond2_n <- sum(!is.na(cond2_data))
cond2_mean <- mean(cond2_data, na.rm = TRUE)
cond2_sd <- sd(cond2_data, na.rm = TRUE)
# T-test comparing attrition by condition
tt <- t.test(attrition ~ condition, data = purrble_wide)
# Print results
cat("Attrition Analysis by Condition (prior to Week 11):\n\n")## Attrition Analysis by Condition (prior to Week 11):cat("Means and SD by Condition:\n")## Means and SD by Condition:cat(cond_levels[1], ": N =", cond1_n, "Mean =", round(cond1_mean, 3), "SD =", round(cond1_sd, 3), "\n")## 1 : N = 78 Mean = 0.064 SD = 0.247cat(cond_levels[2], ": N =", cond2_n, "Mean =", round(cond2_mean, 3), "SD =", round(cond2_sd, 3), "\n\n")## 2 : N = 76 Mean = 0.092 SD = 0.291cat("T-test results:\n")## T-test results:print(tt)##
## Welch Two Sample t-test
##
## data: attrition by condition
## t = -0.64343, df = 146.68, p-value = 0.521
## alternative hypothesis: true difference in means between group Waitlist Control and group Purrble Treatment is not equal to 0
## 95 percent confidence interval:
## -0.11401229 0.05800689
## sample estimates:
## mean in group Waitlist Control mean in group Purrble Treatment
## 0.06410256 0.09210526Baseline Descriptive Analyses T-test to check for baseline differences in outcomes of interest by study condition Result: No difference in baseline pre-test measures
library(dplyr)
library(psych)
# Create vectors of pre- and post-test variables
pre_vars <- c("Pre_shqscreener1", "Pre_shqscreener2", "Pre_shqscreener3",
"Pre_DERS8_Sum", "Pre_GAD7_Sum", "Pre_PHQ9_Sum",
"Pre_SHS_Pathways", "Pre_SHS_Agency", "Pre_SHS_TotalHope",
"Pre_ucla_Sum", "Pre_pmerq_Focus_Avg", "Pre_pmerq_Distract_Avg", "Pre_pmerq_AD_Avg")
post_vars <- c("Post_shqscreener1", "Post_shqscreener2", "Post_shqscreener3",
"Post_DERS8_Sum", "Post_GAD7_Sum", "Post_PHQ9_Sum",
"Post_SHS_Pathways", "Post_SHS_Agency", "Post_SHS_TotalHope",
"Post_ucla_Sum", "Post_pmerq_Focus_Avg", "Post_pmerq_Distract_Avg", "Post_pmerq_AD_Avg")
# Identify condition levels (assuming exactly two conditions)
cond_levels <- unique(purrble_wide$condition)
cond_levels <- cond_levels[!is.na(cond_levels)]
if (length(cond_levels) != 2) {
stop("This code expects exactly two conditions for the t-tests.")
}
# Split data by condition
data_cond1 <- purrble_wide[purrble_wide$condition == cond_levels[1], ]
data_cond2 <- purrble_wide[purrble_wide$condition == cond_levels[2], ]
# Initialize a results data frame
results <- data.frame(
Outcome = character(),
Condition1 = character(),
Condition1_N = numeric(),
Condition1_Mean = numeric(),
Condition1_SD = numeric(),
Condition2 = character(),
Condition2_N = numeric(),
Condition2_Mean = numeric(),
Condition2_SD = numeric(),
df = numeric(),
t = numeric(),
p = numeric(),
CI_Lower = numeric(),
CI_Upper = numeric(),
stringsAsFactors = FALSE
)
# Loop through each pre-test variable and run descriptives and t-test
for (var in pre_vars) {
# Describe by condition
desc1 <- describe(data_cond1[[var]])
desc2 <- describe(data_cond2[[var]])
# Check if we have valid data for both conditions
if (desc1$n > 0 & desc2$n > 0) {
# t-test
tt <- t.test(purrble_wide[[var]] ~ purrble_wide$condition)
# Extract info
n1 <- desc1$n
mean1 <- desc1$mean
sd1 <- desc1$sd
n2 <- desc2$n
mean2 <- desc2$mean
sd2 <- desc2$sd
df_val <- tt$parameter
t_val <- tt$statistic
p_val <- tt$p.value
ci_lower <- tt$conf.int[1]
ci_upper <- tt$conf.int[2]
# Add row to results
results <- rbind(results, data.frame(
