President Trump has put a direct and obvious focus on policy issues important to his base, including the current fight over funding for a wall on the border with Mexico. (Mandel Ngan/AFP)

At the top, the new NPR-PBS NewsHour-Marist poll isn’t great news for President Trump’s reelection prospects.

The pollsters asked Americans if they planned to vote for Trump’s reelection in 2020, and only 30 percent of respondents said they definitely planned to. More worrisome for the president: 57 percent said they definitely wouldn’t back his candidacy.

Broken out by party, the results were about what you’d expect: Nearly all Democrats said they would vote against him, and the majority of Republicans said they would back him. But those numbers show more weakness for Trump than you might at first assume.


(Philip Bump/The Washington Post)

In broad strokes, having 57 percent of the country vote against you isn’t necessarily prohibitive, depending on where the votes are distributed. After all, Trump got only 46 percent of the vote when he won last time, thanks to slipping past Hillary Clinton in key states to earn an electoral college victory. Bill Clinton won election in 1992 with only 43 percent of the vote, but of course, that was a function of a strong third-party candidate.

So why are the numbers above bad for Trump? Consider the distribution of support for his candidacy by party in 2016, as determined by exit polls.


(Philip Bump/The Washington Post)

Notice first that the Republican and Democratic bars showing support for Clinton and Trump respectively are about the same height, unlike in the first chart above. That’s not really that big a surprise, as shown in the chart below, comparing 2016 results to the Marist poll.


(Philip Bump/The Washington Post)

Across the board, Trump has a lower percentage of people saying they’ll definitely vote for him than he got in 2016. Part of that may be a general drop in support for Trump — but part of it, too, is that in the Marist poll we’re comparing Trump to an unknown Democratic candidate. It’s understandable that there would be some uncertainty about making that choice.

What’s much more important is the discrepancy marked with the outlined box above. Trump got 46 percent of the independent vote in 2016 to Clinton’s 42 percent. Now, 62 percent of independents already say that they won’t vote for him.

Trump’s win among independents (and overall) was in part a function of a kind of unpopularity he shared with Clinton. Trump was the preferred candidate among voters who disliked both major-party candidates, many of whom were independents. In 2020, he may get an opponent who ends up as unpopular as he is, but even if he does, it seems less likely that he’ll get the benefit of the doubt from people skeptical about both of them.

So now we come to the coup de grace. Among Trump supporters, only about three-quarters say they will definitely vote for him. Nearly 20 percent of his supporters say they’re unsure whether they’ll back him, about the same as the percentage of Republicans who say that.


(Philip Bump/The Washington Post)

This is amazing in part because of how steadfastly Trump has sought to appease his base. He’s put a direct and obvious focus on policy issues important to his base, including the current fight over funding for a wall on the border with Mexico. And yet a quarter of his supporters are still wavering.

It’s January 2019, more than 22 months before the election. We have no idea who Trump will be running against; we have no idea what might happen in the interim. This poll was conducted at a time when Trump’s support is slipping in part because of his handling of the fight over the wall. (A fifth of Trump supporters, for example, say they think Trump should do more to compromise with Democrats on that issue.)

But, still. This is not the sort of poll you want to see if you’re Donald Trump.