Experimental Data Analysis
Kristina Becvar
2023-05-20
Analyzing Experimental Data
##Initial Data Work
First assessment of the data structure to understand what variables are numerical, factor, character, etc.
#Import Data
library(haven)
experimental_data <- read_sav("grad_experiments23_set1_May+3,+2023_10.56.sav")
#Removing the row with data from the person who opted out of the survey, the person who declined to identify gender, and those that chose independent when asked for lean
experimental_data_1 <- experimental_data %>%
slice(-c(48,130,131,135,152,156,192,211))
#Variable Review
names(experimental_data_1)## [1] "StartDate" "EndDate"
## [3] "Status" "IPAddress"
## [5] "Progress" "Duration__in_seconds_"
## [7] "Finished" "RecordedDate"
## [9] "ResponseId" "RecipientLastName"
## [11] "RecipientFirstName" "RecipientEmail"
## [13] "ExternalReference" "LocationLatitude"
## [15] "LocationLongitude" "DistributionChannel"
## [17] "UserLanguage" "Q_RecaptchaScore"
## [19] "Q_RelevantIDDuplicate" "Q_RelevantIDDuplicateScore"
## [21] "Q_RelevantIDFraudScore" "Q_RelevantIDLastStartDate"
## [23] "Q1.1" "Q2.1"
## [25] "Q2.2" "Q2.4"
## [27] "Q2.5_1" "Q2.5_2"
## [29] "Q2.5_3" "Q2.5_4"
## [31] "Q2.5_5" "Q2.5_DO_1"
## [33] "Q2.5_DO_2" "Q2.5_DO_3"
## [35] "Q2.5_DO_4" "Q2.5_DO_5"
## [37] "Q3.11_1" "Q3.11_2"
## [39] "Q3.11_3" "Q3.11_4"
## [41] "Q3.11_5" "Q3.11_DO_1"
## [43] "Q3.11_DO_2" "Q3.11_DO_3"
## [45] "Q3.11_DO_4" "Q3.11_DO_5"
## [47] "Q3.16_4" "Q3.17_14"
## [49] "Q4.4_1" "Q4.5_1"
## [51] "Q5.5" "Q5.6"
## [53] "Q6.1_1" "Q6.1_2"
## [55] "Q6.1_3" "Q6.1_4"
## [57] "Q6.1_5" "Q6.1_6"
## [59] "Q6.1_7" "Q6.1_8"
## [61] "Q6.1_9" "Q6.2"
## [63] "Q6.3" "Q6.4_1"
## [65] "Q4.6" "Q4.7"
## [67] "Random_ID" "kristina_DO_Q3.9"
## [69] "kristina_DO_Q3.8" "kristina_DO_Q3.7"
## [71] "kristina_DO_Q3.6" "kristina_DO_Q3.5"
## [73] "kristina_DO_Q3.4" "kristina_DO_Q3.3"
## [75] "kristina_DO_Q3.2" "kristina_DO_Q3.1"
## [77] "matthew_DO_Q4.3" "matthew_DO_Q4.2"
## [79] "matthew_DO_Q4.1" "justin_DO_Q5.4"
## [81] "justin_DO_Q5.3" "justin_DO_Q5.2"
## [83] "justin_DO_Q5.1"#Unweighted summary values for numeric variables
skim(experimental_data_1)| Name | experimental_data_1 |
| Number of rows | 276 |
| Number of columns | 83 |
| _______________________ | |
| Column type frequency: | |
| character | 12 |
| numeric | 67 |
| POSIXct | 4 |
| ________________________ | |
| Group variables | None |
Variable type: character
| skim_variable | n_missing | complete_rate | min | max | empty | n_unique | whitespace |
|---|---|---|---|---|---|---|---|
| IPAddress | 0 | 1 | 11 | 15 | 0 | 276 | 0 |
| ResponseId | 0 | 1 | 17 | 17 | 0 | 276 | 0 |
| RecipientLastName | 0 | 1 | 0 | 0 | 276 | 1 | 0 |
| RecipientFirstName | 0 | 1 | 0 | 0 | 276 | 1 | 0 |
| RecipientEmail | 0 | 1 | 0 | 0 | 276 | 1 | 0 |
| ExternalReference | 0 | 1 | 0 | 0 | 276 | 1 | 0 |
| LocationLatitude | 0 | 1 | 5 | 7 | 0 | 82 | 0 |
| LocationLongitude | 0 | 1 | 6 | 9 | 0 | 82 | 0 |
| DistributionChannel | 0 | 1 | 9 | 9 | 0 | 1 | 0 |
| UserLanguage | 0 | 1 | 2 | 2 | 0 | 1 | 0 |
| Q_RelevantIDDuplicate | 0 | 1 | 0 | 0 | 276 | 1 | 0 |
| Random_ID | 0 | 1 | 5 | 5 | 0 | 275 | 0 |
Variable type: numeric
| skim_variable | n_missing | complete_rate | mean | sd | p0 | p25 | p50 | p75 | p100 | hist |
|---|---|---|---|---|---|---|---|---|---|---|
| Status | 0 | 1.00 | 0.00 | 0.00 | 0.0 | 0.00 | 0.0 | 0.00 | 0 | ▁▁▇▁▁ |
| Progress | 0 | 1.00 | 100.00 | 0.00 | 100.0 | 100.00 | 100.0 | 100.00 | 100 | ▁▁▇▁▁ |
| Duration_in_seconds | 0 | 1.00 | 383.22 | 793.08 | 66.0 | 128.00 | 202.5 | 355.50 | 11214 | ▇▁▁▁▁ |
| Finished | 0 | 1.00 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q_RecaptchaScore | 10 | 0.96 | 0.94 | 0.13 | 0.1 | 0.90 | 1.0 | 1.00 | 1 | ▁▁▁▁▇ |
| Q_RelevantIDDuplicateScore | 0 | 1.00 | 0.00 | 0.00 | 0.0 | 0.00 | 0.0 | 0.00 | 0 | ▁▁▇▁▁ |
| Q_RelevantIDFraudScore | 0 | 1.00 | 26.41 | 25.06 | 0.0 | 0.00 | 30.0 | 30.00 | 135 | ▇▇▁▁▁ |
| Q1.1 | 0 | 1.00 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q2.1 | 0 | 1.00 | 1.75 | 0.74 | 1.0 | 1.00 | 2.0 | 2.00 | 3 | ▇▁▇▁▃ |
| Q2.2 | 49 | 0.82 | 1.10 | 0.30 | 1.0 | 1.00 | 1.0 | 1.00 | 2 | ▇▁▁▁▁ |
| Q2.4 | 227 | 0.18 | 1.59 | 0.50 | 1.0 | 1.00 | 2.0 | 2.00 | 2 | ▆▁▁▁▇ |
| Q2.5_1 | 0 | 1.00 | 2.21 | 1.08 | 1.0 | 2.00 | 2.0 | 3.00 | 5 | ▅▇▂▂▁ |
| Q2.5_2 | 0 | 1.00 | 2.12 | 1.06 | 1.0 | 1.00 | 2.0 | 2.25 | 5 | ▅▇▂▁▁ |
| Q2.5_3 | 1 | 1.00 | 2.36 | 1.12 | 1.0 | 2.00 | 2.0 | 3.00 | 5 | ▃▇▂▂▁ |
| Q2.5_4 | 1 | 1.00 | 2.10 | 1.09 | 1.0 | 1.00 | 2.0 | 2.50 | 5 | ▆▇▂▂▁ |
| Q2.5_5 | 0 | 1.00 | 2.02 | 1.02 | 1.0 | 1.00 | 2.0 | 2.00 | 5 | ▅▇▂▁▁ |
| Q2.5_DO_1 | 0 | 1.00 | 2.96 | 1.39 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▆▇▇▆▆ |
| Q2.5_DO_2 | 0 | 1.00 | 3.06 | 1.44 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▇▆▇▇▇ |
| Q2.5_DO_3 | 0 | 1.00 | 2.86 | 1.46 | 1.0 | 1.75 | 3.0 | 4.00 | 5 | ▇▆▆▆▆ |
| Q2.5_DO_4 | 0 | 1.00 | 3.00 | 1.41 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▆▇▆▆▇ |
| Q2.5_DO_5 | 0 | 1.00 | 3.13 | 1.37 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▆▆▇▇▆ |
| Q3.11_1 | 0 | 1.00 | 2.29 | 1.02 | 1.0 | 2.00 | 2.0 | 3.00 | 5 | ▃▇▃▂▁ |
| Q3.11_2 | 0 | 1.00 | 2.63 | 1.22 | 1.0 | 2.00 | 2.0 | 3.00 | 5 | ▃▇▅▃▂ |
| Q3.11_3 | 0 | 1.00 | 2.08 | 1.01 | 1.0 | 1.00 | 2.0 | 3.00 | 5 | ▆▇▃▁▁ |
| Q3.11_4 | 0 | 1.00 | 2.20 | 1.07 | 1.0 | 1.00 | 2.0 | 3.00 | 5 | ▅▇▃▂▁ |
| Q3.11_5 | 0 | 1.00 | 2.15 | 1.11 | 1.0 | 1.00 | 2.0 | 3.00 | 5 | ▆▇▃▂▁ |
| Q3.11_DO_1 | 0 | 1.00 | 2.97 | 1.45 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▇▇▇▆▇ |
| Q3.11_DO_2 | 0 | 1.00 | 2.86 | 1.32 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▆▇▇▆▅ |
| Q3.11_DO_3 | 0 | 1.00 | 3.07 | 1.43 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▇▇▇▇▇ |
| Q3.11_DO_4 | 0 | 1.00 | 3.10 | 1.47 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▇▆▆▇▇ |
| Q3.11_DO_5 | 0 | 1.00 | 3.01 | 1.40 | 1.0 | 2.00 | 3.0 | 4.00 | 5 | ▆▇▆▇▆ |
| Q3.16_4 | 0 | 1.00 | 73.96 | 17.39 | 14.0 | 65.00 | 76.0 | 86.00 | 100 | ▁▁▅▇▇ |
| Q3.17_14 | 0 | 1.00 | 75.39 | 17.93 | 0.0 | 67.75 | 77.5 | 87.00 | 100 | ▁▁▂▇▇ |
| Q4.4_1 | 0 | 1.00 | 2.07 | 0.92 | 1.0 | 1.00 | 2.0 | 2.00 | 5 | ▃▇▂▁▁ |
| Q4.5_1 | 1 | 1.00 | 2.19 | 0.92 | 1.0 | 2.00 | 2.0 | 2.50 | 5 | ▂▇▂▁▁ |
| Q5.5 | 0 | 1.00 | 1.93 | 0.83 | 1.0 | 1.00 | 2.0 | 2.00 | 5 | ▅▇▁▁▁ |
| Q5.6 | 0 | 1.00 | 2.15 | 0.87 | 1.0 | 2.00 | 2.0 | 3.00 | 5 | ▅▇▅▁▁ |
| Q6.1_1 | 258 | 0.07 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_2 | 265 | 0.04 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_3 | 264 | 0.04 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_4 | 267 | 0.03 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_5 | 271 | 0.02 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_6 | 271 | 0.02 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_7 | 29 | 0.89 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_8 | 271 | 0.02 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.1_9 | 270 | 0.02 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| Q6.2 | 0 | 1.00 | 1.56 | 0.50 | 1.0 | 1.00 | 2.0 | 2.00 | 2 | ▆▁▁▁▇ |
| Q6.3 | 1 | 1.00 | 2.58 | 1.13 | 1.0 | 1.00 | 3.0 | 3.00 | 4 | ▆▁▁▇▅ |
| Q6.4_1 | 1 | 1.00 | 17.00 | 8.19 | 2.0 | 13.00 | 15.0 | 18.00 | 52 | ▂▇▁▁▁ |
| Q4.6 | 0 | 1.00 | 1.67 | 0.48 | 1.0 | 1.00 | 2.0 | 2.00 | 3 | ▃▁▇▁▁ |
| Q4.7 | 2 | 0.99 | 1.82 | 0.39 | 1.0 | 2.00 | 2.0 | 2.00 | 2 | ▂▁▁▁▇ |
| kristina_DO_Q3.9 | 241 | 0.13 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.8 | 244 | 0.12 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.7 | 241 | 0.13 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.6 | 242 | 0.12 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.5 | 240 | 0.13 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.4 | 241 | 0.13 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.3 | 241 | 0.13 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.2 | 242 | 0.12 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| kristina_DO_Q3.1 | 0 | 1.00 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| matthew_DO_Q4.3 | 136 | 0.51 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| matthew_DO_Q4.2 | 140 | 0.49 | 2.00 | 0.00 | 2.0 | 2.00 | 2.0 | 2.00 | 2 | ▁▁▇▁▁ |
| matthew_DO_Q4.1 | 0 | 1.00 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| justin_DO_Q5.4 | 208 | 0.25 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| justin_DO_Q5.3 | 210 | 0.24 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| justin_DO_Q5.2 | 212 | 0.23 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
| justin_DO_Q5.1 | 198 | 0.28 | 1.00 | 0.00 | 1.0 | 1.00 | 1.0 | 1.00 | 1 | ▁▁▇▁▁ |
Variable type: POSIXct
| skim_variable | n_missing | complete_rate | min | max | median | n_unique |
|---|---|---|---|---|---|---|
| StartDate | 0 | 1 | 2023-05-03 05:37:44 | 2023-05-03 08:18:35 | 2023-05-03 05:48:44 | 260 |
| EndDate | 0 | 1 | 2023-05-03 05:40:07 | 2023-05-03 08:58:58 | 2023-05-03 05:55:46 | 267 |
| RecordedDate | 0 | 1 | 2023-05-03 05:40:08 | 2023-05-03 08:58:59 | 2023-05-03 05:55:46 | 262 |
| Q_RelevantIDLastStartDate | 0 | 1 | 1582-10-14 00:00:00 | 1582-10-14 00:00:00 | 1582-10-14 00:00:00 | 1 |
#This runs a frequency analysis on values in the variables of interest (Foreign Policy and Terrorism)
freq(experimental_data_1$Q3.11_1) #foreign policy 
## How qualified do you believe the candidate is on the following issues? - Foreign Policy
## Frequency Percent
## 1 58 21.014
## 2 127 46.014
## 3 55 19.928
## 4 25 9.058
## 5 11 3.986
## Total 276 100.000freq(experimental_data_1$Q3.11_2) #terrorism
## How qualified do you believe the candidate is on the following issues? - Terrorism
## Frequency Percent
## 1 47 17.03
## 2 103 37.32
## 3 58 21.01
## 4 40 14.49
## 5 28 10.14
## Total 276 100.00attributes(experimental_data_1$Q3.11_1) #value labels for DV-Foreign Policy## $label
## [1] "How qualified do you believe the candidate is on the following issues? - Foreign Policy"
##
## $format.spss
## [1] "F40.0"
##
## $display_width
## [1] 5
##
## $labels
## Highly Qualified Qualified Somewhat Qualified
## 1 2 3
## Not Very Qualified Not At All Qualified
## 4 5
##
## $class
## [1] "haven_labelled" "vctrs_vctr" "double"attributes(experimental_data_1$Q3.11_2) #value labels for DV-Terrorism## $label
## [1] "How qualified do you believe the candidate is on the following issues? - Terrorism"
##
## $format.spss
## [1] "F40.0"
##
## $display_width
## [1] 5
##
## $labels
## Highly Qualified Qualified Somewhat Qualified
## 1 2 3
## Not Very Qualified Not At All Qualified
## 4 5
##
## $class
## [1] "haven_labelled" "vctrs_vctr" "double"class(experimental_data_1$Q3.11_1) #variable type for DV-Foreign Policy## [1] "haven_labelled" "vctrs_vctr" "double"class(experimental_data_1$Q3.11_2) #variable type for DV-Terrorism## [1] "haven_labelled" "vctrs_vctr" "double"summary(experimental_data_1$Q3.11_1) #central tendencies for DV-Foreign Policy## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.00 2.00 2.00 2.29 3.00 5.00summary(experimental_data_1$Q3.11_2) #central tendencies for DV-Terrorism## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.000 2.000 2.000 2.634 3.000 5.000Transforming variables as needed
#Specifically transforming the variables I need for my analysis
experimental_data_1$age <- as.numeric(experimental_data_1$Q6.4_1)
#Taking the party data and combining self-identified party members with leaners
experimental_data_1$polparty <- as.numeric(experimental_data_1$Q2.1)
experimental_data_1$leaner <- as.numeric(experimental_data_1$Q2.4)
experimental_data_1 <- experimental_data_1 %>%
mutate(party = case_when(
polparty==2 ~ "Republican",
polparty==1 ~ "Democrat",
leaner==1 ~ "Democrat",
leaner==2 ~ "Republican"
))
experimental_data_1$democrat <- ifelse(experimental_data_1$party == "Democrat", 1, 0)
experimental_data_1$republican <- ifelse(experimental_data_1$party == "Republican", 1, 0)
