Figuring out who would vote for Musk using ANES data

main_2024_trump_musk.R

library(tidyverse)
library(survey)

# mutate weight variable
dat = read_csv('anes_timeseries_2024_csv_20250430.csv') %>%
  filter(V241012 == 1) %>%# RVs
  mutate(weight = case_when(!is.na(V240103b) ~ V240103b,
                            !is.na(V240104b) ~ V240104b,
                            !is.na(V240104b) ~ V240104b,
                            !is.na(V240103a) ~ V240103a,
                            !is.na(V240106a) ~ V240106a,
                            !is.na(V240104a) ~ V240104a,
                            T ~ NA_real_
  ),
  weight = ifelse(is.na(weight), mean(weight,na.rm=T), weight),
  party_id = V241227x,
  trump_thermo = V241157,
  harris_thermo = V241156,
  dems_thermo = V241166,
  reps_thermo = V241167,
  gov_spending_scale = V241239, # 1 = anti, 7 = pro
  dem_7pt_libcon = V241183 # 7 is con
  ) %>%
  select(weight, party_id, trump_thermo, harris_thermo, dems_thermo, reps_thermo,
         gov_spending_scale, dem_7pt_libcon)

# sum(is.na(dat$weight))

nrow(dat)

# filter out some obs
dat = dat %>% 
  mutate_all(function(x){ifelse(x < 0, NA, x)})

nrow(dat)

dat = dat %>%
  mutate(typology = 
           case_when(reps_thermo > 50 &
                       dems_thermo < 50 & 
                       gov_spending_scale <= 4  ~ 'Hardcore Rep',
                     
                     reps_thermo < 50 &
                       dems_thermo > 50 & 
                       gov_spending_scale >= 5   ~ 'Hardcore Dem',
                     
                     trump_thermo <= 50 & reps_thermo < 50 &
                       dems_thermo < 50 & 
                       gov_spending_scale < 4 &
                       dem_7pt_libcon < 3 ~ 'America Party',
                     
                     reps_thermo > 50  ~ 'Soft Rep',
                     
                     dems_thermo > 50 ~ 'Soft Dem',
                     
                     gov_spending_scale >= 5 ~ 'Soft Dem',
                     
                     gov_spending_scale < 5 ~ 'Soft Rep'
           )
  )

dat %>% 
  pull(typology) %>% table(.,useNA = 'always') %>% prop.table()
         

svy = svydesign(ids = ~1, data = dat, weights = ~weight)

svymean(~typology, svy, na.rm = T)


# strong rep or dem
n = nrow(dat); n
# dem or rep
(dat %>% filter(party_id %in% c(7,6,1,2)) %>% pull(weight) %>% sum) / n 
# not dem or rep, but pro trump
(dat %>% filter(!party_id %in% c(7,6,1,2)) %>%  
    filter(trump_thermo >= 50) %>% pull(weight) %>% sum) / n 
# not dem or rep, not pro trump, and wants to keep funding
(dat %>% filter(!party_id %in% c(7,6,1,2)) %>% 
    filter(trump_thermo <= 50) %>% 
    filter(gov_spending_scale >= 4) %>% pull(weight) %>% sum) / n 


67.5 + 14 + 15.8