What’s wrong with polling? Some early theories

The survey results among the elderly are another symptom of a deeper failure this year. Unlike in 2016, polls consistently showed Mr. Biden winning by comfortable margins among voters 65 and over. A recent poll by NBC and The Wall Street Journal showed Biden had a 23-point lead among the group. A recent Times / Siena poll showed it had a lead of 10. In the latest novel, there’s no reason to believe any of it was real.

This is a deeper kind of error than the mistakes of 2016. It indicates a fundamental measure of the attitudes of a large demographic group, not merely an underestimation of its share of the electorate. In other words, the primary survey data has worsened over the past four years, canceling the changes that the pollsters made to address the mistake that occurred in 2016.

Helps explain why national surveys are worse than they were in 2016; They achieved weight in terms of education four years ago and have made virtually no changes since then and it also helps explain why the error is so closely related to what happened in 2016: It focuses on the same demographic, even if the primary source of error between the group is very different. .

Voting appears to have some serious challenges. The industry has long relied on statistical adjustments to ensure that each group, such as white voters without a score, accounts for their appropriate share of the sample. But this only helps if the respondents you reach represent those you are not reaching. In 2016, they seemed to be actor enough for many purposes. In 2020, they weren’t.

So how have opinion polls got worse over the past four years? This is mainly speculation, but consider some possibilities:

The president (and opinion polls) harmed the elections. There was no real indication of a “hidden Trump” vote in 2016. But there may have been a vote in 2020. For years, the president has attacked the media and polls, among other institutions. Polls themselves lost a great deal of credibility in 2016.

It’s hard not to wonder if the president’s supporters have become less likely to respond to the polls as their suspicions mount in the institutions, leaving the polls in a worse place than they were four years ago.

“We now have to take seriously some version of Trump’s shy hypothesis,” said Patrick Ruffini, a Republican pollster at Echelon Insights. The problem with opinion polls would be simply not reaching large elements of the Trump coalition, which would make them underestimate Republicans across the board when he’s on the ballot.