The Walker Recall: Is the Past Prologue?
May 24, 2012
In a previous Walker recall election analysis the Wisconsin True Vote Model indicated that Barrett would win a fair election with 53-54%. The purpose of this analysis is to determine what it took for Walker to win in 2010. This information may provide insight into what we can expect in the recall.
The 2010 Election
Approximately 69% of 2008 voters turned out in the 2010 Wisconsin Governor race. Walker defeated Barrett by 125,000 recorded votes (52.2-46.6%). The exit poll was forced to match the recorded vote. That is standard operating procedure.
In order to force the exit poll to match the vote, it required that 49% of the 2010 electorate were returning Obama voters and 43% returning McCain voters. The 6% spread is 8% below Obama’s recorded margin and a whopping 22% below his exit poll margin. The spread implies that 66% of Obama voters and 77% of McCain voters returned in 2010 – a net 11% turnout advantage to Walker.
In the poll, Barrett had 83% of Obama voters and 7% of McCain voters – a net 10% defection of returning Obama voters. Walker had 16% of Obama voters and 93% of McCain voters. In addition, 3% were returning third-party and 5% did not vote in 2008 – but the vote shares were n/a. A simple calculation shows that in order to match the recorded vote, Walker needed to win new and returning third-party voters by a 20% margin.
To summarize, in order to match the recorded vote, the adjusted Final 2010 Wisconsin exit poll assumed…
1) There was ZERO fraud in 2008.
2) McCain returning voter turnout exceeded Obama turnout by 11%.
3) 16% of Obama voters defected to Walker and 7% of McCain voters defected to Barrett.
4) Walker had a 20% margin among new and returning 2008 third-party voters.
The 2012 True Vote Model
The base case assumption in the 2012 Wisconsin Recall True Vote Model that Obama had a 60% vote share is conservative. He had 63.3% in the Wisconsin exit poll (2545 respondents) but just 56.2% recorded – triple the 2.4% margin of error. There is a virtual 100% probability that Obama’s True share exceeded 60%. In other words, the 2008 election was likely extremely fraudulent, but not so fraudulent as to cause Obama to lose Wisconsin.
Unlike final national and state exit polls that are adjusted to conform to the recorded vote (and implicitly assume zero fraud), the True Vote Model is based on a feasible estimated turnout of previous election voters and best vote share estimates of returning and new voters.
The model calculates various scenarios (“sensitivity analysis”) of 2008 election voter turnout in 2012 based on the 2008 a) recorded vote, b) unadjusted exit poll or 3) estimated True Vote.
What does this portend for the recall?
1) Fraud: Walker wins by a similar margin as he did in 2010 (125,000 votes)
2) Fraud: But not enough to steal the election. Barrett wins by 70,000.
3) No fraud. Barrett wins by at least 160,000.
1988-2008: the 8% unadjusted exit poll margin discrepancy
Unadjusted state and national exit poll data is available on the Roper website. The Democrats won the aggregate 1988-2008 presidential unadjusted exit polls by 52-42% -an 8% discrepancy in margin from the 48-46% recorded vote. Bloggers, pollsters and academics are apparently unaware that the data even exists. After all, the NY Times and CNN never reported that fact. But they do show exit polls adjusted to conform to bogus recorded votes on their websites.
How many voters are aware that Obama won the Unadjusted National Exit Poll (17,836 respondents) by a 61-37% margin? Or that he won the state aggregate exit polls (82,388 respondents) by 58-40.5%? The Final 2008 National Exit Poll was forced to match the recorded vote by implying an impossible 103% turnout of living Bush 2004 voters and 12 million more returning Bush than Kerry voters.
In every election the pollsters force state and national exit polls to match the recorded vote. They accomplish this by adjusting all demographic crosstabs that are displayed in the various mainstream media election sites. As a result of the forced match to the recorded vote, the “pristine” demographic percentages are contaminated. In other words, by matching to the recorded vote, final exit polls disguise the true intent of various classes of voters.