Joshua Holland’s editor at The Nation apparently did not think much of his work to debunk election fraud allegations in the contest between Hillary Clinton and Bernie Sanders. Instead, Holland put the half-baked column out at Raw Story after giving a softball interview, replete with a textbook example of circular reasoning, to the sole exit pollster for this election cycle. As with his since debunked debunking of a federally coordinated crackdown on the Occupy movement, Holland afterward went on his merry mocking way. Here, even though he acknowledged that he’d been alerted by dozens, if not hundreds of people, that a citizens group gave sworn testimony to irregularities in Chicago’s audit of electronically cast ballots between Hillary Clinton and Bernie Sanders, Holland did not bother with so much as a phone call to Chicago’s Board of Elections to ask about those allegations.
Where conspiracy theories can be nipped in the bud, they should be. But bad reporting helps no one. Maniacal people have viciously attacked the survivors of victims of the Sandy Hook shooting in part because unscrupulous media initially reported multiple shooters then did nothing to follow-up with the very simple explanation for why they got things wrong.
CounterPunch has a history, even if I do not personally agree in every case, of rejecting big United States conspiracy theories from those surrounding John F. Kennedy’s assassination to 9/11 trutherism. I have spent weeks investigating election fraud claims including interviewing multiple exit pollsters, veteran hackers, academic experts on United States elections, and elections officials and workers in multiple states. These include the Edison exit pollster interviewed by Holland along with a spokesperson for the Chicago Board of Elections. Over the course of several articles, I will be attempting to debunk election fraud allegations where they can be, or, alternatively, to provide the best explanations for why things seem amiss in so many different primaries and caucuses in the race between Clinton and Sanders.
I began with a set of factors to help determine which states required most attention for investigation:
1-Candidate Beats the First Full Wave of Exit Polls by 7.0% or more + 1 Other Factor
2-Evidence Suggesting Substantial Election Fraud Without Significant Rebuttal
3-3 or more of Factors 4-9 Below
4-Candidate Outperforms First Full Wave of Exit Polls by between 3.5% and 7.0%
5-Actual Evidence of Substantial Election Fraud With Some Rebuttal
6-Actual Evidence of Some Election Fraud Without Significant Rebuttal
7-Substantial Areas of State Use Voting or Tabulation Machines Proven to be Hackable or to Miscount
8-Candidate Strongly Outpolls Pre-Election Polls with No Obvious Explanation
9-State Uses All or Majority Voting or Tabulation Machines Greater Than 10 Years Old
Any one of the factors in numbers one through three means a particular state requires significant attention to explain what went wrong, though, for instance, a single good explanation for why exit polling is so bad would knock out most cases.
I will be saying more about the various factors over the course of the series. For now, the first factor, one that has drawn a decent amount of attention in some spheres already, is comparison to exit polls. I’ve already noted that when countries with governments the United States does not like hold elections, exit polling is taken as fairly authoritative in establishing the possibility of election fraud.
“But the polling is different in the United States!” runs a certain logic.
Simply making this assertion without backing it up with data and sound argumentation does not work. I will be looking at those exit polling differences, where they matter, and at exit polling for primaries as a whole in Part 3. For now, I will grant that, in and of themselves, exit polls cannot make the entire case. After all, when exit polls sponsored by the United States government missed badly in Venezuela, the Carter Center negotiated a rigorous audit that would not be possible in many U.S. locations because of poor election machine design. The Carter Center then observed the details of the paper audit from beginning to end. The previous results were confirmed.
But that does not mean exit poll differences can be ignored altogether. When the recall election for Scott Walker in Wisconsin showed a 7% deviation from exit polling, FiveThirtyEight’s Harry Enten addressed what seemed like a outrageous difference, only to argue that it was not so outrageous. Part 3 will address arguments like Enten’s, his colleague Nate Silver’s “Ten Reason’s Why You Should Ignore Exit Polls,” and similar suggestions that exit polling in the U.S. is just bad and we should accept it without question. In short, none of them weigh the evidence for and against manipulation of electronic voting machines a vote tabulators seriously.