#Checking variable attributes for gender data question
attributes(experimental_data_1$Q6.2)## $label
## [1] "How do you typically describe yourself when it comes to your gender?"
##
## $format.spss
## [1] "F40.0"
##
## $display_width
## [1] 5
##
## $labels
## Male Female
## 1 2
## Non-binary Transgender
## 3 4
## Identify in some other way Prefer not to answer
## 5 6
##
## $class
## [1] "haven_labelled" "vctrs_vctr" "double"#Creating easier to work with gender variables
experimental_data_1$gender <- as.numeric(experimental_data_1$Q6.2)
experimental_data_1$female <- ifelse(experimental_data_1$gender == "2", 1, 0)
experimental_data_1$male <- ifelse(experimental_data_1$gender == "1", 1, 0)
#Taking the gender data and creating variables
experimental_data_1 <- experimental_data_1 %>%
mutate(Gender = case_when(
gender==2 ~ "Female",
gender==1 ~ "Male"
))
experimental_data_1$pt_guns <- as.numeric(experimental_data_1$Q2.5_1) #Creates Pre-test gun law evaluation
experimental_data_1$pt_climate <- as.numeric(experimental_data_1$Q2.5_2) #Creates Pre-test climate change evaluation
experimental_data_1$pt_militarization <- as.numeric(experimental_data_1$Q2.5_3) #Creates Pre-test police militarization evaluation
experimental_data_1$pt_guns_order <- as.numeric(experimental_data_1$Q2.5_DO_1) #Indicates where in 5-question pretest this question was viewed
experimental_data_1$pt_climate_order <- as.numeric(experimental_data_1$Q2.5_DO_2) #Indicates where in 5-question pretest this question was viewed
experimental_data_1$pt_militarization_order <- as.numeric(experimental_data_1$Q2.5_DO_3) #Indicates where in 5-question pretest this question was viewed
#Taking the random treatments and assigning a group label
experimental_data_1 <- experimental_data_1 %>%
mutate(group = case_when(
kristina_DO_Q3.2==2 ~ "F,C,D",
kristina_DO_Q3.3==2 ~ "F,C,R",
kristina_DO_Q3.4==2 ~ "F,V,D",
kristina_DO_Q3.5==2 ~ "F,V,R",
kristina_DO_Q3.6==2 ~ "M,C,D",
kristina_DO_Q3.7==2 ~ "M,C,R",
kristina_DO_Q3.8==2 ~ "M,V,D",
kristina_DO_Q3.9==2 ~ "M,V,R",
))
#Taking the random treatments and assigning a numeric group label
experimental_data_1 <- experimental_data_1 %>%
mutate(group_number = case_when(
kristina_DO_Q3.2==2 ~ "1",
kristina_DO_Q3.3==2 ~ "2",
kristina_DO_Q3.4==2 ~ "3",
kristina_DO_Q3.5==2 ~ "4",
kristina_DO_Q3.6==2 ~ "5",
kristina_DO_Q3.7==2 ~ "6",
kristina_DO_Q3.8==2 ~ "7",
kristina_DO_Q3.9==2 ~ "8",
))
#Recoding the treatment groups as in-group for political and gender comparisons
experimental_data_1 <- experimental_data_1 %>%
mutate(treatment_dem = case_when(
kristina_DO_Q3.2==2 ~ "1",
kristina_DO_Q3.3==2 ~ "0",
kristina_DO_Q3.4==2 ~ "1",
kristina_DO_Q3.5==2 ~ "0",
kristina_DO_Q3.6==2 ~ "1",
kristina_DO_Q3.7==2 ~ "0",
kristina_DO_Q3.8==2 ~ "1",
kristina_DO_Q3.9==2 ~ "0",
))
#Recoding the treatment groups as in-group for political and gender comparisons
experimental_data_1 <- experimental_data_1 %>%
mutate(treatment_rep = case_when(
kristina_DO_Q3.2==2 ~ "0",
kristina_DO_Q3.3==2 ~ "1",
kristina_DO_Q3.4==2 ~ "0",
kristina_DO_Q3.5==2 ~ "1",
kristina_DO_Q3.6==2 ~ "0",
kristina_DO_Q3.7==2 ~ "1",
kristina_DO_Q3.8==2 ~ "0",
kristina_DO_Q3.9==2 ~ "1",
))
#Recoding the treatment groups as in-group for political and gender comparisons
experimental_data_1 <- experimental_data_1 %>%
mutate(treatment_female = case_when(
kristina_DO_Q3.2==2 ~ "1",
kristina_DO_Q3.3==2 ~ "1",
kristina_DO_Q3.4==2 ~ "1",
kristina_DO_Q3.5==2 ~ "1",
kristina_DO_Q3.6==2 ~ "0",
kristina_DO_Q3.7==2 ~ "0",
kristina_DO_Q3.8==2 ~ "0",
kristina_DO_Q3.9==2 ~ "0",
))
#Recoding the treatment groups as in-group for political and gender comparisons
experimental_data_1 <- experimental_data_1 %>%
mutate(treatment_male = case_when(
kristina_DO_Q3.2==2 ~ "0",
kristina_DO_Q3.3==2 ~ "0",
kristina_DO_Q3.4==2 ~ "0",
kristina_DO_Q3.5==2 ~ "0",
kristina_DO_Q3.6==2 ~ "1",
kristina_DO_Q3.7==2 ~ "1",
kristina_DO_Q3.8==2 ~ "1",
kristina_DO_Q3.9==2 ~ "1",
))
#Recoding the treatment groups as in-group for veteran vs. civilian comparisons
experimental_data_1 <- experimental_data_1 %>%
mutate(treatment_veteran = case_when(
kristina_DO_Q3.2==2 ~ "0",
kristina_DO_Q3.3==2 ~ "0",
kristina_DO_Q3.4==2 ~ "1",
kristina_DO_Q3.5==2 ~ "1",
kristina_DO_Q3.6==2 ~ "0",
kristina_DO_Q3.7==2 ~ "0",
kristina_DO_Q3.8==2 ~ "1",
kristina_DO_Q3.9==2 ~ "1",
))
#Recoding the treatment groups as in-group for veteran vs. civilian comparisons
experimental_data_1 <- experimental_data_1 %>%
mutate(treatment_civilian = case_when(
kristina_DO_Q3.2==2 ~ "1",
kristina_DO_Q3.3==2 ~ "1",
kristina_DO_Q3.4==2 ~ "0",
kristina_DO_Q3.5==2 ~ "0",
kristina_DO_Q3.6==2 ~ "1",
kristina_DO_Q3.7==2 ~ "1",
kristina_DO_Q3.8==2 ~ "0",
kristina_DO_Q3.9==2 ~ "0",
))
#Transforming the Likert scale to foreignpolicy variable increasing in value
experimental_data_1 <- experimental_data_1 %>%
mutate(foreignpolicy = case_when(
Q3.11_1==1 ~ "5",
Q3.11_1==2 ~ "4",
Q3.11_1==3 ~ "3",
Q3.11_1==4 ~ "2",
Q3.11_1==5 ~ "1",
))
#Transforming the Likert scale to terrorism variable increasing in value
experimental_data_1 <- experimental_data_1 %>%
mutate(terrorism = case_when(
Q3.11_2==1 ~ "5",
Q3.11_2==2 ~ "4",
Q3.11_2==3 ~ "3",
Q3.11_2==4 ~ "2",
Q3.11_2==5 ~ "1",
))
#Transforming the Likert scale to economy variable increasing in value
experimental_data_1 <- experimental_data_1 %>%
mutate(economy = case_when(
Q3.11_2==1 ~ "5",
Q3.11_2==2 ~ "4",
Q3.11_2==3 ~ "3",
Q3.11_2==4 ~ "2",
Q3.11_2==5 ~ "1",
))
#Transforming the Likert scale to healthcare variable increasing in value
experimental_data_1 <- experimental_data_1 %>%
mutate(healthcare = case_when(
Q3.11_2==1 ~ "5",
Q3.11_2==2 ~ "4",
Q3.11_2==3 ~ "3",
Q3.11_2==4 ~ "2",
Q3.11_2==5 ~ "1",
))
#Transforming the Likert scale to education variable increasing in value
experimental_data_1 <- experimental_data_1 %>%
mutate(education = case_when(
Q3.11_2==1 ~ "5",
Q3.11_2==2 ~ "4",
Q3.11_2==3 ~ "3",
Q3.11_2==4 ~ "2",
Q3.11_2==5 ~ "1",
))
experimental_data_1$foreignpolicy <- as.numeric(experimental_data_1$Q3.11_1)
#Create variable for evaluation of foreign policy competence
experimental_data_1$terrorism <- as.numeric(experimental_data_1$Q3.11_2)
#Create variable for evaluation of terrorism competence
experimental_data_1$economy <- as.numeric(experimental_data_1$Q3.11_3)
#Create variable for evaluation of economy competence
experimental_data_1$healthcare <- as.numeric(experimental_data_1$Q3.11_4)
#Create variable for evaluation of healthcare competence
experimental_data_1$education <- as.numeric(experimental_data_1$Q3.11_5)
#Create variable for evaluation of education competence
experimental_data_1$feeling <- as.numeric(experimental_data_1$Q3.16_4)
#Creates variable for 100 point feeling thermometer for candidate
experimental_data_1$support <- as.numeric(experimental_data_1$Q3.17_14)
#Creates variable for 100 point support scale for candidate
experimental_data_1$weight<-1 #Creates a weight for use in later table creation Graphing results of support for education by party support
#Graph results on Education by party of treatment candidate
#Transforming the group to by-party variables
experimental_data_1 <- experimental_data_1 %>%
mutate(group_by_party = case_when(
group=="F,C,D" ~ "Dem_Candidate",
group=="F,C,R" ~ "Rep_Candidate",
group=="F,V,D" ~ "Dem_Candidate",
group=="F,V,R" ~ "Rep_Candidate",
group=="M,C,D" ~ "Dem_Candidate",
group=="M,C,R" ~ "Rep_Candidate",
group=="M,V,D" ~ "Dem_Candidate",
group=="M,V,R" ~ "Rep_Candidate"
))#I will create a clean data frame where I can collect my variables of interest
becvar_data <- experimental_data_1[84:116]
#Check the class/attributes of created variables
class(becvar_data$group_number) #variable type group number, which indicates character.## [1] "character"#Change variables to numeric
becvar_data$group_number <- as.numeric(becvar_data$group_number)
becvar_data$treatment_dem <- as.numeric(becvar_data$treatment_dem)
becvar_data$treatment_rep <- as.numeric(becvar_data$treatment_rep)
becvar_data$treatment_female <- as.numeric(becvar_data$treatment_female)
becvar_data$treatment_male <- as.numeric(becvar_data$treatment_male)
becvar_data$treatment_veteran <- as.numeric(becvar_data$treatment_veteran)
becvar_data$treatment_civilian <- as.numeric(becvar_data$treatment_civilian)
write_csv(becvar_data, file = "becvar_data.csv")Descriptive Data
Looking at the central tendencies of the numerical data
#Descriptive statistics allow us to understand our data, which informs our analyses and how we display our results
becvar_data %>% #This code allows you to summarize a subset of your variables
get_summary_stats(
female, male, democrat, republican, # columns to calculate for
type = "common") ## # A tibble: 4 × 10
## variable n min max median iqr mean sd se ci
## <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 female 276 0 1 1 1 0.562 0.497 0.03 0.059
## 2 male 276 0 1 0 1 0.438 0.497 0.03 0.059
## 3 democrat 276 0 1 0.5 1 0.5 0.501 0.03 0.059
## 4 republican 276 0 1 0.5 1 0.5 0.501 0.03 0.059freq(becvar_data$group) #This will be our primary IV in this example analysis 
## becvar_data$group
## Frequency Percent
## F,C,D 34 12.32
## F,C,R 35 12.68
## F,V,D 35 12.68
## F,V,R 36 13.04
## M,C,D 34 12.32
## M,C,R 35 12.68
## M,V,D 32 11.59
## M,V,R 35 12.68
## Total 276 100.00freq(becvar_data$gender) #gender
## becvar_data$gender
## Frequency Percent
## 1 121 43.84
## 2 155 56.16
## Total 276 100.00freq(becvar_data$party) #party
## becvar_data$party
## Frequency Percent
## Democrat 138 50
## Republican 138 50
## Total 276 100freq(becvar_data$education) #education
## becvar_data$education
## Frequency Percent
## 1 87 31.522
## 2 111 40.217
## 3 43 15.580
## 4 20 7.246
## 5 15 5.435
## Total 276 100.000freq(becvar_data$age) #age
## becvar_data$age
## Frequency Percent Valid Percent
## 2 1 0.3623 0.3636
## 4 3 1.0870 1.0909
## 5 6 2.1739 2.1818
## 6 3 1.0870 1.0909
## 7 4 1.4493 1.4545
## 8 11 3.9855 4.0000
## 9 5 1.8116 1.8182
## 10 6 2.1739 2.1818
## 12 16 5.7971 5.8182
## 13 41 14.8551 14.9091
## 14 5 1.8116 1.8182
## 15 48 17.3913 17.4545
## 16 23 8.3333 8.3636
## 17 11 3.9855 4.0000
## 18 28 10.1449 10.1818
## 19 6 2.1739 2.1818
## 20 8 2.8986 2.9091
## 21 3 1.0870 1.0909
## 22 1 0.3623 0.3636
## 23 7 2.5362 2.5455
## 24 3 1.0870 1.0909
## 25 3 1.0870 1.0909
## 26 3 1.0870 1.0909
## 27 4 1.4493 1.4545
## 28 2 0.7246 0.7273
## 29 3 1.0870 1.0909
## 30 1 0.3623 0.3636
## 31 2 0.7246 0.7273
## 33 3 1.0870 1.0909
## 34 2 0.7246 0.7273
## 35 1 0.3623 0.3636
## 38 1 0.3623 0.3636
## 42 5 1.8116 1.8182
## 43 2 0.7246 0.7273
## 45 1 0.3623 0.3636
## 47 1 0.3623 0.3636
## 49 1 0.3623 0.3636
## 52 1 0.3623 0.3636
## NA's 1 0.3623
## Total 276 100.0000 100.0000##Balance Tests
Conducting balance tests to ensure that our randomization worked. There is a slight indication that the randomization underexposed the treatment for the male/veteran/democratic candidate. This may be because of the one opt-out of the survey, or the elimination of the 6 respondents that did not choose a party lean.
