In the wake of an election, large national news organizations rush to make all sorts of pronouncements about demographics. Slicing up the population by race, gender, religion and other factors, they claim to know how various blocs voted.
Though this might meet a voracious desire among audiences to pin credit or blame on certain groups, it’s filled with pitfalls.
In the latest episode of my podcast and newsletter They Stand Corrected, I explain this mess. These demographic claims are based on exit polls, which are notoriously problematic. News agencies know this. At least, some people inside news agencies do.
Nevertheless, these same agencies inundate you with “news” that might turn out to be unfounded.
Credit: Handout
Credit: Handout
In 2016, after Hillary Rodham Clinton lost to Donald Trump, an exit poll claimed that 52% of white women had voted for Trump. That idea spread so quickly that it became part of established lore. However, it was most likely false, as later analyses of that exit poll showed problems with it.
A more robust survey found that Clinton and Trump virtually tied among white women, with neither getting a majority. (It found Trump ahead by two points, but in a poll of that size, that’s basically a tie.) The idea that so many white women voted for Trump still shocked some people, of course.
Four years ago, after President Joe Biden beat Trump, the Washington Post ran a piece with this headline: “Don’t trust the exit polls. This explains why.”
But, as with so many things, large parts of the media don’t seem to learn their lesson.
Now, the Washington Post has a big interactive spread filled with data about how various demographics are purported to have voted this year. Nowhere is there a caveat about previous exit polls.
If you scroll all the way down, you can find a standard line saying, “Like other surveys, exit polls are subject to sampling error and other sources of error.” But that’s after readers have been fed hundreds of data points.
In addition, the polls themselves have some substantially different findings. There are only two that most news agencies use. One, from Edison Research, is backed by a consortium of several major TV networks; the other is from something called AP VoteCast. People behind the latter, incidentally, say that it’s not technically an exit poll, although it “serves a similar purpose.”
For my latest episode, I looked into various reports about how Black Americans allegedly voted this year.
In a live segment, CNN (where I used to be on-air fact checker) said that among this demographic, Trump had the “best performance for a Republican candidate for president in 48 years.” But a written CNN report says, “CNN’s exit polls showed just a two-point shift among Black men toward Trump nationally.” A two-point shift in an exit poll is not even a clear shift at all. Meanwhile, the AP said 16% of Black voters supported Trump in 2024, up from 8% in 2020.
But wait! A Washington Post report said there was no significant change.
“Roughly 9 in 10 Black women backed Harris and Biden, respectively. Although Democratic Party leaders expressed concern about Harris’s preelection poll numbers among Black men, this group continued to support Harris at similar levels when compared with previous presidential elections in both 2024 and 2020.”
So which is it? We don’t know. That’s what the media should tell you.
In my podcast, I report on the rise of antisemitism on both the right and the left. So some listeners asked me about the votes of Jewish people in this election — even though, as just over 2% of the population, we don’t have enough impact to swing any election.
I looked at news reports citing the two different surveys. One says 79% of Jewish voters supported Vice President Kamala Harris; another other says two-thirds. And despite these substantially different findings, reports about both surveys claim that it’s the same as 2020.
As for Latinos, yes, Trump gained support. But we don’t really know how much.
The two surveys found different figures, although both said Harris won a majority of Latino voters. And some exit polls suggested Black and Latino turnout might have been lower this time, which introduces even more questions.
If reading all this makes your head feel like a fruit salad jumble of numbers, I don’t blame you.
Here’s the upshot: Demographic claims, at this stage, are premature. Trump won, but that’s pretty much all we know for now. And, as California continues to count votes, we don’t even know what the final popular vote will be.
When emotions are high, it can be especially difficult to wait for trustworthy information. But understanding what really happened in an election takes time. Making claims that might not hold up doesn’t help anyone.
Josh Levs is host of They Stand Corrected, the podcast and newsletter fact-checking the media. Find him at joshlevs.com.
About the Author
