2010 Florida and Ohio Governor elections: How the unadjusted exit polls were forced to match the final recorded vote

16 Jun

Richard Charnin

June 15, 2012

The 2010 Florida and Ohio Governor exit polls were adjusted to match the vote counts – and red-shifted from the Democrat to the Republican.

In Florida, Sink (D) won the exit poll by 50.8-45.4% but lost the recorded vote to Scott by 48.4-49.6%, a 6.6% margin discrepancy. There were 3,150 respondents. The margin of error was 2.3%. Sink had a 99% win probability.

In Ohio, Strickland (D) won the exit poll by 49.9-47.4% but lost the vote to Kasich by 49.8-47.7%, a 4.6% margin discrepancy. There were 3,305 respondents. The margin of error was 2.2%. Strickland had an 88% win probability.

In order to match the recorded vote, the exit pollsters had to change ALL demographic category weightings from the unadjusted exit poll to the Final.

From this: Roper unadjusted exit polls
to this: CNN 2010 Election Center (final adjusted exit polls)

The final exit polls show the adjusted weightings for the key demographic categories. But keep in mind that similar changes had to have been made in ALL demographic crosstabs.

Since we have the unadjusted weightings from the Roper site, we can estimate the “pristine” unadjusted vote shares by “goal-seeking” (trial and error). Then we can calculate the changes in weightings and vote shares that were required to force the exit poll to match the recorded vote. These are shown on the right side of the Ohio and Florida worksheet screens.

This spread sheet shows the key exit poll crosstabs – and the adjustments:
2010 Midterms Spreadsheet: Ohio and Florida Governor

Faulty exit polling?

Why is it that the pundits always assume that the exit poll discrepancies are always the result of faulty polling? You would think that after 50 years, the exit pollsters would get it right. And they do get it right, but very few know it.

They get it right in the unadjusted exit polls. But they just keep on adjusting the polls anyway – to match the vote. That is what they get paid for. Otherwise, they would no longer be polling for the National Election Pool.

So what is getting it “right”? Is it forcing the exit polls to match the recorded vote – even when the election is rigged? Or is it by declaring that the uncontaminated, unadjusted exit poll stands by itself – and is a close approximation to the True Vote.

2010 Governor True Vote Analysis

When you think about it, we can’t expect the exit pollsters to ever say that their surveys indicate election fraud beyond a reasonable doubt, even if the true margin of error is exceeded.

The question to ask is: why are the category weights and vote shares changed in the first place? But we already know the answer. It’s what the pollsters and the media won’t talk about. It’s because of election fraud. If they didn’t adjust the numbers, the media would have to report them. And the last thing the media wants to do is to discuss is how voting machines are programmed to miscount the votes.

But how do we prove it?

The Ultimate Smoking Gun: 1988-2008 state presidential exit polls

In the 1988-2008 presidential elections there were 274 state exit polls, of which 226 red-shifted from the poll to the vote for the Republican and 48 shifted to the Democrat. Assuming no fraud, approximately 137 shifts would be expected for each party.

The probability P that 252 would red-shift to the Republican is:
P = 5.27E-51 = Binomdist (252, 274, .5, false)

The margin of error was exceeded in 126 of the 274 polls (only 14 would normally be expected at the 95% confidence level). The probability P is ZERO:
P =7.98E-75 =Poisson (126, .05*274, false)

The margin of error was exceeded in 123 of exit polls in favor of GOP (only 7 would be expected). The probability is P= 5E-106 = Poisson (123,.025*274, false)

The following table summarizes a) the number of state elections which there was a Republican red-shift from the exit poll to the vote, b) the number of states (n) in which the margin of error was exceeded in favor of the Republican, c) the probability that n states would red-shift beyond the MoE, d) the Democratic unadjusted aggregate state exit poll share, e) the Democratic recorded share, f) the deviation between the exit poll and recorded vote.

Year RS >MoE Probability.. Exit Vote Diff
1988 46.. 22… 3.5E-20….. 50.3 45.7 4.6
1992 44.. 26… 2.4E-25….. 47.6 43.0 4.6
1996 43.. 16… 4.9E-13….. 52.6 49.3 3.3
2000 34.. 12… 8.7E-09….. 50.8 48.4 2.4
2004 40.. 22… 3.5E-20….. 51.1 48.3 2.8
2008 45.. 36… 2.4E-37….. 58.0 52.9 5.1

Total 252. 134. 5.0E-115… 51.8 47.9 3.9

2010 Unadjusted National Exit Poll to Final (Red-shift)

Voted 2008 (Obama-McCain)
48-45 to 45-45 (3)

Party ID
37D- 35.8R- 27.2I to 35-35-30 (1.2)

20.6% Liberal – 40.9% Moderate – 38.5% Conservative to 20-38-42

2010 State Unadjusted Exit Poll to Final (Red-shift)

Gender (M/F)
OH 45.5-54.5 to 48-52 (5)
FL 42.9-57.1 to 45-55 (4.2)

Party ID
OH 38.4D- 33.8R- 27.8I to 36-36-20 (4.6)
FL 39.3D- 33.9R- 26.8I to 35-36-27 (6.4)

Governor Vote
OH 49.9D-47.4R to 47-50 (recorded 47.8-49.8) (4.5)
FL 50.8D-45.4R to 48-50 (recorded 48.4-49.6) (5.6)

Obama Approval
OH 45.8 to 42 (3.2)
FL 49.2 to 45 (4.2)

Voted 2008 (Obama-McCain)
OH 50.3-45.4 to 44-47 (7.9)
FL 52.2-44.2 to 47-47 (8.0)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: