Yesterday, Democrats ran the table in Wisconsin. They won every statewide race–Governor (+8), Attorney General (+9), Secretary of State (+5), and Treasurer (+4). They flipped both competitive Congressional seats. Rebecca Cooke took down incumbent Derrick Van Orden by a 7-point margin. Mitchell Berman defeated Bryan Steil by 4 points.

I was especially interested in the state legislative races where, for the first time in 16 years, control of both chambers was truly up for grabs. Back in July, I gave myself the task of rating the battleground districts in each chamber. I came up with 14 districts–5 in the senate and 9 in the assembly–which I thought would determine the outcome. Democrats won 13 of them.

In the State Senate, the 17 odd-numbered districts held elections. Democrats needed to flip 2 of them for a majority. They flipped 3–two in the Milwaukee suburbs and one in the Driftless Area. They will have 18 seats to Republicans’ 15 in the next session.

In the State Assembly, Democrats needed 5 flips to win the chamber. Instead, they flipped 8 seats, for a total of 53. The locations of these districts indicate the great reach of this blue wave. Two are in the Milwaukee suburbs, two in the Fox Valley, one in the rural Driftless Are, and three are in the tier of small cities running across the northern half of the state.

It was these last three races that surprised me the most. All of the other seats Democrats flipped were actually places where Kamala Harris won the vote en route to losing the state in 2024. But the three northern districts were Trump seats featuring Republican incumbents. It didn’t end up mattering. In River Falls, along the border with Minnesota, Kevin Knoke defeated Shannon Zimmerman by 7 points, this in a seat that Zimmerman won by 8.5 points two years ago. In Menomonie, near Eau Claire, Jeremiah Frederickson defeated Clint Moses by 7 points, after Moses won in 2024 by 7 points. And in Wausau, John Kroll defeated the incumbent Republican Patrick Snyder by 8 points. Snyder had won the seat by 6 points in 2024.

This was a uniform swing election. Every contested assembly seat saw Democrats improve their margin relative to 2024. The biggest swing came in the 21st district in the Milwaukee County suburbs. Here the Democrat, Oak Creek mayor Daniel Bukiewicz won the seat left open by longtime Republican incumbent Jessie Rodriguez’s retirement by 25 points.

Assembly Democrats Relative to Crowley

Thanks to the 2024 redistricting, the state’s legislative districts closely match the political division of the state itself. The 2024 election returned a split result in Wisconsin, as the state voted for Donald Trump and simultaneously reelected incumbent Democratic Senator Tammy Baldwin. The state’s new assembly districts captured this split, with each candidate converting their respective narrow statewide victories into a majority of the vote in 50 of the state’s 99 districts.

In 2024, actual Democratic candidates only won 45 Assembly seats, meaning they significantly trailed both Tammy Baldwin and Kamala Harris. This time, 38 assembly Democratic candidates ran ahead of Crowley and 52 won less of the vote than he did.

Most incumbent Democrats did better than Crowley; although, 14 of them were still a bit less popular than him. Democrats running in open races or against Republican incumbents tended to be less popular than Crowley.

This time, the blue wave was large enough that small amounts of incumbency advantage were no longer sufficient to save Republicans in marginal seats.

Crowley Performance by Assembly

The rest of this article uses Wisconsin’s 99 assembly districts as a lens through which to understand the gubernatorial election results. It’s more common to do this using the state’s 72 counties, but assembly districts have several advantages for this purpose. There are more of them, so they offer more detail. They are, by definition, about the same size, each holding about 1% of the state’s total population. With caveats, assembly districts tend to follow locally meaningful boundaries. They combine sparsely populated counties into coherent regions. They divide large cities along recognizable neighborhood lines.

Conveniently for elections analysis purposes, the Census Bureau publishes demographic statistics for state legislative districts. For every Assembly district, we know the income, education, race, and age of the citizen, voting age population who lives there.

The next series of graphs uses assembly districts to divide the state into four, equally-populated tiers by income, education, and age. For each quartile, I show the percent of the vote won by the Democratic candidate in the last three elections for governor and president. I use two-party vote, just Democrats and Republicans, to avoid the confusion of third party vote shares, which vary from one election to the next. District boundaries changed twice during this period, so when necessary I reallocated results from former ward boundaries into the current districts based on the share of voters in each overlapping section.

There will be a consistent theme across these charts: Democrats improved in every single kind of district, often by similar amounts. This was a wave election, by any standard you might choose.

by income

Here are the state’s 99 assembly districts, binned into quartiles by median income. Over the last decade, Democratic performance has generally declined in the poorest quarter of the state and increased in the highest quarter. From 2022 to 2026, Democratic performance jumped everywhere. Still, you can see some quite significant longer term trends. In the poorest parts of the state, Crowley did about as well as Evers in 2018. But in the wealthiest fourth of the state, Crowley won 50% of the vote in 2026, compared with Evers’ 43% in 2018.

by education

Education and income go hand-in-hand. This graph bins the state’s assembly districts into quarters based on the percent of adult citizens with a college degree. In the most educated parts of the state, Crowley won 8 percentage points more of the vote in 2026 than Evers did in 2018, while in the least-educated parts of the state, Crowley won 2 points less than Evers eight years ago.

by age

When dividing the state by age, Crowley’s improvement over Evers was lowest in the oldest places and higher in younger districts.

by race

When counting just the adult, citizen population, 91 of Wisconsin’s 99 assembly districts are majority non-Hispanic white, 4 are majority Black, 1 is majority Latino, and three have no majority. All of the non-white majority districts are in Milwaukee. Here we see an interesting divergence in the trends between presidential election and midterms. In presidential years, the share of the vote going to the Democrat has declined in non-majority white districts. The drop has been quite modest among Black voters and much steeper in the state’s sole majority Latino district. However, in the lower-turnout midterm election, the exact opposite has occurred. David Crowley did much better in these non-white districts than Evers, particularly in his 2018 race.

This divergence reflects, I believe, differences in who shows up in a midterm versus a presidential election. With all but tiny Taylor County reporting, turnout in Wisconsin’s 2026 midterm stands at about 2,650,000. That is slightly down from 2022 or 2018, and it is 776,000 fewer voters than showed up in November 2024. I suspect those missing three-quarters of a million voters are mostly disillusioned Trump supporters.

OnOne of the clearest contributors to Trump’s 2024 presidential victory was his newfound support among young voters. Estimates vary, but all agree that, when considered by age, his support grew the most with voters under 30. According to Pew, Trump’s support grew by 4 percentage points among these voters; Catalist put his improvement at 6 points. In Wisconsin, I’ve calculated that Trump’s vote share grew by 3 percentage points in the youngest fifth of the state, while it grew just a single point in every older quintile.

College campus wards offer a useful way to measure a particular slice of the youth vote—students who’ve just entered the electorate. You might expect that Trump’s improvement with young voters was limited to young people who don’t go to college, given the way educational polarization works in the rest of the electorate. But this isn’t what the data show. Trump saw some particularly large gains on certain Wisconsin college campuses.

However, across these campuses, Republican performance has cratered since Trump’s 2024 election. In April 2025, the Republican-endorsed Supreme Court candidate Brad Schimel received 13.6% in the same wards that gave Trump 31.2% earlier in the same school year. This was the lowest share of any conservative candidate for President, Governor, or State Supreme Court Justice since at least 2016. Then, in 2026, the Republican endorsee Maria Lazar set a new record low in college campus vote share for a conservative candidate, with 11.6%.

Turnout fell among college voters in these elections. The campus wards made up 0.91% of the electorate in November 2024, dipping to 0.78% in April 2025 and 0.55% in April 2026. This suggests two plausible interpretations of why conservative support fell so much in those elections. One is that many college voters changed their minds and voted for a liberal court candidate after casting a ballot for Trump in 2024.

The other possibility is that Trump’s college supporters simply didn’t turn out for April elections in the same numbers as their liberal peers. Over the past 10 years, high-turnout elections have resulted in better Republican performances in campus wards, both overall and relative to the rest of the state. Republican reliance on low-propensity voters is notorious, but it’s usually attributed to their strength with less-educated citizens who vote less frequently. The patterns on Wisconsin’s college campuses suggest that, even among currently enrolled college students, those with Republican leanings are just less reliable voters than those with Democratic leanings.

The Most and Least Liberal Campuses

The campuses themselves vary a lot in their political disposition. Some match the stereotype of a leafy left-wing enclave, but others are actually proportionally home to more Republican voters than the rest of their host city.

For this analysis, I define campus wards as those where more than half the population was ages 18-22 in the 2020 census.[i] This includes wards corresponding to 12 universities and one college. Three are private and the rest are members of the UW System. Altogether these wards contain about 1.1% of the state’s total population. Because ward boundaries can change over time, I reallocated election results from old ward boundaries into the current ones based on validated voter records.