#Balance tests with factor/ordinal variables
#Calculates chi-square test statistic for unbalance across experimental conditions
chisq.test(table(becvar_data$group, becvar_data$female))##
## Pearson's Chi-squared test
##
## data: table(becvar_data$group, becvar_data$female)
## X-squared = 5.7865, df = 7, p-value = 0.5649chisq.test(table(becvar_data$group, becvar_data$democrat))##
## Pearson's Chi-squared test
##
## data: table(becvar_data$group, becvar_data$democrat)
## X-squared = 8.8336, df = 7, p-value = 0.2648#Visual inspection of gender and party id variables
freq(becvar_data$democrat)
## becvar_data$democrat
## Frequency Percent
## 0 138 50
## 1 138 50
## Total 276 100freq(becvar_data$female)
## becvar_data$female
## Frequency Percent
## 0 121 43.84
## 1 155 56.16
## Total 276 100.00##CrossTable code from 'gmodels' package
#Calculates both chi-square tests as well as percentages per cell
CrossTable(becvar_data$female, becvar_data$group, expected = FALSE, chisq=TRUE, prop.c=TRUE, prop.r=FALSE, prop.t=FALSE, prop.chisq = FALSE)##
##
## Cell Contents
## |-------------------------|
## | N |
## | N / Col Total |
## |-------------------------|
##
##
## Total Observations in Table: 276
##
##
## | becvar_data$group
## becvar_data$female | F,C,D | F,C,R | F,V,D | F,V,R | M,C,D | M,C,R | M,V,D | M,V,R | Row Total |
## -------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
## 0 | 14 | 14 | 11 | 20 | 14 | 15 | 17 | 16 | 121 |
## | 0.412 | 0.400 | 0.314 | 0.556 | 0.412 | 0.429 | 0.531 | 0.457 | |
## -------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
## 1 | 20 | 21 | 24 | 16 | 20 | 20 | 15 | 19 | 155 |
## | 0.588 | 0.600 | 0.686 | 0.444 | 0.588 | 0.571 | 0.469 | 0.543 | |
## -------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
## Column Total | 34 | 35 | 35 | 36 | 34 | 35 | 32 | 35 | 276 |
## | 0.123 | 0.127 | 0.127 | 0.130 | 0.123 | 0.127 | 0.116 | 0.127 | |
## -------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
##
##
## Statistics for All Table Factors
##
##
## Pearson's Chi-squared test
## ------------------------------------------------------------
## Chi^2 = 5.786498 d.f. = 7 p = 0.5648885
##
##
## CrossTable(becvar_data$democrat, becvar_data$group, expected = FALSE, chisq=TRUE, prop.c=TRUE, prop.r=FALSE, prop.t=FALSE, prop.chisq = FALSE)##
##
## Cell Contents
## |-------------------------|
## | N |
## | N / Col Total |
## |-------------------------|
##
##
## Total Observations in Table: 276
##
##
## | becvar_data$group
## becvar_data$democrat | F,C,D | F,C,R | F,V,D | F,V,R | M,C,D | M,C,R | M,V,D | M,V,R | Row Total |
## ---------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
## 0 | 14 | 14 | 21 | 18 | 18 | 14 | 16 | 23 | 138 |
## | 0.412 | 0.400 | 0.600 | 0.500 | 0.529 | 0.400 | 0.500 | 0.657 | |
## ---------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
## 1 | 20 | 21 | 14 | 18 | 16 | 21 | 16 | 12 | 138 |
## | 0.588 | 0.600 | 0.400 | 0.500 | 0.471 | 0.600 | 0.500 | 0.343 | |
## ---------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
## Column Total | 34 | 35 | 35 | 36 | 34 | 35 | 32 | 35 | 276 |
## | 0.123 | 0.127 | 0.127 | 0.130 | 0.123 | 0.127 | 0.116 | 0.127 | |
## ---------------------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|-----------|
##
##
## Statistics for All Table Factors
##
##
## Pearson's Chi-squared test
## ------------------------------------------------------------
## Chi^2 = 8.833613 d.f. = 7 p = 0.2648238
##
##
## ##Average Treatment Effect Analysis
Now that we have conducted balance tests and feel confident that the randomization worked as expected, we can start to analyze the average treatment effect (ATE) of exposure to a candidate profile across 8 conditions varying gender, veteran status, and political party of candidate.
All we are looking for is significant differences between the experimental groups on whatever dependent variable of interest we are examining. We do not need control variables with a strictly randomized experimental design.
My DVs are ordinal (with 5 scale points), so I can calculate the ATE using any technique designed for a numerical DV.
To test my established hypotheses, I’ll need to calculate the average treatment effect for exposure to the different candidate profiles on the evaluation of the candidate’s ability to handle foreign policy and terrorism as a member of congress.
