ct_selection_untitled-0DF773D1.R

gg <-cces %>%
  filter(year >= 2020) %>% 
  mutate(pew_religimp = pew_importance) %>% 
  mutate(imp = frcode(pew_religimp == 4 ~ "Not\nat All",
                      pew_religimp == 3 ~ "Not too",
                      pew_religimp == 2 ~ "Somewhat",
                      pew_religimp == 1 ~ "Very")) %>% 
  cces_attend(pew_attendance) %>% 
  group_by(imp) %>% 
  ct(att, wt = weight, show_na = FALSE) %>% 
  na.omit() %>% 
  mutate(pct = n/142261) %>% 
  mutate(pct = round(pct, 2))

gg %>% 
  ggplot(., aes(x= imp, y = att)) +
  geom_tile(aes(fill = pct), color = "black") +
  scale_fill_gradient2(low = "#1fa2ff", mid = "#12d8fa", high = "#a6ffcb", midpoint = .07) +
  theme_rb() + 
  theme(plot.subtitle = element_text(size = 24)) +
  geom_text(aes(x= imp, y = att, label = paste0(pct*100, '%')), size = 11, family = "font") +
  labs(x= "Religious Importance", y = "Religious Attendance", title = "Religious Importance versus Religious Attendance", caption = "@ryanburge\nData: Cooperative Election Study, 2020-2022") 
save("heat_imp_att2022.png", wd = 6, ht = 6)


gg <- cces %>% 
  mutate(pew_religimp = pew_importance) %>% 
  mutate(imp = case_when(pew_religimp == 1 ~ 1, TRUE ~ 0)) %>% 
  mutate(att = case_when(pew_attendance == 5 | pew_attendance == 6 ~ 1, TRUE ~ 0)) %>% 
  mutate(both = imp + att) %>% 
  mutate(both = case_when(both == 2 ~ 1, TRUE ~ 0)) %>% 
  group_by(year) %>% 
  mean_ci(both, wt = weight, ci = .84)


gg %>% 
  mutate(lab = round(mean, 2)) %>% 
  filter(year > 2007) %>% 
  ggplot(., aes(x = year, y = mean, fill = mean)) +
  geom_col(color = "black") +
  theme_rb() +
  error_bar() +
  scale_fill_gradient2(low = "#f1f8fa", mid = "#64a6b9", high = "#2c414d", midpoint = .07) +
  geom_text(aes(y = .0075, label = paste0(lab*100, '%')), position = position_dodge(width = .9), size = 8, family = "font")  +
  geom_text(aes(y = .0075, label = ifelse(mean > .075, paste0(lab*100, '%'), "")), position = position_dodge(width = .9), size = 8, family = "font", color = "white")  +
  scale_x_continuous(breaks = c(2008, 2010, 2012, 2014, 2016, 2018, 2020, 2022)) +
  scale_y_continuous(labels = percent) +
  labs(x = "", y = "", title = "The Share Who Say Religion is Very Important But Don't Attend", caption = "@ryanburge\nData: Cooperative Election Study, 2008-2022")
save("low_att_high_importance.png")


gg <- cces %>% 
  cces_pid3(pid7) %>% 
  mutate(pew_religimp = pew_importance) %>% 
  mutate(imp = case_when(pew_religimp == 1 ~ 1, TRUE ~ 0)) %>% 
  mutate(att = case_when(pew_attendance == 5 | pew_attendance == 6 ~ 1, TRUE ~ 0)) %>% 
  mutate(both = imp + att) %>% 
  mutate(both = case_when(both == 2 ~ 1, TRUE ~ 0)) %>% 
  group_by(year, pid3) %>% 
  mean_ci(both, wt = weight, ci = .84) %>% filter(pid3 != "NA")


gg %>% 
  filter(year > 2007) %>% 
  ggplot(., aes(x = year, y = mean, color = pid3, group = pid3)) +
  geom_point(stroke = 1, shape = 21, fill = "white") + 
  geom_labelsmooth(aes(label = pid3),  method = "loess", formula = y ~ x, family = "font", linewidth = 1, text_smoothing = 30, size = 7, linewidth = 1, boxlinewidth = 0.3, hjust = .75) +
  pid3_color() +
  theme_rb() + 
  scale_y_continuous(labels = percent, limits = c(.03, .12)) +
  labs(x = "", y = "", title = "The Share Who Say Religion is Very Important But Don't Attend", caption = "@ryanburge\nData: Cooperative Election Study, 2008-2022")
save("pid3_low_att_high_importance.png")

