Dear Mr. Holland: After studying and assessing your work this semester, it is with deep regret that I have to inform you that you failed Social Science Statistics 101.

As you know, you have characterized us as “conspiracy theorists” because in our STRIP & FLIP SELECTION OF 2016: FIVE JIM CROWS & ELECTRONIC ELECTION THEFT, Harvey Wasserman and I have suggested that exit polls matter. You have also publicly denounced our colleague Richard Charnin, who has two separate Master’s degrees in Applied Mathematics, for his analysis of this year’s primary exit poll results versus election results.

Since you show so little interest in statistical analysis, let me briefly go over what you should know:

First of all, exit polls are the accepted international standard for indications of election fraud and vote tampering. Here I refer you to Eric Bjornlund and Glenn Cowan’s 2011 pamphlet Vote Count Verification: a User’s Guide for Funders, Implementers and Stakeholders. Their work, done under the auspices of Democracy International for the US Agency for International Development (USAID), outlines how exit polling is used to ensure free and fair elections.

“U.S-funded organizations have sponsored exit polls as part of democracy assistance programs in Macedonia (2005), Afghanistan (2004), Ukraine (2004), Azerbaijan (2005), the West Bank and Gaza Strip (2005), Lebanon (2005), Kazakhstan (2005), Kenya (2005, 2007), and Bangladesh (2009),among other places,” the pamphlet states.   

When election results do not match exit poll results, we should not simply accept these results. What Charnin does, and has been doing for many years, is studying improbable election results that fall statistically outside the “margin of error” (MoE).

For example, Ohio primary exit polls indicated that Clinton would win 51.4% to Sanders’ 47.6%. She was expected to win by 3.8%. The actual vote indicated she won 56.5% to 43%. Clinton won the election by 13.8% which was 10 percentage points more than the exit polls indicated. 

Statistics tell us that the correct Ohio MoE was 3.12% based on N=1670 respondents. There is a 0.1% probability that the 5.1% exit poll discrepancy from the recorded vote was due to chance. Therefore there is a 99.9% probability that the official Ohio primary results were improbable.

These results should trigger further investigative analysis. 

Take a look at Charnin’s statistics at this post for Democratic primary MoE and probability calculations. https://richardcharnin.wordpress.com/category/2016-election/

Please tell me, Mr. Holland, where Charnin’s wrong – and show your math.

If these results happened in the Ukraine in 2004, the U.S. State Department would be denouncing the election as fraudulent and demanding an investigation or a revote. By the way, this did happen in 2004. You may recall the Orange Revolution when Ukrainian people took to the streets to protest the fraudulent election. Unlike Mr. Holland and some of his peers, Ukrainians care about stolen elections.

Why is this important? Obviously, if you don’t understand or accept exit polling and can’t analyze poll results versus election results, you won’t recognize election tampering. And that’s what those who would tamper with elections are counting on.

What you have failed to do is study and attend classes where we went over the basic acronym HISMISTER which provides basic guidelines to determine if a poll is valid.

H = Historical intervention. Did an unexpected historical event affect the results? In an exit poll, voters are asked how they voted as they leave the polls. There is generally no time for a historical intervention to happen between casting a vote and telling the pollsters outside the polling site how you voted.

I = Instrumentation. Were the correct instruments used? If the pollsters asked who people voted for in the presidential race, we assume they were using the correct survey instrument. However, if the voting machines did not record the correct votes, that could cause a discrepancy in the corresponding election results. See “R” below.

S = Sample. Was the sample of voters polled randomized and representative of the demographics of the state? The problem with your work is that you never offer any example for why the exit polls are wrong. You have never taken issue with the pollster’s sample.

Also, while you denounce Charnin’s analysis, you never point out how or why his math is wrong based on standard statistical probability.  

M = Measurement. Were the same identifiers studied in the comparison? In this case, the pollsters asked voters who they voted for in the president race and those numbers were compared to presidential election results. This would not appear to be a problem.

I = Implementation. Were the procedures used to collect, organize and analyze the data done correctly? If the exit pollsters went to the right precincts and used the correct sample based on the right demographics, we assume the implementation was correct.

S = Survey. Did the survey, or poll, reflect the correct methodology? There has been no challenge to the exit pollsters’ methodology.

T = Technique. Was the poll conducted in an unbiased professional manner? Generally polls are conducted by well-paid professional pollsters.

E = Errors. Are there unexplained unintentional human errors? For example, a poll worker or election official incorrectly enters voting data.

R = Recording. If applicable, did the instruments record the information correctly?  When voting machines and central tabulators record different information than what voters are telling exit pollsters, then we must check to make sure our instruments recorded the vote correctly. And we must check to make sure that no one tampered with or rigged the recorded numbers.

On this section, I must give you a zero. There is a more likely than not probability that either faulty instrumentation or election tampering played a role in the improbable official election results. The United States fails basic standards of transparency because there’s no way to verify the actual vote total. Your blind faith in nontransparent voting equipment that is manufactured and programmed by private, partisan, for-profit corporations using secret proprietary software is inexcusable.

Here’s the correct answer: “We can’t verify the vote because seven states – Texas, Florida, Pennsylvania, Indiana, Kentucky, Tennessee and Virginia use Direct Recording Electronic (DRE) voting machines, i.e., computer voting without paper trails. In 18 other states, DREs are used with paper trails – generally a small piece of paper not designed to be recounted. In virtually all other states, private tabulators secretly record the vote.

I was an international election observer for El Salvador’s 1994 presidential election. Had the ARENA Party said that the votes would be counted on machines built by their friends in private industry and tabulated on computers belonging to major donors of their Party, my report would have been simple to write: The election is assumed to be fraudulent due to non-transparency and lack of auditing accountability.

Corporate for-profit media – CNN, the New York Times, Washington Post, ABC, etc. – claim that the official vote count must always be right. As a journalist you should be skeptical of these claims. What these media gatekeepers are telling us are that the universal laws of statistics apply everywhere in the world except in the United States. They would have us believe that when Clinton repeatedly beat Sanders by implausible numbers we are expected to lack an understanding of basic statistics and accept the improbable. They tell us not to pay attention to the technicians behind the curtain secretly programming the DREs and central vote tabulators.

With a legal suit looming that demanded both adjusted and unadjusted exit polls be made public after the California primary, we just learned the Edison group has canceled exit polls for all the remaining primaries. Edison has a monopoly on all exit polling for the mainstream media consortium.

We have now lost what Bjornlund and Cowan call “…an effective method for projecting election results.”

Professor Bob Fitrakis is co-author, with Harvey Wasserman, of the newly-published STRIP & FLIP SELECTION OF 2016: FIVE JIM CROWS & ELECTRONIC ELECTION THEFT.  Fitrakis has a Ph.D. in Political Science and a J.D., and has taught Political Science for 35 years He was an international election observer in El Salvador’s 1994 presidential election. He co-wrote and edited the International Election Observer report to the United Nations.