Democratic state assembly candidates were less popular than either Kamala Harris or Tammy Baldwin this past November. Of the 99 Assembly seats, Baldwin won 50, Harris 49, and actual Democratic candidates 45.
This, in itself, was not a surprise. We know that incumbents usually enjoy a slight popularity boost, and there are more incumbent Republicans than Democrats. The 2024 assembly races featured 57 incumbent Republicans and 27 Democrats in contested races.
But does incumbency advantage explain all of the Republican assembly majority? A careful accounting of the evidence suggests not.
Subscribe to John’s Blue Book newsletter to receive his blog posts in your inbox.
Measuring two kinds of incumbency advantage
The 2024 election is the first to use maps drawn by Democratic Governor Tony Evers, rather than previous maps drawn by Republican legislators. When the new state assembly maps were first released, I estimated that the 2022 assembly elections would’ve resulted in 46 Democratic seats, had the election taken place under the new maps. This, despite the fact that Governor Evers won reelection in that year by 3 percentage points.
Incumbency advantage comes from at least two sources. In places where they’ve run before, incumbents likely enjoy higher name recognition than an opponent. Apart from that, they might still run better campaigns simply by dint of greater experience.
Redistricting offers a unique opportunity to decompose these two sources of incumbent strength. Most incumbents ran in new districts which only partially overlapped with their previous seat. The benefits of name recognition should primarily exist just in those overlapping areas, while the benefits of experience should appear everywhere.
To measure this, I created a statistical model predicting the assembly vote in each ward based on the following variables:
Donald Trump is the most famous person in, arguably, the world. His name was googled more than any other in 2024, 2023, 2021, and 2019. To better understand how American adults feel about this ubiquitous figure, we asked each of the 1,063 respondents in the Marquette Law School Poll national survey of adults, December 2-11, 2024, to answer two questions in their own words.
What do you like about Donald Trump?
What do you dislike about Donald Trump?
The order of the two questions was randomized. Since the survey was conducted online, respondents could write as much as they wished. (These “open-ended questions” were part of a more traditional survey, the subject of separate news releases.)
In the 2024 election, 36% of our sample voted for Kamala Harris, 38% for Donald Trump, 3% for a third-party candidate, and 23% didn’t vote. The sample gets the mix of Harris and Trump supporters right, though it overrepresents voters as a whole. Early estimates suggest that about 36% of eligible voters didn’t participate in the 2024 election.
We classified each of the respondents by whether or not they answered both questions. A majority—51% of respondents—listed at least one thing they liked and disliked about Trump. (We classified clearly sarcastic responses as non-answers.) Listing things they dislike about Trump but including nothing positive were 35%. Fewer adults, 12%, listed positive things about Trump but nothing negative.
Perceptions of Trump and vote choice
The table below shows how each of those groups voted in 2024. Trump and Harris each won equal shares of the adults (76%) with, respectively, only positive or only negative views of the former president. Among those listing both likes and dislikes of Trump, 55% voted for Trump, 17% for Harris, 4% for a third party, and 24% did not vote.
In other words, just over a third of adults held wholly negative views of Trump and just over 10% held wholly positive views. Their votes reflect those views. But Trump won the lion’s share of the vote among people with mixed views of him.
2024 vote of U.S. adults by their view of Trump
About Donald J. Trump
n
pct of total
Harris
Trump
Third party
Nonvoter
Can name likes and dislikes
545
51%
17%
55%
4%
24%
Doesn’t dislike anything
126
12%
1%
76%
3%
20%
Doesn’t like anything
373
35%
76%
1%
1%
22%
no answer
19
2%
31%
18%
6%
46%
Total
1,063
100%
36%
38%
3%
23%
These data also shed some light on how Trump managed to defeat Harris after losing to Joe Biden in 2020. We asked respondents about their participation in the 2020 presidential election. Thirty-five percent remembered voting for Biden, 31% for Trump, 3% for third parties, and 31% said they didn’t vote. Recalled vote may be error-prone, but in this case it closely matches Joe Biden’s 4.5-point margin of victory in 2020.
Harris won about the same share as Biden had among the 35% of adults who couldn’t name anything they like about Trump. Like Biden, she won practically none of the vote among those who dislike nothing about Trump. And Harris won about the same fraction of the vote as Biden among those naming both likes and dislikes about Trump.
While Harris’ vote share changed little from Biden’s in each group, Trump’s vote share grew among those with mixed feelings. Trump won 55% of adults with mixed feelings, up from 46% in 2020. This was possible because the share that did not vote in this group fell from 33% in 2020 to 24% in 2024. Likewise, Trump in 2024 won 76% of the vote among those expressing no negative views of him, up from 59% among these same adults in 2020. And only 20% of these adults didn’t vote in 2024, down from 38% in 2020.
2020 vote of U.S. adults by their view of Trump in Dec. 2024
among adults surveyed Dec. 2-11, 2024
About Donald J. Trump
n
pct of total
Biden
Trump
Third party
Nonvoter
Can name likes and dislikes
545
51%
18%
46%
3%
33%
Doesn’t dislike anything
126
12%
2%
59%
1%
38%
Doesn’t like anything
373
35%
71%
1%
2%
25%
no answer
19
2%
23%
18%
6%
54%
Total
1,063
100%
35%
31%
3%
31%
In our December 2024 sample, Harris defeated Trump by 3 percentage points among 2024 voters who also voted in 2020. Among 2024 voters who didn’t vote in 2020, she lost by 12 points.
Our survey is just one data point in a sense, but it adds to the emerging body of evidence that Trump’s campaign successfully turned out infrequent, “low-propensity” voters who like Trump but often stay home.
