Evaluation of party positions in the CHES dataset

ches_positions.R

# load the data
library(tidyverse)

# write functions to calculate confidence intervals
lower_ci <- function(mean, se, n, conf_level = 0.95){
  lower_ci <- mean - qt(1 - ((1 - conf_level) / 2), n - 1) * se
}

upper_ci <- function(mean, se, n, conf_level = 0.95){
  upper_ci <- mean + qt(1 - ((1 - conf_level) / 2), n - 1) * se
}

# get the 1999-2019 chapel hill expert survey (CHES) trend file
ches <- read.csv("https://www.chesdata.eu/s/1999-2019_CHES_dataset_meansv1.csv")

# name the party families
par_fam <- ches %>% 
  select(year, party_id, family) %>% 
  mutate(family = recode_factor(family, `1` = "Radical Right", `2` = "Conservatives", `3` = "Liberal",
                         `4` = "Christian-Democratic", `5` = "Socialist", `6` = "Radical Left",
                         `7` = "Green", `8` = "Regionalist", `10` = "Confessional", 
                         `11` = "Agrarian/Center", `9` = "No Family"))


# select the position variables to be used
par_pos <- ches %>% 
  select(year, party_id, vote, lrgen, galtan, immigrate_policy)


# start preparing the data
ches %>% 
  
  # keep only the large families
  filter(family < 7) %>% 
  
  # create party combinations
  select(year, party_id, family, lrgen, galtan, immigrate_policy) %>% 
  
  mutate(family = recode_factor(
         family, `1` = "Radical Right", `2` = "Conservatives", `3` = "Liberal",
                 `4` = "Christian-Democratic", `5` = "Socialist", 
                 `6` = "Radical Left", `7` = "Green", `8` = "Regionalist", 
                 `10` = "Confessional", `11` = "Agrarian/Center", 
                 `9` = "No Family")) %>% 
  
  # switch from wide to long data
  pivot_longer(c("lrgen", "galtan", "immigrate_policy"), 
               names_to = "area", values_to = "values") %>%
  
  # calculate statistics by family and position
  group_by(year, family, area) %>% 
  summarise(mean_v = mean(values, na.rm = TRUE),
            sd_v = sd(values, na.rm = TRUE),
            count = n()) %>%
  mutate(se = sd_v / sqrt(count),
         lower_ci = lower_ci(mean_v, se, count),
         upper_ci = upper_ci(mean_v, se, count)) %>% 
  
  # plot the data
  ggplot(aes(x = year, y = mean_v, color = area)) +
  geom_errorbar(aes(ymin = lower_ci, ymax = upper_ci), 
                width = .1, size  =  0.8) +
  geom_line(size  =  1.2) +
  geom_point(size  =  1.3) +
  facet_wrap(. ~ family) +
  theme_bw() +
  theme(legend.position = "bottom",
        legend.title = element_blank()) +
  scale_x_continuous(breaks = c(1999, 2002, 2006, 2010, 2014, 2019)) +
  scale_color_discrete(breaks = c("lrgen", "galtan", "immigrate_policy"),
                       labels = c("Left-Right", "GAL-TAN", "Immigration")) +
  ylab("Mean position\n") + xlab("")