CR Influence and State Polarization

CR_influence.R

## Read in the CCES 2016


## This is calculating the weeklys and the nevers for each state
fff <- cces16 %>% 
  mutate(att = car::recode(pew_churatd, "1:2=1; else = 0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(att) %>% 
  mutate(state = to_factor(inputstate)) %>% 
  select(state, week = mean) 

fff2 <- cces16 %>% 
  mutate(att = car::recode(pew_churatd, "6=1; else = 0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(att) %>% 
  mutate(state = to_factor(inputstate)) %>% 
  select(state, never = mean)

gg <- left_join(fff, fff2)

## Need to make a column of who won each state in 2016

votes <- read_csv("D://votes1216.csv")


win <- votes %>% 
  group_by(state_abbreviation) %>% 
  summarise(gop_sum = sum(votes_gop_2016), dem_sum = sum(votes_dem_2016)) %>% 
  mutate(total = gop_sum - dem_sum) %>% 
  mutate(winner = case_when(total > 0 ~ "GOP",
                            total < 0 ~ "Dem")) %>% 
  select(state_abbreviation, winner)

## Turn those states abb into full state names to join 
# library(openintro)
win$state <- abbr2state(win$state_abbreviation)

win <- win %>% 
  select(state, winner)

gg <- left_join(gg, win) 

## Alaska Got left out for some reason. Fix that. 
gg[2,6] <- "GOP"

gg %>% 
  ggplot(., aes(x = reorder(state, -diff), y = diff, fill = winner)) +
  geom_col(color = "black") +
  coord_flip() +
  theme_gg("Abel") +
  scale_fill_manual(values = c("#1565C0", "#b92b27")) +
  scale_y_continuous(labels = percent) +
  theme(legend.position = c(.85,.45)) +
  labs(y = "<-- More Never Than Weekly : More Weekly Than Never -->", x = "", title = "Which States Have More Regular Attenders Than Never Attenders?", caption = "Data: CCES 2016") +
  annotate("text", y=.2, x = 27, label = "2016 Presidential", size = 12, family = "font", fontface = 2) +
  annotate("text", y=.2, x = 26, label = "Election Result", size = 12, family = "font", fontface = 2) +
  ggsave("D://cces/images/state_diff_att.png", height = 8, width =8)

## Making a HexMap 

hex <- cces16 %>% 
  mutate(notimp = car::recode(pew_religimp, "4=1; else =0")) %>% 
  mutate(state = to_factor(inputstate)) %>% 
  group_by(state) %>% 
  mean_ci(notimp) %>% 
  select(state, mean)


## As a reminder to myself. You have to have the entire unzipped list of files for these shape files to read correctly. Not just the .shp one
US_shp <- rgdal::readOGR("D:\\cb_2017_us_state_5m.shp")
US_shp_cropped <- crop(US_shp, extent(-124.848974, -66.885444, 24.396308, 49.384358)) 
US_shp_cropped@data <- left_join(US_shp_cropped@data, hex, by = c("NAME" = "state"))


new_cells_hex <- calculate_grid(shape = US_shp_cropped, grid_type = "hexagonal", seed = 1)
result_hex <- assign_polygons(US_shp_cropped, new_cells_hex)

clean <- function(shape) {
  shape@data$id = rownames(shape@data)
  shape.points = fortify(shape, region="id")
  shape.df = merge(shape.points, shape@data, by="id")
}

result_hex_df <- clean(result_hex)

font_add_google("Abel", "font")
showtext_auto()

ggplot(result_hex_df) +
  geom_polygon(aes(x = long, y = lat, fill = mean, group = group), color = "black") +
  geom_text(aes(V1, V2, label = STUSPS), size = 5, color = "white") +
  scale_fill_gradient(low = "#108dc7", high = "#ef8e38", name = "                Religion Not at All Important (%)", labels = percent) + 
  theme_void() + 
  coord_equal() +
  theme(legend.text=element_text(size=34)) +
  labs(title = "Which States Have Low Levels of Religious Importance?", caption = "Data: CCES 2016") +
  theme(plot.title = element_text(family = "font", size = 40, vjust =2, face = "bold")) +
  theme(plot.caption = element_text(family = "font", size = 36)) +
  theme(legend.title = element_text(family = "font", size = 24, vjust =1.2, hjust = 112, face = "bold"))  +
  theme(legend.text=element_text(family = "font", face = "bold")) +
  theme(legend.position="bottom") +
  ggsave("D://cces/images/hexmap.png")

