Trump is more popular than many of his policies

Many Americans love Donald Trump and even more hate him, but neither of these groups is large enough to win an election by themselves. Except when turnout is low, American elections for the past 9 years have turned with the preferences of those voters whose views of Trump are mixed.

In order to better understand these voters, the Marquette Law School Poll regularly invites a representative sample of American adults to answer the following two simple questions. What do you like about Donald Trump? What do you dislike about him? Respondents can write as much or as little as they want.

The answers to these questions, when paired with traditional multiple-choice items, show a large chunk of the electorate whose attitudes toward the president and broad policy issues, like immigration or trans rights, are malleable. These (potential) voters often hold combinations of views that are rarely found among politicians, making their support for any candidate contingent on issue salience, framing, and whatever ineffable quality makes some candidates seem more trustworthy than the rest.

Our latest poll was in the field in late March, preceding Trump’s “Liberation Day” tariff announcement. Previous installments were fielded a few weeks before and about a month after the inauguration.

At a high level, views of Trump changed little throughout the first three months of his presidency. Shortly before his inauguration, 49% of adults in our polling had a favorable opinion of him. That stood at 44% at the beginning of February and 46% in late March–all changes within the margin of error.

The overall patterns in the open-ended answers haven’t changed much either. In the latest poll, 50% of adults listed something they both like and disliked about Trump. 11% couldn’t name anything they disliked, and 36% couldn’t name anything they liked.

Summary of open-ended survey responses
in the Marquette Law School Poll, national adult sample
Attitude toward Donald Trumpsurvey dates
12/2-11/241/27-2/6/253/17-27/25
Can name likes and dislikes51%47%50%
Doesn’t dislike anything12%14%11%
Doesn’t like anything35%36%36%
no answer2%3%2%

This stability in overall attitude toward Trump doesn’t surprise me. After all, he has been at the center of American politics since his first primary campaign began a decade ago. Few voters lack an opinion of him and much of his behavior is already “priced in.”

But even though attitudes toward Trump himself are fairly stable, if trending a bit downward, opinions toward Trump’s favored policies are all over the place.

In our latest poll, we asked about 10 topics related to Trump’s agenda or recent Supreme Court decisions. The graph below shows the responses to each.

The most popular position across all of these questions was support for the 2020 Supreme Court ruling prohibiting workplace discrimination against “gay and transgender workers.” Eight-two percent of adults agreed with extending federal civil rights law to these workers.

At the same time, 72% of adults hope the Supreme Court upholds a Tennessee law prohibiting “medical providers from prescribing puberty-delaying medication or performing gender transition surgery for youth under 18.”

A large majority, 68%, support the deportation of undocumented immigrants when asked “Do you favor or oppose deporting immigrants who are living in the United States illegally back to their home countries?” When the wording is changed to include, “even if they have lived here for a number of years, have jobs and no criminal record?” support falls to 41% and opposition rises to 59%.

A decisive share of voters are not consistently “pro” or “anti” trans rights or deportation. Rather, their answers depend on the specific facts included in each question.

graph showing support and opposition for various policies

For each of the above questions, I coded a respondent as “1” if they supported the Trump/conservative position, “-1” if they chose the Democratic/liberal position, and “0” if they declined to take a side. A respondent receives a score of -10 if they took every liberal position and +10 if they always took the conservative side. The graph below shows the distribution of scores for all adults.

Few respondents fell into the most liberal or conservative categories. Forty-two percent are in the most liberal third and 31% are in the most conservative third of possible scores. Twenty-six percent of adults fell in the middle, with scores reflecting a mixture of support for conservative or liberal policies.

It is the open-ended answers from this last group that give insight into the views and beliefs of the most persuadable section of the electorate.

net ideological score of support and opposition for various policies

Click here to access our tool for viewing randomized responses to our open-ended questions. The tool allows you to filter responses by the respondent’s degree of support or opposition to Trump’s policies.

screenshot of interactive tool
Continue ReadingTrump is more popular than many of his policies

The Partisan Implications of ‘Low Turnout’ Have Flipped in Wisconsin

There’s a growing conventional wisdom that the two parties have flipped in their relationship to voter turnout. Now, it seems, Democrats are strongest in lower-turnout elections and Republicans do best when turnout is highest.

This is a real paradigm shift from not too long ago. During the Obama years, Democrats enjoyed a clear majority among potential voters broadly defined, but this majority depended on the adults least likely to participate. Republicans, on the other hand, had great strength with the most regular voters. For this reason, Obama could handily win Wisconsin (and the nation) in 2008 and 2012, but the Republican Tea Party wave dominated in 2010.

Here are a few more interesting data points in support of that emerging conventional wisdom.

Turnout always drops from a presidential election to the following gubernatorial election two years later, but the size of the decline varies from place to place. I was curious: does the decline in voter turnout correlate with changes in vote margin?

To answer this, I ran a regression comparing each municipality’s change in voter turnout with the change in vote margin between elections for president and governor.

The results are striking. In 2002, 2006, and 2010, a 1% decline in voter turnout from the previous presidential election predicted a more than 0.1 increase in the Republican vote margin for governor. This advantage dwindled in 2014 and reversed in 2018 and 2022.

