Group Interest Partisans?

readme.md

Are People Likelier to Support Candidates that Advance their Racial Group Interests More?

To test this, we did a survey experiment on the CCES

John Bowen, a candidate seeking your party's nomination for Congress in another state has made a new jobs plan the centerpiece of his campaign. His plan has received a lot of media attention. The bipartisan Congressional Budget Office has evaluated his plan and concluded it will help create jobs, but that the benefits will not be the same for all Americans. Specifically, the report predicts that the average black (white) family making \$50,000 will see a 7 percent increase in income and the average white (black) family making $50,000 will see a 5 percent increase in income.

Based on this information, how likely would you be to vote for John in this primary? --- Extremely Likely, Very Likely, Somewhat Likely, Somewhat Unlikely, Very Unlikely, Extremely Unlikely

Note: There seems to be a boo-boo in one of the conditions with the last part of last sentence of the White benefit condition reading as: "the average white black family making $50,000 will see a 5 percent increase in income."

race	UCMjobstreat	mean_support	n	se
White	Black 7	0.414	588	0.0111
White	White 7	0.477	572	0.0110
Black	Black 7	0.347	86	0.0235
Black	White 7	0.607	82	0.0334

Authors

Gaurav Sood and Alex Theodoridis

group_interest.R

# Pareto Party

# Load libs
library(tidyverse)
library(readstata13)
library(car)
library(dplyr)
library(xtable)
library(dplyr)

# Read in data
# Data here: https://github.com/soodoku/pareto_partisan/tree/main
cces <- readstata13::read.dta13("CCES18_UCM_unmatched_OUTPUT.dta")

# Jobs/Race

cces$jobsn <- car::recode(cces$UCMjobs, "6 = 1; 5 = .8; 4 = .6; 3 = .4; 2 = .2; 1 = 0; else = NA")

cces %>%
	group_by(UCMjobstreat) %>%
	summarize(mean_support = mean(jobsn, na.rm = T), n = n(), se = sd(jobsn, na.rm = T)/sqrt(n))

# 1 = black
# 2 is white

cces %>%
  group_by(race, UCMjobstreat) %>%
  summarize(mean_support = mean(jobsn, na.rm = T), n = n(), se = sd(jobsn, na.rm = T)/sqrt(n))

cces %>%
	group_by(UCMjobstreat, UCMjobsparty) %>%
	summarize(mean_support = mean(jobsn, na.rm = T), n = n(), se = sd(jobsn, na.rm = T)/sqrt(n))

cces %>%
	filter(race == 1) %>%
	group_by(UCMjobstreat, UCMjobsparty, pid3lean) %>%
	filter(!is.na(pid3lean) & pid3lean != 0) %>%
	summarize(mean_support = mean(jobsn, na.rm = T), n = n(), se = sd(jobsn, na.rm = T)/sqrt(n)) %>%
	arrange(desc(pid3lean, UCMjobstreat))


cces %>%
	filter(race == 1) %>%
	group_by(UCMjobstreat, pid3_lab) %>%
	summarize(mean_support = mean(jobsn, na.rm = T), n = n(), se = sd(jobsn, na.rm = T)/sqrt(n))