Searching for Clues in the Polls About a Trump Conviction

For almost a decade, Donald J. Trump has done, said and survived things that would have doomed any other politician.

He even saw his support increase after four sets of criminal indictments last year — including the charges for falsifying business records that he was ultimately found guilty of Thursday.

The polls cannot tell us how voters will respond to the unprecedented verdict. Most voters weren’t even paying close attention to the trial, and asking voters about hypotheticals is always fraught. With his track record of political resilience, there’s surely little reason to expect his loyal MAGA base to suddenly collapse after a guilty verdict — or even imprisonment. It’s possible he won’t lose any support at all.

But in a close election in a closely divided country, any losses could be pivotal. While Mr. Trump has survived many controversies, he has also suffered a political penalty for his conduct. He did lose re-election, after all. And this cycle, there is one reason to wonder whether Mr. Trump might now be more vulnerable: He depends on the support of many young and nonwhite voters who haven’t voted for him in the past, and who might not prove as loyal as those who have stood by his side from the start.

In the last six months, many pollsters have asked voters to consider the hypothetical scenario where Mr. Trump was convicted at trial. It’s important to emphasize that these poll results shouldn’t be interpreted as simulations of how voters will behave after a real-world conviction. The questions don’t replicate how voters will react to the full context and facts of the case, or to statements of support from Republicans, or to the coverage on Fox News. Instead, they put a hypothetical conviction right in the face of the respondent.

Nonetheless, the results do show that a meaningful number of Mr. Trump’s supporters are understandably uncomfortable with the idea of supporting a felon. This is a line that Mr. Trump hasn’t crossed before, and a sliver of his supporters were even willing to tell a pollster they would vote for President Biden if Mr. Trump were found guilty.

In New York Times/Siena College battleground polls in October, about 7 percent of Mr. Trump’s supporters said they would vote for Mr. Biden if Mr. Trump were found guilty in an unspecified criminal trial. This may not seem like a huge number, but anything like it would be decisive in our era of close elections. Much more recently, a Marquette Law School poll taken during the hush-money trial found that a modest lead for Mr. Trump among registered voters nationwide became a four-point Biden lead if Mr. Trump were found guilty.

To repeat: These results should not be interpreted as indicative of what will happen after this conviction. And even if his numbers fall, many voters might ultimately come back around to Mr. Trump — especially Republicans, or those who can be convinced that the proceedings were “rigged” against him. In the Times/Inquirer/Siena battleground polls earlier this month, voters were divided on whether Mr. Trump could get a fair trial. His allies will do everything they can to convince voters that he did not get one.

But Mr. Trump doesn’t just count on the support of Republicans and MAGA loyalists in the conservative information ecosystem. His strength in the polls increasingly depends on surprising strength among voters from traditionally Democratic constituencies, like young, nonwhite and irregular voters. Many of these voters are registered as Democrats, back Democrats in races for U.S. Senate and may have even backed Mr. Biden in the last election. This is not Mr. Trump’s core of proven support. This is a group of voters whose loyalty hasn’t yet been established — let alone tested.

The Times/Siena and Marquette Law polls both suggest that these young and nonwhite voters might be especially prone to revert to their traditional partisan leanings in the event of a conviction, with Mr. Biden getting back to a far more typical lead among young and nonwhite voters. In fact, almost all of the unusual demographic patterns among young, nonwhite and irregular voters disappear when voters are asked how they would vote if Mr. Trump were convicted.

In the Times/Siena poll, 21 percent of Mr. Trump’s young supporters said they’d back Mr. Biden if there were a conviction. In comparison, only 2 percent of 65-and-older Trump supporters said the same. Similarly, 27 percent of Black voters who backed Mr. Trump flipped to Mr. Biden, compared with just 5 percent of white respondents.

In the real world, the verdict may or may not revitalize Mr. Biden’s support among young and nonwhite voters. But with Mr. Trump counting on the support of so many voters who wouldn’t ordinarily be expected to support him, the conditions for it to help Mr. Biden may be in place.

For one, voters did not see this coming. In Times/Siena polling during the trial earlier this month, just 35 percent of voters in the battleground states expected Mr. Trump to be found guilty. A majority, 53 percent, expected him to be found not guilty.

And voters had not been paying much attention. Only 29 percent of voters said they were paying “a lot” of attention to the trial, and they were disproportionately Biden supporters. Just 10 percent of young voters (18 to 29) said they were paying close attention.

With so many voters doubtful of a conviction and tuned out altogether, the verdict may come as surprising news to millions. This doesn’t mean that young and nonwhite traditionally Democratic voters will snap back to support Mr. Biden, but it seems likelier than if they were already paying attention and expecting it.

One of the better explanations for Mr. Trump’s strength among disengaged voters is that he has benefited from being out of the news — that his political liabilities had faded from the minds of voters.

That might not be true anymore. It might not be clear for some time whether those voters will shift away from Mr. Trump and whether such a shift will last. But in such a close race, anything could be enough to make a difference.