Donald Trump has had a rough couple of weeks. He said he supported punishing women for abortion and then walked it back; his campaign manager was arrested on a charge of battery; he retweeted an unflattering picture of his main opponent’s wife, Heidi Cruz.
And the polls show him trailing by a wide margin in Wisconsin, which holds its primary on Tuesday. For some people, it’s a sign that Mr. Trump is finally losing ground.
But his problem in Wisconsin is mainly about the state’s demographics, not self-inflicted wounds. Even a 10-percentage-point loss there wouldn’t necessarily indicate any shift against him.
The state has always looked as if it would be one of Mr. Trump’s worst. This was true even before the primaries began. Polling by Civis Analytics and Marquette Law both showed Mr. Trump faring much worse there than nationally, just as was the case in neighboring Iowa, where he wound up losing to Ted Cruz.
The results of other primaries have consistently pointed toward Mr. Trump’s weakness in Wisconsin. We’ve frequently published articles using models to predict the results of contests based on demographics, and both models — one after the Super Tuesday primaries and one after the March 15 primaries — projected Wisconsin to be one of Mr. Trump’s worst remaining states.
Our estimate remains virtually unchanged from what it was after the Arizona and Utah contests March 22: a five-point edge in Wisconsin for Mr. Cruz, almost exactly aligned with the polls. The model does not account for the possible redistribution of Marco Rubio’s former support.
Why is Wisconsin such a problematic state for Mr. Trump? It’s not the worst possible state for him (Utah was), but it does force him to confront some major weaknesses.
Lower educational attainment among white adults is one of the strongest correlates of Mr. Trump’s support. It’s true at every level: He does better in places where fewer whites have high school diplomas, college degrees or postgraduate degrees. The lack of a high school education is the biggest correlation.
Wisconsin is average or above average in basically every educational category. It’s also unusual because many of its most strongly Republican areas are well-educated suburbs, so more of the Republican vote will be cast in well-educated areas than in, say, California.
Religion and Family
With the exception of white Roman Catholics, Mr. Trump fares worse in areas where larger shares of the population are reported to be religious adherents in the 2010 U.S. Religion census. (He won a majority of the Catholic vote in Massachusetts, and it bodes well for him in New York and New Jersey).
You might not think of Wisconsin as an especially religious state. But it’s a bit above average for religious adherence across all categories: the total religious rate, evangelicals and Catholics. The net effect is slightly negative for Mr. Trump.
He fares worst in areas with strong traditional families, especially ones with more children, more married couples and fewer single mothers. The state also has an above-average number of married couples — and most of the state’s most Republican districts are in the top quartile for marriage rates.
As with education, Wisconsin isn’t at the top of the list for these characteristics. But there aren’t many states that are consistently average to above average in education, religion and marriage rates. Often, they cut in the opposite directions: The most religious states aren’t usually highly educated, and many have lower marriage rates. Where these factors do converge, though, Mr. Trump struggles, as he did in Utah, Kansas and Iowa.
Ancestry and Culture
Iowa, Utah, Kansas — and Wisconsin — have something else in common: a large population who report their ancestry from predominantly Protestant countries in Northern Europe.
These voters represented the base of the Republican Party for the century after the Civil War, whether it’s the old-stock “Yankees” who spread west from New England, or the German, Scandinavian and Dutch immigrants who generally settled over the same stretch of the northern part of the United States later in the 19th century.
These voters are probably the biggest problem for Mr. Trump that you haven’t heard of: He would fare about 30 points worse in counties where all of the white residents reported their ancestry from Protestant countries in Northern Europe than he would in a place where none did, according to our model. It’s the type of thing that helps separate Northern Virginia — where Mr. Trump struggled greatly — from the Boston area, where he excelled.
Exactly why this is so important is a little speculative.
â– They’re the “old Republicans” — the former base of the party. Perhaps they’re likelier to be offended by Mr. Trump’s breaking with traditional Republican ideology. Mr. Trump fares best among new Republicans — including white Southerners and white Catholics — who might have less consistently conservative views.
â– Racism could play a role (Mr. Trump has made appeals to racial, ethnic and religious resentment). Most measures of racism — like Google searches for racially charged terms or scores on the Implicit Association Test — tend to show lower levels of racism in the Upper Midwest, West and rural New England than in the South and industrial North. It’s essentially the same region settled by people of Protestant Northwest European ancestry.
Now, that might just be a coincidence — there aren’t very many minorities in these same regions, since the same economic forces that drew working-class Catholics to manufacturing jobs in the industrial North also drew African-Americans from the South. Or it might represent a deeper cultural tradition: It was the old-stock Yankees who led the fight for abolition; a century later it was roughly the same area that swung against the fiery conservative Barry Goldwater when he fought and won the nomination against more moderate and liberal Republicans.
They’re also the same areas where Barack Obama defeated Hillary Clinton among white voters in the 2008 Democratic primary.
Momentum vs. Demographics
All considered, the model-based estimates make Mr. Trump a five-point underdog, with a 40-35 edge for Mr. Cruz.
Mr. Cruz could certainly outperform this number because the model doesn’t allocate any of the vote for Mr. Rubio (estimated around 8 percent on March 15), who has suspended his campaign. If the Rubio vote were split evenly between Mr. Cruz and John Kasich, which is about what happened in Utah, then Mr. Cruz might win, 44-35, over Mr. Trump.
To get a sense of whether Mr. Trump really lost ground after his controversial remarks, the key number to watch is 35 percent: the expected share of the vote for Mr. Trump based on the results so far. If he falls short, he has lost ground because of momentum, not demographics.