Last Tuesday, the good folks at the Pew Research Center released their report on America’s Religious Landscape. Predictably, the internet went crazy. Some argued that the growth in the nonaffiliated marked the end of Christianity. Others argued that this was actually good news for evangelicals because they didn’t suffer losses are great as other religious groups. Still others used the data to continue the never-ending saga of “mainline hemorrhage”.
I watched all this from a bit of a remove because it was finals week and I had a pile of grading. But I submitted grades for my last class this morning, which freed me to explore these questions for myself. I don’t have access to the 2014 Pew data (if someone wants to give me access, I’d be thrilled!), so I played around with the 2007 Pew Religious Landscape data to test some assumptions.
What I’ve explored below are the key questions we ought to be asking before writing opinion pieces that simply read a narrative into the data. When I do get access to the 2014 data, I’ll repeat the analysis I did today.
1. The Problem with Pie Charts
One of the first things I did was examine various “religious families” across the two survey periods. This comparison has been the basis of many of the blog posts about the 2014 data. The percentage of respondents who are nonaffiliated jumped from 16% to 23% while other groups took small losses. It is true that the mainline took a bigger hit than evangelicals, reflecting an actual loss of population over time.
But this kind of analysis is what happens when we rely on percentages. The pie charts have to add to 100%, so if the percentage of nonaffiliated goes up, other percentages must go down. It’s not religion; it’s just math.
An alternative interpretation that relies less on parsing changes to pie slices would look at the percentage of respondents who represent the four primary Christian families. In 2007, those families made up 75% of the total, which fell to 66% in 2014. However we look at this, dropping from 3 of 4 Christians to 2 of 3 Christians doesn’t mean Christianity is dying by any stretch.
2. The Challenge of Self-Identification
The Pew Survey asks people about their religious group identification (in denominational terms) and then collapses those into the religious families shown above (a variable they call RELTRAD). In doing my analysis today, I only focused on the primary four families: Evangelicals, Mainlines, Black Protestant, and Catholic.
Pew also asked whether respondents claimed to be “born again”. As commenters on the 2014 data have reported, a substantial percentage of respondents identify with the label.
In my analysis of the 2007 data, slightly less than half (46%) are born again. Most of these are Evangelicals. But four in ten of those “born agains” come from the other three families, with 15% of Catholics and 28% of Mainlines agreeing. If over one in four Mainline respondents say they’re born again, the “mainline doesn’t stand for anything” narrative might need to go.
This may suggest that there are certain cultural dynamics related to labels that evangelicals like to claim as their own. This cultural identification may be consistent with those other surveys that show attitudes toward the historicity of the virgin birth. It may simply be that “that’s what we say” in certain situations. What people mean by born again will need much more analysis.
3. The Problem of Attendance
Things get more complicated when we look at attendance patterns. Since the Mainline Hemorrhage thesis depends on a simple cultural identification that now isn’t needed, it’s important to see what’s really happening in congregations. If one needed to go to church in the past to prove you’re a good community member and religious non-affiliation is now more accepted, we’d expect a lot of members on paper but not in real life.
There is some truth to this, but it cuts across religious families. I broke the attendance data into two sets; those who attended once a month or more and those who attended less than once a month (I used the “once a month” cutoff in my dissertation research as the minimum level of engagement in the congregation). Here are the percentages of each by religious family.
If the “Cultural Christian” thesis holds then we’d get data like we see for Mainlines. But that narrative fails to account for the fact that nearly 6 in 10 Mainline respondents are active in their congregations. Nor can it explain the 1 in 4 Evangelicals who rarely attend church.
A related story told in commentary involves the aging of the Mainline church. We can call that the “Blue Hair Thesis”. If that were to be supported, we’d see a gradual pattern of an aging population that fails to generate sufficient replacement populations to handle losses through death. Related to this pattern is the differential birthrates by religious family (which limits replacement in some traditions).
I was able to examine attendance patterns using the four age cohorts that Pew used in reporting the 2014 data: 18-29, 30-49, 50-64, and 65 plus. It is true that the senior category has a higher attendance percentage than other groups, but what surprised me was the relative stability of the 18-64 groups.
Again, the data from the 2007 Religious Landscape study raises questions about our preferred narratives. While it’s true that attendance patterns run higher among Evangelicals and Black Protestants, every age cohort within every religious family shows a majority attending church at least once a month.
Working out my logic this morning, I played with a Baylor Religion Survey, also done in 2007 (thanks to the folks at The Association of Religion Data Archives). Attendance may also be a necessary qualifier in making sense of “switching data”. Those questions (which are in Pew) compare childhood religious family with current religious family. But the Baylor survey also asks about attendance at age 12. Nearly 1 in 5 respondents didn’t make the once a month attendance threshold (a pattern with surprisingly little variation by tradition). To treat infrequent attenders as “switchers” seem like a distortion of the data.
In the Pew data, I was able to compare the “born again” data to the attendance data without separating the four religious families. I found 20% claiming to be born again and attending church less than once a month. Not everything is as we so easily suspect.
4. Religion is Important
Another of the popular narratives is that religion is become increasingly irrelevant to modern society. This may be true in the sense of lessened hegemony over cultural dynamics but it doesn’t show up in the data for those who regularly attend church. (And data on the non-affiliates show some curious patterns in reporting religion is important.)
As the earlier data showed, there are differences across the four religious traditions but these differences pale in light of the importance of religion to those who attend.
If we take the “very” and “somewhat” options together, the patterns on religious importance for those who regularly attend range from 98% to 99%. On the other hand, the nonattenders show the cultural dynamic of arguing that religion is very important in spite of their non-attendance. (This isn’t an artifact of seniors who simply can’t get out; it cuts across age categories.)
Sometimes it seems that the sociology of religion moves very slowly. It hasn’t been that long ago that we stopped dividing everything into Will Herberg’s Protestant/Catholic/Jew. We understand that there are larger dynamics of religious tradition in play.
But these patterns are clearly mitigated by attendance. We would do far better in understanding the role of religion in postmodern society if we paid more attention to the legitimate faith of those who regularly attend church instead of perpetuating our favorite version of why our particular tribe is better.