A Preliminary Probability Analysis of the Wisconsin Primary

06 Apr

Richard Charnin
April 6, 2016
Updated: May 29

Matrix of Deceit: Forcing Pre-election and Exit Polls to Match Fraudulent Vote Counts
Proving Election Fraud: Phantom Voters, Uncounted Votes and the National Exit Poll 

This is a preliminary analysis of the   Wisconsin Democratic primary exit poll and recorded vote discrepancies.

Bernie Sanders had 563,127 votes (56.5%) and Hillary Clinton 429.738 (43.1%). But the 4:24 pm exit poll indicates that Bernie  did even better. The first 1100 respondents indicated that Sanders had 68% of the white vote. The Public Policy Polling (PPP) pre-election poll  indicated that Sanders was leading black voters by 51-40%.

The final adjusted CNN Exit Poll   indicates that white voters comprised  83% of the electorate (Sanders had  59%). The probability of the 9% decline from 68% to 59% is 1 in 4000. But Census tables showed that Whites comprise 92% of the Wisconsin electorate. I split the difference and assumed 88% were whites. Blacks  comprised 10% of the vote (Sanders had 31%) .

Assuming  Sanders had  31% of the non-white vote, he won WI by  63.6-36.4%.

Why the 20% decline in Sanders 51% pre-election poll share of blacks? Why the 9% decline in Sanders 68%  exit poll share of whites? The probabilities of the discrepancies are way beyond the  margin of error.

The poll shows that 7% of voters were Latino (3%), Asian (2%), Other (2%). But the vote shares are NA.  Why are vote shares not available for 124 (7%) minority respondents?Was it because their respective turnout rates were too low? The pollsters could have combined the 7% as Other Non-whites. Apparently they used the same vote shares for the 7% as they did for blacks in matching the recorded vote.

The CNN final exit poll (1774 respondents) is shown in this spreadsheet.  Vote shares are calculated for each category.  It is standard operating procedure to force the exit poll to match the recorded vote in all categories  by adjusting  the voter turnout percentages and/or the vote shares. 

Who still believes that in Wisconsin, the recorded vote represents the true vote? 

Gender Pct Clinton Sanders Other
Men 43.0% 35.0% 64.0% 1.0%
Women 57.0% 49.0% 50.0% 1.0%
Total 100.0% 43.0% 56.0% 1.0%
Age Pct Clinton Sanders Other
18-29 19.0% 18.0% 82.0% 0.0%
30-44 24.0% 33.0% 66.0% 1.0%
45-64 39.0% 54.0% 46.0% 0.0%
65+ 18.0% 62.0% 37.0% 1.0%
Total 100.0% 43.6% 56.0% 0.4%

Assume vote shares of Latino, Asian, Other  non-white groups are equal to the Blacks

Race Pct Clinton Sanders Other
White 83.0% 40.0% 59.0% 1.0%
Black 10.0% 69.0% 31.0% 0.0%
Latino (na) 3.0% 69.0% 31.0% 0.0%
Asian (na) 2.0% 69.0% 31.0% 0.0%
Other (na) 2.0% 69.0% 31.0% 0.0%
Total 100.0% 44.9% 54.2% 0.9%

Given  a) the 7.9% discrepancy between Sanders’ 56.7% recorded 2-party vote share and his estimated 64.6% exit poll share, and b) assuming a 3.0% exit poll margin of error, there is a 95% probability  that his True share was in the range 62-68%. The probability is 97.5%  that he had at least 62%.

The probability of the 7.9% discrepancy between Sanders’ estimated 64.6% poll and 56.9% vote share is P= 1.14E-07 or 1 in 8.7 million.

The sensitivity analysis shows Sanders vote share for increments   around his base case shares of whites and non-whites.

Sanders had 61.4% in the worst case scenario, assuming he had 65% of whites and 35% of non-whites. He had 67.9% in the best case scenario, assuming he had 71% of whites and 45% of non-whites.


