". . . and having done all . . . stand firm." Eph. 6:13

Newsletter

The News You Need

Subscribe to The Washington Stand

X

Study Showing Evangelical Voting Wave for Dems Has Serious Methodological Flaws

Article banner image
Print Icon
November 20, 2026
News Analysis

Classic American novelist Mark Twain first popularized the maxim that there are three kinds of liars — “Liars, Damned Liars and Statistics” — but scholars are uncertain if the origin was British Parliamentarian Arthur Balfour or British Prime Minister Benjamin Disraeli.

Whoever said it first, the three liars’ saying clearly has application to a new voting data analysis claiming to prove that the November 3 blue wave from these midterms was greatly helped by a flood of normally Republican-voting evangelical Christians who instead crossed over to the Democratic Party.

“The more Christian and Protestant a county was, the more it moved toward Democrats. In the most Christian 5% of counties, Democrats gained 20.5 points on [former Vice President Kamala] Harris’s 2024 margin, and 22.9 points in the 10 most Christian counties with 5,000+ voters, against a national average of 14.5. That’s more than twice the 9.4-point gain in the least Christian 5%. Every one of the 158 most evangelical counties moved toward Democrats,” claims the study entitled “Christians 2026: 2026 Dem Movement by County.” The analysis also claimed that “in the 208 counties where evangelicals are a majority of residents, all 208 moved toward Democrats, with an average gain of 16.1 points, 11% more than the 14.5-point national average. They are in 20 states.” 

The study was produced by Evangelicals for a Better Future (EBF), a political action committee (PAC) that contributed more than $1 million in the form of independent expenditures to the successful campaign by Senator-elect James Talarico, who is set to become the first Democrat representing Texas in the U.S. Senate since 1993.

“This study demonstrates that Christian voters are increasingly swing voters. This isn’t about a re-emergence of the Christian Left. It’s about the emergence of a truly Christian middle,” Lee Scott told The Christian Post, which identified him as EBF’s director and a Presbyterian pastor.

The EBF study is likely to receive significant coverage in the mainstream media, as political observers, journalists covering campaigns, and academic figures seek to understand why the American electorate went from overwhelmingly electing Donald Trump president to a second term two years ago and giving him Republican majorities in Congress to handing Democrats their biggest congressional majorities in decades in 2026.

It is not evident, for example, how the EBF figures can be reconciled with reporting by The Washington Stand’s Joshua Arnold showing evangelical voters’ share of the overall American electorate plummeted from 23% in 2024 to 17% in 2026.

There are also significant methodological questions that call into question the accuracy of the EBF study, according to artificial intelligence analyses sought by TWS.

“First, the obvious correlation vs. causation questions matter. The vote count in states with higher religious scores and increased Democratic votes this election does not mean that religion had anything to do with the increased Dem voting. It may be that every single religious person voted exactly like they did in previous elections and persons with no religious inclination voted more Democratic. In that case, the religious scoring is not even relevant,” economist William Beach told TWS.

Before his appointment as commissioner of the U.S. Bureau of Labor Statistics by President Donald Trump, Beach served as chief economist for the Senate Budget Committee, and before that he was the founder and longtime director of the Center for Data Analysis (CDA) at the Heritage Foundation. He is now executive director of the Fiscal Lab on Capitol Hill. His PhD in Economics was awarded in 2014 by the University of Buckingham.

“Second, there’s the compositional issue. Even if religious voters turned more to Democrats than Republicans this election, that turn does not mean there was a religious root to the voting decision. In fact, religiously inclined voters may also be ones that suffer more from the slowdown in wages and the increase in prices,” Beach continued.

“Third, there are source data problems. Scores of religious attendance or practices or population are almost always based on surveys and not population counts. These surveys are notorious for having issues associated with representation. In other words, the sample is not drawn from a reliable census of religious people (one does not exist, as far as I know) and instead use diocese counts, estimates of large Protestant denominations, and other fragmented sources,” he went on.

“We thus don’t have a complete count or framework from which to draw the sample, which means we do not know if the estimates made based on the sample are representative of the population. In short, we really have statistical trouble definitively stating that one state is more religious than another. Unless you can do that, you can’t claim that Democrats enjoyed a shift in voters in religious states,” Beach added.