In both of Tony Evers’ elections, a 1% decline in voter turnout predicted a significant increase in support for Evers, relative to Trump in the same municipality two years earlier.

graph showing the influence of a 1% decline in voter turnout from the previous presidential election on gubernatorial vote margins

The same dynamic affects Supreme Court races. The people most likely to show up in an April nonpartisan election are older, highly educated, and more wealthy. These demographics used to lean Republican; now they lean Democratic.

In April 2025, the liberal candidate Susan Crawford won 55% of the vote to conservative Brad Schimel’s 45%. Recall that in November 2024, Trump received 50% of the vote to Harris’ 49% in Wisconsin.

All the evidence I’ve seen shows that Crawford’s improvement over Harris is mostly due to who showed up. A survey from Blueprint Research found that 52% of voters in April 2025 had voted for Harris the previous November, and 46% had voted for Trump. Likewise, the researchers at Split Ticket analyzed ward-level election results and concluded, “roughly 70% of Susan Crawford’s win margin was attributable to changes in who was voting, rather than changes in how people voted.”

Here’s an example of all these trends taken from my hometown, the City of Milwaukee.

This graph shows that in the early 2000s, Democrats did best in presidential elections, a little worse in gubernatorial elections, and much worse in elections for Wisconsin Supreme Court.

In 2002, the Democratic candidate for governor won Milwaukee by 39 points, and in 2004 the Democratic presidential candidate won it by 44. Right in between those two elections, in 2003, the conservative candidate for Wisconsin Supreme Court outright won the City of Milwaukee by 5 points.

line graph showing margins among city of Milwaukee voters in races for president, governor, and WI supreme court

Since the early 2000s, things have changed. Democratic presidential margins in the city topped out at 60 points in 2012. Since then, they’ve dwindled slightly. Democratic candidates for governor have just kept climbing. Evers’ margin in 2018 matched Clinton’s share in 2016. But Evers’ Milwaukee margin of victory in 2022 reached heights not even achieved by Barack Obama.

The increase in support for liberal supreme court candidates among Milwaukee voters has been even more spectacular. Liberal candidates were consistently winning the City of Milwaukee by the 2010s, but in 2016, the liberal candidate still trailed Hillary Clinton by 34 points. In 2020, the liberal Court candidate trailed Biden by just 7 points among Milwaukee voters. In 2025, the liberal judicial candidate’s margin of victory exceeded Harris’ 2024 margin by 11 points.

Something fundamental changed in the years following Trump’s first election. Now, the smaller the electorate in Milwaukee, the more liberal it seems to be.

Continue ReadingThe Partisan Implications of ‘Low Turnout’ Have Flipped in Wisconsin

Examining the “System” in Criminal Justice Reform, Part 2: Measuring Justice with Primitive Scales

Wayne McKenzie
Wayne McKenzie

In my immediate previous post, I highlighted some of the motivating inspirations for creating a particular platform in 2007: the Milwaukee County Community Justice Council. The desire was to examine our criminal justice system and to invite external partnerships to help us identify efforts from different systems that might help inform our desire to improve the Milwaukee “product” of justice.

Hidden in the request for help was a perhaps naive presumption that some system somewhere was “doing justice the right way”—such that our need was to discover it, adopt or adapt it, and make it our own. The reality in 2007 (and today) is that there are approximately 2,330 state-level criminal justice systems representing diverse populations and operating in myriad legal and cultural systems sometimes very different from Milwaukee. All are presumably trying in good faith to justly serve their particular communities. And while many of the dynamics of “the criminal justice system” are similar everywhere in the United States, you will find important nuances just by traveling outside your home county.

In all events, given the complexity and deeply structural challenges of the American legal system, how do you objectively identify a problem in your ecosystem, assess what might fix the problem, implement a reform, measure the impact of the effort, and then demonstrate a narrative of progress? Such a process comes with abundant loaded assumptions, each one challenging enough to derail any effort at reform (which helps explain why so few jurisdictions even try).

But perhaps the biggest issue confronting reform-minded practitioners can be distilled to this essence: the challenge of adequately and accurately capturing meaningful data.

Unifying all criminal justice systems in the past and no less in the present are grossly inadequate information management systems combined with sparse analytical capacity. One of the guiding principles adopted early in the Milwaukee reform process is captured by the phrase “You can’t effectively change what you don’t effectively measure,” and while the information collection process has been revolutionized in a short time, effective analysis remains a challenge to most systems.

The Milwaukee County justice ecosystem circa 2006 was predominately an analog, paper-based system. If you practiced criminal law in the 1960s and returned for a day as late as 2010, you would still recognize all the processes and procedures required to represent a client or prosecute a case. Data and information processing systems (including software at the later date) were a hodgepodge of commercial and proprietary products, with the police departments, sheriff’s office, prosecutors, courts, and corrections system all using different means to capture and store the information needed in their respective sphere, but rarely with any interoperability with other agencies. In a pre-Cloud, pre-AI world, a researcher needed to physically enter the space where the work unfolded to even attempt to capture data, and nonetheless he or she would be disappointed in the quality of the information.

Despite the obstacles, in 2005 the Milwaukee County District Attorney’s Office opened itself to outside, independent researchers. This occurred when the Vera Institute of Justice reached out and asked to be allowed to enter the complicated and risk-averse space of the elected prosecutor. It did so on a topic that was and is considered a third rail of police and prosecution controversy: race.

Continue ReadingExamining the “System” in Criminal Justice Reform, Part 2: Measuring Justice with Primitive Scales