Base Case

Race Pct Sanders Clinton
White 88.0% 68.0% 32.0%
Non-white 12.0% 40.0% 60.0%
TOTAL 100.0% 64.6% 35.4%

The sensitivity analysis matrix shows Sanders vote shares for various increments  around his base case shares of white (68%) and non-white voters (40%). He had 64.6% in the base case.

Bernie had 55.6% in the worst case (recorded vote) scenario (59% of whites and 31% of non-whites). He had 68.8% in the best case  (72% of whites and 45% of non-whites). 

Sanders % White voters

Sanders 59.0% 64.0% 68.0% 72.0%
 % Non-white Sanders Total%
45.0% 57.3% 61.7% 65.2% 68.8%
40.0% 56.7% 61.1% 64.6% 68.2%
35.0% 56.1% 60.5% 64.0% 67.6%
31.0% 55.6% 60.0% 63.6%  (base)  67.1%



Posted by on April 6, 2016 in 2016 election, Uncategorized


Tags: , ,

18 responses to “A Preliminary Probability Analysis of the Wisconsin Primary

  1. dersh1234

    April 6, 2016 at 1:04 am

    I’m afraid Sanders polled far worse that 40% with AAs (your assumption) Rather, he received 27%. As for the non-white (non AA vote) they don’t even have stats–too few people to represent..

    • Jared Diamond

      April 6, 2016 at 9:51 am

      do you have a link?

    • Jared Diamond

      April 6, 2016 at 9:56 am

      Also, by my Math he wins at Least 59.8% of the vote Even if he gets 0% of the non-white vote, the fact that the white population makes up 88% of the 100% voting population, he still wins larger than he actually did. Assuming he wins 20% of the non-white vote, he should get 62% of the total vote. Math: .68*.88 = 59.84 || 59.84 + ( .2 * .12 ) = 62.24%

    • Richard Charnin

      April 8, 2016 at 7:31 pm

      He had 51% in the PPP pre-election poll and 31% in the final ADJUSTED poll. Does that seem right to you?
      Do you believe the adjusted exit poll? If you do, then you believe the election was fraud-free.

  2. oscodaartist2013

    April 6, 2016 at 11:00 pm

    Last thing I saw before the Wisconsin election was that Bernie was leading Clinton with Black voters by 11 points. Here’s one link

  3. oscodaartist2013

    April 6, 2016 at 11:02 pm

    It was PPP poll from March 31. Here’s the direct link

    • Richard Charnin

      April 8, 2016 at 7:29 pm

      Thanks for that. I included it in my post. Bernie at 51% of blacks in the PPP pre-election poll but just 31% in the adjusted final exit poll? No way.
      That is how the exit pollsters at Edison Research (who work for the National Election Pool) cover up the fraud – by adjusting the poll numbers to match the bogus recorded vote.

  4. nightengrrl

    April 7, 2016 at 6:45 am

    I’d be interested in your thoughts on the Kloppenberg race. Exit polling had her projected to win, but ended up not.

  5. Christo

    April 7, 2016 at 5:23 pm

    Just wondering, Richard, what you think the delegate count ought to be, had there been no vote tampering. Where would we be had the actual results reflected the unadjusted exit polls? Would Bernie be ahead, and by how much. What are the probabilities?

    • Richard Charnin

      April 8, 2016 at 6:47 pm

      I estimated that Bernie would have at least 100 more delegates

  6. Carol

    April 9, 2016 at 1:58 pm

    Richard, I remember you well as TIA at DU and have followed you. Have the Sanders people been informed of your analysis?

  7. Devon Delsole

    April 15, 2016 at 7:21 am

    Richard, My name is Devon Baldwin/DelSole. I am a part of an Election Justice organization. Due to the public nature of this comment, I can’t say much more here. I would like to talk with you about your findings. We are pursuing this legally and I’d love to talk about how your data can help. Please feel free to add me on facebook or email me at


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: