Gilbert: Wisconsin's Donald Trump-Tammy Baldwin split election explained

Senate Democrat Tammy Baldwin won re-election Tuesday because she outperformed her party's standard-bearer, Kamala Harris, across most of Wisconsin, but especially in smaller counties where former President Donald Trump made his biggest gains and in election wards with lower incomes and lower rates of college education.

Does that mean that there were a lot of Baldwin-Trump voters in Trump Country?

"A lot" would be stretching it.

The difference between these two contests in Wisconsin was very small: Republicans won the presidential race here by nine-tenths of a point and Democrats won the Senate race by nine tenths of a point. Fewer than 2 percentage points and 60,000 votes separated the outcomes of these two races.

According to exit polls, 4% of Trump voters in Wisconsin voted for Baldwin, and 3% of voters for Vice President Kamala Harris voted for Republican Senate candidate Eric Hovde. That suggests that there were very few ticket-splitters, but Baldwin got a little more help from them than Hovde did.

Another clue about the two elections is that Hovde got about 54,000 fewer votes than Trump, his party's presidential nominee, while Baldwin exceeded the Harris vote by a much smaller margin. She got about 4,500 more votes statewide than Harris.

That suggests the modest gap between these two races had a lot to do with the small proportion of Trump votes that Hovde didn't get. Where did those Trump votes go?

There are only three possibilities. Some Trump voters crossed over and voted for Baldwin. Some voted for a third-party candidate for Senate (the two non-major party candidates, Phil Anderson and Thomas Leaguer, combined to get about 71,000 votes — 2% of the total). And some Trump voters skipped over the Senate race. There were about 21,000 fewer votes cast for Senate than for President.

We can't tell from the election results, however, what the exact breakdown is between these three groups.But we can say some things about the kinds of places where Baldwin outperformed Harris the most, and where Hovde underperformed Trump the most.

One good source for this is an analysis Marquette Law School fellow John Johnson did in the aftermath of the election, combining demographic data from the census with election returns at the ward level to show how the two parties did in each of these two races across different kinds of communities.

Where did Baldwin do better than Harris?

Baldwin did three points better than Harris in wards (known as election "reporting units" in Wisconsin) with lower rates of college education. She lost these places overall but lost them by a smaller margin than Harris did. By contrast, she did no better than Harris in places with the highest shares of college grads, places that both Democrats won handily.

Income followed a similar pattern. Baldwin out-performed Harris the most in lower-income places and performed no better than Harris in places with the highest incomes, where both Democrats also did well.

Baldwin performed slightly better than Harris in majority-white wards, where the vast majority of Wisconsinites live, and in majority-Black wards.

But her edge over Harris was the biggest in majority Latino wards, which shifted significantly toward Trump in this election. These wards make up a very small share of the Wisconsin vote, however, so they don't account for much of the difference between the two races.

Finally, Baldwin outperformed Harris a little bit more in wards with younger populations. These kinds of places voted heavily Democratic but saw a bigger shift toward Trump (about three points) than wards with older populations.

To generalize about all these patterns, the gap between how Baldwin did and how Harris did was bigger in communities where Trump was either strong or where Trump made his biggest inroads in this election.

And that also rings true when you look at how these two races varied across Wisconsin's 72 counties.

The counties where Baldwin outperformed Harris the most (and where Hovde lagged Trump the most), were small counties like Lafayette, Crawford and Buffalo in western Wisconsin and Ashland and Forrest and Clark in northern Wisconsin.

In these places, the Democratic margin in the Senate race was 5 to 8 points better than the Democratic margin for president. For example, Democrats lost Lafayette by 20 points for president but by just 12 points for Senate.

They lost Crawford County by 14 points for president but just 8 points for Senate. These are also counties where Trump made some of his biggest gains in 2024. Lafayette had the biggest increase in Trump's point margin over 2020 of any Wisconsin county, from just under 14 to just over 20. Crawford had the third biggest increase, from an 8-point edge to a 14-point edge. Both counties voted Democratic for president before 2016.

One possibility in these places is that Trump attracted new voters to the polls or won over some traditional Democrats who weren't as committed to voting Republican down-ballot. In fact, Republican congressman Derrick Van Orden, who won a close re-election fight in western Wisconsin's Third Congressional District, also ran well behind Trump in the region.

The counties where Baldwin did not outperform Harris and Hovde did not lag behind Trump look a lot different. Hovde's margin was actually better than Trump's in just four counties: the Menominee Indian Reservation, and the three suburban "WOW counties" outside Milwaukee: Waukesha, Ozaukee and Washington.

The WOW counties all voted for Trump, but they were among the few places in Wisconsin where Trump did a little worse in 2024 than in 2020. They are also examples of Republican places where Trump has a history of performing worse than previous GOP nominees. Ozaukee and Waukesha contain a lot of suburban communities, with lots of college grads, that have been trending away from the GOP in the Trump Era.

It makes sense that these were the places where a slightly larger share of Republican-leaning voters backed Hovde but would not vote for Trump.

Meanwhile, Trump outperformed Hovde the most in the more rural north, west and center of the state, places where he has fueled the biggest vote shifts in Wisconsin over the past decade, just as he outperformed fellow Republican Ron Johnson in those places when Johnson shared the ballot with Trump in 2016.

The end result of all this was a historically unusual outcome in Wisconsin, different parties winning for president and Senate on the same ballot, which hadn't happened in 56 years.

But there was nothing unusual about the size of the gap between these two races. In fact, it was relatively small in historical terms. Ticket-splitting did not make a big comeback in this election. But when elections are this close, it takes very little ticket-splitting to produce a split outcome.

Wisconsin wasn't the only state this year where the losing party for president was the winning party for Senate. It happened in Michigan, too, where Democrats won the Senate race by less than half a point but lost the presidential race by a little more than a point.

It appears to have happened in Nevada, where it looks like Trump will win for president and Democrats will win for Senate. And it may yet happen, too, in Arizona, where hundreds of thousands of voters are still being counted.

Eva Wen of the Journal Sentinel and Yoonserk Pyun of USA TODAY contributed.

Craig Gilbert provides Wisconsin political analysis as a fellow with Marquette University Law School's Lubar Center for Public Policy Research and Civic Education. Prior to the fellowship, Gilbert reported on politics for 35 years at the Journal Sentinel, the last 25 in its Washington Bureau. His column continues that independent reporting tradition and goes through the established Journal Sentinel editing process. : .