#Visual examination of the IV and DV
freq(becvar_data$group)
## becvar_data$group
## Frequency Percent
## F,C,D 34 12.32
## F,C,R 35 12.68
## F,V,D 35 12.68
## F,V,R 36 13.04
## M,C,D 34 12.32
## M,C,R 35 12.68
## M,V,D 32 11.59
## M,V,R 35 12.68
## Total 276 100.00freq(becvar_data$foreignpolicy)
## becvar_data$foreignpolicy
## Frequency Percent
## 1 58 21.014
## 2 127 46.014
## 3 55 19.928
## 4 25 9.058
## 5 11 3.986
## Total 276 100.000freq(becvar_data$terrorism)
## becvar_data$terrorism
## Frequency Percent
## 1 47 17.03
## 2 103 37.32
## 3 58 21.01
## 4 40 14.49
## 5 28 10.14
## Total 276 100.00freq(becvar_data$support)
## becvar_data$support
## Frequency Percent
## 0 2 0.7246
## 18 2 0.7246
## 25 1 0.3623
## 28 2 0.7246
## 30 1 0.3623
## 32 2 0.7246
## 35 1 0.3623
## 38 1 0.3623
## 39 1 0.3623
## 40 3 1.0870
## 41 1 0.3623
## 43 1 0.3623
## 44 2 0.7246
## 46 2 0.7246
## 47 1 0.3623
## 49 1 0.3623
## 50 2 0.7246
## 51 1 0.3623
## 52 1 0.3623
## 53 1 0.3623
## 54 1 0.3623
## 55 1 0.3623
## 56 2 0.7246
## 57 3 1.0870
## 58 1 0.3623
## 59 2 0.7246
## 60 1 0.3623
## 61 3 1.0870
## 62 1 0.3623
## 63 3 1.0870
## 64 3 1.0870
## 65 9 3.2609
## 66 4 1.4493
## 67 6 2.1739
## 68 5 1.8116
## 69 1 0.3623
## 70 11 3.9855
## 71 7 2.5362
## 72 8 2.8986
## 73 9 3.2609
## 74 9 3.2609
## 75 9 3.2609
## 76 7 2.5362
## 77 3 1.0870
## 78 5 1.8116
## 79 3 1.0870
## 80 10 3.6232
## 81 4 1.4493
## 82 10 3.6232
## 83 7 2.5362
## 84 10 3.6232
## 85 9 3.2609
## 86 9 3.2609
## 87 3 1.0870
## 88 9 3.2609
## 89 2 0.7246
## 90 7 2.5362
## 91 4 1.4493
## 92 4 1.4493
## 93 6 2.1739
## 94 3 1.0870
## 95 3 1.0870
## 96 1 0.3623
## 97 3 1.0870
## 98 1 0.3623
## 99 2 0.7246
## 100 23 8.3333
## Total 276 100.0000freq(becvar_data$feeling)
## becvar_data$feeling
## Frequency Percent
## 14 1 0.3623
## 18 1 0.3623
## 22 1 0.3623
## 30 2 0.7246
## 32 2 0.7246
## 35 1 0.3623
## 37 2 0.7246
## 38 1 0.3623
## 39 1 0.3623
## 40 2 0.7246
## 42 1 0.3623
## 43 2 0.7246
## 44 1 0.3623
## 45 1 0.3623
## 47 2 0.7246
## 48 1 0.3623
## 49 3 1.0870
## 50 3 1.0870
## 51 3 1.0870
## 53 3 1.0870
## 54 1 0.3623
## 55 9 3.2609
## 56 3 1.0870
## 57 1 0.3623
## 58 1 0.3623
## 59 3 1.0870
## 60 4 1.4493
## 61 3 1.0870
## 62 2 0.7246
## 63 5 1.8116
## 64 2 0.7246
## 65 12 4.3478
## 66 2 0.7246
## 67 4 1.4493
## 68 4 1.4493
## 69 5 1.8116
## 70 7 2.5362
## 71 8 2.8986
## 72 8 2.8986
## 73 3 1.0870
## 74 10 3.6232
## 75 5 1.8116
## 76 6 2.1739
## 77 4 1.4493
## 78 8 2.8986
## 79 7 2.5362
## 80 12 4.3478
## 81 8 2.8986
## 82 5 1.8116
## 83 7 2.5362
## 84 1 0.3623
## 85 6 2.1739
## 86 8 2.8986
## 87 5 1.8116
## 88 7 2.5362
## 89 2 0.7246
## 90 6 2.1739
## 91 7 2.5362
## 92 5 1.8116
## 94 2 0.7246
## 95 4 1.4493
## 96 2 0.7246
## 97 1 0.3623
## 98 2 0.7246
## 99 3 1.0870
## 100 22 7.9710
## Total 276 100.0000#Descriptive statistical examination of the foreignpolicy DV
sd(becvar_data$foreignpolicy)## [1] 1.024804mean(becvar_data$foreignpolicy)## [1] 2.289855#Descriptive statistical examination of the terrorism DV
sd(becvar_data$terrorism)## [1] 1.215117mean(becvar_data$terrorism)## [1] 2.634058For the foreign policy and terrorism variables, we see that for both evaluations, the most common choice was “2”, or “Qualified”. The ratings range from “5” (Highly Qualified) to “1” (Not at all Qualified).
Graphing the Average Treatment Evaluation by Party
####Graphical Approach to ATE - Get Average Value by Treatment Assignment
####Then graph the average value with Confidence Intervals
std.error <- function(x) sd(x)/sqrt(length(x))
graph_fp_data <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean_fp = mean(foreignpolicy),
sd_fp = sd(foreignpolicy),
n = as.numeric(n()),
se_fp = std.error(foreignpolicy),
ub_fp = mean_fp+(1.96*se_fp),
lb_fp = mean_fp-(1.96*se_fp))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.graph_ter_data <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean_ter = mean(terrorism),
sd_ter = sd(terrorism),
n = as.numeric(n()),
se_ter = std.error(terrorism),
ub_ter = mean_ter+(1.96*se_ter),
lb_ter = mean_ter-(1.96*se_ter))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_fp_data, aes(x=group, y=mean_fp, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb_fp, ymax=ub_fp),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Foreign Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 1## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.
ggplot(graph_ter_data, aes(x=group, y=mean_ter, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb_ter, ymax=ub_ter),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Terrorism") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every .5
Graphing the policy ownership issues by evaluation of groups only (non-DV outcome variables)
####Graphical Approach to ATE - Get Average Value by Treatment Assignment
####Then graph the average value with Confidence Intervals
std.error <- function(x) sd(x)/sqrt(length(x))
graph_education_data <- becvar_data %>%
group_by(group) %>%
dplyr::summarise(
mean = mean(education),
sd = sd(education),
n = as.numeric(n()),
se = std.error(education),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
graph_economy_data <- becvar_data %>%
group_by(group) %>%
dplyr::summarise(
mean = mean(economy),
sd = sd(economy),
n = as.numeric(n()),
se = std.error(economy),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
graph_healthcare_data <- becvar_data %>%
group_by(group) %>%
dplyr::summarise(
mean = mean(healthcare),
sd = sd(healthcare),
n = as.numeric(n()),
se = std.error(healthcare),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
ggplot(graph_education_data, aes(x=group, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Education Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_economy_data, aes(x=group, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Economic Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_healthcare_data, aes(x=group, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Healthcare Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
Graphing the Average Treatment Evaluation by Party by Candidate Party
####Graphical Approach to ATE - Get Average Value by Treatment Assignment
####Then graph the average value with Confidence Intervals
std.error <- function(x) sd(x)/sqrt(length(x))
graph_fp_data2 <- becvar_data %>%
group_by(group_by_party) %>%
dplyr::summarise(
mean = mean(education),
sd = sd(education),
n = as.numeric(n()),
se = std.error(education),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
graph_terr_data2 <- becvar_data %>%
group_by(group_by_party) %>%
dplyr::summarise(
mean = mean(education),
sd = sd(education),
n = as.numeric(n()),
se = std.error(education),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
graph_education_data2 <- becvar_data %>%
group_by(group_by_party) %>%
dplyr::summarise(
mean = mean(education),
sd = sd(education),
n = as.numeric(n()),
se = std.error(education),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
graph_economy_data2 <- becvar_data %>%
group_by(group_by_party) %>%
dplyr::summarise(
mean = mean(economy),
sd = sd(economy),
n = as.numeric(n()),
se = std.error(economy),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
graph_healthcare_data2 <- becvar_data %>%
group_by(group_by_party) %>%
dplyr::summarise(
mean = mean(healthcare),
sd = sd(healthcare),
n = as.numeric(n()),
se = std.error(healthcare),
ub = mean+(1.96*se),
lb = mean-(1.96*se))
ggplot(graph_fp_data2, aes(x=group_by_party, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Foreign Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_terr_data2, aes(x=group_by_party, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Terrorism Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_education_data2, aes(x=group_by_party, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Education Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_economy_data2, aes(x=group_by_party, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Economic Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_healthcare_data2, aes(x=group_by_party, y=mean)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Healthcare Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
graph_education_data2## # A tibble: 2 × 7
## group_by_party mean sd n se ub lb
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 Dem_Candidate 2.22 1.16 135 0.100 2.42 2.03
## 2 Rep_Candidate 2.08 1.06 141 0.0889 2.25 1.90Graphing the policy ownership issues by evaluation of groups by party of evaluator (non-DV outcome variables)
####Graphical Approach to ATE - Get Average Value by Treatment Assignment
####Then graph the average value with Confidence Intervals
std.error <- function(x) sd(x)/sqrt(length(x))
graph_education_data2 <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean = mean(education),
sd = sd(education),
n = as.numeric(n()),
se = std.error(education),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.graph_economy_data2 <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean = mean(economy),
sd = sd(economy),
n = as.numeric(n()),
se = std.error(economy),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.graph_healthcare_data2 <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean = mean(healthcare),
sd = sd(healthcare),
n = as.numeric(n()),
se = std.error(healthcare),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_education_data2, aes(x=group, y=mean, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Education Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_economy_data2, aes(x=group, y=mean, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Economic Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_healthcare_data2, aes(x=group, y=mean, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Healthcare Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
Graphing the Average Treatment Evaluation by Gender
####Graphical Approach to ATE - Get Average Value by Treatment Assignment
####Then graph the average value with Confidence Intervals
std.error <- function(x) sd(x)/sqrt(length(x))
graph_fp_data2 <- becvar_data %>%
group_by(group, Gender) %>%
dplyr::summarise(
mean_fp = mean(foreignpolicy),
sd_fp = sd(foreignpolicy),
n = as.numeric(n()),
se_fp = std.error(foreignpolicy),
ub_fp = mean_fp+(1.96*se_fp),
lb_fp = mean_fp-(1.96*se_fp))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.graph_ter_data2 <- becvar_data %>%
group_by(group, Gender) %>%
dplyr::summarise(
mean_ter = mean(terrorism),
sd_ter = sd(terrorism),
n = as.numeric(n()),
se_ter = std.error(terrorism),
ub_ter = mean_ter+(1.96*se_ter),
lb_ter = mean_ter-(1.96*se_ter))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_fp_data2, aes(x=group, y=mean_fp, fill=Gender)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb_fp, ymax=ub_fp),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Foreign Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
ggplot(graph_ter_data2, aes(x=group, y=mean_ter, fill=Gender)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb_ter, ymax=ub_ter),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Terrorism") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
Starting to look specifically at the effects of individual treatments on policy evaluations
#Calculate the average candidate evaluation score by experimental group per hypothesis
#H1: Military experience <> stronger evaluation on foreign policy and terrorism
#H2: Male candidates <> stronger evaluation on foreign policy and terrorism
#Uses linear regression to evaluate "foreign policy" variable by treatment group
lm_1<-lm(foreignpolicy ~ group, data=becvar_data)
summary(lm_1)##
## Call:
## lm(formula = foreignpolicy ~ group, data = becvar_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.7647 -0.7647 -0.1143 0.6667 2.9143
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.5000 0.1738 14.383 <2e-16 ***
## groupF,C,R -0.3857 0.2440 -1.581 0.1152
## groupF,V,D -0.3286 0.2440 -1.346 0.1793
## groupF,V,R -0.1667 0.2424 -0.688 0.4923
## groupM,C,D 0.2647 0.2458 1.077 0.2825
## groupM,C,R -0.4143 0.2440 -1.698 0.0907 .
## groupM,V,D -0.4062 0.2496 -1.627 0.1048
## groupM,V,R -0.2429 0.2440 -0.995 0.3206
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.013 on 268 degrees of freedom
## Multiple R-squared: 0.04686, Adjusted R-squared: 0.02196
## F-statistic: 1.882 on 7 and 268 DF, p-value: 0.0726#Uses linear regression to evaluate "terrorism" variable by treatment group
lm_2<-lm(terrorism ~ group, data=becvar_data)
summary(lm_2)##
## Call:
## lm(formula = terrorism ~ group, data = becvar_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2.0000 -0.8823 -0.2188 0.7812 2.7812
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.85294 0.20655 13.812 <2e-16 ***
## groupF,C,R -0.48151 0.29001 -1.660 0.0980 .
## groupF,V,D 0.14706 0.29001 0.507 0.6125
## groupF,V,R -0.18627 0.28802 -0.647 0.5184
## groupM,C,D 0.02941 0.29211 0.101 0.9199
## groupM,C,R -0.28151 0.29001 -0.971 0.3326
## groupM,V,D -0.63419 0.29664 -2.138 0.0334 *
## groupM,V,R -0.36723 0.29001 -1.266 0.2065
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.204 on 268 degrees of freedom
## Multiple R-squared: 0.04259, Adjusted R-squared: 0.01758
## F-statistic: 1.703 on 7 and 268 DF, p-value: 0.1083#H3: Ingroup partisans <> stronger evaluation on foreign policy and terrorism
#Uses linear regression to evaluate "foreign policy" evaluated by in-party treatment group and out-party treatment group
lm_3<-lm(foreignpolicy ~ democrat, data=becvar_data)
summary(lm_3)##
## Call:
## lm(formula = foreignpolicy ~ democrat, data = becvar_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.3551 -0.3551 -0.2246 0.6449 2.7754
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.22464 0.08722 25.507 <2e-16 ***
## democrat 0.13043 0.12335 1.057 0.291
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.025 on 274 degrees of freedom
## Multiple R-squared: 0.004065, Adjusted R-squared: 0.0004298
## F-statistic: 1.118 on 1 and 274 DF, p-value: 0.2912#Uses linear regression to evaluate "terrorism" evaluated by party
lm_4<-lm(terrorism ~ democrat, data=becvar_data)
summary(lm_4)##
## Call:
## lm(formula = terrorism ~ democrat, data = becvar_data)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.6812 -0.6812 -0.5870 0.4130 2.4130
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.5870 0.1036 24.983 <2e-16 ***
## democrat 0.0942 0.1464 0.643 0.521
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.216 on 274 degrees of freedom
## Multiple R-squared: 0.001508, Adjusted R-squared: -0.002136
## F-statistic: 0.4138 on 1 and 274 DF, p-value: 0.5206#Use tidy broom package
tidy_lm1 <- tidy(lm_1, conf.int = TRUE)
tidy_lm2 <- tidy(lm_2, conf.int = TRUE)
tidy_lm3 <- tidy(lm_3, conf.int = TRUE)
tidy_lm4 <- tidy(lm_4, conf.int = TRUE)
options(scipen = 999)
summary(tidy_lm1)## term estimate std.error statistic
## Length:8 Min. :-0.41429 Min. :0.1738 Min. :-1.6976
## Class :character 1st Qu.:-0.39085 1st Qu.:0.2436 1st Qu.:-1.5922
## Mode :character Median :-0.28571 Median :0.2440 Median :-1.1707
## Mean : 0.10255 Mean :0.2360 Mean : 0.9407
## 3rd Qu.:-0.05882 3rd Qu.:0.2445 3rd Qu.:-0.2465
## Max. : 2.50000 Max. :0.2496 Max. :14.3834
## p.value conf.low conf.high
## Min. :0.0000 Min. :-0.8977 Min. :0.06620
## 1st Qu.:0.1013 1st Qu.:-0.8733 1st Qu.:0.09238
## Median :0.1472 Median :-0.7662 Median :0.19478
## Mean :0.1982 Mean :-0.3621 Mean :0.56714
## 3rd Qu.:0.2920 3rd Qu.:-0.5377 3rd Qu.:0.42006
## Max. :0.4923 Max. : 2.1578 Max. :2.84221summary(tidy_lm2)## term estimate std.error statistic
## Length:8 Min. :-0.63419 Min. :0.2066 Min. :-2.1379
## Class :character 1st Qu.:-0.39580 1st Qu.:0.2895 1st Qu.:-1.3648
## Mode :character Median :-0.23389 Median :0.2900 Median :-0.8087
## Mean : 0.13484 Mean :0.2804 Mean : 0.9673
## 3rd Qu.: 0.05882 3rd Qu.:0.2905 3rd Qu.: 0.2023
## Max. : 2.85294 Max. :0.2966 Max. :13.8123
## p.value conf.low conf.high
## Min. :0.00000 Min. :-1.2182 Min. :-0.05016
## 1st Qu.:0.08187 1st Qu.:-0.9668 1st Qu.: 0.17520
## Median :0.26955 Median :-0.8029 Median : 0.33514
## Mean :0.34016 Mean :-0.4173 Mean : 0.68695
## 3rd Qu.:0.54190 3rd Qu.:-0.5153 3rd Qu.: 0.63291
## Max. :0.91987 Max. : 2.4463 Max. : 3.25961summary(tidy_lm3)## term estimate std.error statistic
## Length:2 Min. :0.1304 Min. :0.08722 Min. : 1.057
## Class :character 1st Qu.:0.6540 1st Qu.:0.09625 1st Qu.: 7.170
## Mode :character Median :1.1775 Median :0.10528 Median :13.282
## Mean :1.1775 Mean :0.10528 Mean :13.282
## 3rd Qu.:1.7011 3rd Qu.:0.11431 3rd Qu.:19.394
## Max. :2.2246 Max. :0.12335 Max. :25.507
## p.value conf.low conf.high
## Min. :0.00000 Min. :-0.1124 Min. :0.3733
## 1st Qu.:0.07281 1st Qu.: 0.4289 1st Qu.:0.8790
## Median :0.14561 Median : 0.9703 Median :1.3848
## Mean :0.14561 Mean : 0.9703 Mean :1.3848
## 3rd Qu.:0.21842 3rd Qu.: 1.5116 3rd Qu.:1.8906
## Max. :0.29123 Max. : 2.0529 Max. :2.3963summary(tidy_lm4)## term estimate std.error statistic
## Length:2 Min. :0.0942 Min. :0.1035 Min. : 0.6433
## Class :character 1st Qu.:0.7174 1st Qu.:0.1143 1st Qu.: 6.7283
## Mode :character Median :1.3406 Median :0.1250 Median :12.8132
## Mean :1.3406 Mean :0.1250 Mean :12.8132
## 3rd Qu.:1.9638 3rd Qu.:0.1357 3rd Qu.:18.8982
## Max. :2.5870 Max. :0.1464 Max. :24.9832
## p.value conf.low conf.high
## Min. :0.0000 Min. :-0.1941 Min. :0.3825
## 1st Qu.:0.1301 1st Qu.: 0.4502 1st Qu.:0.9846
## Median :0.2603 Median : 1.0945 Median :1.5866
## Mean :0.2603 Mean : 1.0945 Mean :1.5866
## 3rd Qu.:0.3904 3rd Qu.: 1.7388 3rd Qu.:2.1887
## Max. :0.5206 Max. : 2.3831 Max. :2.7908aug_lm1 <- augment(lm_1, se_fit = TRUE)
aug_lm2 <- augment(lm_2, se_fit = TRUE)
graph_lm1 <- aug_lm1 %>%
group_by(group) %>%
summarise_at(vars(.fitted),
list(mean_fp = mean))
graph_lm2 <- aug_lm2 %>%
group_by(group) %>%
summarise_at(vars(.fitted),
list(mean_ter = mean))
