The August edition of the Storytelling with Data challenge #SWDchallenge stars the dot plot. Here is a simple plot of the gender gap in voting in national elections using the most recent 8th Round of the European Social Survey, ESS.
Getting and reshaping the data
library(essurvey) # getting European Social Survey data library(tidyverse) # data cleaning and reshaping library(countrycode) # converting country codes to names library(ggplot2) # plots
essurvey package the ESS data can be downloaded directly to R.
set_email("firstname.lastname@example.org") # login email to the ESS data
ess8 <- import_rounds(8) # imports Round 8 of ESS
Custom theme to make the dot plot more clear:
theme_dotplot <- theme_bw(14) + theme(axis.text.y = element_text(size = rel(1.1)), axis.ticks.y = element_blank(), axis.text.x = element_text(size = rel(1.1)), panel.grid.major.x = element_blank(), panel.grid.major.y = element_line(size = 0.5), panel.grid.minor.x = element_blank())
Thanks to the
tidyverse data manipulation and plotting can be done in a single block of code.
The gender gap in voting is a ratio of two proportions: (1) the (weighted) proportion of positive answers to the question about voting in the last general election among women divided by (2) the (weighted) proportion of positive answers to the question about voting in the last election among men.
recode_missings(ess8) %>% # recode missing values to NAs select(idno, cntry, gndr, vote, pspwght) %>% # select the needed variables mutate( vote1 = plyr::mapvalues(vote, from = c(1,2,3,NA), # recode "not eligible" to NA to = c(1,0,NA,NA))) %>% filter(!is.na(gndr)) %>% group_by(cntry, gndr) %>% summarise(vote = weighted.mean(vote1, pspwght, na.rm = TRUE)) %>% # weighted mean spread(gndr, vote) %>% mutate(ratio = `2` / `1`, country_name = countrycode(cntry, 'iso2c', 'country.name')) %>% ggplot(., aes(x = ratio, y = reorder(country_name, ratio))) + # sort country name by ratio theme_dotplot + # add custom theme geom_point(col = "chartreuse3", size = 2.5) + # format dots scale_x_continuous(limits = c(0.85,1.2), breaks = seq(0.85,1.2,0.05)) + geom_vline(xintercept = 1, color = "chocolate1") + # add vertical line at 1 xlab("") + ylab("") + ggtitle("Women-to-men election turnout ratio") + labs(caption="Source: European Social Survey Round 8") + annotate("text", x = 0.927, y = 22.8, # add custom text label = "women's turnout < men's turnout", col = "chocolate1", size = 3.1) + annotate("text", x = 1.06, y = 22.8, label = "women's turnout > men's turnout", col = "chocolate1", size = 3.1)
The Dot Plot
In the plot below countries are sorted according to the descending order of the turnout ratio. Values above 1 indicate that more women went to vote than men. This is the case in Russia, where declared turnout among women is 13% higher than among men. In Czechia turnout is about 10% higher among men than among women.
Overall women are more likely to vote than men in 11 out of 23 of the surveyed countries. The ratio is almost exactly 1 in Spain, while in the remaining 11 countries men tend to vote more often. There is a mix of western and eastern European countries in both groups. Who will explain this pattern?