What About White Evangelicals Who Aren't Conservative?

Every once in a while I will get an email from someone who doesn’t really fit the mold. An atheist who has voted for Donald Trump numerous times. A Black Protestant who is increasingly becoming more aligned with conservative politics. And, what feels like an increasing number of white evangelicals who are not so closely aligned with the Republican party. Let me be clear that the amount of communication I get from Bernie Sanders supporting white evangelicals is not huge. The more common sentiment is a white evangelical who sees themselves as politically moderate or really pushed out by the modern MAGA Republican party.

What’s nice about having large datasets is that I can explore these niche-y type of subgroups. Let me describe what I mean with actual numbers. The 2020 Cooperative Election Study has a total of 61,000 respondents. Of that, 11,067 are white self-identified evangelicals. If I subdivide that into five point political ideology (very liberal to very conservative), here is the N size for each subgroup:

• Very Liberal: N = 386, .6% of the sample.

• Liberal: N = 749, 1.2% of the sample.

• Moderate: N = 3219, 5.3% of the sample.

• Conservative: N = 4324, 7.1% of the sample.

• Very Conservative: N = 4443, 7.3% of the sample.

Before we got into the era of ‘big survey data’, the largest publicly available surveys would be 3,000. Which means that a scholar may be lucky to get 50 liberal white evangelicals and another 150 moderate white evangelicals. Neither subgroup is large enough to do much with in terms of breaking it down by other demographic variables (gender, education, age, etc.).

But because we have huge sample sizes, this allows folks like me to get a much sharper picture of groups that aren’t necessarily that large in the general public. So, let’s take advantage of this possibility by taking a look at not conservative white evangelicals.

Just to get a sense of the ‘lay of the land’ here’s the ideological composition of white evangelicals over the last fifteen years or so.

Yeah, there’s not a whole lot to write home about here. There are a bunch of conservative white evangelicals. In 2008, when Barack Obama won the White House, about two-thirds of white evangelicals described themselves as conservative or very conservative. Another quarter said that they were moderate. That leaves a bit less than ten percent who were liberal.

The only thing that is really worth talking about in terms of shifting is the mix between conservative and very conservative. For instance, in 2016, only 16% of white evangelicals said that they were very conservative and another 48% were just conservative. That changed in 2020, when the right wing of white evangelicalism was evenly divided between conservative (33%) and very conservative (34%). I don’t really know what to make of that, by the way. But it’s basically stayed that way for the last four surveys. So, it’s not just a blip.

But I wanted to poke under the hood of those ‘moderate’ evangelicals. There’s a lot of really good political science work about these type of respondents. For instance, 538 published a piece entitled, “The Moderate Middle is a Myth.” Ezra Klein wrote a piece for Vox, “No one’s less moderate than moderates.” I think David Broockman, at UC Berkeley does a good job of making the academic case in this YouTube video, too.

You get the point here - people like to see themselves as these really thoughtful middle of the road voters, when in actuality they lean pretty hard to one side and often take on some pretty extreme positions on certain public opinion issues. Let me just show you the partisan composition of moderate white evangelicals.

Now, this is interesting. In 2008 and 2012, a moderate white evangelical was actually more aligned with the Democrats than they were with the Republicans. It was about a ten point margin for the Democrats. In 2016, the partisan composition of moderate white evangelicals was evenly split - 40% Republicans and Democrats. That was somewhat of the ‘hinge’ moment in this data. In both 2020 and 2022, moderate white evangelicals clearly favored the GOP (by 20 points in 2020 and 13 points in 2022). But in the most recent data from 2023, that Republican advantage narrowed to just seven points. That’s worth watching going into the 2024 election cycle.

But what about actual votes, though? Are white evangelical moderates casting ballots for Republicans in a pattern that differs from all white moderates in the data? I looked at the voting patterns of both of these subgroups over the last four presidential elections to try and figure out how these moderates are making their decisions when they actually cast a ballot.

Okay, so this is a pretty interesting finding that is worth some reflection. In 2008, among moderate white voters, Obama bested McCain by seventeen percentage points (58% vs 41%). But among white evangelical moderates, McCain did eight points better. That becomes a recurring theme in this data. In 2012, white moderates were D+18, while white evangelical moderates were D+6. In 2016, It was D+7 for all white moderates, but Trump got 57% of white evangelical moderates and Clinton only netted 37% of this subgroup. In 2020, the story was largely similar: D+9 among white moderates, R+16 among white evangelical moderates.

Do you see what’s happening here? It used to be the case that the Republicans had a small but noticeable advantage among white evangelical moderates (6-8 percentage points). But in the last two election cycles that jumped to a 15-20 point gap. In other words, white evangelical moderates have become more aligned with the GOP over the last five years. But it’s also the same thing happening with all white moderates too, they were D+17 in 2008. In 2020, it was only D+9.

Is this impacted by church attendance at all? Not really. Here’s what I did - I calculated Trump’s vote share in 2020 among white evangelicals. But I broke the sample down into liberal, moderate, and conservative and then looked at all six levels of church attendance running from never to more than once a week. Church attendance has almost zero explanatory power when it comes to vote choice once you control for ideology.

For instance, Trump got 95% of the votes of conservative white evangelicals who never attend religious services. He got 98% of the votes of white evangelical conservatives who attend church multiple times a week. Among liberal white evangelicals, there’s really no substantive difference across the attendance levels that range from never to weekly attendance. There is a pretty big bump among liberal white evangelicals who attend multiple times a week, but let’s be honest here - that’s almost no one. It’s 129 folks in a sample of 61,000. That’s 0.2%.

Now, the moderate category should show a little variation, right? Maybe white evangelicals who say that they are in the middle of the road ideologically may be more likely to vote for Trump if they attend a lot. But, really, this may be the thing that sticks with me - that line is basically flat. Among white evangelical moderates who never attend religious services, Trump got 64% of their votes. He got 61% among moderates who attend multiple times a week. The overall conclusion from this is simply: once you control for ideology, church attendance doesn’t matter among white evangelicals.

Let me pivot away from just voting now and focus on some public opinion issues that could help us understand if moderate white evangelicals are different than moderate white voters of all religious affiliations. I looked at four areas that seem to be discussed a lot right now - abortion, gender transition for minors, forgiving student loan debt, and a pathway to citizenship for people who have been in the country for a while but have not run afoul of the law.

On abortion, white evangelicals are less supportive than all white voters, even after controlling for ideology. For instance, 90% of white liberal voters support a woman’s right to choose. It’s only 71% of liberal white evangelicals. But that pretty much tracks, in my mind. That’s where the evangelical identity really comes to the surface. On gender identity it’s not such a clean result, though. There’s a huge gap in white liberals and white evangelical liberals. While only 30% of the former group would favor a ban on minors going through a gender transition, it’s 63% of the latter group. But among moderates and conservatives, the gap between those two groups is relatively small.

What is really stunning to me is the bottom two questions about forgiving student loan debt and a pathway to citizen for illegal immigrants. The differences between the entire white sample and the white evangelical subsample is incredibly small and in many times not statistically significant. When it comes to student loan forgiveness and a pathway to citizenship, once you control for ideology, the gaps between those groups are never more than 6-7 percentage points. These are areas in which the evangelical identity is much less impactful than the ideological label.

To reiterate this one more time - politics is the master identity now. Knowing how someone answers the ideology question on a survey is a way more valuable piece of information than anything we can ask about religion. The only exception to that is in the area of abortion. But when it comes to non-culture war issues, there’s just not a whole lot of difference between a white evangelical and a white America once you control for ideology.

Code for this post can be found here.