Taking the average of the past 12 elections for President, Governor, and State Supreme Court Justice, the most liberal campuses are Lawrence University, UW-Madison, and UW-Milwaukee. Trump won between 12% and 22% in these campuses’ wards in 2024. Lazar won 3%-8% in April 2026.

On average, the most conservative campuses are UW-Platteville, UW-Whitewater, and UW-Green Bay. Here, Trump won 39%-47% in 2024 and Lazar 13%-22% in April 2026.

Trump did especially well across many of these campuses, winning 50% of the vote in the wards around UW-La Crosse and more than 40% in the UW campuses at Oshkosh, River Falls, Whitewater, and Platteville.

Several campuses—even some very liberal ones—are home to noticeably more conservatives than the rest of their home municipalities. This shouldn’t be surprising. Many schools draw a wide mix of students from across the state, and those students reflect—at least to some degree—the political diversity of their homes.

A good example of this is the state’s flagship university, UW-Madison. Among conservatives, Madison is a bit of a punchline. Lee Dreyfus called it “30 square miles surrounded by reality” in his 1978 gubernatorial campaign.

In UW-Madison’s campus wards, Trump received 21% of the vote in 2016, 18% in 2020, and 22% in 2024. But in the rest of the City of Madison, Trump received 15%, 14%, and 15%, respectively. There are wards on the east side of Madison’s isthmus where Trump received fewer than 1 in 20 votes in 2024. But across UW-Madison’s student-occupied wards, more than 1 in 5 voted for Trump.

This solid Trump-voting minority on UW-Madison’s campus has done conservatives little good during April Supreme Court elections. Since 2020, Republican-endorsed Justice candidates have never gotten out of the single digits in the campus wards, actually lagging their performance in the rest of the city.

UW-La Crosse also stands out as a place with a noticeable (and growing) town-gown political divide. In La Crosse, the campus wards have grown much more conservative—giving Trump 35% in 2016, 41% in 2020, and 50% in 2024. These wards have always been more conservative than the rest of the city (at least during presidential elections), but the gap has grown from 2 percentage points in 2016 to 14 in 2024.

The town-gown divide runs in the opposite direction in Appleton, home to Lawrence University, a private liberal arts school and the state’s most Democratic campus. Here, Republican voters are less than half as common as in the rest of the city.

Milwaukee is home to two campuses with somewhat different political personalities.

Voters in the UW-Milwaukee campus wards have grown much more Democratic over the past decade, and they are now less Republican than the city as a whole. Trump won 25% in these wards in 2016, dropping to 22% in 2020 and 19% in 2024. Meanwhile, Trump support in the rest of the city grew from 18% to 20% to 21% across his three campaigns.

Apart from April elections, voters in the Marquette campus wards have consistently voted more conservatively than voters in the rest of Milwaukee. In 2024, this gap surged. Trump won 27% in the Marquette wards in 2016, 23% in 2020, and 33% in 2024.

Implications for November 2026

The basic dynamic in Wisconsin politics is that, while there are more Republicans than Democrats, Democrats dominate among the most reliable voters and Republicans depend on turning out people who only sometimes show up in November. This latter group apparently includes a large cohort of conservative-leaning but politically unengaged college students. To the extent that the past few elections hold any lessons, they might be this: potential conservative voters exist in meaningful numbers on Wisconsin’s college campuses, but Republicans face an uphill battle turning them out.

[i] The census is officially conducted on April 1. By that date in 2020 many college students had been sent home in response to the COVID-19 pandemic. However, the Census Bureau went to considerable effort to count these students where they would’ve been under normal circumstances.

Wisconsin votes a lot. I mean that in two ways: voter turnout is often high and opportunities to vote arrive frequently. In 2024, the state’s voter turnout was the highest in the nation (76.66% of eligible voters). And statewide elections occur almost every year, often more than once. The total number of voters participating in these contests routinely varies by millions, with inescapable consequences for the parties. With maximum turnout, Republicans enjoy a slight edge, but Democrats have opened a seemingly insurmountable advantage in lower-turnout contests.