A TWS analysis of the EBF study turned up additional problems, including confusion in identifying who within a voting jurisdiction are credibly classified as “Christian” in any meaningful sense and knowing how each of such individuals actually voted. The EBF study relied primarily on the 2020 U.S. Religion Survey, which is based on identification of people connected to a congregation within a county.

There are multiple reasons that approach is problematic in claiming, as the EBF study does, that some percentage of “Christian” voters voted for a particular candidate or party. For example, there are followers of Jesus Christ who are accurately defined as Christians, but who for any number of reasons are not connected with a particular congregation at the time of a survey. In such cases, they would not be included in the overall “Christian” tabulation, even though they may well have voted.

Religious identity surveys also suffer from a problem like that which plagues Uniform Crime Statistics compiled by the Department of Justice (DOJ). Just as different law enforcement agencies in cities and counties vary in how they define crime categories, thus making comparisons across jurisdictions difficult. The government has guidelines to address this problem, but it persists.

Similarly, there is wide variance in how congregations report attendance and membership among denominations and independent churches that may or may not be evident to data analysts. There is also a further complication in the lack of distinctions within religious surveys to account for members who are only nominally connected to a specific congregation.

This flaw is especially relevant to the EBF’s basic claim that “Christians moved toward Democrats.” It would be somewhat more accurate, though not conclusive, to say “some counties with higher reported congregational membership totals may correlate” with particular voting outcomes. Even so, there remains the fundamental obstacle of not being able to link specific individuals with a concrete religious affiliation with the actual votes they cast.

Open Secrets, which tracks campaign contributions and organizations, describes EBF as “Democratic/Liberal,” and lists Christopher Koob as the group’s treasurer. Koob is a political compliance professional with more than 20 years of experience in campaigns and political committees at the federal, state, and local levels, according to MBA Consulting, of which he is a principal. No EBF officers are named on the group’s website, and the study authors are also absent.

The EBF is registered with the Federal Election Commission (FEC) as a hybrid PAC from September 2024 and given the ID COO887968. There is no known street address for the organization, according to the FEC records. Those records only indicate a Washinton, D.C. post office box for the organization. The EBF raised more than $1.6 million during the 2026 midterm election campaign.

When TWS shared its findings with EBF’s Director Scott, he responded thoughtfully, including an acknowledgement of the validity of at least one of the TWS criticisms.

“Thanks for digging in. Your numbers match ours and we published all the data to allow for these conversations. So, thank you. You are right that ‘church attending’ is a generalized label. A more precise one is members of Christian congregations as a share of residents, which is also why there are 30 small counties where members outnumber residents from the tiers. We flagged this in our Q&A section, and the 30 counties don’t meaningfully shift the outcomes, but for readers who want to dig in, understanding precisely what is being measured and counted is helpful. So we welcome your clarifying that point,” Scott said.

Scott disagreed, though, with the suggestion that county results should be compared with results from the same state as a control factor in the analysis.

“The most Christian quarter of states moved 1.7 times as much as the least Christian quarter (+18.1 vs. +10.7). And the top 5% of counties hold under 1% of the population, so 98% of Christians live in the ‘rest of the state’ you’d be subtracting. If Christians moved across Texas, for example, a Texas benchmark absorbs the vast majority of that movement, and comparing the top tier with it can’t detect it. So, the control isn’t neutral or independent of what we are measuring,” Scott explained.

He suggested an alternative approach to a statistical control, pointing to comparing “each county with counties in the same state that voted similarly in 2024. On that test, the most Christian 5% (not counting the small ones from the 30 with more members than residents) moved 5 points more than their look-alikes, and the least Christian 5% moved 2.8 points less.”

Finally, Scott pointed out that “where turnout fell by about a third, Democrats got more votes than Harris in nearly half of America’s most Christian counties. That’s 2.4 times as often as in the least Christian counties. Where there was a Senate race, Democrats got 101% of Harris’s votes in the most Christian counties, against 83% in the least Christian ones.” 

The TWS research for this report included assessments by artificial intelligence of the EBF study’s methodological approach.

Support the work of TWS with a gift to FRC

Mark Tapscott
Mark Tapscott is senior congressional analyst at The Washington Stand.


RELATED



Support the work of TWS with a gift to FRC