In my last post, I looked at U.S. religious affiliation, and it was based on the General Social Survey question that asked people about their current denominational affiliation. There's another General Social Survey question that asks in which religion people were raised. Assuming that people can remember this information correctly, which I think that I can, this gives us affiliation data going back to the early 1900s. Why? The General Social Survey began in the 1970s, and some of its elderly respondents then were recalling their religious experiences from many decades before.
Using these data, here are estimated affiliation rates for about the last century. As you can see, the decline of Mainline Protestantism has been going on for some time now, as has the increase in the religiously unaffiliated. Also, the percentage of Catholics increased off-and-on during the whole time period, and the percentage Evangelicals hovered steadily around 25%.
I created this graph by using General Social Survey data to determine how many respondents were in each religion during the decade of their sixteenth birthday. This is divided by the number of Americans alive during that century as per Census data. Unfortunately, the retrospective religion question in the GSS doesn’t ask about church attendance rates in youth, so I wasn’t able to implement fully Steensland et al.’s coding scheme for nondenominational Christians. I therefore split the nondenominational Christians between Evangelicals and Mainline Protestants based of on the proportion of each among respondents who identified their denomination.