── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr 1.1.4 ✔ readr 2.1.5
✔ forcats 1.0.0 ✔ stringr 1.5.1
✔ ggplot2 4.0.0 ✔ tibble 3.2.1
✔ lubridate 1.9.4 ✔ tidyr 1.3.1
✔ purrr 1.0.2
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag() masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
liss <-read_csv("/Users/davidsanchezperez/Documents/Doctorado/Guerra de Ucrania/LISS/Clean data/liss_wave14.csv")
Rows: 5626 Columns: 18
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (1): educ_level
dbl (16): ID, treatment, treatment1, women, age, house_income, intereff_cap...
date (1): quest_end_2
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.
Quiero ver de qué manera un threat externo puede aumentar/disminuir la polarización afectiva.
A priori la teoría dice que el efecto rally round the flag debería generar una disminución de la polarización afectiva.
Sin embargo, esto no está tan claro. Algunos papers dicen que la respuesta frente a un threat externo depende de los desacuerdos percebidos entre los partidos, es decir, que trasladan su posición de la política exterior a desacuerdos en su política interna (Myrick, 2021)
Un experimento en abril de 2022 mostró que hacer saliente la invasión rusa de Ucrania condujo a una modesta reducción de la polarización afectiva en EEUU
Contribución: aplicar este marco a un semi-experimento natural en Europa con mayor validez externa
Siguientes pasos: Utilizar panel GLES alemán y analizar el efecto de la postura de los partidos con el apoyo al aumento al gasto militar que se estaba debatiendo como proxy??
A priori lo iba a hacer con una encuesta finlandesa, pero solo hay 15 observaciones para después de la invasión
model1 <-lm(affective_polar ~ treatment, data = treat_5)summary(model1)
Call:
lm(formula = affective_polar ~ treatment, data = treat_5)
Residuals:
Min 1Q Median 3Q Max
-2.8999 -0.9882 -0.2091 0.7586 5.2162
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.8999 0.1695 17.108 <2e-16 ***
treatment -0.4072 0.2142 -1.901 0.0584 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 1.661 on 255 degrees of freedom
Multiple R-squared: 0.01398, Adjusted R-squared: 0.01011
F-statistic: 3.616 on 1 and 255 DF, p-value: 0.05837
model2 <-lm(affective_polar ~ treatment + days + treatment * days, data = treat_5)summary(model2)
Call:
lm(formula = affective_polar ~ treatment + days + treatment *
days, data = treat_5)
Residuals:
Min 1Q Median 3Q Max
-3.0679 -1.0432 -0.1678 0.7132 5.2398
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.45058 0.50041 4.897 1.74e-06 ***
treatment 0.72467 0.68302 1.061 0.290
days -0.15857 0.16623 -0.954 0.341
treatment:days -0.04919 0.21464 -0.229 0.819
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 1.657 on 253 degrees of freedom
Multiple R-squared: 0.02649, Adjusted R-squared: 0.01495
F-statistic: 2.295 on 3 and 253 DF, p-value: 0.07836
t.test(affective_polar ~ treatment, data = treat_5)
Welch Two Sample t-test
data: affective_polar by treatment
t = 1.8457, df = 181.67, p-value = 0.06656
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-0.02810199 0.84252810
sample estimates:
mean in group 0 mean in group 1
2.899852 2.492639
2.2 Diferencias de medias en el resto de variables para ver si la muestra está sesgada
t.test(women ~ treatment, data = treat_5)
Welch Two Sample t-test
data: women by treatment
t = -1.6656, df = 194.37, p-value = 0.09741
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-0.23259895 0.01960723
sample estimates:
mean in group 0 mean in group 1
0.5208333 0.6273292
t.test(age ~ treatment, data = treat_5)
Welch Two Sample t-test
data: age by treatment
t = 0.24373, df = 195.73, p-value = 0.8077
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-3.569159 4.575759
sample estimates:
mean in group 0 mean in group 1
45.03125 44.52795
t.test(house_income ~ treatment, data = treat_5)
Welch Two Sample t-test
data: house_income by treatment
t = 0.86781, df = 197.36, p-value = 0.3866
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-339.9393 874.2411
sample estimates:
mean in group 0 mean in group 1
4097.555 3830.404
model_all <-lm(affective_polar ~ treatment1, data = treat_all)summary(model_all)
Call:
lm(formula = affective_polar ~ treatment1, data = treat_all)
Residuals:
Min 1Q Median 3Q Max
-2.7196 -0.8724 -0.2313 0.6339 5.7921
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.71957 0.02225 122.242 <2e-16 ***
treatment1 -0.22693 0.12082 -1.878 0.0604 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 1.507 on 4746 degrees of freedom
Multiple R-squared: 0.0007428, Adjusted R-squared: 0.0005323
F-statistic: 3.528 on 1 and 4746 DF, p-value: 0.0604
model_all2 <-lm(affective_polar ~ treatment1 + days + treatment1 * days, data = treat_all)summary(model_all2)
Call:
lm(formula = affective_polar ~ treatment1 + days + treatment1 *
days, data = treat_all)
Residuals:
Min 1Q Median 3Q Max
-2.8304 -0.8720 -0.2277 0.6350 5.7795
Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 2.836033 0.090433 31.361 <2e-16 ***
treatment1 0.339219 0.432249 0.785 0.4326
days 0.002807 0.002113 1.329 0.1840
treatment1:days -0.210559 0.123483 -1.705 0.0882 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 1.506 on 4744 degrees of freedom
Multiple R-squared: 0.00171, Adjusted R-squared: 0.001079
F-statistic: 2.709 on 3 and 4744 DF, p-value: 0.04358
t.test(affective_polar ~ treatment1, data = treat_all)
Welch Two Sample t-test
data: affective_polar by treatment1
t = 1.7894, df = 170.28, p-value = 0.07533
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-0.02340831 0.47727040
sample estimates:
mean in group 0 mean in group 1
2.719570 2.492639
3.2 Diferencias de medias en el resto de variables para ver si la muestra está sesgada
t.test(women ~ treatment1, data = treat_all)
Welch Two Sample t-test
data: women by treatment1
t = -2.4504, df = 172.1, p-value = 0.01527
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-0.17223074 -0.01855147
sample estimates:
mean in group 0 mean in group 1
0.5319381 0.6273292
t.test(age ~ treatment1, data = treat_all)
Welch Two Sample t-test
data: age by treatment1
t = 7.4644, df = 175.9, p-value = 3.717e-12
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
6.97973 11.99707
sample estimates:
mean in group 0 mean in group 1
54.01635 44.52795
t.test(house_income ~ treatment1, data = treat_all)
Welch Two Sample t-test
data: house_income by treatment1
t = -1.2442, df = 186.47, p-value = 0.215
alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
95 percent confidence interval:
-622.6769 141.0224
sample estimates:
mean in group 0 mean in group 1
3589.576 3830.404