The philosophical and empirical cases for ignoring bad or biased pollsters

… I am admittedly a bit late to this, but I have a few comments to make on FiveThirtyEight’s recently published retrospective on their forecasts of the 2022 midterm elections.

I am writing this as someone who, like Nate Silver, makes a lot of election forecasts, so knows the struggle of communicating about polls in an environment where people are always misunderstanding you — but also as someone who wrote a book about public opinion polling and democracy and has different attitudes about which polls we should use, how we should use them, and what the whole purpose of forecasting is, anyway. I view reacting to Silver’s biannual forecasting post-mortem not as an opportunity to join the chorus of people complaining about his models — indeed, I believe much of that criticism has gone way too far, to the point where it borders on misleading — but to reflect on the challenges of polling in the year 2023 and comment on how popular methods need to be updated for the future. CONTINUED

G. Elliott Morris, Politics by the Numbers


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

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.