Like many people, I found things to like and dislike in Ross Douthat’s piece over the weekend. Reflecting on issues related to the Alabama Senate race, he suggested that we may be approaching an “evangelical crisis” (which he sees as not altogether bad). It’s possible, he writes, that what we’re seeing is White Christian Tribalism.
When I shared his article on Facebook and raised a number of questions about how little we know about “rank and file” evangelicals, as opposed to national and social media thought leaders, it prompted a wonderful dialogue among my friends. I suggested that perhaps what Douthat was getting at was the tremendous overlap between region of the country, rural culture, Republicanism, and Evangelicalism. I represented that idea through the following Venn Diagram (created very inartfully via PowerPoint).
While I don’t swear by the specific location of the circles, it did convey what I was pondering. Baylor’s Elesha Coffman pointed me to county level religious data that was done by the Association of Statisticians of America’s Religious Bodies. Looking closer, I found that the data had been gathered by a friend of mine, Rich Houseal, who serves as lead researcher for the Church of the Nazarene. I further realized that the actual data was available through the Association of Religious Data Archives (a marvelous site).
Naturally, I downloaded the data. Then I found 2010 Census data and matched the Counties so that I could measure the percentage rural within each county. Finally, I located 2008 presidential election results and determined the percentage of each county that voted for John McCain over Barack Obama. Using around 3,000 counties, this would allow a rough analysis of my Venn Diagram. If i was right, there would be correlations between the rate of evangelical membership in the county per 1,000 population, the percent rural, and the percent voting Republican. Furthermore, I expected to find the pattern stronger in the south.
It turns out that the percent rural pretty much washed out in every analysis I did. But the relationship between the evangelicals and voting McCain showed itself to be at least mildly correlated at a rate of .37 (.60 is strong, 1.00 is perfect).
Controlling for region showed some different patterns. Using what the Census department calls East South Central (AL, KY, MS, TN) didn’t change the correlation at all. So I started playing around by looking at the correlations within individual states. The relationship completely disappears in the Mountain region (AZ, CO, ID, MT, NV, NM, UT, WY).
When I did just Alabama, the correlation jumps to .72. That means that counties with high rates of evangelical adherents were more likely to vote for McCain, even in a strong red state. The quarter of Alabama’s 67 counties that had the highest Evangelical rate (over 520 per 1,000) voted Republican at a rate 15% higher than those below 520 (72% compared to 57%).
This is some very preliminary analysis, but it’s yet another instance where i’ve tried to find other rationales for the linkage between evangelicals and being Republican but it’s just really tough to debunk. As Myriam Renaud observed at the time of Trump’s inauguration, this is not new.
The cultural tribalism, at least in some areas seems pretty vibrant. The generational shifts may result in fewer people identifying as evangelicals, as data has regularly shown. But in the short term, that will only increase the homogeneity of the population that continues to identify as evangelical.