graph_lm1## # A tibble: 8 × 2
## group mean_fp
## <chr> <dbl>
## 1 F,C,D 2.50
## 2 F,C,R 2.11
## 3 F,V,D 2.17
## 4 F,V,R 2.33
## 5 M,C,D 2.76
## 6 M,C,R 2.09
## 7 M,V,D 2.09
## 8 M,V,R 2.26aug_lm1## # A tibble: 276 × 9
## foreignpolicy group .fitted .se.fit .resid .hat .sigma .cooksd .std.resid
## <dbl> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 1 F,V,R 2.33 0.169 -1.33 0.0278 1.01 6.36e-3 -1.33
## 2 5 M,V,R 2.26 0.171 2.74 0.0286 1.00 2.77e-2 2.75
## 3 2 F,C,R 2.11 0.171 -0.114 0.0286 1.02 4.81e-5 -0.114
## 4 2 M,V,R 2.26 0.171 -0.257 0.0286 1.02 2.44e-4 -0.257
## 5 3 F,C,D 2.50 0.174 0.500 0.0294 1.01 9.50e-4 0.501
## 6 2 M,C,R 2.09 0.171 -0.0857 0.0286 1.02 2.71e-5 -0.0858
## 7 1 F,C,R 2.11 0.171 -1.11 0.0286 1.01 4.57e-3 -1.12
## 8 3 F,V,D 2.17 0.171 0.829 0.0286 1.01 2.53e-3 0.829
## 9 4 F,C,D 2.50 0.174 1.50 0.0294 1.01 8.55e-3 1.50
## 10 2 M,C,R 2.09 0.171 -0.0857 0.0286 1.02 2.71e-5 -0.0858
## # ℹ 266 more rowsNow I’m revisiting the outcome variables of the feeling thermometer and the support thermometer.
####Graphical Approach to ATE - Get Average Value by Treatment Assignment
####Then graph the average value with Confidence Intervals
std.error <- function(x) sd(x)/sqrt(length(x))
graph_feeling_data <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean_feeling = mean(feeling),
sd_feeling = sd(feeling),
n = as.numeric(n()),
se_feeling = std.error(feeling),
ub_feeling = mean_feeling+(1.96*se_feeling),
lb_feeling = mean_feeling-(1.96*se_feeling))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.graph_support_data <- becvar_data %>%
group_by(group, party) %>%
dplyr::summarise(
mean_support = mean(support),
sd_support = sd(support),
n = as.numeric(n()),
se_support = std.error(support),
ub_support = mean_support+(1.96*se_support),
lb_support = mean_support-(1.96*se_support))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_feeling_data, aes(x=group, y=mean_feeling, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb_feeling, ymax=ub_feeling),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Warmth Toward Candidate") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 100, 10)) # Ticks from 0-5, every 1
ggplot(graph_support_data, aes(x=group, y=mean_support, fill=party)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb_support, ymax=ub_support),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Support for Candidate") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 100, 10)) # Ticks from 0-5, every 1
Breaking this down further using various things…
fp_data_2 <- becvar_data %>%
group_by(group) %>%
get_summary_stats(foreignpolicy, type = "mean_sd")
fp_data_2## # A tibble: 8 × 5
## group variable n mean sd
## <chr> <fct> <dbl> <dbl> <dbl>
## 1 F,C,D foreignpolicy 34 2.5 1.14
## 2 F,C,R foreignpolicy 35 2.11 0.867
## 3 F,V,D foreignpolicy 35 2.17 1.10
## 4 F,V,R foreignpolicy 36 2.33 0.894
## 5 M,C,D foreignpolicy 34 2.76 1.10
## 6 M,C,R foreignpolicy 35 2.09 0.981
## 7 M,V,D foreignpolicy 32 2.09 0.856
## 8 M,V,R foreignpolicy 35 2.26 1.12ter_data_2 <- becvar_data %>%
group_by(group) %>%
get_summary_stats(terrorism, type = "mean_sd")
ter_data_2## # A tibble: 8 × 5
## group variable n mean sd
## <chr> <fct> <dbl> <dbl> <dbl>
## 1 F,C,D terrorism 34 2.85 1.23
## 2 F,C,R terrorism 35 2.37 1.19
## 3 F,V,D terrorism 35 3 1.21
## 4 F,V,R terrorism 36 2.67 1.26
## 5 M,C,D terrorism 34 2.88 1.27
## 6 M,C,R terrorism 35 2.57 1.12
## 7 M,V,D terrorism 32 2.22 1.10
## 8 M,V,R terrorism 35 2.49 1.22becvar_data %>%
group_by(group) %>%
identify_outliers(foreignpolicy)## # A tibble: 25 × 35
## group age polparty leaner party democrat republican gender female male
## <chr> <dbl> <dbl> <dbl> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 F,C,D 18 1 NA Democrat 1 0 2 1 0
## 2 F,C,D 8 1 NA Democrat 1 0 2 1 0
## 3 F,C,D 52 2 NA Republic… 0 1 2 1 0
## 4 F,C,R 18 2 NA Republic… 0 1 1 0 1
## 5 F,C,R 13 2 NA Republic… 0 1 1 0 1
## 6 F,C,R 5 1 NA Democrat 1 0 2 1 0
## 7 F,C,R 13 2 NA Republic… 0 1 2 1 0
## 8 F,C,R 10 1 NA Democrat 1 0 2 1 0
## 9 F,C,R 12 2 NA Republic… 0 1 1 0 1
## 10 F,C,R 26 1 NA Democrat 1 0 2 1 0
## # ℹ 15 more rows
## # ℹ 25 more variables: Gender <chr>, pt_guns <dbl>, pt_climate <dbl>,
## # pt_militarization <dbl>, pt_guns_order <dbl>, pt_climate_order <dbl>,
## # pt_militarization_order <dbl>, group_number <dbl>, treatment_dem <dbl>,
## # treatment_rep <dbl>, treatment_female <dbl>, treatment_male <dbl>,
## # treatment_veteran <dbl>, treatment_civilian <dbl>, foreignpolicy <dbl>,
## # terrorism <dbl>, economy <dbl>, healthcare <dbl>, education <dbl>, …becvar_data %>%
cor_test(foreignpolicy, terrorism)## # A tibble: 1 × 8
## var1 var2 cor statistic p conf.low conf.high method
## <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <chr>
## 1 foreignpolicy terrorism 0.47 8.77 1.98e-16 0.370 0.555 PearsonCreating new variable tables until I figure this out
becvar_data2 <- becvar_data %>%
mutate(Respondents = case_when(
female==1 & democrat==1 ~ "FemaleDemocrat",
female==1 & republican==1 ~ "FemaleRepublican",
male==1 & democrat==1 ~ "MaleDemocrat",
male==1 & republican==1 ~ "MaleRepublican"
))
#Descriptive variables of Respondents
freq(becvar_data2$Respondents)
## becvar_data2$Respondents
## Frequency Percent
## FemaleDemocrat 77 27.90
## FemaleRepublican 78 28.26
## MaleDemocrat 61 22.10
## MaleRepublican 60 21.74
## Total 276 100.00#H3: Ingroup partisans <> stronger evaluation on foreign policy and terrorism
#Uses linear regression to evaluate "foreign policy" evaluated by in-party treatment group and out-party treatment group
lm_5<-lm(foreignpolicy ~ Respondents, data=becvar_data2)
summary(lm_5)##
## Call:
## lm(formula = foreignpolicy ~ Respondents, data = becvar_data2)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.4426 -0.4426 -0.2500 0.7143 2.7949
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.28571 0.11701 19.535 <0.0000000000000002 ***
## RespondentsFemaleRepublican -0.08059 0.16494 -0.489 0.626
## RespondentsMaleDemocrat 0.15691 0.17599 0.892 0.373
## RespondentsMaleRepublican -0.03571 0.17680 -0.202 0.840
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.027 on 272 degrees of freedom
## Multiple R-squared: 0.007203, Adjusted R-squared: -0.003747
## F-statistic: 0.6578 on 3 and 272 DF, p-value: 0.5787#Uses linear regression to evaluate "terrorism" evaluated by in-party treatment group and out-party treatment group
lm_6<-lm(terrorism ~ Respondents, data=becvar_data2)
summary(lm_6)##
## Call:
## lm(formula = terrorism ~ Respondents, data = becvar_data2)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.7377 -0.7133 -0.4333 0.5667 2.5667
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.63636 0.13864 19.016 <0.0000000000000002 ***
## RespondentsFemaleRepublican 0.06876 0.19544 0.352 0.725
## RespondentsMaleDemocrat 0.10134 0.20853 0.486 0.627
## RespondentsMaleRepublican -0.20303 0.20950 -0.969 0.333
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.217 on 272 degrees of freedom
## Multiple R-squared: 0.008539, Adjusted R-squared: -0.002396
## F-statistic: 0.7809 on 3 and 272 DF, p-value: 0.5055Creating Variable for In and Out Group Evaluations paired with appropriate Treatments
becvar_data3 <- becvar_data2 %>%
mutate(PartyGroup = case_when(
group_number==1 & Respondents=="FemaleDemocrat" ~ "InGroup",
group_number==1 & Respondents=="FemaleRepublican" ~ "OutGroup",
group_number==1 & Respondents=="MaleDemocrat" ~ "InGroup",
group_number==1 & Respondents=="MaleRepublican" ~ "OutGroup",
group_number==2 & Respondents=="FemaleDemocrat" ~ "OutGroup",
group_number==2 & Respondents=="FemaleRepublican" ~ "InGroup",
group_number==2 & Respondents=="MaleDemocrat" ~ "OutGroup",
group_number==2 & Respondents=="MaleRepublican" ~ "InGroup",
group_number==3 & Respondents=="FemaleDemocrat" ~ "InGroup",
group_number==3 & Respondents=="FemaleRepublican" ~ "OutGroup",
group_number==3 & Respondents=="MaleDemocrat" ~ "InGroup",
group_number==3 & Respondents=="MaleRepublican" ~ "OutGroup",
group_number==4 & Respondents=="FemaleDemocrat" ~ "OutGroup",
group_number==4 & Respondents=="FemaleRepublican" ~ "InGroup",
group_number==4 & Respondents=="MaleDemocrat" ~ "OutGroup",
group_number==4 & Respondents=="MaleRepublican" ~ "InGroup",
group_number==5 & Respondents=="FemaleDemocrat" ~ "InGroup",
group_number==5 & Respondents=="FemaleRepublican" ~ "OutGroup",
group_number==5 & Respondents=="MaleDemocrat" ~ "InGroup",
group_number==5 & Respondents=="MaleRepublican" ~ "OutGroup",
group_number==6 & Respondents=="FemaleDemocrat" ~ "OutGroup",
group_number==6 & Respondents=="FemaleRepublican" ~ "InGroup",
group_number==6 & Respondents=="MaleDemocrat" ~ "OutGroup",
group_number==6 & Respondents=="MaleRepublican" ~ "InGroup",
group_number==7 & Respondents=="FemaleDemocrat" ~ "InGroup",
group_number==7 & Respondents=="FemaleRepublican" ~ "OutGroup",
group_number==7 & Respondents=="MaleDemocrat" ~ "InGroup",
group_number==7 & Respondents=="MaleRepublican" ~ "OutGroup",
group_number==8 & Respondents=="FemaleDemocrat" ~ "OutGroup",
group_number==8 & Respondents=="FemaleRepublican" ~ "InGroup",
group_number==8 & Respondents=="MaleDemocrat" ~ "OutGroup",
group_number==8 & Respondents=="MaleRepublican" ~ "InGroup",
))
#H3: Ingroup partisans <> stronger evaluation on foreign policy and terrorism
