Trump may be bouncing back
Welcome to YouGov's weekly newsletter The Surveyor, with new polling data, insights, and charts on politics, life, and other topical issues — from our U.S. News team.
This week, we're spotlighting surveys about Donald Trump, Trump's policies, neighborhood politics, people who don't like taking political polls, MAGA, sexual assault, cheating, addiction, and vaccines.
Rising Trump support in recent weeks
For several weeks now, Donald Trump has been registering modest week-over-week improvements on several trending poll questions. For example, the share of Americans who approve of Trump's job performance has gone up, as has the share of Americans with favorable views of Trump personally:
Trump remains unpopular, but after taking a battering throughout much of April he appears to have leveled out and even improved a little. Since mid-April, Trump has gained a few percentage points of approval with Republicans and Independents who lean Republican, going from 84% approve / 13% disapprove in the April 19 - 22, 2005 Economist / YouGov Poll (a net approval of +71) to 86% / 10% (net +76) in the May 9 - 12 poll. He's also gained among Independents who lean to neither party, from 25% / 60% in the April 19 - 22 poll (-35) to 33% / 52% (-20).
Opinion of Trump's handling of economic issues has not improved over this time; his net approval on handling jobs and the economy is -9, and for inflation and prices it's -22. But more Americans now approve of how Trump is handling crime (+9) and immigration (+4) than his negative net disapproval for those areas in April.
Policies, whales, and political weirdos
Your YouGov America team has spent most of this week at the annual conference of the American Association for Public Opinion Research (AAPOR) — seeing (and giving) lots of fascinating reports on innovative survey research. So in a slight twist for this newsletter, we wanted to share a few of those interesting findings with you!
Note that all these come from conference papers, which are often a way for researchers to share preliminary findings and get feedback from their peers. So some of these findings are not yet published and are still being refined.
250 policies (and we polled each one)
YouGov's Taylor Orth, our team member and a contributing author to this newsletter, presented the results of a massive study based on polling Americans about 250 Trump policies.
Overall, 61 of the 250 policies had more support than opposition, while the other 189 had net opposition. Policies concerning health care, the environment and energy, and executive power were particularly likely to be unpopular; policies concerning LGBT issues were the only area with greater average support than opposition.
Most of the issues showed lots of partisan polarization — Republicans much more likely to support policies and Democrats more likely to oppose them. Some did have bipartisan support, such as requiring ticket companies to fully disclose prices, designating drug cartels as foreign terrorist organizations, and eliminating federal taxes on Social Security benefits. Others had bipartisan opposition, including removing building codes intended to protect against natural disasters, allowing automakers to withhold reports of certain non-fatal self-driving car crashes, and deleting a government database tracking police misconduct.
Policies that involved spending cuts were more likely to be unpopular than other policies, both overall and among Democrats, Independents, and Republicans.
You can read more, and see the results for all 250 policies, in Orth's full article.
YouGov hopes to continue polling about these and other policies as Trump's second term progresses.
Won't you be my neighbor?
News junkies might remember a story from late in the 2024 election: The so-called "Trump Whale" who placed massive multimillion dollar bets that Trump would win — and enjoyed a massive payday when this bet paid off.
What didn't come out until after the election was that this unnamed bettor built the confidence to make these massive bets in part by commissioning his own private polls of battleground states. Those polls, which were run by a division of YouGov, were calibrated to test whether there was a "shy Trump voter" effect.
To do this, Bradley Jones, Ashley Grosse, and Alexis Essa of YouGov's Scientific Research Group (separate from YouGov's U.S. News Team that produces this newsletter, which had no involvement) asked residents of Michigan, North Carolina, Pennsylvania, and Wisconsin not just who they were voting for, but who they thought their neighbors were voting for. The theory was that people who were hesitant to share their own true views might be more honest when talking about other people.
The experiment showed majorities backing Trump in each of the four states using the neighbor-based method, while traditional polling had Harris in the lead in two of the four states. With this evidence in hand, the Trump Whale made his bets.
But as Jones, Grosse, and Essa shared in their presentation, the irony is that the neighbor-based measure ended up being farther from the actual results than the traditional polling questions were! Here's the key graphic from their presentation:
This experiment gets at an open debate in election polling: Is a poll more accurate if it's as close as possible to the final percentages but gets the winner wrong, or if it gets the winner right but is farther away from the final percentages? Arguments can be made for both sides. In this case, $50 million turned out to be a very persuasive argument.
