A few more thoughts on the Fox News poll, which has traditionally been quite a good poll. 1/ https://twitter.com/psephologist/status/1248406662639054848">https://twitter.com/psepholog...
I was curious at the time whether this was also the death knell for independence at the Fox News poll, but generally it& #39;s remained quite good. Yesterday& #39;s is the first time I can remember shaking my head. 4/
They& #39;re pretty transparent, and even provide a useful guide to margin of sampling error for the subgroups they& #39;re showing in their tabs. And they report stuff like this trendline on the partisan makeup of their sample 5/
You& #39;ll notice a few things there, starting with the fact that over their previous 11 polls, partisanship averaged D+7, and yesterday is the first time in that stretch they& #39;ve had things closer than D+5 6/
There are bigger debates over whether / how much to let partisanship float in your sample, and easy / innocent explanations for why an unweighted survey might show D+9 one week and D+0 the next 7/
It& #39;s mostly a case of caveat emptor -- don& #39;t breathlessly report big movement in the vote when what a survey really shows is big movement in who they interviewed 8/
It& #39;s something to keep an eye on going forward. Fox News poll has traditionally been very good but Trump/Barr haven& #39;t been shy about putting the arm on them and as the election gets closer there will be more be pressure to put a thumb on the scale. /fin
You can follow @psephologist.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: