We are no stranger to the Cooperative Election Study. This is a super rich survey with lots of cool items. Let’s see what we can cook up in this lab!
Let’s load the data.
library(tidyverse)
library(magrittr)
library(rio)
library(broom)
# Let's load the data -----------------------------------------------------
##thanks to Thomas J Wood for uploading this file to github -- gorgeous work and means we don't have to upload files independently to our machines
t1 <- "https://github.com/thomasjwood/constraint/raw/main/ces/ces_22_c.rds" %>%
rio::import(
setclass = "tbl_df"
)## Warning: Missing `trust` will be set to FALSE by default for RDS in 2.0.0.Now that we have the data, let’s make a codebook and poke around at a few variables.
# build a codebook:
cb <- tibble(
nms = t1 %>%
map(
\(i)
i %>%
attr(., "label")
) %>%
unlist %>%
names,
labs = t1 %>%
map(
\(i)
i %>%
attr(., "label")
) %>%
unlist
)
##does the codebook work?
cb## # A tibble: 697 × 2
## nms labs
## <chr> <chr>
## 1 caseid Case ID
## 2 tookpost Took post-election survey
## 3 commonweight Weight - Common
## 4 commonpostweight Weight - Common Post
## 5 CCEStake Consent to participate
## 6 add_confirm Confirm address
## 7 inputzip Zip code
## 8 birthyr Birth Year
## 9 gender4 Gender
## 10 gender4_t Gender - Other
## # ℹ 687 more rows#we can treat the codebook like a list and use map on it
# then we look at the specific items we’re interested in
c("Executive Orders Pre") %>%
map(
\(i)
cb %>%
filter(
labs %>%
str_detect(i)
)
)## [[1]]
## # A tibble: 5 × 2
## nms labs
## <chr> <chr>
## 1 CC22_355a Executive Orders Pre -- The United States re-joins the Paris Climat…
## 2 CC22_355b Executive Orders Pre -- The United States re-joins the World Health…
## 3 CC22_355c Executive Orders Pre -- Order all federal agencies to buy clean ene…
## 4 CC22_355d Executive Orders Pre -- Increase the minimum wage paid to federal c…
## 5 CC22_355e Executive Orders Pre -- Require that all employees at large compani…#that is so nifty and cool, we could even look at the first few responses
c("Executive Orders Pre ") %>%
map(
\(i)
cb %>%
filter(
labs %>%
str_detect(i)
) %>%
use_series(nms)
) %>%
map(
\(i)
t1 %>%
select(
any_of(i) #any_of is a selector helper, which lets us select variables as a function of characters, in contrast with all_of
)
)## [[1]]
## # A tibble: 60,000 × 5
## CC22_355a CC22_355b CC22_355c CC22_355d CC22_355e
## <fct> <fct> <fct> <fct> <fct>
## 1 Support Support Support Support Support
## 2 Oppose Oppose Oppose Support Oppose
## 3 Support Support Oppose Support Oppose
## 4 Support Support Support Support Support
## 5 Support Support Support Support Oppose
## 6 Oppose Oppose Oppose Oppose Oppose
## 7 Oppose Oppose Oppose Oppose Oppose
## 8 Support Support Support Support Oppose
## 9 Support Support Oppose Support Support
## 10 Support Support Support Support Support
## # ℹ 59,990 more rowsExecutive orders get lots of attention from the media. In this lab, I want us to look at support for Executive Orders.
##first let's select out the variables that we are interested in working with for this exercise.
t2 <- t1 %>%
select(
caseid,
commonweight,
CC22_355a:CC22_355e,
educ,
pid3,
ideo5,
newsint,
race
) %>%
na.omit
t2## # A tibble: 59,180 × 12
## caseid commonweight CC22_355a CC22_355b CC22_355c CC22_355d CC22_355e educ
## <dbl> <dbl> <fct> <fct> <fct> <fct> <fct> <fct>
## 1 1.98e9 3.65 Support Support Support Support Support Post…
## 2 1.98e9 0.780 Oppose Oppose Oppose Support Oppose Some…
## 3 1.98e9 0.892 Support Support Oppose Support Oppose 4-ye…
## 4 1.98e9 1.10 Support Support Support Support Support Post…
## 5 1.98e9 0.543 Support Support Support Support Oppose Post…
## 6 1.98e9 0.114 Oppose Oppose Oppose Oppose Oppose 4-ye…
## 7 1.98e9 0.899 Oppose Oppose Oppose Oppose Oppose High…
## 8 1.98e9 0.633 Support Support Support Support Oppose Post…
## 9 1.98e9 0.900 Support Support Oppose Support Support 4-ye…
## 10 1.98e9 0.725 Support Support Support Support Support Post…
## # ℹ 59,170 more rows
## # ℹ 4 more variables: pid3 <fct>, ideo5 <fct>, newsint <fct>, race <fct>If we just want to test relationships between pairs of variables, we can use map2 in the following way:
c("CC22_355a",
