clock menu more-arrow no yes mobile

Filed under:

Presidential candidates leading polls at this point in the campaign have almost always won

Andrew Prokop is a senior politics correspondent at Vox, covering the White House, elections, and political scandals and investigations. He’s worked at Vox since the site’s launch in 2014, and before that, he worked as a research assistant at the New Yorker’s Washington, DC, bureau.

As you’re surely aware by now, most polls taken after the conventions ended have shown Hillary Clinton surging to a sizable lead over Donald Trump. But write-ups of these polls have often been accompanied by the caveat that polling is volatile in the immediate wake of the conventions. A bounce for a candidate might not end up lasting.

So when, you might wonder, should we really start trusting the polls again?

The answer: right about now.

We are two weeks out from the final day of the Democratic convention. And according to historical research by political scientists Robert Erikson and Christopher Wlezien, this is usually the time that the convention fog finally lifts and the polls become more predictive.

"Although the convention season is the time for multiple bounces in the polls, one party ends up with an advantage when the dust clears. And this gain is a net convention bump rather than a bounce," Erikson and Wlezien write in their book The Timeline of Presidential Elections.

That is, once the volatility dies down, one candidate usually emerges from the convention chaos with a durable lead. And almost always, the political scientists find, that candidate goes on to win the election.

Polling two weeks after the convention tends to look a lot more like the eventual outcome

According to Erikson and Wlezien’s research and analysis of historical data, two important things tend to happen around this point in the campaign.

First, the post-convention polling chaos settles down, and the polls become more stable. And second, the polls that start to come out now tend to be much more predictive of the eventual outcome than pre-convention polls were.

Take a look at this graph from the authors’ 2012 book, The Timeline of Presidential Elections.

The language here is technical, but here’s what you’re seeing:

  • On the Y-axis is, essentially, a measure of how closely the polling at a certain point is to the eventual outcome. The higher a point is, the more predictive polling at that date is.
  • Then, the X-axis displays when the polls were taken in relation to when the convention season begins and ends. -30 is 30 days before the first convention starts, +30 is 30 days after the second convention finishes.
  • The gap in the middle is because polls during and immediately after the conventions are left out, since the authors found they’re messy and not particularly predictive. (They examine polling in this period more closely in other charts in their book.)

So what you see in the image is that polling two weeks after convention season wraps up (+14 on the X-axis) is dramatically more predictive of the eventual outcome than any polling beforehand.

That is — whatever happens during the convention period tends to last. "Most influences on vote intentions during these periods leave permanent imprints that survive to Election Day," the authors write.

Whoever’s leading the polls at this point usually ends up winning

Indeed, the authors looked at general election contests going back to 1952, and found that the candidate who was in the lead two weeks after the conventions ended went on to win the popular vote every single time.

Yup, you read that right. In all 16 of the most recent elections, the popular vote winner was the candidate who was ahead around this point in the campaign season.

Now, it would be a mistake to treat this as an iron law, or to conclude that nothing else in the campaign will matter, for a few reasons:

  • Some of those races did end up getting much closer in the closing months, and could have conceivably tipped the other way with a few more points of change.
  • Al Gore ended up losing the presidency despite winning the popular vote.
  • The data set starts right after the infamous polling miss in 1948, when (cruder) polls showed Thomas Dewey defeating Harry Truman.
  • And there’s always the possibility, however small, for some historically unusual major news event — a scandal, a terror attack, an economic crisis — to scramble things in a way we don’t usually see.

Still, Erikson and Wlezien’s research shows that conventions tend to have a consistent and profoundly important impact of the type that’s hardly ever observed at any other brief phase of the campaign, even those much-hyped fall debates.

"Once the conventions are over, further campaign events — even presidential debates— rarely result in dramatic change," they write in a 2012 update to their book. And if Hillary Clinton’s current lead holds up in the next few days of polls, that will be comforting news for her and Democrats.

The bad map we see every presidential election