New poll: Trump and deportations unpopular, Dems up 8 in House vote

More: 60% see ethics/corruption problems in Trump administration, and the "Abundance Agenda" is popular (except zoning reform)

Summary:

Americans broadly disapprove of the job Donald Trump is doing as president and favor Democratic U.S. House candidates for the 2026 midterms by 8 points, a new Strength In Numbers/Verasight poll finds. The president is underwater on 10 out of 11 issues we tested with voters.

In a survey experiment, Democratic messaging on the economy does not outperform messaging on immigration. And: The Abundance Agenda as formulated by Ezra Klein and Derek Thompson is broadly popular, except for local zoning reforms.

This poll contains results for two topics submitted by Strength In Numbers readers. Subscribe to our newsletter to stay up to date on our July survey, and have a chance to ask your own question to the American people.

---

Democrats have an eight-point edge on the House generic ballot

If the 2026 midterms were held today, 45% of U.S. adults say they would back the Democratic candidate in their local congressional district, versus 37% for the Republican.

It’s still early, so an expectedly high 17% of respondents say they are not sure who they would vote for.

These results are statistically indistinguishable from our May survey, when Democrats held a 6-point lead on the generic ballot.

Historically, a generic ballot lead of this size typically translates into a gain of around 20 House seats for the leading party in the next congressional election. This advantage for the Democrats highlights another finding from our survey:

Most Americans say we are on the wrong track

Just 31% think the country is on the right track, while 58% say it's on the wrong track, according to our poll. Such high dissatisfaction can reflect poorly on the party in control of the White House, and usher in a majority for the opposition in Congress.

This is a small increase in pessimism relative to our May survey, when 33% of adults said the country was on the right track and 55% said wrong.

Trump's approval mostly stable from May

Trump was at -16 net approval in our May poll, with 40% approving and 56% disapproving of the job he is doing as president.

In June, his approval rating is 42%, with the same 56% disapproval.

Issue by issue, Trump remains unpopular on almost every topic we track:

Border security (+11) is the one issue on which Trump still enjoys a cushion. But Trump is deeply unpopular on health care (-19) and social spending (-17), issues that have been highlighted by Congress's debate over the GOP budget bill.

Our polling data suggests Trump’s position on these issues is not meaningfully changed since May:

---

---

What news is breaking through?

For this month's poll, we began asking Americans what news they saw, read, or heard about over the last month. We asked them in two ways. First, we adapted an open-ended question from an earlier University of Michigan/CNN/Verasight survey that was used to track news agendas during the 2016, 2020, and 2024 elections.

Respondents were able to answer this question in their own words. In the word cloud below, we present a visual summary of all the words used by respondents to describe the recent news they have seen:

The word cloud indicates a high occurrence of terms related to deportations, immigration, inflation, and trade, highlighting how deeply these specific topics have penetrated public consciousness in recent news. Our poll was conducted as national news increased coverage on the immigration raids, protests, and federal law-enforcement response in Los Angeles.

To analyze these responses on an issue-by-issue basis, we used a custom AI program to programmatically process each response and put it into one of ten (mostly) exclusive categories. We created several of those categories manually based on a small sample of responses, and had an AI model create and refine further categories with the full dataset.

The bar chart below shows the frequency with which respondents’ answers were sorted into each category:

I have excluded the bar for "other", which includes responses such as

Dude what is he actually doing. He's essentially killing people whilst not giving a flying f bomb about the people keeping him in power .

and

Clusterf**k

As a check on the model, we compared the aggregate category totals to a different question about topics in the news that required respondents to pick one of 11 pre-set options. The top 5 options were:

1. Deportations (28%)

2. Trade with foreign countries (21%)

3. Immigration more broadly (17%)

4. Prices/inflation (8%)

5. Government funding and social programs (7%)

In terms of rank order, this is a pretty good match. We are going to keep refining our AI-driven approach to categorizing open-ended questions. The ability to process a large amount of data quickly and flexibly is very good for a small team, but the differences between the self-reported category and AI categories illustrate the need to keep a human in the loop. (It also highlights some of the nuance a pollster loses when they present respondents with a pre-baked set of categories, compared to letting them answer freely.)

Politically, the increasing focus on deportations may be a warning sign for Trump, whose approval ratings on the issue are low and have recently fallen in ours and other surveys.

---

---

Message test: Economics or immigration — or both?

Following up on this newsletter's recent analysis of immigration, in this month's survey, we wanted to add some fresh data to the national conversation about whether Democrats would do better electorally by messaging on economics or immigration (or both!). We conducted a survey experiment testing how different messages impacted how voters felt about the major parties and Donald Trump.

That experiment worked like so: First, our 1,500 respondents were randomly split into one of four roughly equally sized groups. Each respondent was assigned to one of the following four experimental conditions:

• A control group: Respondents saw no message.

• Economic populism message: Respondents saw the text "Some Democrats say the party should focus more on pocketbook issues such as raising wages, reducing prices for housing and health care, and reining in corporate power. They argue that working people have been left behind while big corporations and the rich have gotten tax breaks and special treatment. A populist approach to economics can help everyday Americans get ahead."

• Immigration message: "Some Democrats argue that America needs more immigrants, but that the system should be both fair and secure. That means securing the border and enforcing the law, but also creating a pathway to citizenship for immigrants who have been living and working in America for at least 5 years. They say immigration strengthens the economy and helps America stay competitive."

