:format(webp)/cdn.vox-cdn.com/uploads/chorus_image/image/50395513/GettyImages-588631096.0.jpg)
Conservative outlet Breitbart News, which has earned a reputation as one of the most pro–Donald Trump sources in existence, has been dissatisfied with poll after poll that shows the Republican nominee losing to Hillary Clinton. So it commissioned its own poll — only to find that Clinton is in fact beating Trump. By 5 points.
The Breitbart/Gravis poll published Sunday found Clinton leading Trump 42 to 37 percent in a four-way contest with Green Party candidate Jill Stein (3 percent) and Libertarian Gary Johnson (9 percent).
There’s been a lot of noise this summer over general election polling and whether to trust the numbers. Both Trump and Clinton received what many called an artificial bump after their conventions. But with the general election in full swing, polling has begun to settle into how it will likely look until Election Day.
Trump and his allies — including Breitbart’s editor in chief, Alex Marlow — have attacked polls that put Clinton comfortably in the lead, attributing them to the media’s pro-Clinton bias. Breitbart’s interest is in unskewing polls with Clinton ahead.
"It’s an open secret that polls are often manipulated and spun to create momentum for a particular candidate or issue," Marlow said in a statement. "Breitbart News Network’s first national poll marks the start of a major initiative to give our readers an accurate assessment on where the American people stand on the key topics and people of the day — without the mainstream media filter."
But Breitbart's findings weren’t so dissimilar from general election polling by CNN, Quinnipiac, Bloomberg, and NBC, which all have Clinton in a comfortable lead. Breitbart’s poll puts Trump about 2 points higher than the RealClearPolitics average. The similarities are likely in the methodology: Breitbart/Gravis polled 2,832 likely voters through automated telephone calls with a 1.8 percentage margin of error. As credible political polling operations go, that’s pretty normal.
Trump can claim media bias, but this is a good time to start paying attention to polls
Polls haven’t been looking very good for Trump for some time now. Right after the Democratic convention, a Fox News poll had Clinton beating Trump nationally by 10 points. A McClatchy/Marist poll had Clinton up by a sizable 15 points.
While two weeks ago much of that Clinton hype could be attributed to the post-convention bump, a period that historically has very volatile polling, political research shows that the numbers start to settle out right about now.
As Vox’s Andrew Prokop reported last week, historically the post-convention bounce seems to become more predictive about two weeks after the conventions, according to political scientists Robert Erikson and Christopher Wlezien:
"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…
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.
Even while accounting for bias, Breitbart's slightly smaller margin between Trump and Clinton is still a lead for Clinton. Because however intent it is on publishing unskewed polls, Breitbart has to follow the same polling rules as everyone else. And based on past elections, Clinton’s lead could likely stick around until November.
