Lessons from Wisconsin’s 2018 Democratic Gubernatorial Primary

Lessons from Wisconsin’s 2018 Democratic Gubernatorial Primary

Tony Evers won, but the detailed results could providence guidance for the 2026 race.

On Tuesday, Wisconsin’s Democrats will choose their nominee to replace Governor Tony Evers. Primaries like this one provide a unique glimpse of divisions within a party, but they don’t happen very often. The last time Democrats held a seriously contested statewide primary was actually the 2018 selection of Tony Evers to challenge then-incumbent Republican Governor Scott Walker.1

The 2026 primary features frontrunner Francesca Hong, a democratic socialist, vs. a field of candidates representing other tendencies in the Democratic party. As of this writing, four candidates remain actively campaigning while a further three have dropped out but will nonetheless remain on the ballot.

The field was even larger in 2018, when 10 candidates appeared on the ballot, only two of whom had withdrawn their campaigns ahead of Election Day. While we wait for the results of the 2026 primary, now is a good time to revisit the results from eight years ago. What lessons does that election hold for Wisconsin’s Democrats today?

A fractured electorate made Evers’ plurality look dominant

After eight years as governor, Evers is now among the best-known politicians in Wisconsin, but he was far from the obvious choice to many Democratic movers and shakers in 2018. His only previous electoral experience was running five times (losing twice) for state superintendent of public instruction. Still, in this role, he was the most prominent elected Democrat in state government, and he drew endorsements from the state’s two former Democratic senators, Russ Feingold and Herb Kohl.

Most people, 58%, voted for someone other than the winner, Evers. However, the non-Evers voters were so fractured that his 42% of the vote represented a dominant victory over any other individual candidate. No single candidate consolidated non-Evers support.

The second-place candidate, Mahlon Mitchell, had the most support from labor unions. He was endorsed by the state AFL-CIO and SEIU, along with AFSCME, the Amalgamated Transit Union and IUOE locals. In the end, he won 16% of the vote.

Kelda Roys was endorsed by many progressive organizations and celebrities, including EMILY’s List, NARAL, NOW, Bon Iver, and Sarah Silverman. She won 13%.

Nobody else cracked 10%, but four more candidates still received at least 5%, or 1 in 20 votes cast. Kathleen Vinehout, a state senator from rural western Wisconsin, won 8%. Mike McCabe, the former director of the Wisconsin Democracy Campaign, took 7%. Matt Flynn, an attorney from Milwaukee who had served as state Democratic Party chair in the early 1980s, ran a campaign largely distinguished by the strength of his opposition to Walker’s then-recently announced Foxconn deal. He won 6%. Paul Soglin, a fixture of Madison politics who at that point had served as mayor for 21 of the preceding 45 years, won 5%.

Rounding out the field, the two withdrawn candidates, Andy Gronik and Dana Wachs, nonetheless received about 1% each. This was still more than the last candidate, a little-known attorney from Racine named Josh Pade. His campaign remained active through Election Day, when he won 0.4%.

Evers won (almost) everywhere, while other candidates were popular in only certain kinds of places

Evers won 67 of the state’s 72 counties. Support for the other candidates often followed regional lines. Vinehout exhibited this tendency to the greatest degree. Despite finishing in a distant fourth, she won a plurality in four counties, all in her rural home region. Mitchell, while taking second place, won a plurality in just one county: Milwaukee. Here, as elsewhere in the state, his support was concentrated among Black voters and in traditional union strongholds.

None of the other candidates won a county, but their support was still often regionally defined. To show this, I’ve created a map with layers for each candidate. Click here to access the interactive version. For each reporting unit, the map plots the votes for a given candidate as a circle whose radius corresponds to the number of votes that candidate received. The map allows you to toggle on-and-off layers for each candidate. When multiple candidates are selected, only the winner of a given ward is shown. Click on any ward to see its complete election results for all candidates.

Below are two screenshots, illustrating these patterns.

In the first map, only the winner of each ward is shown. Evers’ support (red) came from nearly everywhere. Notice the small number of purple dots in the west-central part of the state. These are wards won by Vinehout. Also notice the dark blue circles emerging from southeastern Wisconsin. These are the wards in Milwaukee carried by Mitchell.

The second map shows the results when Evers’ support is removed. Remember that he won only 42% of the vote, so this map actually shows a substantial majority of the votes cast. Here many more patterns are evident.

Vinehout’s territory (purple) grows to include most of western Wisconsin, even stretching all the way into the Madison suburbs. She received scarcely any support in the eastern half of the state, with the interesting exception of Door County’s upper half.

Mitchell drew support (blue) throughout southeastern Wisconsin and the state’s more industrial centers, including Milwaukee and its mostly working-class southern suburbs, Racine, Kenosha, Beloit, Wisconsin Rapids and Superior. This latter city is noteworthy. In Milwaukee, Mitchell’s support strongly correlated with race. He was, after all, the only Black candidate, and he did best in the majority-Black wards on the north side. But fewer than 2% of Superior’s residents were Black in the 2020 census. While the city overall voted for Evers, Mitchell was the second choice in every single Superior ward, a pattern that does not occur anywhere else in northwestern Wisconsin. The most plausible explanation is that his strong union support made the difference in this city, where port-related heavy industry still plays a large role in the economy.

The most liberal candidate in 2018 was Kelda Roys (green). She drew support from a different set of cities, including most of Madison, Appleton, Stevens Point, Wausau, much of Green Bay, and Milwaukee’s wealthiest neighborhoods and suburbs. In a general election, many of these cities in the Fox Valley and north-central Wisconsin are politically quite purple, but judging by the primary results, the Democrats who do live there are among the more liberal in the state.

Mike McCabe (orange), who ran an outsider, populist campaign, was particularly popular in places like La Crosse, Waupaca, Eau Claire and much of the Northwoods. He came in a strong second to Evers in rural northern communities like Minocqua, Park Falls, Ashland, Washburn and Bayfield.

Despite his long mayoral tenure, Paul Soglin (brown) finished lower than second in all but a handful of Madison wards. However, he showed surprising popularity in many surrounding communities, taking second in places like Janesville, Monroe, and Baraboo. Perhaps many of the residents in these towns were former Madisonians with memories of Soglin’s heyday.

Evers’ support was spread across most demographic groups

To better understand the sources of support for each candidate, I allocated the election results into census tracts and used ecological inference to estimate the share of major demographic groups supporting the top candidates. These are just estimates, and there is a fair degree of uncertainty around them. (An important source of this uncertainty is that the model uses statistics for the citizen voting age population in each tract, because we don’t know the demographics of the actual Democratic primary voters themselves.)

For the most part, Evers drew relatively even levels of support across these different categories. I estimate that he won 47% of the vote among those with no