Appendix F — Skill Assignment 6: Logistic Regression Analysis

Author

Methods Student

Published

November 13, 2024

Find the sample assignment here.

F.1 Overall Discussion

F.1.1 Load Packages

library(tidyverse)
library(haven)
library(hrbrthemes)
library(survey)
library(srvyr)
library(labelled)
library(sjmisc)
library(sjPlot)
library(gmodels)
library(gtsummary)
library(skimr)
library(ggblanket)

F.1.2 Load Your Dataset

Need help? Go to chapter x in the webbook.

load("another_anes_pilot_smaller_2.RData")

F.1.3 Manage your data as needed

Need help? Go to chapter x in the webbook.

F.1.4 Final Regression Model

Need help? Go to chapter 4 in the webbook.

roe_glm <- glm(roe_recode ~ ideo5 + urbanicity2_recode + pew_religimp, data = another_anes_pilot_smaller_2, family = "binomial", weights = another_anes_pilot_smaller_2$weight)
summary(roe_glm)


Call:
glm(formula = roe_recode ~ ideo5 + urbanicity2_recode + pew_religimp, 
    family = "binomial", data = another_anes_pilot_smaller_2, 
    weights = another_anes_pilot_smaller_2$weight)

Coefficients:
                                 Estimate Std. Error z value Pr(>|z|)    
(Intercept)                       1.41299    0.34234   4.127 3.67e-05 ***
ideo5Liberal                      0.09176    0.41695   0.220    0.826    
ideo5Moderate                    -0.98755    0.34309  -2.878    0.004 ** 
ideo5Conservative                -2.86078    0.35038  -8.165 3.22e-16 ***
ideo5Very conservative           -3.48971    0.39194  -8.904  < 2e-16 ***
ideo5Not sure                     0.25586    0.74398   0.344    0.731    
urbanicity2_recodeSuburb         -0.06706    0.20903  -0.321    0.748    
urbanicity2_recodeRural          -0.16370    0.21899  -0.747    0.455    
pew_religimpSomewhat important    0.94584    0.20458   4.623 3.78e-06 ***
pew_religimpNot too important     1.54525    0.27888   5.541 3.01e-08 ***
pew_religimpNot at all important  1.47602    0.26265   5.620 1.91e-08 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 1334.70  on 1055  degrees of freedom
Residual deviance:  873.07  on 1045  degrees of freedom
  (79 observations deleted due to missingness)
AIC: 976.48

Number of Fisher Scoring iterations: 5

tbl_regression(roe_glm, exponentiate = TRUE) |>
  add_glance_source_note(
    include = c(AIC))

Characteristic	OR¹	95% CI¹	p-value
Profile: Ideology
Very liberal	—	—
Liberal	1.10	0.48, 2.48	0.8
Moderate	0.37	0.18, 0.71	0.004
Conservative	0.06	0.03, 0.11	<0.001
Very conservative	0.03	0.01, 0.06	<0.001
Not sure	1.29	0.34, 6.99	0.7
Profile: Urban-rural status
City	—	—
Suburb	0.94	0.62, 1.41	0.7
Rural	0.85	0.55, 1.31	0.5
Profile: Importance of religion (Pew version)
Very important	—	—
Somewhat important	2.57	1.73, 3.86	<0.001
Not too important	4.69	2.74, 8.21	<0.001
Not at all important	4.38	2.64, 7.40	<0.001
AIC = 976
¹ OR = Odds Ratio, CI = Confidence Interval

tab_model(roe_glm)

	Favor/oppose-overturn Roe v Wade
Predictors	Odds Ratios	CI	p
(Intercept)	4.11	2.17 – 8.37	<0.001
Profile:Ideology: Liberal	1.10	0.48 – 2.48	0.826
Profile:Ideology: Moderate	0.37	0.18 – 0.71	0.004
Profile:Ideology: Conservative	0.06	0.03 – 0.11	<0.001
Profile:Ideology: Very conservative	0.03	0.01 – 0.06	<0.001
Profile:Ideology: Not sure	1.29	0.34 – 6.99	0.731
Profile:Urban-rural status: Suburb	0.94	0.62 – 1.41	0.748
Profile:Urban-rural status: Rural	0.85	0.55 – 1.31	0.455
Profile:Importance of religion(Pew version): Somewhat important	2.57	1.73 – 3.86	<0.001
Profile:Importance of religion(Pew version): Not too important	4.69	2.74 – 8.21	<0.001
Profile:Importance of religion(Pew version): Not at all important	4.38	2.64 – 7.40	<0.001
Observations	1056
R² Tjur	0.012

F.1.5 Save your updated dataset?

Need help? Go to chapter 4 in the webbook.