There are a lot of questions being raised about polling in the wake of Tuesday’s election results. Most polls gave Hillary Clinton a big chance of winning, but that’s not what happened.
So was Tuesday a failure of data or, as Wired puts it, “A failure of people using the data?” Here & Now‘s Robin Young looks at polling with GOP pollster Whit Ayres (@whitayres), president of North Star Opinion Research.
On incumbents vs. challengers in polling
“There’s a well-established principle in polling, with incumbents running for re-election, that what you see is what you get. In other words, if you’re at 48 as an incumbent, and your opponent is 45, we will frequently tell our incumbent candidates that they’re in trouble… The reason is that frequently incumbents get the number at the polls that they have on the final survey.
Their opponents are generally not saddled with the image of incumbency, so frequently, undecided voters go disproportionately to the challenger. And the issue here is whether or not Hillary Clinton was, if not technically an incumbent, effectively an incumbent running for the third term of Barrack Obama… It seems like more than a coincidence that the number she had in the average of polls at the end of the race is remarkably similar to where she ended up on the final ballot. But Trump made substantial gains, as frequently challengers do.”
On lying in polling
“If that were a factor, our own polls wouldn’t have been as accurate as they were… I don’t believe that there’s a consistent pattern of lying, but I do believe that many of us — and I include myself in this — interpreted the polls at the end as leading to a Clinton victory, which she did get in the popular vote, but she did not get in the Electoral College.
Trump put the pieces together the right way in terms of a map — he didn’t necessarily win more white voters, but he won the right white voters, and that allowed him to win numerous large states by one percentage point: Florida, Michigan, Pennsylvania, Wisconsin.”