• A combined economics + immigration message: "Some Democrats say the party should focus on rebuilding an economy that works for working people — by raising wages, lowering costs for housing and health care, and reining in corporate power. They also argue that immigration policy should be both fair and secure. That means securing the border and enforcing the law, while creating a pathway to citizenship for immigrants who have lived and worked in America for at least five years. They say this approach will strengthen the middle class and keep America competitive in a changing world."

Then, respondents were immediately asked to place the Democratic Party, Republican Party, and Donald Trump on a scale from 0 to 100, with 0 representing a "cold," unfavorable feeling and 100 a "warm," favorable feeling. If favorability ratings are significantly higher or lower for one of the four experimental groups, we can attribute effects to the messages respondents saw.

However, in general, this experiment did not meaningfully shift evaluations of either party or the president. While the combined messaging condition produced the largest gap between the Democrats and Republicans/Trump, in the Democrats' favor, the differences in average feeling thermometer ratings across the samples were not statistically significant, especially after controlling for party identification.

One possible explanation for the null results is that our sample was not large enough to pick up small effects via randomization. With 375 respondents per bucket, we would need pretty large effects of messaging to overcome the margin of error among respondents. If, on the other hand, we expect small effects from messaging (and in an era of high polarization and party sorting there's good reason to), then we'd need a larger sample to pick them up. It's also plausible that our wording did not capture the strongest party messages on each issue, and other messages would elicit larger attitudinal shifts.

But the results of our experiment do match the suggestions from other data. Our moderate immigration message performed better than a populist, more left-leaning economics message, and messaging on both issues simultaneously produces a larger effect against Trump/Republicans than either in isolation. Moderates tend to outperform in elections, since most voters occupy the center of the ideological spectrum, so our result on the moderate immigration message makes sense. Notably though, our immigration condition achieves moderation by combining a right-leaning position on immigration with a left-leaning one on amnesty and the value of immigrants — highlighting the fact that voters don’t look to parties to occupy the center on every single issue.

We also know voters (can) care about more than one issue at a time. So we should expect larger impacts on evaluations of the parties when multiple issues are messaged about simultaneously. Indeed, when it came to impacting feelings toward Trump and Republicans, the pro-immigrant, pro-border security message performed best for Democrats when it was combined with an economically populist message.

But again, these effects are small — so small that we cannot distinguish them from zero with the sample size available to us. We would welcome others to replicate our study with a larger sample size. The small/null effects might be a warning to people who are overly prescriptive about party strategy based on the supposed effects of party messaging.

One finding that is statistically significant, though: Democrats are rated higher on the feeling thermometer than Republicans and Trump regardless of experimental condition. This is contrary to recent polling questions that use a favorable/unfavorable framing and find Democrats doing worse than their competition.

Trump and charges of corruption

This topic was suggested by a subscriber.

Americans also say they are worried about corruption in the White House.

A clear majority — 62% — say that Donald Trump’s promotion of the USD $TRUMP cryptocurrency (which his family financially benefits from) represents a conflict of interest. 20% of adults say it does not constitute a conflict of interest. Among the respondents who said yes, 86% said they were also worried about the influence of money on the president’s decisions.

We also asked about Trump's pardons, which have been criticized as corrupt. Even when we told respondents that the power to pardon criminals is written in the Constitution, 29% of adults said they approve of the way Trump is using his pardons, while 60% said they disapprove.

Americans were similarly worried about the president dismissing federal inspectors general. By a 55%-27% margin, adults said dismissing inspectors general weakened oversight.

Finally, 29% of adults said it was appropriate for the president to accept a luxury jet from Qatar as a gift. 54% said it was inappropriate.

How the "Abundance Agenda" plays with voters

This topic was suggested by a subscriber.

In our poll, the so-called “Abundance Agenda” — a policy platform focused on building more housing and infrastructure, among other public goods, by relaxing bureaucratic and other constraints on supply — finds broad support among voters, with one exception.

Majorities back efforts to speed up construction of clean energy projects, expand high-speed rail, and make it easier for skilled workers to immigrate to the U.S.

However, support drops sharply and uncertainty rises when respondents are asked about proposals that would override local zoning laws to allow more housing construction. Just 28% favored curtailing local control, while 43% are opposed.

The data suggest that while Americans are open to ambitious, pro-growth policies, they remain wary of changes that would reduce local control over development. Small-"d" democratic opposition to zoning reform remains one of the key challenges for Abundance-oriented policymakers.

---

With a year and a half until the midterms, Democrats maintain a sizable advantage in the race for the U.S. House of Representatives. Meanwhile, voters dislike Trump’s agenda on nearly every front except the border. His position on economic issues, key to his victory in 2024, is still deeply lacking.

The electorate is open to a fast-building, pro-growth “Abundance Agenda,” and deeply concerned about ethics in the executive branch

---

Methodology note: Verasight conducted this poll among 1,500 U.S. adult residents from June 6-12, 2025. It has a margin of error of 2.6%. The survey was weighted to match the political and demographic characteristics of the U.S. adult population according to the April 2025 Current Population Survey, as well as recent benchmarks for partisanship and past vote.

Verasight uses mail, SMS text, and the internet to recruit a sample using both probability-based and non-probability techniques. Verasight handled recruitment, interviewing, and weighting. Strength In Numbers had input on questions but did not participate in other methodological decisions, and conducted all analysis, including creating the topline document.

You can download a PDF of the poll results and a full methodology statement at the bottom of this page.

Download the complete topline document and a full methodological statement below.

June 2025 Strength In Numbers:verasight Poll

254KB ∙ PDF file

Download

Download