## Scatter of Nones and Evangelicals

ddd1 <- cces16 %>% 
  mutate(bagain = car::recode(pew_bornagain, "1=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(bagain) %>% 
  mutate(state = to_factor(inputstate)) %>% 
  dplyr::select(state, evanmean = mean)

ddd2 <- cces16 %>% 
  group_by(inputstate) %>% 
  mean_ci(none) %>% 
  mutate(state = to_factor(inputstate)) %>% 
  dplyr::select(state, nonemean = mean)

graph <- left_join(ddd1, ddd2)

tot <- cces16 %>% 
  group_by(inputstate) %>% 
  count() %>% 
  ungroup(inputstate) %>% 
  mutate(state = to_factor(inputstate)) %>%
  dplyr::select(state, total = n)

graph <- left_join(graph, tot)
graph <- left_join(graph, win)

graph[2,5] <- "GOP"


graph %>% 
  ggplot(., aes(x=evanmean, y = nonemean, size = total, fill = winner)) +
  geom_point(shape = 21, stroke = 2) +
  geom_smooth(method =lm, aes(color = winner), alpha = .15, linetype = "dashed") +
  theme_gg("Abel") +
  geom_text_repel(aes(evanmean, y = nonemean, label = state), size = 10, nudge_y = .005) +
  scale_fill_manual(values = c("dodgerblue3", "firebrick3")) +
  scale_color_manual(values = c("dodgerblue3", "firebrick3")) +
  scale_y_continuous(labels = percent) +
  scale_x_continuous(labels = percent) +
  labs(x = "Percent of Born-Again", y = "Percent of Nones", title = "Relationship between Born-Agains and Nones", caption = "Data: CCES 2016") +
  ggsave("D://cces/images/scatter_none_evan.png", width = 8, height = 8)
  
  
## This is making the animated gif

library(haven)
library(labelled)
library(gganimate)

aaa1 <- cces08 %>% 
  mutate(bagain = car::recode(V215, "1=1; else =0")) %>% 
  group_by(V206) %>% 
  mean_ci(bagain) %>% 
  mutate(year = 2008) %>% 
  select(state = V206, bamean =mean, year)

aaa2 <- cces08 %>% 
  mutate(none = car::recode(V219, "9:11=1; else =0")) %>% 
  group_by(V206) %>% 
  mean_ci(none) %>% 
  mutate(year = 2008) %>% 
  select(state = V206, nonemean =mean, year)

aaa3 <- cces08 %>% 
  group_by(V206) %>% 
  count() %>% 
  select(state = V206, total =n)

add08 <- left_join(aaa1, aaa2) 
add08 <- left_join(add08, aaa3)

aaa4 <- cces08 %>% 
  filter(CC307a < 8) %>% 
  group_by(V206) %>% 
  mean_ci(CC307a) %>% 
  select(state = V206, pid = mean)
  
add08 <- left_join(add08, aaa4)

aaa1 <- cces10 %>% 
  mutate(bagain = car::recode(V215, "1=1; else =0")) %>% 
  group_by(V206) %>% 
  mean_ci(bagain) %>% 
  mutate(year = 2010) %>% 
  select(state = V206, bamean =mean, year)

aaa2 <- cces10 %>% 
  mutate(none = car::recode(V219, "9:11=1; else =0")) %>% 
  group_by(V206) %>% 
  mean_ci(none) %>% 
  mutate(year = 2010) %>% 
  select(state = V206, nonemean =mean, year)

aaa3 <- cces08 %>% 
  group_by(V206) %>% 
  count() %>% 
  select(state = V206, total =n)

add10 <- left_join(aaa1, aaa2)
add10 <- left_join(add10, aaa3)

aaa4 <- cces10 %>% 
  filter(V212d < 8) %>% 
  group_by(V206) %>% 
  mean_ci(V212d) %>% 
  select(state = V206, pid = mean)

add10 <- left_join(add10, aaa4)

aaa1 <- cces12 %>% 
  mutate(bagain = car::recode(pew_bornagain, "1=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(bagain) %>% 
  mutate(year = 2012) %>% 
  select(state = inputstate, bamean = mean, year)

aaa2 <- cces12 %>% 
  mutate(none = car::recode(religpew, "9:11=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(none) %>% 
  mutate(year = 2012) %>% 
  select(state = inputstate, nonemean = mean, year)

aaa3 <- cces12 %>% 
  group_by(inputstate) %>% 
  count() %>% 
  select(state = inputstate, total = n)

add12 <- left_join(aaa1, aaa2)
add12 <- left_join(add12, aaa3)

aaa4 <- cces12 %>% 
  filter(pid7 < 8) %>% 
  group_by(inputstate) %>% 
  mean_ci(pid7) %>% 
  select(state = inputstate, pid = mean)

add12 <- left_join(add12, aaa4)

aaa1 <- cces14 %>% 
  mutate(bagain = car::recode(pew_bornagain, "1=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(bagain) %>% 
  mutate(year = 2014) %>% 
  select(state = inputstate, bamean = mean, year)