Outcome = var,
Condition1 = cond_levels[1],
Condition1_N = n1,
Condition1_Mean = mean1,
Condition1_SD = sd1,
Condition2 = cond_levels[2],
Condition2_N = n2,
Condition2_Mean = mean2,
Condition2_SD = sd2,
df = df_val,
t = t_val,
p = p_val,
CI_Lower = ci_lower,
CI_Upper = ci_upper,
stringsAsFactors = FALSE
))
}
}
cat("Baseline differences in study outcomes by condition\n\n")## Baseline differences in study outcomes by conditionprint(results)## Outcome Condition1 Condition1_N Condition1_Mean
## df Pre_shqscreener1 Waitlist Control 77 2.426407
## df1 Pre_shqscreener2 Waitlist Control 77 1.805195
## df2 Pre_shqscreener3 Waitlist Control 77 1.374459
## df3 Pre_DERS8_Sum Waitlist Control 77 28.350649
## df4 Pre_GAD7_Sum Waitlist Control 77 13.649351
## df5 Pre_PHQ9_Sum Waitlist Control 77 14.675325
## df6 Pre_SHS_Pathways Waitlist Control 74 13.783784
## df7 Pre_SHS_Agency Waitlist Control 74 11.202703
## df8 Pre_SHS_TotalHope Waitlist Control 74 24.986486
## df9 Pre_ucla_Sum Waitlist Control 74 7.270270
## df10 Pre_pmerq_Focus_Avg Waitlist Control 74 2.834459
## df11 Pre_pmerq_Distract_Avg Waitlist Control 74 4.313514
## df12 Pre_pmerq_AD_Avg Waitlist Control 74 3.573986
## Condition1_SD Condition2 Condition2_N Condition2_Mean Condition2_SD
## df 0.8837774 Purrble Treatment 76 2.307018 0.7750998
## df1 0.8339483 Purrble Treatment 76 1.712719 0.7371611
## df2 0.6130619 Purrble Treatment 76 1.267544 0.3840129
## df3 4.3003625 Purrble Treatment 76 27.918860 5.1017421
## df4 3.7141645 Purrble Treatment 76 13.780702 4.2495499
## df5 4.2195673 Purrble Treatment 76 15.390351 4.9000796
## df6 3.9940719 Purrble Treatment 75 12.806667 4.5268667
## df7 5.0962778 Purrble Treatment 75 10.220000 4.7414503
## df8 8.2237432 Purrble Treatment 75 23.026667 8.3649419
## df9 1.3578518 Purrble Treatment 74 6.918919 1.8339026
## df10 1.0800788 Purrble Treatment 75 2.641667 1.0369714
## df11 1.0979837 Purrble Treatment 75 4.132000 1.1549143
## df12 0.9636318 Purrble Treatment 75 3.386833 0.8766801
## df t p CI_Lower CI_Upper
## df 148.9466 0.8886754 0.3756105 -0.1460795 0.3848583
## df1 149.2051 0.7269566 0.4683916 -0.1588886 0.3438396
## df2 127.9451 1.2944955 0.1978257 -0.0565081 0.2703381
## df3 146.1480 0.5656812 0.5724780 -1.0767605 1.9403399
## df4 147.8107 -0.2034630 0.8390530 -1.4071046 1.1444024
## df5 147.1651 -0.9666385 0.3353107 -2.1768400 0.7467876
## df6 145.2059 1.3975805 0.1643707 -0.4047063 2.3589405
## df7 145.9327 1.2182135 0.2251087 -0.6115726 2.5769780
## df8 146.9982 1.4421041 0.1514002 -0.7258820 4.6455216
## df9 134.5433 1.3245408 0.1875690 -0.1732729 0.8759756
## df10 146.5693 1.1112036 0.2683002 -0.1500902 0.5356758
## df11 146.7989 0.9832994 0.3270779 -0.1832962 0.5463232
## df12 145.3132 1.2395598 0.2171364 -0.1112543 0.4855606ANCOVA Of all outcomes of interest: Examining Intervention Outcome Controlling for Baseline Covariates
# Assuming pre_vars and purrble_wide are defined
pre_vars <- c("Pre_shqscreener1", "Pre_shqscreener2", "Pre_shqscreener3",
"Pre_DERS8_Sum", "Pre_GAD7_Sum", "Pre_PHQ9_Sum",
"Pre_SHS_Pathways", "Pre_SHS_Agency", "Pre_SHS_TotalHope",
"Pre_ucla_Sum", "Pre_pmerq_Focus_Avg", "Pre_pmerq_Distract_Avg", "Pre_pmerq_AD_Avg")
# Identify condition levels
cond_levels <- unique(purrble_wide$condition)
cond_levels <- cond_levels[!is.na(cond_levels)]
if (length(cond_levels) != 2) {
stop("This code expects exactly two conditions for the t-tests.")
