Is Tom Barrett really up 6 points on Walker? Could be

A big headline this week was the monthly Marquette Law School poll which showed that Tom Barrett had caught Scott Walker, with the race between those two being a 1-point race either way. The same poll indicated that Walker couldn't get over 50 percent against any of the 4 real Dems in Tuesday's primary, and that Barrett led Walker by 1.5% among registered voters,(47.2-45.7). This was quite a reversal from 2 months prior, when Charles Franklin's poll for the law school showed that Walker with a lead of 6-11 points over all opponents.

So I was curious to see if Franklin had taken the advice of people like me and others with a clue and actually modified his poll to more closely resemble who will actually vote in the election on June 5. Let's go inside the numbers and see.

First big check to make is the liberal/ moderate/ conservative breakdown, because we know that Franklin and PPP have skewed these numbers this to the right before. Remember that the accurate 2010 CBS exit poll breakdown was 21% liberal, 43% moderate, and 36% conservative. Well, the trend of oversampling righties continues with this poll, albeit with a few more liberals thrown in as well. Instead of 21-43-36, those polled who gave a preference came out as 23-33-43. So let's adjust all the numbers for the Walker-Barrett race to match that 2010 CBS exit poll, and see what the heads-up results become, and you'll see Barrett jumps quite a bit ahead.

Walker vs. Barrett- Barrett 49.79- 44.04 (Barrett +5.75!)

This is in no small part to Barrett's 55-38 advantage over Walker with moderates, and if moderates come out like they did in 2010, Walker ends up in big trouble. 5.75% is a helluva lot different than the 1.5% that Barrett led by in Franklin's overly-conservative poll. Walker's lead also drops by about 4 points over the other candidates when you use the 2010 exit poll filter.

Walker vs. Falk- Walker 47.27-43.88 (Walker +3.39, MU poll - Walker +7.5%)
Walker vs. Vinehout Walker 47.79-41.98 (Walker +5.81, MU poll - Walker +9.8%)
Walker vs. La Follette Walker 46.62-41.77 (Walker +4.85, MU poll - Walker +8.5%)

Hmmm, the "Walker is even with Barrett" result sent right-wing radio into a scared frenzy on Friday (Sly had a good topic on this, it should be up on his site in the next 2 days or so). Can you imagine if Frankin had done a legitimate ideology breakdown and had those numbers be even worse for Scotty?

There's another reason to believe Walker's real-life numbers are even worse than the MU Law poll. Check out the poll weighting of this question:

Who did you vote for in the 2010 election?
Scott Walker 50
Tom Barrett 39
Someone else 4
Don't know 1 (???)
Refused 6

But we know the final result in 2010 was Walker-52.25, Barrett-46.48, other-1.27. So that's clearly biased in favor of 2010 Walker voters. What I'm going to do is to use the actual 2010 vote to replace the MU sample, and then use the MU poll's results from the 17% responding that didn't vote in the 2010 election but are registered now. Once we do that, we see Walker slip here as well.
Walker vs. Barrett- Barrett 49.03- 46.03 (Walker loses 1.50%)
Walker vs. Falk- Walker 48.88-43.84 (Walker loses 2.46%)
Walker vs. Vinehout- Walker 49.22-40.84 (Walker loses 1.42%)
Walker vs. La Follette- Walker 48.91-42.26 (Walker loses 1.85%)

Scotty's still under 50% in all cases, and trails Barrett by 3. Walker also loses to all 4 Dems among people who didn't vote in 2010, with the biggest gap being with Barrett, who wins among 2010 non-voters 62-38. Guess we know why GOPs wanted to supress the vote, eh?

Lastly, I'll leave with a positive on the MU poll. I'm glad to see that Marquette actually asked if someone had signed a recall petition, because this has been a pet peeve of mine in all Wisconsin polls. The poll shows that 38% of respondents signed a recall petition in 2011 or 2012. If anything, this number is a bit low, as 38% of 2010 voters would be a little over 821,000 people. The actual number should probably be closer to 42%, as 41.79% would be the 903,000 GAB-confirmed signers against Walker, and you also should pick up a few people who signed Senate recalls that may not have signed for Walker's recall. I couldn't find a crosstab on how petition signers would vote, but given that they are disproportionately liberal (not surpringly), they'd be more like to vote against Walker, so I feel safe in saying that increasing the recall petition signers in the poll would also have hurt Walker's numbers.

So while it is true that April's Marquette Law poll showed movement away from Scott Walker and toward Dems (especially Tom Barrett), it also still continue to overstate Walker's voter share in the real Wisconsin. For those of us who want to end the age of Fitzwalkerstan, this poll is good news indeed, and no wonder the Walker folks are shaking in their boots. Because it is becoming quite likely that if a normal Wisconsin voting populace shows up on June 5, Scott Walker will be removed from office.