Are you not entertained?
One issue survey researchers have to grapple with is the degree to which people willing to take polls about politics are representative of the broader population. Clearly to some degree survey respondents are weird (and we love them for it), but how weird?
Michael Bailey of Georgetown University and Kabir Khanna of CBS News, a YouGov polling partner, used YouGov survey data to test this with an interesting experiment. About half of their sample (the "control group") were given a standard political survey. The other half were given a choice: they could take a survey about politics, or a survey about entertainment news. Each group of respondents, after taking the survey they chose, were asked whether they planned to vote in the 2024 election, from which Bailey and Khanna calculated an overall estimate of turnout. After the election, they compared their projected turnout in 10 states with actual turnout.
Turnout projections from the surveys of the control group turned out to be consistently higher, by an average of 9 percentage points, than actual turnout — suggesting that the survey respondents were more politically engaged than the overall population.
The group of respondents who chose a political survey over a non-political survey showed an even bigger difference: turnout estimates that were an average of 16 points higher than actual turnout.
But the respondents who opted for an entertainment survey instead of a political survey produced a turnout estimate that almost entirely matched actual turnout — an average difference in the 10 states of less than 1 point.
Bailey and Khanna conclude that "election polls clearly over-sample people interested in politics," caution that these findings "could be contingent on political and survey context," and note that randomization experiments like theirs could be a way to better estimate the views of all Americans, including those who are less responsive to political surveys. When it came to respondents' presidential vote choice, however, the differences were much smaller.
Studies like this one can help pollsters including YouGov map results from people who take surveys onto those who don't.
Charting opinions
Between September 2022 and January 2024, less than half of Republicans said they identified with MAGA. A large share weren’t sure about their attachment to MAGA. Identification as MAGA has risen for most of 2024 and 2025, remaining mostly above 50% this year and reaching a peak of 60% in mid-March. But Republican identification as MAGA dropped again after that, including falling below 50% several times in recent weeks. In the latest Economist/YouGov Poll, it has gone up a bit. 53% of Republicans now describe themselves as MAGA Republicans, 35% do not, and 12% are unsure. Among the entire population of adult citizens, the share of MAGA supporters has never risen above 20%. (Kathy Frankovic)
45% of Americans say they know someone who has been sexually assaulted
45% of Americans say they know someone who has been sexually assaulted, and 21% say they know someone who has been accused of sexual assault. Overall, 51% of Americans say the rate of sexual assault has gone up in the past 20 years, while 20% say it has stayed the same and 9% say it has decreased. 74% of Americans say the sexual assault allegations against Sean "Diddy" Combs are true, a higher figure than for any of 22 other figures asked about who have been accused of sexual assault.
Quick takes
Cheating: 32% of Americans say that they cheated in high school either frequently (4%) or occasionally (28%), while 59% say they didn't cheat and 3% preferred not to answer
Addiction: 11% of Americans say they have been addicted to drugs or alcohol, while 50% say they have not been but know someone who has been; 30% say they don't know anyone who has been addicted and 5% prefer not to say
Shots: 73% of Americans say they have ever received a COVID-19 vaccine, including 88% of Democrats and 61% of Republicans. Just 25% of Americans say they have received a COVID vaccine in the past year, while 48% of Americans received one but not in the past year
Posting: 43% of Americans say Donald Trump posts too often on social media, while 24% say he posts the right amount and 5% say he doesn't post enough
Elsewhere
Polling partnerships
The Economist + YouGov on international respect for and from Trump, trade relationships, the new pope, and space exploration
Polling abroad
Polling in the press
Is the market up or down? Republicans and Democrats disagree (Economist)
What the slow accumulation of data tells us about the 2024 election (Washington Post)
Unpopular Democrats Can Still Win the Midterms. Here's Why. (Bloomberg)
Swift, Carpenter, more: Is this the worst-era for US pop culture? (WHYY)
The Popularity of True Crime: Why We Can't Stop (Psychology Today)
Learn more and get in touch
Sign up to participate in our polls and share your opinions on the people, companies, and issues that shape your life and the world
For more polling data and insights, visit our website or follow us on X, Threads, and Bluesky.
Have questions or feedback?
If you have a press or media inquiry, email uspress@yougov.com
If you have an idea for a future survey, share it here
If you are a panelist and have questions, contact support here
For information on privacy, please read our privacy notice
This newsletter is compiled by David Montgomery and Carl Bialik