This AAPOR report is amazingly comprehensive. It's based on a massive data collection effort that compiled polling across elections at various levels and geographies extending back to 2000.

Major conclusions:

And here are party affiliation trends from six years of NPORS data www.pewresearch.org/politics/fac...

Here's a great starter pack filled with some of my wonderful colleagues at @pewresearch.org Give 'em a follow!

Thanks Elliott. These both appear to be bespoke mail push to web (and phone option, for Franklin and Marshall). Although Dartmouth reinterviewed people from an early Oct survey, neither is a typical prob panel. Plus, F&M's LV estimate was Trump +1 (election result Trump+2), their RV was Harris +4.

I'm puzzled by that last line. I assumed that "online prob panel" (bias = 2.8 D) is the line for most of the probability panels used in this cycle. I'm also unaware of errors from any probability panels that were as large as 13 points. @gelliottmorris.bsky.social help us out here!

The party ID is from NPORS and can be filtered on registration in the survey, while the registration parameter itself comes from the CPS

it's not a slam dunk decision, but we think it's better to use a presidential year estimate than a more recent off-year estimate

And of course there is a quiz www.pewresearch.org/methods/quiz...

There is a "Worm Moon" coming on March 25, so I suppose there is also a....

Relatively few Americans (22%) say they are closely following news about the Israel-Hamas war, and only about half can correctly report that more Palestinians than Israelis have died since the war’s start. On many attitude questions about the war, sizable numbers express no opinion

We look at measures of religious affiliation, family income, political party affiliation, gender identity and sexual orientation, race and ethnicity, and age and generation. Many of these measures have evolved over time in response to growing diversity and changing attitudes among the U.S. public.