#Uses linear regression to evaluate "foreign policy" evaluated by in-party treatment group and out-party treatment group
lm_7<-lm(foreignpolicy ~ PartyGroup, data=becvar_data3)
summary(lm_7)##
## Call:
## lm(formula = foreignpolicy ~ PartyGroup, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.3050 -0.3050 -0.2741 0.6950 2.7259
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.27407 0.08835 25.74 <0.0000000000000002 ***
## PartyGroupOutGroup 0.03089 0.12361 0.25 0.803
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.027 on 274 degrees of freedom
## Multiple R-squared: 0.0002279, Adjusted R-squared: -0.003421
## F-statistic: 0.06245 on 1 and 274 DF, p-value: 0.8029#Uses linear regression to evaluate "terrorism" evaluated by in-party treatment group and out-party treatment group
lm_8<-lm(terrorism ~ PartyGroup, data=becvar_data3)
summary(lm_8)##
## Call:
## lm(formula = terrorism ~ PartyGroup, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.6454 -0.6454 -0.6222 0.3778 2.3778
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.62222 0.10477 25.029 <0.0000000000000002 ***
## PartyGroupOutGroup 0.02317 0.14658 0.158 0.875
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.217 on 274 degrees of freedom
## Multiple R-squared: 9.117e-05, Adjusted R-squared: -0.003558
## F-statistic: 0.02498 on 1 and 274 DF, p-value: 0.8745#Graph results on Foreign Policy
graph_group <- becvar_data3 %>%
group_by(group, PartyGroup) %>%
dplyr::summarise(
mean = mean(foreignpolicy),
sd = sd(foreignpolicy),
n = as.numeric(n()),
se = std.error(foreignpolicy),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_group, aes(x=group, y=mean, fill=PartyGroup)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Foreign Policy") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
#Graph results on Terrorism
graph_group2 <- becvar_data3 %>%
group_by(group, PartyGroup) %>%
dplyr::summarise(
mean = mean(terrorism),
sd = sd(terrorism),
n = as.numeric(n()),
se = std.error(terrorism),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_group2, aes(x=group, y=mean, fill=PartyGroup)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Evaluation on Terrorism") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 5, 0.5)) # Ticks from 0-5, every 0.5
#H4: Ingroup partisans <> stronger evaluation on feelings thermometer and support for candidate
#Uses linear regression to evaluate "feelings thermometer" evaluated by in-party treatment group and out-party treatment group
lm_8<-lm(feeling ~ PartyGroup, data=becvar_data3)
summary(lm_8)##
## Call:
## lm(formula = feeling ~ PartyGroup, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -59.369 -9.585 1.523 12.631 26.631
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 74.585 1.498 49.78 <0.0000000000000002 ***
## PartyGroupOutGroup -1.216 2.096 -0.58 0.562
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 17.41 on 274 degrees of freedom
## Multiple R-squared: 0.001227, Adjusted R-squared: -0.002418
## F-statistic: 0.3367 on 1 and 274 DF, p-value: 0.5622#Uses linear regression to evaluate "terrorism" evaluated by in-party treatment group and out-party treatment group
lm_9<-lm(support ~ PartyGroup, data=becvar_data3)
summary(lm_9)##
## Call:
## lm(formula = support ~ PartyGroup, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -73.943 -6.943 1.089 12.057 26.057
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 76.911 1.541 49.925 <0.0000000000000002 ***
## PartyGroupOutGroup -2.968 2.155 -1.377 0.17
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 17.9 on 274 degrees of freedom
## Multiple R-squared: 0.006872, Adjusted R-squared: 0.003248
## F-statistic: 1.896 on 1 and 274 DF, p-value: 0.1696#Graph results on Feeling
graph_group3 <- becvar_data3 %>%
group_by(group, PartyGroup) %>%
dplyr::summarise(
mean = mean(feeling),
sd = sd(feeling),
n = as.numeric(n()),
se = std.error(feeling),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_group3, aes(x=group, y=mean, fill=PartyGroup)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Warmth Toward Candidate") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 100, 10)) # Ticks from 0-5, every 1
#Graph results on Support
graph_group4 <- becvar_data3 %>%
group_by(group, PartyGroup) %>%
dplyr::summarise(
mean = mean(support),
sd = sd(support),
n = as.numeric(n()),
se = std.error(support),
ub = mean+(1.96*se),
lb = mean-(1.96*se))## `summarise()` has grouped output by 'group'. You can override using the
## `.groups` argument.ggplot(graph_group4, aes(x=group, y=mean, fill=PartyGroup)) +
theme_classic(base_size = 15) +
geom_bar(position=position_dodge(), stat="identity",
colour="black", # Use black outlines,
size=.3) + # Thinner lines
geom_errorbar(aes(ymin=lb, ymax=ub),
size=.3, # Thinner lines
width=.2,
position=position_dodge(.9))+
xlab("Treatment Condition") +
ylab("Support For Candidate") +
ggtitle("") +
theme(plot.title = element_text(hjust = 0.5)) +
scale_fill_manual(values=c("grey", "grey32")) +
scale_y_continuous(breaks=seq(0, 100, 10)) # Ticks from 0-5, every 1
write_csv(becvar_data3, file = "becvar_data3.csv")###More formal test of significance using linear regression
#Foreign policy evaluations by in-vs-outgroup and gender status
mlm1<-lm(foreignpolicy ~ PartyGroup*gender, data=becvar_data3)
summary(mlm1)##
## Call:
## lm(formula = foreignpolicy ~ PartyGroup * gender, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.3636 -0.3636 -0.2907 0.6727 2.8116
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.5389 0.2818 9.009 <0.0000000000000002 ***
## PartyGroupOutGroup -0.1750 0.4107 -0.426 0.670
## gender -0.1752 0.1771 -0.990 0.323
## PartyGroupOutGroup:gender 0.1387 0.2508 0.553 0.581
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.028 on 272 degrees of freedom
## Multiple R-squared: 0.00397, Adjusted R-squared: -0.007016
## F-statistic: 0.3614 on 3 and 272 DF, p-value: 0.781stargazer(mlm1, type = "text")##
## =====================================================
## Dependent variable:
## ---------------------------
## foreignpolicy
## -----------------------------------------------------
## PartyGroupOutGroup -0.175
## (0.411)
##
## gender -0.175
## (0.177)
##
## PartyGroupOutGroup:gender 0.139
## (0.251)
##
## Constant 2.539***
## (0.282)
##
## -----------------------------------------------------
## Observations 276
## R2 0.004
## Adjusted R2 -0.007
## Residual Std. Error 1.028 (df = 272)
## F Statistic 0.361 (df = 3; 272)
## =====================================================
## Note: *p<0.1; **p<0.05; ***p<0.01#Candidate evaluations by in-vs-outgroup and gender status
mlm2<-lm(support ~ PartyGroup*gender, data=becvar_data3)
summary(mlm2)##
## Call:
## lm(formula = support ~ PartyGroup * gender, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -79.873 -7.174 1.988 12.918 29.849
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 83.804 4.813 17.411 <0.0000000000000002 ***
## PartyGroupOutGroup 5.791 7.014 0.826 0.410
## gender -4.561 3.024 -1.508 0.133
## PartyGroupOutGroup:gender -5.160 4.283 -1.205 0.229
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 17.56 on 272 degrees of freedom
## Multiple R-squared: 0.05068, Adjusted R-squared: 0.04021
## F-statistic: 4.84 on 3 and 272 DF, p-value: 0.00268stargazer(mlm2, type = "text")##
## =====================================================
## Dependent variable:
## ---------------------------
## support
## -----------------------------------------------------
## PartyGroupOutGroup 5.791
## (7.014)
##
## gender -4.561
## (3.024)
##
## PartyGroupOutGroup:gender -5.160
## (4.283)
##
## Constant 83.804***
## (4.813)
##
## -----------------------------------------------------
## Observations 276
## R2 0.051
## Adjusted R2 0.040
## Residual Std. Error 17.564 (df = 272)
## F Statistic 4.840*** (df = 3; 272)
## =====================================================
## Note: *p<0.1; **p<0.05; ***p<0.01#Candidate evaluations by party and gender status
mlm3<-lm(support ~ PartyGroup*female, data=becvar_data3)
summary(mlm3)##
## Call:
## lm(formula = support ~ PartyGroup * female, data = becvar_data3)
##
## Residuals:
## Min 1Q Median 3Q Max
## -79.873 -7.174 1.988 12.918 29.849
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 79.2424 2.1620 36.652 <0.0000000000000002 ***
## PartyGroupOutGroup 0.6303 3.2068 0.197 0.844
## female -4.5613 3.0242 -1.508 0.133
## PartyGroupOutGroup:female -5.1603 4.2828 -1.205 0.229
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 17.56 on 272 degrees of freedom
## Multiple R-squared: 0.05068, Adjusted R-squared: 0.04021
## F-statistic: 4.84 on 3 and 272 DF, p-value: 0.00268stargazer(mlm3, type = "text")##
## =====================================================
## Dependent variable:
## ---------------------------
## support
## -----------------------------------------------------
## PartyGroupOutGroup 0.630
## (3.207)
##
## female -4.561
## (3.024)
##
## PartyGroupOutGroup:female -5.160
## (4.283)
##
## Constant 79.242***
## (2.162)
##
## -----------------------------------------------------
## Observations 276
## R2 0.051
## Adjusted R2 0.040
## Residual Std. Error 17.564 (df = 272)
## F Statistic 4.840*** (df = 3; 272)
## =====================================================
## Note: *p<0.1; **p<0.05; ***p<0.01