Friday, September 2, 2016

Marquette and other Poll Shit

I suppose you may have heard that this week's Marquette Law School poll indicated that the big statewide elections in Wisconsin were tightening up, with both Hillary Clinton and Russ Feingold seeing their double-digit leads dwindle to only 3 points with likely voters over Donald Trump and Ron Johnson, respectively. I figured those races may be closer from the huge post-DNC Convention leads both Clinton and Feingold enjoyed in early August, but that was quite a narrowing.

Those topline numbers were surprising enough, and made me want to look further to see what was going on. Then another number in that Marquette Poll caused my BS detector to go off.

Scott Walker approval- Marquette Law
Aug 4-7 38%
August 25-28 43%

If anything, the headlines have been negative when it comes to Walker and the state of Wisconsin’s government over the last 3 weeks, so how did he gain 5 points in approval? Combined with the uniform shift to Republicans in the presidential and Senate races, something stunk about these numbers. And the answer to one question solved much of the mystery - “Which party do you identify with?”

For reference, the 2012 presidential election’s exit polls had it at Democrats +5 in Wisconsin, when Obama beat Romney by 7 points. So let's look at the last two Marquette Polls, and use matchups that include the 3rd-Party figures, if those polled were asked about them.

Marquette Law Poll, Voter ID with Leaners
Likely Voters
Aug 4-7 poll- Dem +5.2
Aug 25-28 poll- GOP +1.1 (swing- GOP +6.3)
Change in Clinton-Trump race- Clinton +13 to Clinton +3
Change in Feingold-Johnson race- Feingold +11 to Feingold +3

Registered Voters
Aug 4-7 poll- Dem +3.1
Aug 25-28 poll- Dem +0.5 (swing- GOP +2.6)
Change in Clinton-Trump race- Clinton +10 to Clinton +5
Change in Feingold-Johnson race- Feingold +9 to Feingold +4

The Republican “surge” was in the sample. Gee, Republicans do better when more Republicans are sampled? NO SHIT, and this factor alone accounts for half of the tightening in both of those races. Now maybe that’s legit, and that a lot of Wisconsinites changed from ID’ing as Democratic to Republican in the last 3 weeks. But I severely doubt it.

I’m just wondering why the sample was so different this time. Is that just who picked up the phones right before school was to start? Or is it something worse- that there needs to be the appearance of a closer race to have more people care about what the Marquette Law Poll and their associates at the Journal-Sentinel have to say about it? And because TV and radio stations need the extra ad money that a “closer” race would provide? Yes, I am that cynical- having an adult life over the last 20 years in this country will do that to you.

What’s also noteworthy is that other Wisconsin polls have come out this week that largely disagree with Marquette’s finding of a tighter race.

Clinton +5
Feingold +13
(I don’t understand the gap on that one).

Clinton +7
Feingold +7

Internal Poll from the League of Conservation Voters
Clinton +12
Feingold +18 (!)

So if you average the margins of these 4 polls together, you get this

Clinton +7
Feingold +10

That seems plausible, especially given that neither Clinton or Trump is advertising much in Wisconsin, and because the Kochs and other outside groups have been relatively quiet in the Senate race compared to the drenching of ads that are happening in other places. The ad flows to me are the real tell, because that gives an indication what the internal polls are telling these candidates, and the interest groups that back them.

So as Labor Day hits, the bottom line for Wisconsin Dems is best said in a song recorded on the east side of Madison in 1998.

"Don't Worry Bay-bee", and PUSH IT

(PS- Happy belated 50th Shirley. Yes, that fact just made us all feel inferior. Not only cause it means we are all older, but for the fact that none of us will ever look so good or be as cool at 50).

No comments:

Post a Comment