}
cat("Baseline differences in study outcomes by condition\n\n")## Baseline differences in study outcomes by condition# Initialize a results data frame
results <- data.frame(
Outcome = character(),
Condition1 = character(),
Condition1_N = numeric(),
Condition1_Mean = numeric(),
Condition1_SD = numeric(),
Condition2 = character(),
Condition2_N = numeric(),
Condition2_Mean = numeric(),
Condition2_SD = numeric(),
df = numeric(),
t = numeric(),
p = numeric(),
CI_Lower = numeric(),
CI_Upper = numeric(),
stringsAsFactors = FALSE
)
for (var in pre_vars) {
# Extract data for both conditions
cond1_data <- purrble_wide[purrble_wide$condition == cond_levels[1], var, drop = TRUE]
cond2_data <- purrble_wide[purrble_wide$condition == cond_levels[2], var, drop = TRUE]
# Compute descriptives
cond1_n <- sum(!is.na(cond1_data))
cond1_mean <- mean(cond1_data, na.rm = TRUE)
cond1_sd <- sd(cond1_data, na.rm = TRUE)
cond2_n <- sum(!is.na(cond2_data))
cond2_mean <- mean(cond2_data, na.rm = TRUE)
cond2_sd <- sd(cond2_data, na.rm = TRUE)
# Only run t-test if both groups have data
if (cond1_n > 0 & cond2_n > 0) {
tt <- t.test(purrble_wide[[var]] ~ purrble_wide$condition)
df_val <- tt$parameter
t_val <- tt$statistic
p_val <- tt$p.value
ci_lower <- tt$conf.int[1]
ci_upper <- tt$conf.int[2]
results <- rbind(results, data.frame(
Outcome = var,
Condition1 = cond_levels[1],
Condition1_N = cond1_n,
Condition1_Mean = cond1_mean,
Condition1_SD = cond1_sd,
Condition2 = cond_levels[2],
Condition2_N = cond2_n,
Condition2_Mean = cond2_mean,
Condition2_SD = cond2_sd,
df = df_val,
t = t_val,
p = p_val,
CI_Lower = ci_lower,
CI_Upper = ci_upper,
stringsAsFactors = FALSE
))
}
}
print(results)## Outcome Condition1 Condition1_N Condition1_Mean
## df Pre_shqscreener1 Waitlist Control 77 2.426407
## df1 Pre_shqscreener2 Waitlist Control 77 1.805195
## df2 Pre_shqscreener3 Waitlist Control 77 1.374459
## df3 Pre_DERS8_Sum Waitlist Control 77 28.350649
## df4 Pre_GAD7_Sum Waitlist Control 77 13.649351
## df5 Pre_PHQ9_Sum Waitlist Control 77 14.675325
## df6 Pre_SHS_Pathways Waitlist Control 74 13.783784
## df7 Pre_SHS_Agency Waitlist Control 74 11.202703
## df8 Pre_SHS_TotalHope Waitlist Control 74 24.986486
## df9 Pre_ucla_Sum Waitlist Control 74 7.270270
## df10 Pre_pmerq_Focus_Avg Waitlist Control 74 2.834459
## df11 Pre_pmerq_Distract_Avg Waitlist Control 74 4.313514
## df12 Pre_pmerq_AD_Avg Waitlist Control 74 3.573986
## Condition1_SD Condition2 Condition2_N Condition2_Mean Condition2_SD
## df 0.8837774 Purrble Treatment 76 2.307018 0.7750998
## df1 0.8339483 Purrble Treatment 76 1.712719 0.7371611
## df2 0.6130619 Purrble Treatment 76 1.267544 0.3840129
## df3 4.3003625 Purrble Treatment 76 27.918860 5.1017421
## df4 3.7141645 Purrble Treatment 76 13.780702 4.2495499
## df5 4.2195673 Purrble Treatment 76 15.390351 4.9000796
## df6 3.9940719 Purrble Treatment 75 12.806667 4.5268667
## df7 5.0962778 Purrble Treatment 75 10.220000 4.7414503
## df8 8.2237432 Purrble Treatment 75 23.026667 8.3649419
## df9 1.3578518 Purrble Treatment 74 6.918919 1.8339026
## df10 1.0800788 Purrble Treatment 75 2.641667 1.0369714
## df11 1.0979837 Purrble Treatment 75 4.132000 1.1549143
## df12 0.9636318 Purrble Treatment 75 3.386833 0.8766801
## df t p CI_Lower CI_Upper
## df 148.9466 0.8886754 0.3756105 -0.1460795 0.3848583
## df1 149.2051 0.7269566 0.4683916 -0.1588886 0.3438396
## df2 127.9451 1.2944955 0.1978257 -0.0565081 0.2703381
## df3 146.1480 0.5656812 0.5724780 -1.0767605 1.9403399
## df4 147.8107 -0.2034630 0.8390530 -1.4071046 1.1444024
## df5 147.1651 -0.9666385 0.3353107 -2.1768400 0.7467876
## df6 145.2059 1.3975805 0.1643707 -0.4047063 2.3589405
## df7 145.9327 1.2182135 0.2251087 -0.6115726 2.5769780
## df8 146.9982 1.4421041 0.1514002 -0.7258820 4.6455216
## df9 134.5433 1.3245408 0.1875690 -0.1732729 0.8759756
## df10 146.5693 1.1112036 0.2683002 -0.1500902 0.5356758
## df11 146.7989 0.9832994 0.3270779 -0.1832962 0.5463232
## df12 145.3132 1.2395598 0.2171364 -0.1112543 0.4855606# Define pre-test and post-test variables
pre_vars <- c("Pre_shqscreener1", "Pre_shqscreener2", "Pre_shqscreener3",
"Pre_DERS8_Sum", "Pre_GAD7_Sum", "Pre_PHQ9_Sum",
"Pre_SHS_Pathways", "Pre_SHS_Agency", "Pre_SHS_TotalHope",
"Pre_ucla_Sum", "Pre_pmerq_Focus_Avg", "Pre_pmerq_Distract_Avg", "Pre_pmerq_AD_Avg")
post_vars <- c("Post_shqscreener1", "Post_shqscreener2", "Post_shqscreener3",
"Post_DERS8_Sum", "Post_GAD7_Sum", "Post_PHQ9_Sum",
"Post_SHS_Pathways", "Post_SHS_Agency", "Post_SHS_TotalHope",
"Post_ucla_Sum", "Post_pmerq_Focus_Avg", "Post_pmerq_Distract_Avg", "Post_pmerq_AD_Avg")
# Run ANCOVAs for each post-test variable controlling for the corresponding pre-test variable
for (post_var in post_vars) {
# Identify the corresponding pre variable
pre_var <- sub("^Post_", "Pre_", post_var)
# Ensure that both variables exist in the dataset
if (!(pre_var %in% names(purrble_wide))) {
cat("No corresponding pre-test variable found for:", post_var, "\n")
next
}
# Fit the ANCOVA model
# Model: Post_Test_Outcome ~ condition + Pre_Test_Outcome
formula_str <- paste(post_var, "~ condition +", pre_var)
model <- aov(as.formula(formula_str), data = purrble_wide)
# Calculate Means and SDs by condition for the post-test variable
cond_means <- tapply(purrble_wide[[post_var]], purrble_wide$condition, mean, na.rm = TRUE)
cond_sd <- tapply(purrble_wide[[post_var]], purrble_wide$condition, sd, na.rm = TRUE)
# Print the header
cat("\n---------------------------------------\n")
cat("ANCOVA for:", post_var, "controlling for", pre_var, "\n")
cat("---------------------------------------\n")
# Print Means and SDs by condition
cat("Condition-level Descriptives for", post_var, ":\n")
cond_levels <- names(cond_means)
for (lvl in cond_levels) {
cat(lvl, ": Mean =", round(cond_means[lvl], 2), "SD =", round(cond_sd[lvl], 2), "\n")
}
cat("\n")
# Print the model summary
print(summary(model))
}##
## ---------------------------------------
## ANCOVA for: Post_shqscreener1 controlling for Pre_shqscreener1
## ---------------------------------------
## Condition-level Descriptives for Post_shqscreener1 :
## Waitlist Control : Mean = 2.11 SD = 0.94
## Purrble Treatment : Mean = 2.04 SD = 0.92
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.17 0.17 0.342 0.56
## Pre_shqscreener1 1 51.71 51.71 102.074 <2e-16 ***
## Residuals 138 69.91 0.51
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_shqscreener2 controlling for Pre_shqscreener2
## ---------------------------------------
## Condition-level Descriptives for Post_shqscreener2 :
## Waitlist Control : Mean = 1.72 SD = 0.78
## Purrble Treatment : Mean = 1.7 SD = 0.8
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.01 0.01 0.014 0.905
## Pre_shqscreener2 1 33.44 33.44 86.447 2.87e-16 ***
## Residuals 138 53.38 0.39
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_shqscreener3 controlling for Pre_shqscreener3
## ---------------------------------------
## Condition-level Descriptives for Post_shqscreener3 :
## Waitlist Control : Mean = 1.3 SD = 0.53
## Purrble Treatment : Mean = 1.27 SD = 0.53
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.06 0.063 0.273 0.602
## Pre_shqscreener3 1 7.94 7.938 34.587 2.93e-08 ***
## Residuals 138 31.67 0.230
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_DERS8_Sum controlling for Pre_DERS8_Sum
## ---------------------------------------
## Condition-level Descriptives for Post_DERS8_Sum :
## Waitlist Control : Mean = 28.58 SD = 6.48
## Purrble Treatment : Mean = 25.26 SD = 7.8
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 410 410.4 13.07 0.000419 ***
## Pre_DERS8_Sum 1 2779 2778.5 88.53 < 2e-16 ***
## Residuals 138 4331 31.4
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_GAD7_Sum controlling for Pre_GAD7_Sum
## ---------------------------------------
## Condition-level Descriptives for Post_GAD7_Sum :
## Waitlist Control : Mean = 13.22 SD = 4.43
## Purrble Treatment : Mean = 12 SD = 5.47
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 54 54.0 3.498 0.0636 .
## Pre_GAD7_Sum 1 1311 1311.2 84.884 4.68e-16 ***
## Residuals 138 2132 15.4
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_PHQ9_Sum controlling for Pre_PHQ9_Sum
## ---------------------------------------
## Condition-level Descriptives for Post_PHQ9_Sum :
## Waitlist Control : Mean = 15.16 SD = 5.89
## Purrble Treatment : Mean = 13.44 SD = 6.66
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 110.5 110.5 6.195 0.014 *
## Pre_PHQ9_Sum 1 3034.1 3034.1 170.110 <2e-16 ***
## Residuals 138 2461.4 17.8
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_SHS_Pathways controlling for Pre_SHS_Pathways
## ---------------------------------------
## Condition-level Descriptives for Post_SHS_Pathways :
## Waitlist Control : Mean = 14.84 SD = 4.2
## Purrble Treatment : Mean = 14.47 SD = 4.46
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 9.1 9.1 0.642 0.424
## Pre_SHS_Pathways 1 602.4 602.4 42.711 1.45e-09 ***
## Residuals 125 1763.0 14.1
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 26 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_SHS_Agency controlling for Pre_SHS_Agency
## ---------------------------------------
## Condition-level Descriptives for Post_SHS_Agency :
## Waitlist Control : Mean = 12.72 SD = 5.06
## Purrble Treatment : Mean = 12.52 SD = 5.42
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 3.5 3.5 0.172 0.679
## Pre_SHS_Agency 1 922.2 922.2 44.821 6.53e-10 ***
## Residuals 125 2571.8 20.6
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 26 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_SHS_TotalHope controlling for Pre_SHS_TotalHope
## ---------------------------------------
## Condition-level Descriptives for Post_SHS_TotalHope :
## Waitlist Control : Mean = 27.57 SD = 8.63
## Purrble Treatment : Mean = 26.98 SD = 9.03
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 24 23.9 0.421 0.518
## Pre_SHS_TotalHope 1 2786 2786.1 49.027 1.37e-10 ***
## Residuals 125 7103 56.8
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 26 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_ucla_Sum controlling for Pre_ucla_Sum
## ---------------------------------------
## Condition-level Descriptives for Post_ucla_Sum :
## Waitlist Control : Mean = 6.91 SD = 1.57
## Purrble Treatment : Mean = 6.68 SD = 1.84
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 1.76 1.76 1.092 0.298
## Pre_ucla_Sum 1 169.72 169.72 105.241 <2e-16 ***
## Residuals 124 199.98 1.61
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_pmerq_Focus_Avg controlling for Pre_pmerq_Focus_Avg
## ---------------------------------------
## Condition-level Descriptives for Post_pmerq_Focus_Avg :
## Waitlist Control : Mean = 2.89 SD = 1.18
## Purrble Treatment : Mean = 3.1 SD = 1.21
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.92 0.92 1.128 0.29
## Pre_pmerq_Focus_Avg 1 77.79 77.79 95.442 <2e-16 ***
## Residuals 124 101.06 0.82
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_pmerq_Distract_Avg controlling for Pre_pmerq_Distract_Avg
## ---------------------------------------
## Condition-level Descriptives for Post_pmerq_Distract_Avg :
## Waitlist Control : Mean = 4.26 SD = 1.05
## Purrble Treatment : Mean = 4.42 SD = 1.06
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.64 0.64 0.77 0.382
## Pre_pmerq_Distract_Avg 1 35.83 35.83 42.88 1.39e-09 ***
## Residuals 124 103.64 0.84
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## ---------------------------------------
## ANCOVA for: Post_pmerq_AD_Avg controlling for Pre_pmerq_AD_Avg
## ---------------------------------------
## Condition-level Descriptives for Post_pmerq_AD_Avg :
## Waitlist Control : Mean = 3.58 SD = 0.97
## Purrble Treatment : Mean = 3.76 SD = 0.93
##
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.78 0.78 1.492 0.224
## Pre_pmerq_AD_Avg 1 49.29 49.29 94.794 <2e-16 ***
## Residuals 124 64.47 0.52
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingnesslibrary(effectsize)##
## Attaching package: 'effectsize'## The following object is masked from 'package:psych':
##
## phi# Define pre-test and post-test variables
pre_vars <- c("Pre_shqscreener1", "Pre_shqscreener2", "Pre_shqscreener3",
