Wednesday, May 23, 2012

St Norbert poll- Walker by 5 = Barrett up 3?

In what continues to be an infuriating trend, once again we saw a Wisconsin recall poll come out, and once again it oversampled voters favorable to Scott Walker, resulting in a Walker lead. This time it's the Wisconsin Public Radio/ St. Norbert College poll, which has a topline of Walker 50, Barrett 45. However, once you go inside the numbers, it's pretty obvious why Walker leads.

Who did you vote for in the 2010 election? St. Norbert poll
Walker 52
Barrett 38

2010 Governor's race results
Walker 52
Barrett 46

Well no kidding Walker leads if you have a sample that voted +14 for Walker in 2010 while the real Wisconsin voted Walker +5.77. In fact, that should make Walker backers worry, because if Walker gets an 8 point advantage in the sample and he's only up 5, isn't it logical to say if we get a similar turnout to 2010 that this race becomes Barrett +3? And Dems and lefty voters are a whole lot more fired up than we were in 2010.

That being said, when you get into the "attitudes" questions, they're very enlightening, as it shows that the respondents have serious problems with Walker's character, and think Barrett's best attributes are his ability to unify people and his integrity. The poll also shows that a majority of this pro-Walker sample of Wisconsinites favor collective bargaining rights and they see Walker's moves as one to hurt unions over helping the budget, so a combination of negative ads on Walker's character and positive ads on Barrett's decency and desire to restore collective bargaining rights to public employees seems like a winning strategy to me.

St. Norbert's not the only poll that has oversampled GOPs in the last 10 days. The faulty polls from last week set a incorrect media narrative of "Walker ahead" into motion. It's tough enough to battle against talk radio and Walker ad propaganda, we don't need the rest of the media working against the good guys as well.

Now another poll that blunted that agenda was the We Are Wisconsin internal poll from the weekend, which showed Barrett within the margin of error, basically a toss-up. However, we should also be skeptical of this poll, not just because We Are Wisconsin has an obvious leaning, but because there are few cross-tabs released with the poll. However, the cross tab that was released was intriguing.

We Are Wisconsin poll, by party ID
Democrat- Barrett 93-7
Independent- Barrett 50-44
Republican- Walker 96-3

So if you assumed each group was 1/3 of the population, you'd come up with it dead even, 49-49. That indicates the sample was probably GOP +1 or +2, and while that's not the GOP +7 absurdity that we saw in the PPP poll last week, it still goes against the 2010 CBS exit poll from Wisconsin, which was Dem +1.

Bottom line, it's a turnout game right now, and if Dem voters turn out in Wisconsin (where they are the majority), Barrett would win. But this thing is far from being in the barn either way. I just wish the pollsters would recognize that reality and make their samples match Wisconsin's voting population.

7 comments:

  1. Every poll is showing Walker winning now

    They're ALL wrong?

    You're letting your partisanship get in the way of your critical thinking

    Barrett will be beaten, and likely badly beaten...

    ReplyDelete
  2. Well if you can name me a poll that has respondents in the neighborhood of actual Wisconsin voter turnout (which is approximately even to Dem+2, and a 20-40-40 lib-mod-con breakdown), and has Walker up by more than the margin of error, then I'll say he's ahead.

    But NONE OF THOSE POLLS EXIST. This is the point. This St. Norbert poll was Walker +14 for 2010 voters when it was Walker +6 in the election. PPP had GOP +7 for respondents, Marquette Law had 47% conservative turnout. None of this is close to what the real turnout is and will be.

    Please if you have a legitimate source that fits what I have above, bring it on. But to me, its obvious that this race is a toss-up if not a slight Barrett lead when you take into account a real Wisconsin turnout.

    ReplyDelete
  3. Well it seems to me that the nature of this election has changed the normal dynamics of a race in the state of Wisconsin

    Surely there are voters who are Democrats, oppose Walker, but feel that the recall is inappropriate under these circumstances

    Surely there are also voters turned off by the stridency of the unions and their behavior who would normally lean and vote Democratic

    Again, ALL of these pollsters are wrong in their turnout assumptions?

    ALL of them...?

    ReplyDelete
  4. MFG- But Dems aren't being adequately represented in these polls and conservatives are being overrepresented, so your theory goes out the window. Plus, the crosstabs of these polls show that almost as many Dems back Barrett as GOPs back Walker, and Barrett usually leads with independents.

    So yes, from what I can see and what I've dissected, these pollsters are DEAD WRONG in their turnout assumptions, and last summer's recall turnouts show it. If anything, there'll be even more Dems out than there were in 2011.

    Not surprisingly, we're starting to see internal polls get released showing the race to be a dead heat. Once PPP and others adjust their turnout models to fit the Wisconsin reality since 2000, and they should do that in the next few days, I think you'll see Barrett tied or possibly leading in at least a few of them.

    ReplyDelete
  5. I have a potentially dumb question:

    One of the results of the 5/30 Marquette poll has been blowin' up my Twitter, and that is the question about collective bargaining. The poll found that 55% of respondents favored collective bargaining limitations for public employees whereas 41% opposed curtailing collective bargaining rights for public employees. In the St. Norbert poll, the response was, well, pretty much the opposite (59% favor CB, 35% oppose CB). My question is, what gives? Do you have any thoughts on what would cause this kind of discrepancy?

    ReplyDelete
  6. Wording of the question. The St. Norbert poll asks if people support collective bargaining rights, and people say they do.

    Then Franklin asks about "limits" on collective bargaining and the "limits" put on union workers in the Marquette poll. This is intentional weasel words that make it sound like there is room for negotiations, and that Act 10 only allowed for higher pension and health insurance contributions. This isn't true, but the Marquette poll doesn't mention that. Tells you where they lean right there.

    ReplyDelete
  7. That's what I suspected. Thank you!

    ReplyDelete