Friday, August 16, 2013

Adjusting Exit Polls? Assumptions Make All The Difference

I was sent this post by a friend of mine from Healthcare 4 All PA on how the unadjusted exit polls in the Presidential Elections from 1988-2008 show voter fraud in favor of the Republican Candidate. It argues that the polls were adjusted to reflect the official election results.  The post can be read here:

Election Fraud: An Introduction to Exit Poll Probability Analysis

Blog author mathematician Richard Charnin claims that the graph above proves that there is massive election fraud as the size of the margin of error in local election exit polls seems to have decreased over the last 20 years.  There were a total of 126 exit polls that exceeded the margin of error over this period and 123 of them were won by the GOP.  He also argues that the popular vote percentage is consistently greater for the Republican winner than the Democrat in the exit poll.  

He relies on Richard Roeper's Roper Center Public Opinion Archives for the raw poll data which are unadjusted.  He seems to assume that the raw data are more reliable are more accurate than the adjusted numbers.  Charnin does a decent job of discussing the normal, binomial, and Poisson probability distributions and uses them to argue that the Democrat should have won every Presidential election since 1988-2008.   The author does not seem to consider that the totals are adjusted for population differences and/or sampling error.  For some this sounds like fudging of the numbers but there are theoretically valid ways to adjust sample values to estimate population values.

I have written many posts which have taken exit poll results at face value except for one in the Healthcare 4 All PA/PUSH where I noticed an inconsistency between national, PA exit poll results in 2012 and previous poll results on whether the public wants a better health care law.  Does this support Charnin's claim that exit polls are skewed to mask voter fraud?  Not necessarily,  It takes a lot more data to prove that there is systematic skewing of the data.  The assumptions one makes can invalidate the best of statistical methods and the most beautiful of graphics.

Dean Chambers of argues exactly the opposite of Charnin with Romney winning with 51% and now insinuates that there was massive voter fraud against him.  His method page was taken down now but I saw it before and critiqued it here.  He made similar assumptions to Charnin and used methods to fit his beliefs.  All of this may confirm to the layman the adage "there are lies, damn lies, and statistics."  A more careful reading of the numbers can separate the signal from the noise (as Nate Silver would say)  to find the relevant info.

**Related Posts**


  1. Paul,

    You are making a number of errors:
    1. You shuld not compare my exit poll analysis to the bogus Dean Chamber pre-election forecasts

    2. And you should not quote Nate Silver. To Nate and apparently to you, Election Fraud does not exist. Well, I have proved it is systemic by my independent True Vote Model which is consistently close to the unadjusted exit polls.

    3. FWIW, I exactly forecast Obama's RECORDED 2008 and 2012 Electoral votes (365,332) in my Election Forecat Model. But I also forecast his True Votes which were much better. The TVM exactly matched the unadjusted 2008 state exit poll aggregate: Obama had 58%. Check out my track record here:

    Regarding exit polls, I debunked Nate Silver (and now you) here:

    Now, to prove exit poll fraud, one only needs to look at Bush's stolen 2004 election: There were 5-6 million phantom returning Bush 2000 voters, based on the adjusted National Exit Poll which was forced (as all exit polls) to match the recorded vote. Read about it here:

    I have written two books on proving election fraud:

    You should go to my website which links to my blog. But scroll down the Website to view the True Vote Model results for all elections since 1968:

    I invite you to run the National 1968-2012 True Vote Model yourself:

  2. This is an incorrect statement.
    "Blog author mathematician Richard Charnin claims that the graph above proves that there is massive election fraud as the size of the margin of error in local election exit polls seems to have decreased over the last 20 years".

    I never said that. I do show the number of state and national PRESIDENTIAL exit poll respondents in this 1988-2008 exit poll spreadsheet:

    The post:

    By the way, if you want to check the State and National True Vote Model for 1988-2008:

    You can run it for any state in any election using 5 different methods. Go to the INPUT sheet to enter your assumptions. You need only enter the state code, election year and calc method (1-4). THE RESULTS WILL BE DISPLAYED IN THE 'MAIN' SHEET.

    In cell C26, the default (no entry) will display the EXIT POLL ADJUSTED TO MATCH THE RECORDED VOTE. Enter 1 to view the UNADJUSTED exit poll.

    1. Thanks for your comments Richard. I wasn't saying that there was absolutely no voter fraud in recent elections. I was questioning your assumption that the raw data is necessarily more accurate than the adjusted data. You and Chambers seem to assume that there is an attempt to skew the numbers in favor of the other party. There is always sampling error in surveys and problems of non-response in raw data. The survey firms have sophisticated methods dealing with this error. I did discuss an anomaly that I found in a healthcare question in the exit poll from 2012 which may support your argument but more data would be needed than a probabilistic model based on raw exit poll data to prove fraud. A paper trail for the actual votes would be better.

      Fraud on such a large scale at the national level would be hard to cover up. Fraud at the local level (such as certain counties in Florida) is easier to cover up.

  3. You are sadly mistaken. It is a fact that unadjusted exit polls are ALL adjusted to match the vote count. The 8% marginal discrepancies are much too large to attribute to anything but fraud.

    It is also a fact that my True Vote Model has confirmed the unadjusted exit polls to within 1%. Have you bothered to look at the models I linked to? And I mean really look.

    What you are saying in effect is that the exit polls are always wrong and the recorded counts are always right. That indicates that you are unfamiliar with or choose to ignore the factual data and the models. It is quite amazing that you would make that statement.

    I mentioned the 2004 election and the 6 million phantom returning Bush 2000 voters. I gave you the link to the model and a descriptive post. Apparently you have not read it. Or is it that you cannot explain the anomaly? IT IS ABSOLUTE MATHEMATICAL PROOF THAT BUSH STOLE MILLIONS OF VOTES - IN BOTH 2000 AND 2004. If you choose to reply, please discuss the data ans assumptions specifically.

  4. This post destroys the myths about 2004.

  5. Mr. Ricci, I'm asking you to think about your statement, "Fraud on such a large scale at the national level would be hard to cover up. Fraud at the local level is easier to cover up."

    I'm not familiar (sorry!) with your areas of expertise, but it doesn't sound as if election technology and election fraud strategies are among them. Fraud sufficient to swing a national election would most likely consist of targeted local-scale interventions, which is easier than you might think because of the privately controlled and centralized programming of election computers. If you can envision manipulation of election tabulators on a small scale, you should be able to envision how a statewide or national election could be manipulated.

    Normally prudent people like you and me and most voters tend to assume that prudent voting-machine audits routinely verify that pre-election security measures worked as intended. But such assumptions are wrong.

    Our elections are on the functional equivalent of an honor system. Election officials make very little effort after each election to verify the machines counted correctly--in most jurisdictions after most elections, none. Most typically, they test the computers before Election Day and then, after the polls open, do nothing more than trust that no hackers got past their (usually amateurish) security. Many states don't retain enough record of the votes even to allow a post-election audit. Other states have an auditable election record but never actually audit it.

    If election officials could produce post-election audits that prove the machines counted accurately, they could easily refute people like Mr. Charnin and Mr. Chambers. But they cannot produce that evidence because it does not exist.

    Most of us--including you, it seems--find the idea of stolen elections so alarming that we don't ask even basic, obvious questions about what our election officials are doing to verify the machines' counts. Would-be hackers already know. When you ask those questions, you will, too.

    1. Thank you for your critiques Richard and Karen. I agree that independent verification is essential to detect fraud and abuse (see post The Need for Exactness linked to in **related posts** above). The problem of our times is having access to raw data. I did write a post on a 2012 exit poll question on health care (linked to in the post above) which seems to be consistent with Charnin's argument but is that necessarily proof?

    2. Thanks for the reply, Paul. Focusing purely on the statistical questions, your earlier post makes sense. From the point of view of a citizen who wants to be sure our election results consistently reflect the will of the electorate, those suggestions seem irrelevant at best, distracting at worst. With all due respect, the statistical challenges in protecting our elections are the easy ones.

      Remembering that it takes vote-flipping in only a few key precincts to steal a winner-take-all election, think it through the way a would-be election manipulator would.

      Let's assume, for the sake of this thought exercise, that you are a major campaign donor who seeks to install elected officials who will use our government to benefit your business, and who has enough power and money to control one or more of the voting-machine-company employees who service the voting machines. (What do you suppose are the odds that such a person exists?) One effective approach: Take the precincts that are 'safe' for your candidate and make those results overwhelming, so that people will shrug their shoulders at the news that your candidate did well there.

      Will the jurisdiction(s) in which you are interested in flipping votes have the will and ability to conduct credible exit polls? How big of a difference between the exit poll results and the machine tallies will be enough to cause the electoral officials to believe the exit poll and reject the machine results? (Do you know of any laws--I don't--that would require the results of a miscounted election to be overturned even if they did?)

      Will the jurisdictions in which you want to flip votes have the ability and willingness to spend tax dollars on the sort of independent post-election software analysis required reliably to detect malicious software? And if they do, how will the forensic software analysts be certain they are looking at the actual software that was operating in the voting machine on Election Day? And if the analysts find a problem, will the elected officials (whose campaigns you financed or who hope you do not finance their next opponent) want to brave the disruption and discord that would follow if people believe the analysis--or will those elected officials find it easier to discredit and disavow the post-election analysis?

      Next time someone asks you, "How can voting machines with no paper trail be checked for accuracy?" please answer: "By taking them to a world in which election officials have the independence, ability, and funds to conduct redundant elections from which the results can be used to verify or nullify the results reported by the paperless voting machines, and in which the voters have the patience to vote in at least two elections and the sophistication to read and evaluate the results of the post-election reports on the accuracy of the voting machines, and to insist that effective action be taken if problems are discovered. In that world, the following statistical methods will be useful..."

      Paul, I for one do not think it necessary that Richard Charnin ever proves past-tense fraud with his analyses. It might help to wake people up, but it's not necessary. Anyone who cares to look can see with his or her own eyes that our elections are vulnerable and that this world does in fact contain people with the resources, motive, and opportunity to steal our money and national resources by stealing our government by stealing our elections. That's not paranoia; that's basic knowledge of history and patriotic prudence.

      Please join us in promoting the cause of true election integrity, which can come only when we successfully deter and detect election manipulation by using voter-marked paper ballots and verifying election results in publicly observable hand counts. It's less expensive and more effective than layering on even more computer security.