“We knew it would be record-high turnout, but we had to figure out who the extra voters would be,” said Patrick Murray, director of the Monmouth University Polling Institute. “Still, the polls were very good indicators of what was going to happen. The misses weren’t huge.”
So where were the polls off the mark?
• In Florida, Republican turnout exceeded expectations. Nonpartisan polls in the last two weeks had Andrew Gillum, the Democrat, leading by between 4 and 7 points. He lost by half a point (barring a recount).
• In Indiana, polls for the Senate race since October showed results ranging from a big win for the Democratic incumbent Joe Donnelly, to a tie, to a narrow loss. In the end, he was crushed by Mike Braun, losing by more than 8. (Indiana, it’s worth noting, is notoriously difficult to poll because of state laws restricting automatic phone dialing.)
• In the Missouri Senate race between Claire McCaskill and Josh Hawley, different polls in the last week had both of them with 3-point leads, and others showed a tie. Mr. Hawley, the Republican challenger, won the race by 6.
Pollsters this year made more of an effort to try and be transparent about their process. At The New York Times, we conducted our surveys publicly, in real time.
“You saw a definite tamping down on the expectations that polls would produce a clear outcome,” said Courtney Kennedy, director of survey research at the Pew Research Center.
So with all that in mind, and their good track record this year, why is it that polls always get such a bad rap after elections?