Reltrad in the United States, 1972 - 2016

Religious Affiliation According to Reltrad

How the United States has Changed Religiously 1972-2016

In a previous post I generated the code to create the reltrad variable in the General Social Survey. In this post, I use that code to plot how religious affiliation has changed in the United States.

In the R code below, I use the package srvyr to use dplyr-like syntax to create a weighted survey object.

The data are available on Github and were created with my reltrad script.

#The recoded data are loaded
gss=read.csv("https://github.com/thebigbird/R_Stata_Reltrad/blob/master/gss7216_reltrad.csv?raw=true")
library(tidyverse)
library(srvyr)

gssSvy <-  gss %>%  as_survey_design(ids = 1, weights = wtssall)

#Grouping by year and reltrad. survey_mean calculates proportions of factor variables
#when an argument is not used.
relTable = gssSvy %>% group_by(year, reltrad) %>%
  summarize(proportion = survey_mean(na.rm=T))
#for the error bars
relTable = relTable %>% mutate(ymin = proportion - 1.96*proportion_se,
                               ymax = proportion + 1.96*proportion_se)
#Reorder the factors so they appear in same order on the legend and the plot 
relTable$reltrad <- factor(relTable$reltrad, levels = 
                             c("Roman Catholic", "Mainline Protestant",
                               "Conservative Protestant", "Black Protestant", "None",
                               "Other", "Jewish"),
                            labels =
                             c("Roman Catholic", "Mainline Protestant",
                               "Conservative Protestant", "Black Protestant","None",
                               "Other", "Jewish"))

#Get rid of NA's in reltrad
relTable = relTable[!is.na(relTable$reltrad),]

Now, using the magic of ggplot2 I generate a plot of how religious affiliation has changed in the United States from 1972 to 2016.

ggplot(data = relTable, aes(x=year, y=proportion, ymin = ymin, ymax = ymax, color = reltrad)) + 
  geom_ribbon(alpha = .2, linetype = 0) + geom_point() + geom_line() + 
  ylab("") + xlab("") +
  scale_x_continuous(limits=c(1970, 2020)) + 
  scale_y_continuous(limits=c(0, .35)) + 
  scale_color_discrete("Reltrad")+
  theme_light() + 
 labs(title = "1972-2016, Nones are on the rise, Christians on the Decline", 
      subtitle = "(Proportion)",
      caption = "Source: General Social Survey, 1972-2016")  

A few notable trends are evident in this plot. First of all, the proportion of Roman Catholics in the United States has stayed relatively steady over this period. While Latin American immigration has brought a lot of new Catholics to the United States, this has been offset by losses among the non-hispanic Catholic population.

Second, there has been a very steep rise in the number of people claiming no religious affiliation. Non-affliates were a pretty steady 5-8% of the US population from the early 1970s to the early 1990s. The proportion of nones has climbed steeply since then, reaching a high of 25% of the US population, making them the second largest “religious” group in the United States. They could very well surpass Catholics by the time the next wave of GSS data are released.

The higher proportion Black Protestant in 1972 is probably due to some sort of artifact in the GSS data. There seems to be a lower proportion of Conservative Protestants that year as well. This blip is also found on other plots of GSS reltrad data, here, for example.

Third, Mainline Protestants are in free fall. They are now about 12% of the US population, down about 20 points from their peak in the early 1970s. Conservative Protestants made some gains in the mid-1980s to mid-1990s (probably due to people leaving mainline denominations over cultural issues), but are also declining. They are down about 10 points from their high in the early 1990s.

Also important is that the “Other” religion category is steadily climbing, and has reached a high of about 5% of the US population. The proportion of the population identifying as Black Protestants remained fairly steady over this period, but recently has begun to decline.

Avatar
David Eagle
Associate in Research