Exit polling has been far more accurate on the GOP side, even with multiple candidates, including multiple anti-establishment candidates. There has been just one miss outside the expected margin of error, so far as I can tell, perfectly in keeping with what one expects in terms of scientific polling’s accepted accuracy interval of nineteen times out of twenty. Even where things have been closer to the margin of error, they have not all favored one candidate, least of all the GOP winner. Donald Trump has overperformed near the margin of error, or by more than 3.5%, just one time. Ted Cruz has overperformed four times, once outside the margin of error (Texas), and Marco Rubio underperformed significantly once. “Clinton, meanwhile, has overperformed by at least 3.5% sixteen times, nine of those times outside the margin of error.”. Sanders beat a 3.5% margin versus exit poll expectations just once (Oklahoma), and it was within the margin of error.
Using the criteria above, I initially grouped the states into three categories. Having one factor of numbers four through nine is not enough to move a state into the “moderate” category. Some of the states, as we will note particularly tomorrow in “Debunking Some Election Fraud Allegations (Part 2),” can quickly be downgraded with a little research and reasoning. Others require far greater scrutiny to figure out what has happened. (The word “substantial” matters. I am not concerned in this series with the electioneering hi-jinks reported in every election ever from dogcatcher to the highest offices in the land. Bill Clinton’s election day antics have irritated people and are the subject of a lawsuit, but they are not likely responsible for even one delegate of Sanders’ current 275 pledged delegate deficit. As I’ve written previously, if exit polling had been reasonably accurate, the gap between Clinton and Sanders would likely be well over a hundred delegates closer.):
Strong Potential Indicators of Substantial Election Fraud for Sanders Versus Clinton (12)
South Carolina, Alabama, Georgia, Massachusetts, Tennessee, Texas, Mississippi, Illinois, Ohio, Arizona, New York, Delaware
Moderate Potential Indicators of Substantial Election Fraud for Sanders Versus Clinton (8)
Virginia, Arkansas, Michigan, Missouri, Pennsylvania, Louisiana, Wyoming, Indiana
Little to No Potential Indicators of Substantial Election Fraud for Sanders Versus Clinton (21)
Oklahoma, Wisconsin, Maryland, Vermont, Minnesota, Kansas, North Carolina, Florida, Connecticut, Iowa, Nevada, Rhode Island, New Hampshire, Colorado, Nebraska, Maine, Idaho, Utah, Alaska, Hawaii, Washington
Danger Zone for Upcoming States:
California, New Jersey
Tomorrow’s piece will begin with a look at accusations of fraud for states like Wyoming, Delaware, and Iowa. As the week wears on, we will turn attention to more acute suggestions of fraud. Throughout, I will return to what I call the Alabama Test. I will be taking very seriously the best explanations for why, what people are perceiving as indicators of fraud, may actually not show fraud. But then we will apply the Alabama Test. Could a particular explanation help tell us why the first full wave of exit polling in Alabama missed the eventual outcome by fourteen points?
While Alabama may seem like an unusual place for a test case – Bernie Sanders didn’t even compete there – the proportional nature of Democrats delegate allocation means it matters. What’s more, various characteristics of Alabama’s election and exit polling put particular stressors on various explanations, fraudulent and non-fraudulent, for why exit polling missed there worse than any other place so far. Not all of the allegations of fraud involve exit polling, of course. Part 4, as planned, will look more in depth at the question of switched registrations, especially in Arizona and New York.
My commitment is, as ever, to follow the best evidence wherever it might lead, whether to fraudulent, non-fraudulent, or inconclusive conclusions.
Correction: A previous version of this article incorrectly stated the number of times Clinton has outperformed the first full exit poll by at least 3.5%. The correct number is sixteen rather than nineteen. I apologize for the misstatement.
Part 1: Taking Election Fraud Allegations Seriously
Part 2: Debunking Some Election Fraud Allegations
Part 3: In-depth Report on Exit Polling and Election Fraud Allegations
An Interview With Lead Edison Exit Pollster Joe Lenski
Part 4: Purged, Hacked, Switched
Part 5: Chicago Election Official Admits “Numbers Didn’t Match”
Part 6: Clinton Does Best Where Voting Machines Flunk Hacking Tests