4 ms·
How did polls go so wrong? “gold standard” Ann Selzer predicted +3 Harris in Iowa and it became +14 Trump. That’s an incredible miss from a pollster.
by haunter 2y ago
How did polls go so wrong? “gold standard” Ann Selzer predicted +3 Harris in Iowa and it became +14 Trump. That’s an incredible miss from a pollster.
- acdha 2y agoThat’s one poll and there are always outliers, but the averages were pretty accurate - roughly a tossup based on turnout, with error rates around the 2-3% we saw. As with 2016, those correlated in the same direction so the polling industry still hasn’t figured out how to weight Trump’s impact on turnout.
- athrowaway3z 2y agoMy guess: Goodhart's law > Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes.
- kelnos 2y agoThe polls, in aggregate, were fairly accurate: 50/50 chance for the most part. The Selzer Iowa poll felt like false hope to me immediately.