For the purposes of this analysis, I’m interested in statewide elections when Democrats, or Democratic-endorsed nonpartisan candidates, ran against Republicans, or their endorsees. This includes November general elections, April elections for State Supreme Court Justice or State Superintendent, and February primaries for those same offices. Unlike in partisan primaries, when Democrats and Republicans vote in separate contests, nonpartisan Wisconsin primaries occur only when at least three candidates are on the ballot, and the top two vote-getters both advance to the April general.

These days, the nonpartisan status of statewide April (and February) elections is just a formality. The parties endorse and fund their preferred candidates, and the voters are well aware of even these unofficial candidate affiliations. The correlation between the vote for Wisconsin Supreme Court Justice and the previous presidential election grew from 0.002 in 1978 to 0.99 by 2025.

Since 2020, there have been 11 statewide elections pitting Democrats against Republicans, counting formally nonpartisan races in which the party affiliation of the candidates was clear.

Those elections ranged in turnout from 3.4 million in the November 2024 presidential election to 326 thousand during the February 2021 primary for State Superintendent.

At the Marquette Law School Poll, we draw most of each sample from the list of registered voters in the state, so for those voters, we can see their actual voting history—not who they voted for but if they voted in each election. I identified everyone we’ve interviewed who remained registered to vote in Wisconsin throughout this entire period, 2020–2026. This includes about 78% of current registered voters in the state and 8,275 respondents to our polls.

I used those respondents to measure the party split among voters at every level of engagement, shown in the graph below.

Republicans enjoy a strong advantage in party ID among those who voted in 7 or fewer of the 11 elections for which they were registered. For instance, about 10% of registered voters voted in only 3 elections, most commonly the three November contests during this period. Among these 3-time voters, our poll finds that 50% identify as Republican, 31% Democratic, and 19% Independent.

Democrats have the advantage with those who voted in 8 or more elections, and the advantage grows steadily with frequency of voting. About 5% of these registered voters had a 100% participation rate; among them, 55% were Democrats, 36% Republicans, and just 9% Independents.

Independents are far less likely to vote frequently. These are “pure” Independents who deny leaning toward either party. Among registered voters who vote the least, more than a fifth are pure Independents. That is more than twice the rate of Independents among high frequency voters.

This means that, as the electorate shrinks, it also becomes more predictable and more ideologically extreme. Among those who voted in two or fewer elections—prototypically, someone who only votes for president, if that—45% describe themselves as moderate, 8% very conservative, and 8% very liberal. Contrast that with the ideology of those who voted all 11 times. Just 20% describe themselves as moderate, 15% very conservative, and 26% very liberal.

In other words, during a high-turnout presidential election, the last voters to be convinced to vote often reject both parties along with the terms “liberal” and “conservative.” A successful politician must appeal to these voters to win in Wisconsin. However, in a low-turnout April election, about 90% of the participants identify with one of the parties and few see themselves as “moderate.” Candidates in these elections play to their respective bases because there is little audience for doing anything else.

Of course, knowing turnout alone can only tell you so much. The qualities of the candidates, the national environment, and other dynamics of the race play a role. But the importance of turnout to shaping the outcome can’t be denied. There were 17 individual contests in the 11 statewide elections from 2020 through April 2026. Here are the results of each of those, compared with the total number of voters in each.

Overall, Republican vote share increases with the size of the electorate. It’s no coincidence that Donald Trump’s 2024 victory coincided with record-breaking turnout in the state.

In general, Republicans remain quite competitive whenever turnout is above 2.5 million. They also won the State Treasurer and U.S. Senator races in 2022. However, no Republican or Republican-endorsed candidate has come within 5 points of victory this decade in a race with turnout below 2.5 million.

Things are particularly dire for Republicans in State Superintendent primaries, when voter turnout can fall below half a million. In 2021 and 2025, multiple Democratic-aligned candidates ran, receiving about two-thirds of the vote combined. Although the single Republican-endorsed candidate in those races still managed to advance to the general election, it’s conceivable that a future February primary could result in two Democratic-aligned State Superintendent candidates advancing to the general.

Given the way that low turnout now works to the benefit of Democrats while Republicans need high turnout to win, it may appear surprising that Democrats remain the party promoting voting access while Republicans pursue efforts that would shrink the electorate. It can seem as if each party is accidentally working to make things better for the other. A few long-standing political beliefs contribute. Many Democrats believe that higher turnout is normatively good, and the Obama-era mindset that “When everyone votes, we win” is still common. Likewise, I suspect many Republicans still believe, despite evidence to the contrary, that the marginal voter is more likely to vote for a Democrat than a Republican.