Read on Twitter
Are the polls understating the number of Republicans?
Probably not. https://www.nytimes.com/2020/07/29/upshot/polls-political-party-republicans.html
Probably not. https://www.nytimes.com/2020/07/29/upshot/polls-political-party-republicans.html
In Times/Siena polling last month, registered Republicans were 12 percent likelier to respond to the survey than registered Democrats.
After controlling for factors used in weighting, the effect of partisanship on likelihood to respond was a *zero* (coefficient = .0003)
After controlling for factors used in weighting, the effect of partisanship on likelihood to respond was a *zero* (coefficient = .0003)
Voter file data offers other clues.
Nationwide, Democrats (registered Dems or recent primary Dem voters), outnumber Republicans by 6 points, 36 to 30 percent, in the 45 or so states with either measure.
Doesn't prove advantage in Dem *ID*, but certainly consistent with it
Nationwide, Democrats (registered Dems or recent primary Dem voters), outnumber Republicans by 6 points, 36 to 30 percent, in the 45 or so states with either measure.
Doesn't prove advantage in Dem *ID*, but certainly consistent with it
In our national poll, a slightly smaller share of file-based 'Republicans' identify as such (65%) compared to the 'Democrats' (69%). Unaffiliated voters--disproportionately young and nonwhite--identified as Dems by a 5 point margin (24 to 19).
It added up to D+8 by party ID
It added up to D+8 by party ID
There are some really interesting regional patterns here that I don't go into, btw.
18% of white southern rural Democrats call themselves Republicans
Only 56% of northern metropolitan registered Republicans called themselves GOP.
18% of white southern rural Democrats call themselves Republicans
Only 56% of northern metropolitan registered Republicans called themselves GOP.
And in general, the polls... don't really show a shift in unleaned party identification since earlier in the year. To some extent that still begs the question, but it does imply that Biden's gains aren't attributable to partisan nonresponse
As my note on 'unleaned' implies, Biden's big gains are among independents. There are plenty of polls where he's up 20. There's no plausible party ID target where that's consistent with a close race
And as you may recall, it was Romney who led among independents back when we were doing those 'unskewed polls' fights back in 2012. The Democratic party ID edge was the foundation of Obama's edge. That's just not true now.