"Pre_DERS8_Sum", "Pre_GAD7_Sum", "Pre_PHQ9_Sum",
"Pre_SHS_Pathways", "Pre_SHS_Agency", "Pre_SHS_TotalHope",
"Pre_ucla_Sum", "Pre_pmerq_Focus_Avg", "Pre_pmerq_Distract_Avg", "Pre_pmerq_AD_Avg")
post_vars <- c("Post_shqscreener1", "Post_shqscreener2", "Post_shqscreener3",
"Post_DERS8_Sum", "Post_GAD7_Sum", "Post_PHQ9_Sum",
"Post_SHS_Pathways", "Post_SHS_Agency", "Post_SHS_TotalHope",
"Post_ucla_Sum", "Post_pmerq_Focus_Avg", "Post_pmerq_Distract_Avg", "Post_pmerq_AD_Avg")
# Run ANCOVAs for each post-test variable controlling for the corresponding pre-test variable
for (post_var in post_vars) {
# Identify the corresponding pre variable
pre_var <- sub("^Post_", "Pre_", post_var)
# Ensure that both variables exist in the dataset
if (!(pre_var %in% names(purrble_wide))) {
cat("No corresponding pre-test variable found for:", post_var, "\n")
next
}
# Fit the ANCOVA model
formula_str <- paste(post_var, "~ condition +", pre_var)
model <- aov(as.formula(formula_str), data = purrble_wide)
# Print a header and the model summary
cat("\n---------------------------------------\n")
cat("ANCOVA for:", post_var, "controlling for", pre_var, "\n")
cat("---------------------------------------\n")
print(summary(model))
# Compute partial eta squared
eta_sq_results <- eta_squared(model, partial = TRUE)
# Compute Cohen's f (partial)
# By default, cohens_f() returns partial effects for each term in the model
f_results <- cohens_f(model, partial = TRUE)
cat("\nPartial Eta Squared:\n")
print(eta_sq_results)
cat("\nCohen's f (Partial):\n")
print(f_results)
}##
## ---------------------------------------
## ANCOVA for: Post_shqscreener1 controlling for Pre_shqscreener1
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.17 0.17 0.342 0.56
## Pre_shqscreener1 1 51.71 51.71 102.074 <2e-16 ***
## Residuals 138 69.91 0.51
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------
## condition | 2.47e-03 | [0.00, 1.00]
## Pre_shqscreener1 | 0.43 | [0.33, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ----------------------------------------------------
## condition | 0.05 | [0.00, Inf]
## Pre_shqscreener1 | 0.86 | [0.69, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_shqscreener2 controlling for Pre_shqscreener2
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.01 0.01 0.014 0.905
## Pre_shqscreener2 1 33.44 33.44 86.447 2.87e-16 ***
## Residuals 138 53.38 0.39
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------
## condition | 1.03e-04 | [0.00, 1.00]
## Pre_shqscreener2 | 0.39 | [0.28, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ----------------------------------------------------
## condition | 0.01 | [0.00, Inf]
## Pre_shqscreener2 | 0.79 | [0.63, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_shqscreener3 controlling for Pre_shqscreener3
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.06 0.063 0.273 0.602
## Pre_shqscreener3 1 7.94 7.938 34.587 2.93e-08 ***
## Residuals 138 31.67 0.230
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------
## condition | 1.97e-03 | [0.00, 1.00]
## Pre_shqscreener3 | 0.20 | [0.11, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ----------------------------------------------------
## condition | 0.04 | [0.00, Inf]
## Pre_shqscreener3 | 0.50 | [0.35, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_DERS8_Sum controlling for Pre_DERS8_Sum
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 410 410.4 13.07 0.000419 ***
## Pre_DERS8_Sum 1 2779 2778.5 88.53 < 2e-16 ***
## Residuals 138 4331 31.4
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ---------------------------------------------
## condition | 0.09 | [0.03, 1.00]
## Pre_DERS8_Sum | 0.39 | [0.29, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## -------------------------------------------------
## condition | 0.31 | [0.16, Inf]
## Pre_DERS8_Sum | 0.80 | [0.64, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_GAD7_Sum controlling for Pre_GAD7_Sum
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 54 54.0 3.498 0.0636 .
## Pre_GAD7_Sum 1 1311 1311.2 84.884 4.68e-16 ***
## Residuals 138 2132 15.4
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## --------------------------------------------
## condition | 0.02 | [0.00, 1.00]
## Pre_GAD7_Sum | 0.38 | [0.28, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ------------------------------------------------
## condition | 0.16 | [0.00, Inf]
## Pre_GAD7_Sum | 0.78 | [0.62, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_PHQ9_Sum controlling for Pre_PHQ9_Sum
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 110.5 110.5 6.195 0.014 *
## Pre_PHQ9_Sum 1 3034.1 3034.1 170.110 <2e-16 ***
## Residuals 138 2461.4 17.8
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 13 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## --------------------------------------------
## condition | 0.04 | [0.00, 1.00]
## Pre_PHQ9_Sum | 0.55 | [0.46, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ------------------------------------------------
## condition | 0.21 | [0.07, Inf]
## Pre_PHQ9_Sum | 1.11 | [0.93, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_SHS_Pathways controlling for Pre_SHS_Pathways
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 9.1 9.1 0.642 0.424
## Pre_SHS_Pathways 1 602.4 602.4 42.711 1.45e-09 ***
## Residuals 125 1763.0 14.1
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 26 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------
## condition | 5.11e-03 | [0.00, 1.00]
## Pre_SHS_Pathways | 0.25 | [0.15, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ----------------------------------------------------
## condition | 0.07 | [0.00, Inf]
## Pre_SHS_Pathways | 0.58 | [0.42, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_SHS_Agency controlling for Pre_SHS_Agency
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 3.5 3.5 0.172 0.679
## Pre_SHS_Agency 1 922.2 922.2 44.821 6.53e-10 ***
## Residuals 125 2571.8 20.6
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 26 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ----------------------------------------------
## condition | 1.37e-03 | [0.00, 1.00]
## Pre_SHS_Agency | 0.26 | [0.16, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## --------------------------------------------------
## condition | 0.04 | [0.00, Inf]
## Pre_SHS_Agency | 0.60 | [0.44, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_SHS_TotalHope controlling for Pre_SHS_TotalHope
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 24 23.9 0.421 0.518
## Pre_SHS_TotalHope 1 2786 2786.1 49.027 1.37e-10 ***
## Residuals 125 7103 56.8
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 26 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## -------------------------------------------------
## condition | 3.35e-03 | [0.00, 1.00]
## Pre_SHS_TotalHope | 0.28 | [0.18, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## -----------------------------------------------------
## condition | 0.06 | [0.00, Inf]
## Pre_SHS_TotalHope | 0.63 | [0.46, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_ucla_Sum controlling for Pre_ucla_Sum
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 1.76 1.76 1.092 0.298
## Pre_ucla_Sum 1 169.72 169.72 105.241 <2e-16 ***
## Residuals 124 199.98 1.61
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## --------------------------------------------
## condition | 8.73e-03 | [0.00, 1.00]
## Pre_ucla_Sum | 0.46 | [0.36, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ------------------------------------------------
## condition | 0.09 | [0.00, Inf]
## Pre_ucla_Sum | 0.92 | [0.74, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_pmerq_Focus_Avg controlling for Pre_pmerq_Focus_Avg
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.92 0.92 1.128 0.29
## Pre_pmerq_Focus_Avg 1 77.79 77.79 95.442 <2e-16 ***
## Residuals 124 101.06 0.82
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ---------------------------------------------------
## condition | 9.02e-03 | [0.00, 1.00]
## Pre_pmerq_Focus_Avg | 0.43 | [0.33, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## -------------------------------------------------------
## condition | 0.10 | [0.00, Inf]
## Pre_pmerq_Focus_Avg | 0.88 | [0.70, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_pmerq_Distract_Avg controlling for Pre_pmerq_Distract_Avg
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.64 0.64 0.77 0.382
## Pre_pmerq_Distract_Avg 1 35.83 35.83 42.88 1.39e-09 ***
## Residuals 124 103.64 0.84
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------------
## condition | 6.17e-03 | [0.00, 1.00]
## Pre_pmerq_Distract_Avg | 0.26 | [0.15, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ----------------------------------------------------------
