The question I get asked most often

Who will win the presidential race?

Hello all! It’s good to be back writing Wake Up To Politics after a week off.

One of the things I did while I was away was speak at my old high school in St. Louis, where I gave a talk on political polarization and the media’s role in it. After the presentation, I opened it up to Q&A from the audience and I can’t say the first question surprised me. After all, “Who will win the presidential race?” is — by far — the question I receive most often.

Later in the week, I headed to Chicago for a Bat Mitzvah and, sure enough, more than a few relatives asked me the same question.

Political journalism probably isn’t the only profession where the question you get asked most often is a fundamentally unanswerable one — meteorologists and theoretical physicists are two others that come to mind — but it is still a somewhat difficult-to-navigate part of the job.

Because, the truth is, my answer to the question hasn’t really changed since I wrote the following in April: “The presidential race is — and always has been — tied.”

That might sound odd, considering just how much has changed in the last five months (the Democratic candidate dropped out, the Republican candidate was shot, etc.) But the polls have never really budged from a tossup race that entire time.

To be fair, there have been two times this election when the race no longer felt like a tossup: the period from late June to late July after the first presidential debate (when the race felt firmly in Trump’s grasp) and the period from late July to late August after Kamala Harris launched her campaign (when the contest seemed to have turned in Harris’ favor).

But, even during those two stretches, polling never really supported the narrative that the race was a slam-dunk for either candidate.

Per FiveThirtyEight, the average polling error — dating back to the 2000 cycle — for presidential elections is 4.3 percentage points, which means that since the beginning of the year, no candidate has held a lead in this race larger than the average amount polls end up being wrong by. Trump came closest in the post-debate period, boasting an average lead between 3 and 3.5 percentage points per RealClearPolitics, but still never assumed average-polling-error-proof dominance; Harris’ lead has never grown larger than 2 percentage points since entering the race.

In our divided media climate, few things have the power to shift narratives like the New York Times/Siena poll — empirically the most accurate pollster in America — and you can already feel Harris’ vibes honeymoon ending in the hours after the Grey Lady’s latest found Trump regaining a 1-point lead.

Cue the Democratic freakout, and the Democratic cleanup brigade. “I for one was not expecting zero bad days in the campaign,” Sen. Brian Schatz (D-HI) wrote on X shortly after the poll dropped. “Buck up.”

But, as I explained in April, there is not much of a statistical reason to treat a Trump +1 poll any different than a Harris +2 poll. Both are going to be within the margin of error for that particular poll, and well within the average polling error for a presidential cycle. So neither should adjust your floating mental model for the race, that of a tied contest.

The important thing to remember about polling error is the polls do often end up being systematically wrong in one direction — but we have no way of knowing which direction that will be until the race is over.

The Times itself has a nice illustration of that this morning, showing that the current polls could translate to a 303-electoral-vote victory for Harris (if the 2024 polling error mimics 2022), a 312-electoral-vote victory for Trump (if the 2024 polling error mimics 2020), or, of course, anything in between.

Graphic by The New York Times

Any time you see a poll, you’ll be better off if you mentally add these error bars to the numbers you see. If the poll’s outcome is only a 2020-or-2022-polling-error away from either candidate claiming a comfortable victory, consider the race tied as ever, no matter who happens to be up in that particular survey.

The last two elections came down to fewer than 100,000 votes in three battleground states that pollsters found difficult to poll, and this one probably will too.

That might not make for a satisfying answer at a Bat Mitzvah party, but it remains the truth.

---

However, just because you can safely ignore the topline results in most presidential polls — since anything from roughly Trump +3 to Harris +3 just tells us it’s a tossup — doesn’t mean that there isn’t valuable information to be gleaned from polls.

For example, one stat that jumped out at me from the Times/Siena survey was this pair of questions:

Only 12% of Americans feel they need to learn more about Donald Trump before casting their presidential vote, while nearly triple that amount (31%) say the same about Kamala Harris. These numbers shouldn’t be surprising per se, since Donald Trump has been running for president for eight years and Kamala Harris hasn’t even being doing so for eight weeks, but they are a notable statement about the race.

And they suggest that tomorrow night’s debate in Philadelphia will be a crucial moment for Harris to define herself.

Even if the polls have never supported the idea that Harris suddenly gained an insurmountable edge after announcing her campaign, I do think it’s worth noting that the positive vibes around Harris in the early days of her bid seem to have been conditioned around her commanding a lot of earned media.

That makes it even more surprising that Harris’ campaign has largely chosen to cede the earned media space to Trump in recent weeks, with Trump and Vance giving 39 interviews since late July and Harris and Walz giving only six, according to Fox News. Oftentimes, as Semafor’s Dave Weigel has noted, this aversion to media engagement has even extended to declining shouted questions from reporters on topics the campaign has prepared answers for — thereby depriving their own campaign of snappy soundbites that would only help them drive the news cycle.

Up to this point, for Harris, remaining difficult-to-define has been something of a strategy. As has been amply noted, Harris’ campaign has remained mostly vague on policy since its launch, an attempt to attract as many members of the broad anti-Trump coalition as possible. (Liz and Dick Cheney’s recent announcement of support for Harris is a reflection of this strategy.)

But, when running to be president of the United States, that’s a hard visage to maintain for long, and the Harris campaign debuted did debut an “Issues” page on its website last night in advance of the debate. The page is largely devoid of specifics — it calls for Congress to pass a bill to “restore reproductive freedom nationwide,” without outlining what the bill would or would not allow, and for an increase to the minimum wage, without naming an amount — but could signal a more outward-facing phase of her campaign, as Harris ups her efforts to reach that 31% by proactively defining herself instead of allowing Trump to do it for her.