Thread by @Nate_Cohn

Response rates by party looked fundamentally normal, including in Pennsylvania where we have the best partisanship data out there

https://x.com/Wertwhile/status/1822225193638535599

That said, it is worth noting that the sample does look a little blue. It's about a net-1 pt more Democratic, Dem-leaning and 2020 'Biden compared to the last time we polled these three states in May. This could be an indication of a shift in response patterns, though it could indicate change in attitude as well

I do think the idea of a 'real' 1 pt change in Dem PID/Biden recall, etc., to be fairly plausible. Not only is that a fairly small shift, but no one can deny that the national political environment hasn't changed in a huge way over the last few weeks

There's another question, of course: is it possible that both polls were skewed Dem, even if the change is real? That possibility can't be dismissed, given the nature of polling and the history of polling error in the Midwest. I'm just not sure these measures help identify it

Recall vote, for instance... may not be all its cracked up to be. All of our polls in 2020 showed too many Trump '16 voters, and they were ultimately biased toward Biden. That's why we don't use it: it would have made our polls worse, both in 20 and 22

I'll also note that... basically all of our polls this year have leaned Biden on 2020 recall vote v. the last result, much as our 2020 polls leaned Trump. Our Feb national poll had a Biden+12 recall and Trump+6 among RVs. Maybe Trump really was up double-digits? Maybe not.

Anyway, these questions are tough. Many pollsters do weight on past vote -- by their account, it made their data more accurate in the past. If that was true for us, perhaps we'd be doing it too.

And in case it's unknown: instead, we weight on hard measures of partisanship - party registration, past dem/rep primary participation, or modeled estimates.

Unlike past vote or PID, these are facts - we *know* how many people in PA are reg D/R, and we know the truth about the respondent's party registration based on their voter file record. The poll reflects that hard truth, which we do publish in the tabs for states with reg like PA