People aren’t always good at predicting their own behavior. We know what we’re thinking and feeling right now, but we’re not as good at guessing how we’ll react to different situations in the future.
As a pollster, I see this all the time as I think about how to ask questions that will reveal what people really think and feel. If someone asked me what I was doing right now, I would say I was writing this essay, and that would be an accurate answer. If someone asked me if I was making pasta for dinner later this week, I would say, if My answer probably wasn’t accurate because I knew the grocery store had pasta sauce on sale, would I want pasta?
Similarly, if you ask any voter what they think about Donald Trump today, believe me. They know. Voters have long since formed their attitudes toward Trump and his personality, but asking what voters will do in the future if there is an unprecedented change in circumstances, such as a potential presidential candidate receiving a prison sentence, is fraught with uncertainty.
As a result, one of my biggest pet peeves as a pollster is “more likely or less likely” survey questions. This type of question asks respondents whether a certain hypothetical situation or piece of information makes them “more or less likely” to perform a future behavior. It’s often used to test campaigns and political messaging strategies, and takes the form, “If you knew candidate A voted to increase taxes, would you be more or less likely to vote for him?”
In the Trump legal drama, this kind of question has been a favorite of pollsters, with results like this one from last week: “1 in 10 Republicans less likely to vote for Trump after conviction.” Whenever I see a headline like this, I wonder if 1 in 10 Republicans ever voted for Trump. anywayOr were they “Never Trump” Republicans who were already leaning to President Biden and simply became more entrenched in their defections? These questions will be crucial in determining whether the ruling is having a real impact on the race.
A much better, though still imperfect, way to gauge the impact of new developments is to see whether people have actually changed their voting intentions. We found that so far, not many people have changed their voting intentions after Trump was convicted on 34 counts of falsifying business records. My firm asked about 500 voters across the country the day after the verdict who they would vote for and whether the verdict had changed their vote. The question is, already We had surveyed those people over the past few months and knew what they had told us previously.
When these voters were asked who they intended to vote for in 2024, 97% of those who previously said they would vote for Trump continued to support him. Similarly, 98% of those who previously said they would vote for Biden continued to support Biden. The best news for Biden is that of the very small minority who previously said they were undecided, 40% say they are now leaning toward Biden, while only 3% of previously undecided voters have switched to Trump.
We then asked these respondents specifically whether the ruling had caused them to change their vote, and among those who said they had changed their vote because of the ruling, almost none had actually changed their vote from either Trump or Biden from previous surveys.
The guilty verdict itself is unlikely to sway voters in the short term. Those who believed Trump was a good person and fit to be president still do. Those who disliked him were confirmed. Perceptions of Trump are so firmly entrenched that it would take something more significant to change them than a conviction for falsifying business records.
But I think it’s too early to throw up your hands and say this doesn’t matter politically. I believe if Trump receives a sentence that includes prison time, it could decide the outcome of the election.
“Trump is a convicted felon” is a statement about who he is – corrupt, rule-breaking, dishonest – and people already have a strong opinion about that. “Trump is going to jail” is a development of a different nature that highlights the seriousness of the situation and indicates what a Trump presidency means for the country beyond simply “Trump is a bad guy.”
If Judge Juan Marchand sentences Trump to probation, community service or some other light punishment, the impact of the case will be whether people feel uncomfortable electing someone labeled a felon as president. Early data suggests that a label alone won’t make a difference. Voters who support Trump may conclude that regular checkups with his probation officer won’t hinder his performance as commander in chief.
Prison is different. In reality, most voters don’t expect Trump to actually be jailed. In my data, only one in five think it’s likely that Trump will be jailed, and most of those voters have already voted for Biden. It’s not hard to imagine that voters who are not enthusiastic about Trump but tend to vote for him and who don’t think the trial and sentence are a big deal would have a strong reaction to the possibility of voting for someone who has been sentenced to prison.
The reaction to the sentence could affect both sides. Most obviously, voters who are reluctantly considering Trump could decide that it was too much to send a man facing prison to the White House. Unlikely but not impossible, a harsh sentence could harden the resolve of disgruntled Republicans who dislike Trump but feel wronged in this case.
The fact that Trump might go to prison is something most people are undecided about, and Trump supporters do not have That’s factored into their assumptions about the presidential election, and it makes the situation very unstable.
That raises concerns about the second-order national and racial impacts of a harsh sentence. Trump has issued ominous warnings about a “breaking point” if he receives a prison sentence. Clearly, the sentence alone, without the sentence, has rallied Trump’s most ardent supporters; Trump raised $53 million in donations just 24 hours after the verdict. But it’s unsettling to think about what a prison sentence could mean in a political environment where Trump supporters feel that our very system of government is in danger as a result of the trial and sentence. (Few things have significantly swayed the polls on Trump, but one thing has, at least temporarily: the violent scenes of January 6th.)
Trump’s sentencing comes just days before he is formally nominated as the Republican presidential nominee. While there is historical precedent for knowing how a convention surge might fare in the polls, there is no precedent for knowing how voters might react to a sentencing of a leading presidential candidate.
In the coming weeks, caution is advised against anyone making confident declarations about how a Trump conviction would affect the election: the sentence, not just the verdict, will determine the case’s ultimate impact on the election.
Kristen Soltis Anderson is a contributing opinion writer for The New York Times, a Republican pollster, speaker, commentator and author of “Selfie Vote: Where Millennials Are Leading America (And How Republicans Can Catch Up).”
The Times is committed to publishing Diverse characters To the Editor: Tell us what you think about this article or any other article. Tips. And here is our email: letters@nytimes.com.
Follow The New York Times Opinion section Facebook, Instagram, Tick tock, WhatsApp, X and thread.