"CC22_355b",
"CC22_355c",
"CC22_355d",
"CC22_355e") %>%
map2(
c("pid3",
"ideo5",
"educ",
"race",
"newsint"),
\(i, j)
str_c(
"commonweight ~",
i,
" + ",
j
) %>%
xtabs(
data = t2
) %>%
prop.table(2) %>%
round(2)
) ## [[1]]
## pid3
## CC22_355a Democrat Republican Independent Other Not sure skipped not asked
## Support 0.89 0.30 0.56 0.51 0.60
## Oppose 0.11 0.70 0.44 0.49 0.40
## skipped 0.00 0.00 0.00 0.00 0.00
## not asked 0.00 0.00 0.00 0.00 0.00
##
## [[2]]
## ideo5
## CC22_355b Very liberal Liberal Moderate Conservative Very conservative
## Support 0.96 0.94 0.75 0.37 0.23
## Oppose 0.04 0.06 0.25 0.63 0.77
## skipped 0.00 0.00 0.00 0.00 0.00
## not asked 0.00 0.00 0.00 0.00 0.00
## ideo5
## CC22_355b Not sure skipped not asked
## Support 0.72
## Oppose 0.28
## skipped 0.00
## not asked 0.00
##
## [[3]]
## educ
## CC22_355c No HS High school graduate Some college 2-year 4-year Post-grad
## Support 0.58 0.55 0.59 0.56 0.62 0.65
## Oppose 0.42 0.45 0.41 0.44 0.38 0.35
## skipped 0.00 0.00 0.00 0.00 0.00 0.00
## not asked 0.00 0.00 0.00 0.00 0.00 0.00
## educ
## CC22_355c skipped not asked
## Support
## Oppose
## skipped
## not asked
##
## [[4]]
## race
## CC22_355d White Black Hispanic Asian Native American Two or more races Other
## Support 0.65 0.86 0.80 0.78 0.67 0.75 0.59
## Oppose 0.35 0.14 0.20 0.22 0.33 0.25 0.41
## skipped 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## not asked 0.00 0.00 0.00 0.00 0.00 0.00 0.00
## race
## CC22_355d Middle Eastern skipped not asked
## Support 0.80
## Oppose 0.20
## skipped 0.00
## not asked 0.00
##
## [[5]]
## newsint
## CC22_355e Most of the time Some of the time Only now and then Hardly at all
## Support 0.42 0.45 0.43 0.36
## Oppose 0.58 0.55 0.57 0.64
## skipped 0.00 0.00 0.00 0.00
## not asked 0.00 0.00 0.00 0.00
## newsint
## CC22_355e Don't know skipped not asked
## Support 0.43
## Oppose 0.57
## skipped 0.00
## not asked 0.00The above is useful if we are exploring relationships in the data, but what if we want to be a little more sophisticated. Imagine we want to take the ‘Executive Orders Pre survey items’ CC22_355a:CC22_355e, and ideology, partisanship, education, race, and news interest, and for every combination of item and predictor, report the categorical correspondence.
Well no worries because we can certainly do that!
c("pid3",
"ideo5",
"educ",
"race",
"newsint") %>%
map2(
c("CC22_355a",
"CC22_355b",
"CC22_355c",
"CC22_355d",
"CC22_355e"),
\(i, j)
lm(
as.formula(
str_c
("commonweight ~ ", j, " + pid3 + ideo5 + educ + race + newsint")
),
data = t2
) %>%
glance()
)## [[1]]
## # A tibble: 1 × 12
## r.squared adj.r.squared sigma statistic p.value df logLik AIC BIC
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0.0710 0.0706 1.09 174. 0 26 -89001. 178058. 178309.
## # ℹ 3 more variables: deviance <dbl>, df.residual <int>, nobs <int>
##
## [[2]]
## # A tibble: 1 × 12
## r.squared adj.r.squared sigma statistic p.value df logLik AIC BIC
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0.0709 0.0705 1.09 174. 0 26 -89004. 178063. 178315.
## # ℹ 3 more variables: deviance <dbl>, df.residual <int>, nobs <int>
##
## [[3]]
## # A tibble: 1 × 12
## r.squared adj.r.squared sigma statistic p.value df logLik AIC BIC
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0.0709 0.0705 1.09 174. 0 26 -89005. 178067. 178319.
## # ℹ 3 more variables: deviance <dbl>, df.residual <int>, nobs <int>
##
## [[4]]
## # A tibble: 1 × 12
## r.squared adj.r.squared sigma statistic p.value df logLik AIC BIC
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0.0709 0.0704 1.09 173. 0 26 -89006. 178067. 178319.
## # ℹ 3 more variables: deviance <dbl>, df.residual <int>, nobs <int>
##
## [[5]]
## # A tibble: 1 × 12
## r.squared adj.r.squared sigma statistic p.value df logLik AIC BIC
## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0.0709 0.0705 1.09 174. 0 26 -89004. 178065. 178316.
## # ℹ 3 more variables: deviance <dbl>, df.residual <int>, nobs <int>