## condition | 0.08 | [0.00, Inf]
## Pre_pmerq_Distract_Avg | 0.59 | [0.43, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].
## ---------------------------------------
## ANCOVA for: Post_pmerq_AD_Avg controlling for Pre_pmerq_AD_Avg
## ---------------------------------------
## Df Sum Sq Mean Sq F value Pr(>F)
## condition 1 0.78 0.78 1.492 0.224
## Pre_pmerq_AD_Avg 1 49.29 49.29 94.794 <2e-16 ***
## Residuals 124 64.47 0.52
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 27 observations deleted due to missingness
##
## Partial Eta Squared:
## # Effect Size for ANOVA (Type I)
##
## Parameter | Eta2 (partial) | 95% CI
## ------------------------------------------------
## condition | 0.01 | [0.00, 1.00]
## Pre_pmerq_AD_Avg | 0.43 | [0.33, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].
## Cohen's f (Partial):
## # Effect Size for ANOVA (Type I)
##
## Parameter | Cohen's f (partial) | 95% CI
## ----------------------------------------------------
## condition | 0.11 | [0.00, Inf]
## Pre_pmerq_AD_Avg | 0.87 | [0.70, Inf]
##
## - One-sided CIs: upper bound fixed at [Inf].Post-test Logistic Regression to Investigate Intervention Effects on Self-Harm Outcomes
library(dplyr)
library(gtsummary)
library(broom)
#----------------------------------------------------------
# 1) Logistic regression for SHQ1 at Week 12
# controlling for Week 2 SHQ1 and Condition
#----------------------------------------------------------
model_shq1 <- glm(
SHQ1_12 ~ condition + SHQ1_2,
data = purrble_wide,
family = binomial
)
#----------------------------------------------------------
# 2) Logistic regression for SHQ2 at Week 12
# controlling for Week 2 SHQ2 and Condition
#----------------------------------------------------------
model_shq2 <- glm(
SHQ2_12 ~ condition + SHQ2_2,
data = purrble_wide,
family = binomial
)
#----------------------------------------------------------
# 3) Logistic regression for SHQ3 at Week 12
# controlling for Week 2 SHQ3 and Condition
#----------------------------------------------------------
model_shq3 <- glm(
SHQ3_12 ~ condition + SHQ3_2,
data = purrble_wide,
family = binomial
)
#----------------------------------------------------------
# 4) Logistic regression for SHQ_Any at Week 12
# controlling for Week 2 SHQ_Any and Condition
#----------------------------------------------------------
model_shqAny <- glm(
SHQ_Any_12 ~ condition + SHQ_Any_2,
data = purrble_wide,
family = binomial
)
# Create gtsummary tables for each model, exponentiating for OR
tbl_shq1 <- tbl_regression(model_shq1, exponentiate = TRUE) %>%
bold_labels() %>%
add_significance_stars()
tbl_shq2 <- tbl_regression(model_shq2, exponentiate = TRUE) %>%
bold_labels() %>%
add_significance_stars()
tbl_shq3 <- tbl_regression(model_shq3, exponentiate = TRUE) %>%
bold_labels() %>%
add_significance_stars()
tbl_shqAny <- tbl_regression(model_shqAny, exponentiate = TRUE) %>%
bold_labels() %>%
add_significance_stars()
merged_tbl <- tbl_merge(
tbls = list(tbl_shq1, tbl_shq2, tbl_shq3, tbl_shqAny),
tab_spanner = c("SHQ1 Model", "SHQ2 Model", "SHQ3 Model", "SHQ_Any Model")
)
merged_tbl| Characteristic |
SHQ1 Model
|
SHQ2 Model
|
SHQ3 Model
|
SHQ_Any Model
|
||||
|---|---|---|---|---|---|---|---|---|
| OR1,2 | SE2 | OR1,2 | SE2 | OR1,2 | SE2 | OR1,2 | SE2 | |
| condition | ||||||||
| Waitlist Control | — | — | — | — | — | — | — | — |
| Purrble Treatment | 1.15 | 0.452 | 0.98 | 0.412 | 0.87 | 0.546 | 1.10 | 0.434 |
| SHQ1_2 | 11.6*** | 0.484 | ||||||
| SHQ2_2 | 4.36*** | 0.408 | ||||||
| SHQ3_2 | 3.14* | 0.559 | ||||||
| SHQ_Any_2 | 5.83*** | 0.486 | ||||||
| 1 *p<0.05; **p<0.01; ***p<0.001 | ||||||||
| 2 OR = Odds Ratio, SE = Standard Error | ||||||||
When you save the notebook, an HTML file containing the code and output will be saved alongside it (click the Preview button or press Cmd+Shift+K to preview the HTML file).
The preview shows you a rendered HTML copy of the contents of the editor. Consequently, unlike Knit, Preview does not run any R code chunks. Instead, the output of the chunk when it was last run in the editor is displayed.