Here’s some political advice that most people will, unfortunately, ignore: Don’t fret too much about general election polls until both parties’ conventions are over.
Over the next two weeks, election polls will likely show some wild swings between Donald Trump and Hillary Clinton. (The Republican convention will be held from Monday to Thursday, and the Democratic convention will be held from July 25 to July 28.)
This happens every four years. Polling usually goes haywire during the conventions, since a lot is happening: The primaries have ended, the parties are consolidating, vice presidential candidates are picked, and so on.
But here’s the catch: Polling during the conventions is even less likely to be predictive of the final outcome of the election than polling at other times of the year, says Princeton election guru Sam Wang. So if you’re eager to find out who’s truly ahead in the presidential race, you’re better off looking at earlier polls from the spring — or, if you can resist, waiting until August.
Why polling during the conventions is usually all over the place
Political scientists Robert Erikson and Christopher Wlezien, authors of the book The Timeline of Presidential Elections: How Campaigns Do (and Do Not) Matter, are widely regarded as the top authorities on general election polling.
And one thing their data from 1952 to 2008 clearly shows, according to Wang, is that head-to-head polls from January through April tend to be better predictors of the eventual winner than polls from May through July:
(That’s relatively good news for Clinton, since she led Trump by a substantial margin during this early period, although polls this autumn will give us an even better sense of the race.)
Nobody knows exactly why the polls suddenly became screwy and less predictive during the summer. One possibility is that there are still a lot of bad feelings lingering within each party after primaries: Supporters of the losing candidates sometimes say they’ll refuse to support the party nominee for a spell. These holdouts typically come around in time for the general election, but they can obscure what’s happening in the summer.
Another possible reason for the variability in mid-summer polling is the conventions, which normally produce temporary "bounces" for each of the candidates (more on why below). According to political scientist John Sides, each candidate typically sees a 5- to 6-point bump in the polls after officially receiving his or her party’s nomination.
In 2008, for instance, Barack Obama held a 4-point lead over John McCain going into the Republican convention on September 2. But by the second day of the convention, McCain’s polling numbers had eclipsed Obama’s. They took a brief while to sink back down to where they were before:
In 2012, Mitt Romney also spiked in the polls after the Republican convention, which was held from August 27 to 30 — only to sink back down again.
Similarly, in 2004, John Kerry climbed ahead of George W. Bush after the Democratic convention from July 26 to 29 — but, again, eventually got eclipsed:
So if history is any guide, Trump and Clinton will each enjoy a momentary surge in the polls in late July after their respective conventions. But don’t take either of those as a sign that the candidate has a lock on the race.
What to look at during the conventions instead of general election polls
That said, the polls aren’t entirely useless during the two-week convention period. They won’t necessarily tell us who will win the election, but there are some interesting things to watch.
To understand why, we first have to understand why conventions tend to cause polling bumps in the first place. One explanation political scientists like to offer is the imbalance of news coverage: Attention to each candidate surges during that period.
Here’s John Sides:
Conventions generate a larger "dose" of information than is provided by the daily ebb and flow of news coverage. Thomas Holbrook’s study of campaigns from 1984-92 found that front-page coverage of each presidential candidate rose sharply during his party’s convention.
A second reason is that the news coverage during the convention favors the candidate being nominated. There are typically no increases in the coverage of the Republican candidate during the Democratic National Convention, and vice versa.
But why does the wall-to-wall coverage boost a candidate’s poll numbers? Georgetown professor Jonathan Ladd says that the glut of information in turn serves to remind voters about what party they belonged to in the first place.
Most people aren’t political junkies, and they don’t pay all that much attention during the primaries. So the conventions remind those people which party they’ve traditionally voted for, and why. That’s what helps account for the polling surges.
"Conventions appeal to those with some affiliation or affection for that party," Ladd says. "The conventions are when the parties get ready for the fall."
This is how you should interpret convention polling: by looking to see if Clinton and Trump are able to shore up support within their own parties. If they don’t get much of a bounce and are still struggling to consolidate their party’s voters, that might really be a sign of weakness.
When to start paying attention to the polls again
Despite the weirdness of polls around the conventions, they’ve historically proven pretty accurate as we head into fall and Election Day draws near.
Here’s a chart from Erikson and Wlezien that shows how the polls grow increasingly predictive after August:
"The polling averages just don’t move much more than 2 or 3 percentage points after the conventions," Ladd says. "Especially since the conventions are in July, the person in the lead in August will have a really good chance of winning."
That’s why Kathy Frankovic, the former CBS polling director and analyst at YouGov, says she once tried to follow a simple rule: Don’t check the polls until Labor Day.
"Otherwise, you're spinning wheels. You're worrying — or celebrating — prematurely," she says.