The "worst polls in decades" call key assumptions of surveying, and forecasting, into question

The politicization of public opinion surveys has hurt their ability to reach a representative portrait of Americans. But surveys are still useful, despite critics' claims.

The American Association for Public Opinion Research (AAPOR) just released its quadrennial analysis of what went wrong in the polls during last election. “The 2020 polls featured polling error of an unusual magnitude,” the commission responsible for the report wrote:

It was the highest [error] in 40 years for the national popular vote and the highest in at least 20 years for state-level estimates of the vote in presidential, senatorial, and gubernatorial contests.  Among polls conducted in the final two weeks, the average error on the margin in either direction was 4.5 points for national popular vote polls and 5.1 points for state-level presidential polls. 

That is indeed the worst misfire in generations. The last time polls were that astray was in 1980. That year, the average absolute error was about 6 points on the Democratic candidate’s vote margin.

But one problem with the way AAPOR reported topline polling error is that survey data can still be accurate and have a high absolute…