The Rise of the Non-Christian Evangelical
Some Jews, Muslims, Hindus, and Buddhists are Identifying as Evangelicals - Why?
I teach a graduate course in research methods every Fall semester at EIU. It’s always on Wednesday nights and it lasts for two and a half hours. It’s an absolute slog. I tell the students on the first day of class that this may be the most difficult course that they have ever taken. But, they will probably learn more skills in research methods than all their other classes.
We spend a lot of class time talking about research design - which is the framework that social scientists use to help them answer questions in an empirically rigorous way. I dedicate an entire class session to the challenges of measurement. Students will come to me with a research idea like, “Do more experienced bureaucrats lead to a more effective government?” I then ask them to define, “effective” and they inevitably stammer and stutter, not entirely sure how to proceed.
One exercise that I have them talk through is just a very simple question that is asked on almost every survey: Thinking back over the last year, what was your family’s income? Seems pretty straightforward, right? Well, there are dozens of ways that this single question can induce error into a research design.
A person could make money doing something illegal and they don’t want to report that. Someone may not know how much money their spouse makes and has to guess. A respondent may only report their personal income, not the household’s income. They could just check the wrong box on the web browser when they are taking the survey. There are so many ways that this simple question can get messed up.
Now, imagine if that question was a little less straightforward. Like this one, “Would you describe yourself as a ‘born-again’ or evangelical Christian, or not?” Pew Research first devised that question and it appears on lots of surveys now. But here’s what is really interesting about that question - no matter how one responded to the prior question about religious affiliation, they are still asked if they are evangelical or not. That means that atheists are asked about their born-again status and so are Muslims.
The typical question about religious tradition has about a dozen response options, here’s the share of each of those traditions who answered the question about evangelical identification in the affirmative. The data comes from the Cooperative Election Study in 2022.
There’s only one tradition where a majority say that they are born-again - Protestants. Which makes sense, right? The next highest is “something else.” I can tell you what’s happening there. A lot of people don’t know that they are Protestant. They know that they are Baptist, but that’s not a top level option. As a result, they select "something else" and provide their denomination in a free response.
A quarter of Orthodox Christians say that they are evangelicals, and it’s about the same share of Latter-day Saints. Then, things get really weird. Fourteen percent of Muslims say that they are evangelical - it’s the same percentage as Roman Catholics. Even 12% of those identifying as nothing in particulars and Hindus say that they are evangelicals!
Also, note that at the bottom that about one percent of atheists/agnostics say that they are evangelicals. People see those results and scoff, but for a guy who looks at survey data every day, that’s actually a very good sign. People are imperfect and offer contradictory answers on a regular basis. If I ever see a percentage on a crosstab that is 100% or 0%, I get nervous. And it makes me think that the data is cooked somehow.
Okay, I know what you are thinking. This guy has a crazy survey that makes no sense. The instrument must be flawed, right? Evangelical Catholics can’t exist. Well, let’s just check with another instrument. This time it’s the Nationscape Survey, which had a total sample size of over 477,000 respondents. They use basically the same setup as the CES.
Okay, those non-Christian evangelicals are still here. And these numbers don’t differ that much from one survey to another. For instance, 23% of Mormons were evangelicals in the CES, it’s 28% in Nationscape. The Orthodox number is statistically the same in both, about 25%. Let’s look at some other groups:
CES vs Nationscape - Share Self-ID’ing as evangelical
Hindu: 12% vs 20%
Catholic: 14% vs 18%
Muslim: 14% vs 20%
Buddhist: 9% vs 13%
Jewish: 5% vs 11%
It’s pretty clear that the Nationscape does manage to have a higher percentage of self-identified evangelicals in the sample compared to the Cooperative Election Study. However, the gaps aren’t huge. The biggest one is Hindus at 8 points, but many of the differences are in the 5-point range.
This gives us quite a bit more confidence in the share of these non-Christian groups that identify as evangelical. If there’s a massive amount of survey error, then these numbers would be nowhere close to each other. Many of the gaps in the estimates of self-identified evangelicals could be explained by something as simple as survey weighting differences or random chance.
But, there’s one more way to validate this - let’s look at several waves of the Cooperative Election Study. If there was a lot of survey nonsense happening, then the estimate of 2016 should be way different from the estimate of 2020. We would likely see a see-saw pattern in the data. There would be no discernible trend lines - just a whole lot of noise.
I am showing you the results from 2012, 2016, 2020, and 2022. The sample sizes in those years were: 54K, 65K, 61K, and 60K. This is not your average horse race poll of 1000 people. These are huge instruments and that’s why the error bars visualized are so small. The smallest sample size is among Hindus in 2012, with 93 individuals, followed by Muslims in 2012 with 171. The median bar graph visualized above has a sample of 1880 folks. So, the small N argument just doesn’t apply here.
And notice that many of these estimates over the last decade are not that noisy. The evangelical Catholic share is steadily rising - that’s also the case for Mormons, Orthodox Christians, and Jews. There is a bit of a seesaw pattern for Muslims - 17% in 2016, 12% in 2020, 19% in 2022. But the numbers for Hindus and Buddhists are remarkably consistent.
I have validated this data in two ways now - across two separate survey instruments that were administered by completely different organizations but also across time in the Cooperative Election Study. There’s very little evidence here that the phenomenon of non-Christian evangelicals is a result of survey error. The results are just way too consistent for that.
But, here’s another way to parse through a possible case of survey error. If Jewish evangelicalism is nothing more than a bunch of respondents just mashing random buttons to get through the instrument and get paid, then we should not be able to reliably predict what would make one more likely to self identify as an evangelical. There would be no rhyme or reason behind it.
I can, though. What variable in a survey would be most predictive of being an evangelical? The one that should be jumping out to you now is partisanship. Republicans are more likely to be evangelicals than Democrats. I don’t think I need to go into a long discourse about why this is the case.
So, let’s just test that simple theory out. Are non-Christian Republicans more likely to self-identify as evangelical compared to non-Christian Democrats? The answer from the Cooperative Election Study is below. I binned together the last five years of the survey, by the way. That’s so the sample size can be larger (N = 224,700) and our confidence intervals will tighten up.
The top row shows our four Christian groups. It’s pretty clear that Protestant Republicans are more likely to say that they are evangelicals compared to Democrats (69% vs 45%). But for the other three groups, the differences really aren’t that big. Three points for Catholics and LDS, and six points for Orthodox Christians. But, in each case the Republican share is higher.
But look at the non-Christian groups. Nine percent of Republican Jews self-identify as evangelical, compared to 3% of Democratic Jews. For Muslims, the gap is huge: 32% vs 11%. It’s also fairly large for Buddhists (16% vs 6%) and Hindus (18% vs 10%). You can even see it among nothing in particulars. 19% of the Republicans are evangelicals; it’s just 9% of the Democrats.
However, I can also test that same hypothesis with the Nationscape survey. Remember, this one is over 477,000 respondents. It uses nearly the same setup as the CES.
What we find in the Nationscape is very similar to what I just described from the CES. Republican Protestants are more likely to be evangelical, but the partisan gaps for the other Christian groups are a lot smaller. It’s notable here that Democrat Latter-day Saints are more likely to be evangelical. I don’t know what to do with that one.
But look at the non-Christian groups. A Republican Jew is about three times more likely to self-identify as evangelical compared to a Democratic Jew. It’s the same ratio among Buddhists. One third of Republican Muslims say that they are evangelicals; it’s 16% of Muslim Democrats.
If you look at the last two graphs for Jews, Muslims, Buddhists, and Hindus - in every case a Republican respondent is more likely to say that they are born-again compared to a Democrat. Could all four groups across these two separate surveys have measurement errors that are all in the same direction? Sure, it’s certainly possible. But that’s certainly not very plausible.
There’s a concept in science called Occam’s Razor. Simply put, “the simplest explanation is usually the best one.” Maybe all these graphs I produced are the result of nothing more than simple survey error. Or maybe the data is suggesting that the word evangelical doesn’t mean what it used to mean.
Code for this post can be found here.
Remember that Jewishness is beyond our category of "religion," which is a product of post- Reformation Europe. "Jewish" is an identity category with elements of religion, culture, nationality, ethnicity, and extended family. So there may be people who identify as Jews (by culture, ethnicity, etc.) but who are also straightforwardly Evangelical Christians by religion. Such people may have picked one response to one question and another for the other. I'm not saying that explains all such responses, but it may explain some. (More, though, all of this reminds us indeed that large datasets always contain the weird. I remember when Pew's "Portrait of Jewish Americans" came out in 2013, https://www.pewresearch.org/religion/2013/10/01/jewish-american-beliefs-attitudes-culture-survey/ they found a non-zero number of people who identified as "ultra-Orthodox" Jews but reported that they have a Christmas tree in December! If you know anything about Haredi Judaism, you know that's hilarious.)
The obvious question would be the definition. For many people, evangelical just means evangelizing, spreading the faith. Mormons and Muslims are supposed to spread the faith, while Judaism is a closed and locked club. Atheists are faith-speaders, but they also know the current political meaning of the word (evangelical = TRUMPUTIN) so they wouldn't choose the term.