How polling got its groove back (or fell into a rut)
New Power & Flour podcast on adjustments to polls in 2024
and I have a new episode of the up that I wanted to recommend. John does a really nice job explaining the reasons that most polls understated Trump’s support in 2016 and 2020 and the methods that many polling organizations have adopted to avoid such errors in 2024. Some polling organizations, for example, are attempting to weight their current polls by people’s recall of their vote from four years ago. This, as John notes, could work, but it also has the effect of essentially reproducing the 2020 election in 2024 polls.
One consequence of all this is that there’s just as good a chance that current polls are overstating Trump’s support as that they are understating it. After the past two presidential elections, we’ve gotten used to just adding several points to Trump’s polling numbers in our heads, but that might not be right this year. Kamala Harris leading by three points might literally mean that she’s leading by three points. We won’t know until this is all over, but it’s not like pollsters are just going into this blind, or that they’re just adding points to Trump’s numbers for the hell of it.
Anyway, I hope you’ll check it out. You can find the podcast on Apple Podcasts, Spotify, Overcast, and elsewhere. And here’s the show page.