gg <- cces %>% 
  filter(year >= 2020) %>% 
  mutate(pew_religimp = pew_importance) %>% 
  mutate(imp = case_when(pew_religimp == 1 ~ 1, TRUE ~ 0)) %>% 
  mutate(att = case_when(pew_attendance == 5 | pew_attendance == 6 ~ 1, TRUE ~ 0)) %>% 
  mutate(both = imp + att) %>% 
  mutate(both = case_when(both == 2 ~ 1, TRUE ~ 0)) %>% 
  group_by(age) %>% 
  mean_ci(both, wt = weight, ci = .84)

gg %>% 
  filter(age <= 75) %>% 
  ggplot(., aes(x = age, y = mean)) +
  geom_point(stroke = 1, shape = 21, alpha = .5) + 
  geom_smooth(se = FALSE, color = "firebrick1", linetype = "twodash") +
  scale_y_continuous(labels = percent, limits = c(0, .13)) +
  theme_rb() + 
  labs(x = "Age", y = "", title = "The Share Who Say Religion is Very Important But Don't Attend", caption = "@ryanburge\nData: Cooperative Election Study, 2020-2022")
save("low_att_high_importance_age.png")


gg1 <- cces %>% 
  filter(year >= 2020) %>% 
  mutate(pew_religimp = pew_importance) %>% 
  mutate(imp = case_when(pew_religimp == 1 ~ 1, TRUE ~ 0)) %>% 
  filter(imp == 1) %>% 
  mutate(att = case_when(pew_attendance == 5 | pew_attendance == 6 ~ 1, TRUE ~ 0)) %>% 
  group_by(age) %>% 
  mean_ci(att, wt = weight)


gg1 %>% 
  filter(age <= 75) %>% 
  ggplot(., aes(x = age, y = mean)) +
  geom_point(stroke = 1, shape = 21, alpha = .5) + 
  geom_smooth(se = FALSE, color = "firebrick1", linetype = "twodash") +
  scale_y_continuous(labels = percent, limits = c(0, .30)) +
  theme_rb() + 
  labs(x = "Age", y = "", title = "Among Those Who Say Religion is Very Important, The Share Who Attend Seldom/Never", caption = "@ryanburge\nData: Cooperative Election Study, 2020-2022")
save("low_att_high_importance_age2.png")
  



regg <- cces %>% 
  filter(year >= 2020) %>% 
  mutate(pew_religimp = pew_importance) %>% 
  mutate(imp = case_when(pew_religimp == 1 ~ 1, TRUE ~ 0)) %>% 
  mutate(att = case_when(pew_attendance == 5 | pew_attendance == 6 ~ 1, TRUE ~ 0)) %>% 
  mutate(both = imp + att) %>% 
  mutate(both = case_when(both == 2 ~ 1, TRUE ~ 0)) %>% 
  mutate(nokids = case_when(havekids == 2 ~ 1, 
                            havekids == 1 ~ 0)) %>% 
  mutate(single = case_when(marital_status == 5 ~ 1, 
                            TRUE ~ 0)) %>% 
  mutate(male = case_when(gender == 1 ~ 1, 
                          gender == 2 ~ 0)) %>% 
  mutate(white = case_when(race == 1 ~ 1, 
                           TRUE ~ 0)) %>% 
  mutate(cons = case_when(ideo5 == 4 | ideo5 == 5 ~ 1, 
                         ideo5 <= 3 ~ 0)) %>% 
  select(age, income, educ, male, white, cons, single, nokids, both)


one <- glm(both ~  male + white + age + income + educ + cons, family = "binomial", data = regg)

library(jtools)

coef_names <- c("Age" = "age", 
                "Male" = "male", 
                "White" = "white", 
                "Income" = "income", 
                "Education" = "educ", 
                "Conservative" = "cons")


out <- plot_summs(one,  robust = "HC3", scale = TRUE, coefs = coef_names)


out + 
  theme_rb() +
  add_text(x = .25, y = 4.75, word = "More Likely", sz = 8) +
  add_text(x = -.25, y = 1.75, word = "Less Likely", sz = 8) +
  labs(x = "", y = "", title = "What Factors Predict Someone to Say That Religion is Very Important But Don't Attend", caption = "@ryanburge\nData: Cooperative Election Study, 2020-2022")
save("reg_predict_high_imp_low_attend.png", ht = 4)