aaa2 <- cces14 %>% 
  mutate(none = car::recode(religpew, "9:11=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(none) %>% 
  mutate(year = 2014) %>% 
  select(state = inputstate, nonemean = mean, year)

aaa3 <- cces14 %>% 
  group_by(inputstate) %>% 
  count() %>% 
  select(state = inputstate, total = n)

add14 <- left_join(aaa1, aaa2)
add14 <- left_join(add14, aaa3)


aaa4 <- cces14 %>% 
  filter(pid7 < 8) %>% 
  group_by(inputstate) %>% 
  mean_ci(pid7) %>% 
  select(state = inputstate, pid = mean)

add14 <- left_join(add14, aaa4)


aaa1 <- cces16 %>% 
  mutate(bagain = car::recode(pew_bornagain, "1=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(bagain) %>% 
  mutate(year = 2016) %>% 
  select(state = inputstate, bamean = mean, year)

aaa2 <- cces16 %>% 
  mutate(none = car::recode(religpew, "9:11=1; else =0")) %>% 
  group_by(inputstate) %>% 
  mean_ci(none) %>% 
  mutate(year = 2016) %>% 
  select(state = inputstate, nonemean = mean, year)

aaa3 <- cces16 %>% 
  group_by(inputstate) %>% 
  count() %>% 
  select(state = inputstate, total = n)

add16 <- left_join(aaa1, aaa2)
add16 <- left_join(add16, aaa3)


aaa4 <- cces16 %>% 
  filter(pid7 < 8) %>% 
  group_by(inputstate) %>% 
  mean_ci(pid7) %>% 
  select(state = inputstate, pid = mean)

add16 <- left_join(add16, aaa4)


all <- bind_df("add")

cc <- read_dta("D://cces/data/cces16.dta")

tt <- cc %>% 
  mutate(stname = to_factor(inputstate)) %>% 
  mutate(state = as.numeric(inputstate)) %>% 
  select(state, stname) %>% 
  distinct(state, stname)

# rm(cc)

graph <- merge(all, tt) %>% as.tibble()

pid_check <- graph %>% 
  group_by(stname) %>% 
  mean_ci(pid) %>% arrange(-mean)

pid_check$mark <- rep(1:3, each = 17)

pid_check <- pid_check %>% select(stname, mark )

graph <- merge(graph, pid_check)

p <- graph %>% 
  ggplot(., aes(x= bamean, y = nonemean))+
  geom_point(aes(x = bamean, y = nonemean, size = total, fill = factor(mark)), show.legend = FALSE, shape = 21, stroke = 2) +
  geom_text(aes(x = bamean, y = nonemean, label = stname), size=4, family = "Product Sans", nudge_y = .0075) +
  theme_minimal() +
  # scale_fill_gradient2(low = "dodgerblue3", mid = "azure4", high = "firebrick3") +
  scale_fill_manual(values= c("firebrick3", "azure4", "dodgerblue3")) +
  geom_smooth(se = FALSE) +
  theme(text=element_text(size=22, family="Product Sans")) +
  scale_y_continuous(labels = scales::percent) +
  scale_x_continuous(labels = scales::percent) +
  theme(legend.position = "none") +
  labs(title = 'Relationship Between Born Agains and Nones', subtitle = 'Year: {round(frame_time,0)}', x = 'Percent Born-Again', y = 'Percent of Nones', caption = "Data: CCES (2008-2016)") +
  transition_time(year) +
  ease_aes('linear')


animate(p, width = 1000, height = 600, fps = 25, nframes = 200, end_pause = 15)





conger <- read_dta("D://conger.dta")

conger <- conger %>% 
  dplyr::select(state = var1, CRInfluenceIndex)
  

graph <- left_join(graph, conger)


graph %>% 
  ggplot(., aes(x=evanmean, CRInfluenceIndex)) +
  geom_point() +
  geom_smooth(method = lm)


gg <- lm(CRInfluenceIndex ~ evanmean*nonemean, data = graph)
gg2 <- interact_plot(gg, pred= evanmean, modx = nonemean, interval = T, int.width = .76, modx.values = "plus-minus")

gg2 +
  labs(x = "Born Again Share", y = "Christian Right Influence", title = "Interaction of Nones and Born-Agains on Christian Right Influence") +
  theme_gg("Abel") +
  scale_x_continuous(labels = percent) +
  scale_fill_npg() +
  scale_color_npg() +
  theme(legend.position = c(.75,.25)) +
  theme(plot.title = element_text(size = 38)) +
  annotate("text", y=2, x = .465, label = "% of Nones", size = 12, family = "font", fontface = 2) +
  annotate("text", y=1.8, x =.49, label = "In State Population", size = 12, family = "font", fontface = 2) +
  ggsave("D://cces/images/jtools_interact.png")