What Really Happened to Religion in the 1990s?

Young people abandoned Christianity in larger numbers - did they leave Catholicism, evangelicalism, the mainline?

I’ve got a lot of data visualizations in the post today, and I don’t want to make it overly long. So, let me just get right to the graph that started me down this rabbit hole. It’s from the General Social Survey, and I am just looking at respondents who are between the ages of 18 and 35. For very keen observers of this newsletter, you will notice that this graph has appeared once before in a post titled, “Four of the Most Dramatic Shifts in American Religion Over the Last 50 Years.” Well, it’s updated now with data from the last couple of waves from the GSS.

The highlighted period is one that I wanted to focus all my attention on - it spans from 1991 to 1998. In my estimation this is the most consequential period of American religious history in the past five decades. For the twenty years prior, the share of young Americans who were Christians was about 85%, while the non-religious portion never moved above 10%. But, then in the little seven year window of time, everything changed for young adults.

In 1991, Christians were 87% of the sample. In 1998, it was 73%. While the share who were non-religious went from 8% to 20% during that same time period. How long did it take for the Christian share to drop another 14 points? It didn’t happen until 2018. How about the nones jumping up by a dozen points? Again, it was 2018. But how did this happen? Did evangelicals take a nosedive? Did the Catholic Church hemorrhage a bunch of folks? For some reason, I hadn’t really thought much about investigating that until now. So, let’s get to work.

This is the evangelical share of the 18-35 year sample from the General Social Survey. For the next couple of graphs, I am going to show you the entire time series so that you can get a sense of how the 1991-1998 window fits into the larger picture.

In the early 1970s, about 20% of young adults were evangelical, but that share slowly began to rise over the subsequent decades. By 1990, it was nearly 25% of those 18-35. What’s interesting is that it appears evangelicals were at their peak in the early 1990s. But they slowly began to fall over the next 10-15 years. By 2012, they were in the same place they were 40 years earlier. Currently the share of young adults who are evangelical is about 13%. But during that seven year window of time, we can see that evangelicalism only really dropped by about 3 points. That certainly is not enough to explain the first graph.

While the evangelical graph looks like a rainbow (low on the sides and high in the middle), that’s certainly not the case when it comes to mainline Protestant Christians. It looks like a ski slope. The mainline has been declining for decades, and that clearly comes through in this analysis. In the early 1970s, almost a quarter of young adults were affiliated with the mainline, but that didn’t last long. By 1991, it had dropped to just about 14% and it continued to slide over the next couple of years. By 1998, it was 11% and only continued to erode from there.

The good news, if there is any in this graph, is that the trend line has hit a pretty clear plateau since 2016. The share of adults aged 18 to 35 who are mainline Protestants is just about 5% now. It may have even ticked up slightly. But, to go back to that shaded box - the mainline dropped 3 points between 1991 and 1998. That’s the same numeric decline as evangelicals.

What about Black Protestants? Well, this graph looks distinct from both evangelicals and the mainline. There are no dramatic swings in the trend line in either direction. For decades, Black Protestants were just about 8% of young folks. It is possible to see a really gradual decline beginning around 2000, though. In the most recent data less than 5% of adults between 18 and 35 were members of a Black Church. That’s the lowest in the time series. But what about that critical window between 1991 and 1998? I think we can just admit that the 1991 figure is an aberration. That’s what makes the decline 3 points and not zero points. I just don’t see any clear evidence that the decline in Christianity is directly related to a decline in the Black Church.

There’s one more stone to turn over - Catholics. In the early 1970s, about 30% of young adults said they were Catholic. But, over time that share has eroded. By 2012, it had dropped to about 25% and it continued to quickly slide from there. In the 2022 data, just 20% of young adults said that they were Catholic - that’s a decline of 50% over a five decade period. But it doesn’t look like the 1991-1998 window was a real inflection point for a big defection from Catholicism. During that seven year time span, the decline was a meager two points. Certainly unremarkable compared to the decade before and after the 1990s.

So, what does this look like when thrown together into a single graph? I’ve got to admit that I was really looking for the “aha” moment, but there’s none to be found. If the nones went up 12.5 points, where did the declines come from? A little bit here and a little bit there, really. For evangelicals, the mainline, and the Black Church it was 3 points each. Add the Catholic drop in there and that’s really the ballgame. There’s no specific religious tradition to pin this on - it’s all types of Christianity and in nearly equal measure. It’s not like evangelicals were hanging on while the mainline crashed. That’s not what the data says at all. Both traditions were losing share in equal numbers.

Before I go, I wanted to see if I could find the ripple effects of this religious defection in other metrics, though. So, I looked toward religious attendance.

Pretty amazing how flat these trend lines are between 1972 and 1991, right? The share of monthly attending young adults did creep up above five points during this window, but the other attendance levels just didn’t change a whole lot. Once we enter the shaded box between 1991 and 1998, though, we can see things really starting to move in a significant way. The share of never attenders began rising in the late 1980s but really started to take off through the 1990s. It looks like the trajectory that was established during this time period just carried through for the next three decades. Now, about 40% of young adults say they never attend church services.

You can also clearly see that the share of weekly attendees went down between 1991 and 1998, but that trend line did not continue a strong movement. It just kind of floated slowly downward in the subsequent time period. Looking at this measure of religious behavior, I think it’s possible to point towards the 1990s as being a critical decade. The two decades prior were fairly static and whatever we’ve seen since 2000 seems like it’s largely the result of what happened in the 1990s.

The share of young adults who attended religious services nearly every week was 26% in 1991 and that dropped to 22% by 1998. Four points in seven years is certainly noticeable, but remember that the share of Christians dropped by 14 points during this time period. There were also drops among monthly attenders (down 2 points), and yearly attenders (down 4 points). The only two groups that saw gains were seldom and never attenders. Those two groups were 23% of the sample in 1991 and in 1998 that had increased to 33%. A ten point gain in the share of people who were attending less than once a year.

The headline graph for this piece is certainly eye-catching. Young people lost their religion in droves in the 1990s. When I first put this visualization together, the folks at Religion News Service asked me to write it up for a longer post. It turned into this one, “How America's youth lost its religion in the 1990s.” The arguments I make are not new. They have been also made in Secular Surge by Campbell, Layman, and Green; in Nonverts by Bullivant and in Christian Smith’s forthcoming book Why Religion Went Obsolete. I still believe that we are living in the residue of a tremendous amount of cultural, social, and political changes that began in the 1990s. I often wonder if there will be a confluence of events that reverse this trend in the decades to come. What I am sure of is that we won’t even begin to identify it until years into the future.

Code for this post can be found here.