The partisan makeup of the sample, including those who lean to a party, is 47 percent Republican, 43 percent Democratic and 9 percent independent. The partisan makeup of the sample, excluding those who lean to a party, is 28 percent Republican, 26 percent Democratic, and 44 percent independent.So R+4 (STOP LAUGHING!).
Then compare to the December Marquette Poll
The partisan makeup of the sample, including those who lean to a party, is 45 percent Republican, 44 percent Democratic and 9 percent independent. The partisan makeup of the sample, excluding those who lean to a party, is 30 percent Republican, 29 percent Democratic, and 39 percent independent.OK, R+1.
Well that 3 point change helps explain this "shift" in the poll.
Trump job performance: 48% approve of how he is doing as prez, 49% disapprove. In Dec, it was 47% approve, 50% disapprove. In Jan 2019, it was 44% approve, 52% disapprove. #mulawpoll
— MULawPoll (@MULawPoll) January 15, 2020
Then realize that the December poll was more Republican than the 2016 exit polls of Wisconsin (D+1), and 2018 exit polls of Wisconsin (EVEN).
So MU is assuming a Wisconsin electorate where Dems are going to be voting in big numbers out of anger/disgust against Trump will be…..5 points more Republican than 2016? Not a f’ing chance!
I just wish I knew why Chucky Franklin continues to play stupid with such a slanted poll (COUGH-Bradley funding COUGH). But when you, adjust for that 5-point slant, the numbers make a lot more sense.
MU Law Poll, with results 5 points more Dem
Biden +9
Sanders +6
Warren +2
Buttigieg +3
Trump approval -6
So Dems are leading and Trump is slightly disapproved of in Wisconsin (on the whole). Sounds about right.
Polls are data points, but can also agenda-setters and weapons too. Especially when RW-funded organizations are holding them. Always read the cross-tabs, and don’t fall for the “news” at face value.
Just to be clear: the exit poll pid3 results of D+1 in 2016 and being even in 2018 should apples-to-apples be compared to the R+2 pid3 result in this poll rather than the R+4 grouped pid5 result.
ReplyDeleteDisagree. Because those "leaner" numbers also swing GOP by 2, which tells you the real slant of the poll. People can say all they want, but the fact that "independents" actually slant GOP and conservative is another tell that this isn't on the level.
DeleteI'm not talking about the sampling problems the poll may have, I'm talking about the exit polls asking one question (namely, "Generally speaking, do you think of yourself as a ...?") which generates those D+1 and even results and the January MLSP asking literally the exact same question and getting R+2 from their sample. pid3 phrasing is completely standard in the industry.
DeleteSure, that MLSP ask those who respond to that question with anything but Democrat, Republican, or refusal with the further question of which party they lean towards and get R+4 when grouped might raise an eyebrow (note though that the R+4 amongst leaners alone isn't even half a standard deviation), but that's pid5 apples to the exit polls' pid3 oranges.
If you try to directly compare the MLSP's pid5 to the exit polls' pid3 results then you are necessarily presuming the composition of leaners among the exit polls' independents when those polls didn't ask that question.
Irrelevant point, and I'm not going deeper into it, but most "Independents" ain't independent.
DeleteThe other tell is the absurd number of people describing themself as "conservative" in the MU Poll - typically in the mid-40s when most exit polls in Wisconsin end up in the mid-30s.
Right on the money, Jake, as always! ��
ReplyDelete