How Will the Faithful Vote in 2024?

Well, tomorrow’s the day. Thank God. As Taylor Schumann tweeted recently, “The human body was not built to withstand 3 Trump election cycles.”

The weeks before an election are always some of the busiest of my life. In an average year, I do about 80-100 media interviews. In the run-up to a presidential election, I’ve done three or four in a day. Everyone is looking for another angle on how to cover the same event, and many reporters think that religion and politics is a good topic for a short story.

Here’s a behind the scenes bit of information: I don’t have any raw data at my disposal about how religious groups are intending to vote in the 2024 presidential election. However, I do have a way to back into some information about how things should shake out on November 5th.

For those who have been following along, you know that one of my favorite data sources is the Cooperative Election Study. It’s fielded every year and they have a huge sample. Well, it’s done in a split design in election years with one part being fielded in early October and the rest after the election in November. Well, Brian Schaffner and team managed to do so quick analysis of that October data collection. Their write up of a sample of 78,247 is here.

But they also provided a Shiny Application that shows us some preliminary crosstabs of about 15 different demographics. You can access that with this link. So, here’s what I did - I just replicated anything in that Shiny app that pertains to religion with the data from the prior 4 elections from the CES. For those who are wondering, I am using their ‘likely voter model’ percentages for 2024 and in 2016 and 2020, I am using validated voter data. The 2008 and 2012 datasets did not have voter validation, so I am just showing you the normal survey weights.

Let’s start with a broad look at six religious categories, and this is the vote share for the Republican in the last five election cycles.

Nothing really jumps off the screen right now, which, from a pure data perspective is a good thing. We should always assume relative stability in data from year to year and it’s reassuring to see that there’s nothing incredibly weird in what is happening in 2024. The Protestant share is rock steady and so is the Catholic trend line. Jews have fluctuated a little bit over time but that’s probably due to small sample size. It’s fair to say that about 30% of Jews have voted for the GOP the last five elections.

The bottom row are the three types of non-religious voters, and again, stability is the norm here. I would not be surprised in the slightest if the atheist/agnostic vote in 2024 is exactly the same as it was in 2020. However, if there’s anything worth flagging in this entire graph, it’s the bottom right. The share of Americans who claim no religion in particular is huge (about 23% of the sample in recent years), and therefore any little wiggle there can have huge implications. The fact that Trump has gained 3 points with this group could be consequential.

The real issue though is that nothing in particulars are very low engagement folks. About a third of them have a high school diploma or less and make no more than $50,000/year as a household. For comparison, it’s only 12% of atheists/agnostics. Thus, their voter turnout is low. In 2022, Pew found that just 32% of them voted, compared to 50% of atheists. So, that three percent will only matter if those folks actually show up on election day.

Here’s the same analysis, but focusing only on the white sample.

Okay, now we have something to really think about. The numbers for Protestants and Catholics are pretty catastrophic if you would like to see Donald Trump in the Oval Office in January. According to this data, Trump has lost five points among white Protestants and and five points among white Catholics. That translates to a lot of votes on November 5th, especially in the ‘Blue Wall’ states, Michigan, Wisconsin, and Pennsylvania. If this actually bears out tomorrow, I don’t see a path to victory for Trump.

However, think about this result in comparison to the prior graph where there was basically no movement among all Protestants and all Catholics. The only way for the share of white Christians voting for Trump to drop, but this impact the numbers when looking at all races is that a lot of non-white Christians have shifted right in the last four years. And that could be the biggest story tomorrow.

What else jumps out to me on this one? Eh, not a whole lot. Atheist and agnostics are basically in the same spot in 2024 as they were four years ago. However, those nothing in particular voters are not a good sign for Trump. This has him back to 42% with these folks and that was where Obama was during his reelection campaign in 2012. But, turnout is the key. They have to show up to actually make a difference.

The Shiny App displaying the 2024 data also included a crosstab based on self-identified evangelical/born-again status. It’s a simple yes or no question asked to every respondent.

This data hints at a shift that could be really consequential for future students of religion and politics. Trump’s share among folks who self-ID as evangelical dropped from 70% in 2020 to an estimated 66% in 2024. In the most recent data from the CES, about a third of the sample says that they are evangelical. If this movement is real, that’s about 1.3% of the population that is not supporting Trump this time around. However, what cushions that is the fact that he may be gaining ground with folks who say that they aren’t born-again. This data has him getting two points back. Which is, ironically, also 1.3% of the population. So, maybe this is a wash in terms of actual number of votes on November 5th.

There’s one last graph from the Shiny App that I wanted to replicate. It was a crosstab of: evangelical identification, gender, and education. This is the white only sample. I replicated the categories that were displayed from 2024 in the raw data from the prior four election cycles. Here, there are some striking figures.

The top row really has me scratching my head. It’s men who self-identify as evangelical, with or without a college degree. This indicates among evangelical men who graduated college, Trump has experienced a huge loss in 2024. He went from ~75% support in 2016 and 2020 to just 63% support in the 2024 data. And these are voters who will turnout at a very high rate tomorrow. But then, the trend line among born-again men without a degree is sharply in the other direction. He went from 70% in 2016 to 74% in 2020 to 81% in 2024. That’s a pretty believable trajectory, honestly. So, these two figures may offset each other.

You can also see this same trajectory among evangelical women without a college degree, too. Among this type of voter, McCain only earned 59% of their votes in 2008 and Romney did about the same in 2012. Trump was at 64% among evangelical women with low education in 2016, but then it rose to 68% in 2020 and the current estimate is 79% in 2024. A 15 point gain in just eight years. Unbelievable.

What about the non-evangelical part of the sample? There are a few little data points that seem relevant. For instance, this data says that among non-evangelical women with a college degree that Trump is going to get 31% of their votes. That would be much higher than 2020 (24%) and back to the level that Romney received in 2012. I don’t really know why that would be the case, but maybe inflation is enough to move the needle for social moderates in this category? And maybe it’s just a blip. But the data also says that Trump is making big gains among non-evangelical women without a degree, too. This indicates he’s at 46% in 2024 when he only earned 39% of their votes in 2020.

Let me show you one more bit of analysis, but this time I am pulling in data from the Pew Research Center from a poll that they conducted in late August and early September of 2024. It allows me to show you a couple religious groups that weren’t available above.

They report that about 82% of white evangelicals will vote for Donald Trump tomorrow. That’s pretty much in line with what he got in 2016 and I don’t think there’s anything worth thinking about too much there. However, the data about white non-evangelicals is noteworthy. This trend line is clearly pointing toward the fact that white Christians (regardless of evangelical status) are continuing to move toward the GOP. Just 49% of this group voted for McCain in 2008 and Trump will possibly do ten points better in 2024.

That same general trend is there among white Catholics, too. They are slowly but surely becoming a clear base of support for Republicans. In 2008, 56% of them voted for McCain and Trump has gotten at least 60% of their votes in the last three cycles. Among Hispanics, I don’t think this data drives home any clear narratives about whether they are clearly moving to the right or not. It’s still the case that Hispanic Catholics are 2 to 1 for the Democrats in the voting booth.

The full dataset from the Cooperative Election Study will become available to the general public sometime in March or April of 2025. When that happens, you better believe that this newsletter will be filled with a lot of content about what in the world happened in the 2024 presidential election.

In the next 24 hours, my only advice is a (possibly apocryphal) quote from St. Augustine, “Pray as though everything depends on God. Work as though everything depends on you.”

And let’s hope our democracy endures.

Code for this post can be found here.