… Putting the major polls together, their miss in last year’s presidential election was, on average, 4 percentage points, mainly because they underestimated the Trump vote; they also underestimated the Republican down-ballot votes by about the same margin. (Fivethirtyeight.com’s final averages of polls gave Biden an 8.4-point lead; he ended up winning by 4.4 points.) As presidential elections forecasting in recent decades go, this error was roughly average. …
The polling industry will eventually circle in on the best explanations for the 2020 errors, figure out which tasks they failed at and why. Meanwhile, here are what appear to be the leading candidate explanations roughly in declining order of likelihood and importance. Obviously, more than one distortion could have been at work. Keep in mind that pollsters forecasting elections have to do three things successfully: draw a sample that represents (or can, with adjustments, be made to represent) the population of potential voters; assess which candidate the people who answered the poll are leaning to; and estimate the chances that those people will actually vote. Each step is precarious. CONTINUED
Claude Fischer, UC Berkeley
The OPINION TODAY email newsletter is a concise daily rundown of significant new poll results and insightful analysis. It’s FREE. Sign up here: opiniontoday.substack