Explore the data
There is no substitute for reading the words of voters themselves. Click here to access our web app for viewing responses. The tool allows you to see 5 randomly* selected responses with each click of the button. Some of these responses contain profane language and many contain typos. We present them in unedited form.
Respondent
Sample random responses
Vote in 2024
Vote in 2020
What do you like about Donald Trump?
What do you dislike about Donald Trump?
Male, 34, Hispanic
Donald Trump
Donald Trump
Business plans to bring back production to America, and border control
Grandiose attitude and speaking without thinking about the consequences.
Male, 27, Other/Multiple
Did not vote
Did not vote
Good president for economic purposes. In my opinion, prefer trump over kamala
Tariffs for other countries are too much
Female, 50, White
Donald Trump
Donald Trump
His policies of smaller gov’t and less regulations which will open America up to less restrictions on oil and gas. Also his policies on immigration and closing the border align with my views.
His mouth and some of the unprofessional things he says
Small shifts from one election to another may be random and temporary, but consistent changes across three elections show something more enduring.
Across 29% of Milwaukee County, Trump in 2016 was more popular than Mitt Romney and he grew more popular still in each reelection campaign. In another 26% of the county he was less popular than Romney and continued to decline in each reelection campaign. Throughout the remainder of the county, his popularity has fluctuated—in most places dipping in 2020 and growing in 2024.
For this analysis, I compared the election results from each Milwaukee County suburb along with the 15 City of Milwaukee aldermanic districts (using 2024 boundaries).[i] Each aldermanic district is home to about 38,000 residents, making their population about the same as Oak Creek or Franklin.
Consider the following four regions.
Milwaukee Core
About 269,000 people live in the Milwaukee core region. This is the poorest area, with a per capita income of $22,000. Only 15% of adults over age 24 have a bachelor’s degree. About 14% of adults are white, 56% are black, and 23% Hispanic.
The voters in the Milwaukee core overwhelmingly support Democratic candidates—Kamala Harris won these 7 districts combined by 74 points. But Donald Trump has made significant inroads here. His 2024 performance was 12 points better than Mitt Romney’s in 2012.
Trump’s improvement here has been practically monotonic. Trump in 2016 was 3.9 points more popular than Romney in 2012. Trump in 2020 was 3.4 points more popular than Trump in 2016. And Trump in 2024 was 4.5 points more popular than Trump in 2020.
Northern Suburbs
The opposite political trend is occurring in Milwaukee’s northern suburbs (pop. 116,000). Across these communities the adult population is 78% white, 64% of those over 24 have a college degree, and the per capita income is $58,000.
Voters here have shifted sharply to the left in response to Trump. Romney lost the area by 13 points, Trump 2016 by 32 points, Trump 2020 by 42 points, and Trump 2024 by 45 points.
Southern Suburbs
Milwaukee County’s southern suburbs (pop. 246,000) are more working class than their northern counterpart. About 33% of adults over 24 have a college degree, and the per capita income is $41,000. The same share of population is white (78%) as in the northern suburbs, but more residents are Hispanic and fewer are Black.
The southern suburbs have long been the most conservative part of the county. Trump won them by 2.2 points in 2016, before losing them by 3.4 in 2020 and 2.9 in 2024.
Milwaukee Periphery
This leaves the collection of neighborhoods I’m calling the “Milwaukee periphery,” as they surround the Milwaukee core. These eight aldermanic districts hold 308,000 residents. The adult population is 57% white, 21% Black, and 14% Hispanic. The per capita income ($37,000) is slightly lower than in the southern suburbs, but the share of those ages 25 and up with a college degree (37%) is slightly higher.
Collectively, the Milwaukee Periphery gave Harris a 45-point victory, virtually identical to her margin of victory in the Northern Suburbs. However, the Trump-era political trajectory of the periphery has more closely resembled that of the southern suburbs.
Trump in 2016 declined by 2.8 points relative to Romney in 2012, and Trump in 2020 was 5 points less popular here than in 2016. Then, the trend reversed slightly, with Trump in 2024 improving by 1.2 points over his 2020 defeat.
Educational sorting
The results in Milwaukee are consistent with the national political realignment in the Trump era. If anything, Trump’s increase in support in Milwaukee’s urban core may trail the changes seen in some other cities.
Formal educational attainment remains one of the clearest predictors of a neighborhood’s changing levels of support for Donald Trump. About 32% of Wisconsin adults (ages 25 and older) have earned a bachelor’s degree (or more).
Since 2012, Democrats have gained the most in the parts of Milwaukee county where more than 50% have earned a bachelor’s degree. Democratic gains have been smaller in places where between 30% and 50% of adults have a college degree.
In places where college degree attainment trails the state average, Trump has usually made gains.
There are exceptions. Fewer than 20% of adults in the Village of West Milwaukee have a college degree, but Harris’ margin of victory was 10 points larger than Obama’s in 2012. This increase in Democratic support likely reflects changes to the racial composition of West Milwaukee, where the non-Hispanic white share of the population declined by about 15 percentage points over the past decade.
Keep this point in mind. Even though Black and Latino voters give Republicans more support than they once did, they still vote for Democrats at a higher rate than white voters overall. Consequently, an increase in the Black or Latino population of a community is still likely to increase its Democratic vote share.
Data Note:
See this repository for the data and graphics used in this article.
This table shows the trend in each Milwaukee County suburb and city aldermanic district in 2012 and 2024.
[i] Past elections were held under different ward boundaries, so I aggregated the Milwaukee city results into the current alder districts based on population overlap.
