Thursday, October 4, 2012

Marquette polls skewed? Maybe, and not for the WisDems

This week featured the release of another Marquette University Law School poll, which likely voters picking President Obama over Mitt Romney by double digits, and Tammy Baldwin leading Tommy Thompson by 4 in the Senate race. However, the right-wing bloggers had been whining about Marquette's outlier status (Obama's led in Wisconsin by 5-7 in most other legitimate polls), so even though poll numbers master Nate Silver doesn't list Wisconsin as competitive and has it as a near-lock for Obama next month, let's look into the latest Marquette poll and see if there is some BS there that's making the races tilt more Dem than they are.

First of all, let's look at the partisan breakdown of the poll, which is something GOPs have been complaining about nationwide. The Marquette Law Poll had respondents that identified as follows:

Dem 33%
GOP 26%
Ind 37%
Other 3%

And Dem +7 seems like a pretty big disparity...until you realize a few things. The 2008 exit poll of Wisconsin was Dem +6, and Obama won by 14 over McCain then. So it's reasonably to think it'll be near that again in the next presidential election. And there's been plenty of evidence that people hate Republicans in general, with a recent poll showing 53% view the GOP unfavorably, including 60% of moderates. With that in mind, who'd want to admit to being Republican, even if they're leaning that way? (Independent "leaners" split evenly 17-17 between Dem and GOP in the Marquette poll)

But on the flip side, the two most recent statewide elections have been more even between the parties. The 2010 Wisconsin exit polls were Dems +1, and the adjusted exit poll for June's recall election was GOP +1. So let's humor the GOP and figure there might be a slight shift, and split the difference in party ID to make it Dem +3. So make it Dem 31%, GOP 28%, Indy 37% Other 3%. And we get..

Obama 51.7, Romney 41.3%

So it's Obama +10 by this standard instead of Obama +11. Minor narrowing. We run the same test on Tammy vs. Tommy and the numbers narrow from 47-43 Baldwin to 45-44 Baldwin. So that makes a bit of difference, and shows that Tammy will be more reliant on getting her voters out than Obama might.

But then you look inside a couple of other demographics, and it shows that the Dems might be getting UNDERsold in the Marquette poll, if possible. As I've pointed out before, Charles Franklin's poll consistently overweights conservatives compared to who actually turns out to vote in Wisconsin. And this poll is no different. Among those giving a preference, the ideological breakdown in yesterday's poll was as follows:

Conservative/ Very cons 42%
Moderate 35%
Liberal/Very liberal 23%

However, this flies in the face of recent exit polls. Even in the more-GOP turnouts of Scott Walker's two wins in 2010 and 2012, there have been more moderates than conservatives.

2010-2012 guv election average ideology
Conservative 36%
Moderate 43%
Liberal 21%

And this switch from conservative to moderate makes a big difference, because Obama destroys Romney among moderates 65-26, and Baldwin beats Thompson 58-30. So you plug in those numbers, and the Dems extend their leads- Obama into "boat race" levels.

Adjusted for exit poll ideology
Obama 55.1%, Romney 38.6% (Obama +16.5%)
Baldwin 48.8%, Thompson 41.7% (Baldwin +7.1%)

The other skew Marquette Law (as well as other pollsters) tends to have is a lack of young people responding. It makes sense, these individuals tend to be cell-phone early and are less likely to answer calls from strange numbers. This shows up with Marquette respondents under 30 being at 9%, while Xers like me (30-44) are listed at 31%, and 45+ is 59%. The problem with this is that when Obama was first elected in 2008, the geezers got outnnumbered.

Age of voters, 2008 Wisconsin exit poll
22% 18-29
29% 30-44
35% 45-64
14% 65+

Even in the 2010 and 2012 guv elections, people 18-29 were 15-16% of the vote, not 9%. This is important to bring up because Obama beats Romney among this group 65-23, and Baldwin also beats Thompson among this group (but only 40-34, not sure what's up with that). If you figure the youth vote to be near their presidential-year levels at 20%, make Xers at 30%, and Boomers/Geezers at 50%, and it becomes blowout city for Obama.

Adjusted for age of prez year exit polls
Obama 54.8%, Romney 37.7% (Obama 17.1%)
Baldwin 46.0%, Thompson 41.4% (Baldwin +4.6%)

Oh, but that's OK for WisGOP, because the young people aren't going to vote like it's 2008, so they won't get worked like that in November. The thrill for Barack is totally gone and they won't turn out for him, especially after Obama's weak performance in yesterday's debate.


Uh, maybe not.

You just heard every Republican in Wisconsin go "GULP!" this afternoon when they saw those pictures and heard a more fiery Barack giving props to Baldwin and other Dems. And the demos in the Marquette poll show why- WisGOP is in big trouble with any kind of big turnout from moderates and young voters. No wonder the GOP tried to suppress their vote.

No comments:

Post a Comment