poll position

Harris vs. Trump Polls: Is the Trump Vibe Shift Real?

Photo-Illustration: Intelligencer; Photos: Getty Images

It’s unsurprising that in what may be the closest presidential contest since 2000, political observers (not to mention the campaigns) have grown frustrated with the glacial nature of the polling trends. Big developments like the one debate haven’t changed things much, nor has the unprecedented torrent of paid advertising, nor has Donald Trump’s increasingly violent rhetoric.

So there’s a tendency to inflate trends discernible mostly by microscope, giving us headlines like this from The Hill: “Momentum shifts against Harris in presidential race.” Now to begin with, “momentum” is a suspect term because it suggests a trend that has self-propelling force, whether that’s true or not. But a close examination of the polls suggests the Harris-Trump race is still a dead heat, even if it’s possible to discern distinct trends out of the data.

Using the FiveThirtyEight polling averages, the national race is slightly closer than it’s been in a good while. Kamala Harris leads by 2.0 percent (48.3 to 46.3 percent), a half-point less than her lead a week ago and 0.7 percent less than two weeks ago. On the other hand, her total vote share is at 48.3 percent, near her all-time peak of 48.6 percent on October 6. For what it’s worth, Nate Silver’s averages give Harris a 2.3 percent lead; the New York Times (which rounds off its numbers) has her up by two points.

In the battleground states, the mega-trend has been a tightening of the race across the map, with Trump leading in Arizona, Georgia, and North Carolina and Harris leading in Michigan, Nevada, Pennsylvania, and Wisconsin — but both candidates being within striking distance everywhere. Yet Trump is making slight but measurable progress. At the moment, Trump’s leads in Arizona (2.0 percent, at 48.7 to 46.7 percent) and in Georgia (2.2 percent, at 48.9 to 46.7 percent) are higher than they were a couple of weeks ago, but that may be partially attributable to the mix of pollsters releasing data. The former president’s lead in North Carolina is 0.9 percent (48.2 to 47.3 percent), about where it’s been throughout October. Harris’s lead in Michigan is down to 0.4 percent, at 47.5 to 47.2 percent; this is another state where the recent mix of pollsters has probably had an effect. That’s true also in Wisconsin, where the two candidates are now even at 47.8 percent. In Pennsylvania, Harris is hanging onto a 0.1 percent lead (47.8 to 47.7 percent). And in Nevada, her lead is 0.4 percent (47.6 to 47.2 percent) — pretty much where it’s been all this month.

The bottom line is that if the polling sticks, Harris would have 266 electoral votes to Trump’s 262, with Harris winning the national popular vote by a bit less than Hillary Clinton did in 2016. The (mostly) poll-based prediction models out there agree it remains crazy close. FiveThirtyEight and Nate Silver both now give Trump a 52 percent probability of winning. DecisionDeskHQ projects the Electoral College results as a 50/50 proposition, and the Economist rates Harris a 54 percent favorite to win.

For reasons too complicated to go into here, Republicans tend to be optimistic, if not triumphalist, about their presidential candidates, particular in the three cycles in which that candidate has been the inveterate salesman Trump, who is forever predicting (and retroactively asserting) a massive landslide win (unless, of course, it’s stolen from him). Democrats, on the other hand, tend to be fretters, or as some say, “bed-wetters.” So despite the extremely positive “vibes” associated with the early stages of Harris’s campaign, no one in the long run will ever out-self-promote Donald Trump and his MAGA loyalists. The only anecdote to spin and hype is more, not less, reliance on the objective data we have, which is mostly from public polls. And no matter what you hear in the wind, those numbers are stubbornly even.

This post has been updated.

More on politics

See All
Harris vs. Trump Polls: Is the Trump Vibe Shift Real?