Accusations of election fraud have been rampant throughout the 2016 primary. Recent WikiLeaks revelations about collusion by the Democratic National Committee (DNC) against Bernie Sanders have reignited cries of foul play. As the Democratic National Convention heads to its climax, the question remains, “Did Hillary win fair and square?”
When contesting an election, many want absolute proof of fraud. But unequivocal evidence is hard to find. It often requires painstaking work of many teams and substantial resources. Questions about fraud often begin with just a hunch, but only the dedicated few follow it up with an investigation.
Alas, where politics is involved, attempts are usually made to stifle any inkling of moving to an actual investigation phase. Even widely held, logical hunches are delegitimized by the establishment-biased media as mere “conspiracy theories.” That is how seemingly credible analysis by the likes of Greg Palast are dismissed out of hand, leaving the entire question of fraud — by all accounts a very serious issue — in the hands of comics like Lee Camp, who are apparently a few of the only ones still allowed a voice on the matter.
Stay in the loop
Never miss the news and analysis you care about.
On the other end of the scale, once suspicions creep in, it can be easy to assume everything is evidence of fraud, and that makes it hard to sort the wheat from the chaff.
Yet, whatever one believes about the election, there is too much at stake to dismiss the charges without as much as a nominal consideration of the evidence. With that in mind, and in the hope of bringing some clarity to issues raised, a brief overview is provided of some of the more prominent claims — and counterclaims — that underlie assertions of possible fraud in the Democratic primary.
There has been considerable uproar by Sanders supporters about large disparities between exit polls and actual results in 11 states — all favoring Clinton. However, the legitimacy of exit polls as a tool for verifying elections is a matter of some dispute.
For example, James Carter IV (Jimmy Carter’s grandson) has monitored multiple elections in other countries and supported other monitoring groups from the Carter Center for 17 years. He cannot recall exit polls being mentioned once as a factor in judging elections. Indeed, his experience indicates the US government will accept any election result, regardless of fraud, if the outcome is to their liking.
In contrast to Carter, election specialist Brad Friedman of TheBradBlog says the United States has used exit polls to monitor elections abroad. USAID, International Republican Institute (IRI) and National Democratic Institute (NDI) are some examples.
Friedman feels exit polls here can be used as possible red flags, but believes too much emphasis has been placed on them. His reasoning is that exit polls in the US are not designed to detect fraud like they are in other countries. Rather, they are meant to gather demographic data about who voted and why, so the press will have something to report before early results are in.
Joe Lenski, executive vice president of Edison Research who conducted exit polls for the 2016 primary, agrees. Lenski says that in order to detect errors in vote count, many more respondents would need to be approached to minimize sampling errors, whereas Edison only polled about 200 voters per location. Furthermore, the length of the questionnaire would have to be shortened to maximize response rate. In addition, hundreds of sample polling locations would be needed, rather than the 20 to 45 Edison polled in each state.
Election analyst Josh Mitteldorf disagrees on this last point, stating that a smaller number of precincts may be all that is needed for an election verification poll. He asserts that results could be made more reliable if compared precinct-by-precinct, instead of aggregating them statewide.
Unfortunately, this kind of analysis cannot be done at this time because Edison will not release polling locations or raw data. According to Lenski, Edison adjusts exit poll data throughout the day and reports results three times — in the morning, mid-afternoon and post polls’ closure time. Once all the polls close, the data is further adjusted, sometimes several times.
Edison’s reluctance to disclose the raw data or the precincts polled naturally gives rise to questions about lack of transparency. Lenski cites respondent privacy as the principal reason for maintaining the confidentiality of raw data. Yet, as polling analyst Jonathan Simon asserts, this reasoning by itself is unconvincing since anonymous preference responses and basic demographics would not be traceable out of hundreds of precincts.
As for Lenski’s argument that precincts are not revealed in order to prevent outside groups from trying to unduly influence voters, Simon raises the counter-argument that just by changing polling locations post-election, any extra voter outreach efforts could be avoided as campaigns would not know which precinct to target.
Given their lack of transparency, it is not surprising that a lawsuit has been filed (by attorney Bob Fitrakas) aiming to force Edison Research and The Media Consortium to provide access to the raw data.
Aside from problems doing analysis given Edison’s secrecy around key data, many feel that exit polls here can still be useful indicators of possible fraud. As activist Bev Harris of Black Box Voting quips, “It seems incredibly presumptuous for the media to make predictions and even announce winners based on exit polls, if they are not even designed to predict results.”
Election analyst Richard Charnin also dismisses any notion that Edison polls cannot detect fraud. He points out that any poll designed to detect fraud asks one simple question, “Who did you vote for?” Since Edison asks this question, if we had access to the raw data, he reasons it would tell us what we need to know.
Charnin believes exit polls have been exposed in this election as a major issue, because so many have taken notice of their value in detecting fraud through people like him. Charnin has become rather well-known due to shows like “Redacted Tonight.” Host Lee Camp’s interview of Charnin garnered 50,000 views.
Charnin feels that the spotlight on poll disparities coupled with a racketeering lawsuit recently announced by attorney Cliff Arnebeck is what led to Edison canceling their exit polls in California and other remaining states during the June cycle.
Margin of Error Theories
Although there were a large number of exit poll disparities skewed in Clinton’s favor, some argue that this alone does not prove fraud. There may be many reasons for such disparities, one being early or absentee votes.
Edison combines exit polls and pre-election telephone polls in states with 20-25 percent (or more) early voting. Pre-election telephone polls of absentee and early voters were conducted in eight states (Arkansas, Georgia, Tennessee, Texas, Michigan, Florida, North Carolina and Ohio), using random digital samples where anyone with a landline or cell may be called seven to 10 days prior to an election. Edison interviews those stating they already voted or will definitely vote prior to election day.
As analyst Richard Hayes Phillips explains, incorporating phone polls may contaminate the raw data. Exit polls have often been touted as being more reliable than pre-election polls because they are sampling people who have already voted vs. those who have not voted yet and who may still change their mind. Therefore, combining both into a single sample, effectively muddies the water.
Other factors can skew polls as well. A common theory proposed to explain the numerous disparities between exit polls and actual results, all favoring Clinton, is that Sanders voters were more enthusiastic, thereby increasing the response rate. James Carter also posits that youth voters known for being pro-Sanders were more likely to vote in groups, so if only one in 20 lied due to peer pressure, that would skew results by as much as 10 percent (5 percent in either direction).
Although Lenski agreed with the enthusiasm factor possibly introducing some bias in favor of Sanders, he seemed to discount the likelihood of this occurring by explaining that Edison adjusts for age and tries to factor for enthusiasm as they evaluate the data. At least one analysis found that the enthusiasm theory does not hold up when applied to states with lower youth turnout. In addition, Lenski said that exit polls are not conducted verbally, eliminating the peer pressure factor.
Despite various reasons for disparities beyond the margin of error, many analysts feel exit polls are important for identifying questionable results. Ultimately most agree that it is only by doing hand counts or audits that fraud — or honest mistakes — can be proven. But examining discrepancies provides another layer of checks and balances to help us know where to investigate.
Unfortunately, each state has a different voting system and only some have a paper trail that can be audited or hand counted. It is for this reason that Bev Harris asserts that since computerized voting does not allow verification of vote, it is impossible for anyone to categorically state there was no fraud in the primary. She likened it to doing a financial audit where the auditors say there is no evidence of fraud, but were not allowed to see the books.
Oklahoma, New York, Arizona, California
In addition to the issue of exit poll disparities, there are a number of states that have raised questions about the fairness of the elections.
Oklahoma is one of them. Jonathan Simon points out that Oklahoma was the one state where the disparity was skewed in Sanders’ favor, but also happened to be the only one where state officials rather than private vendors programmed the computers.
Simon has several hypotheses for why this happened:
1.) The votes were counted accurately because a rigger did not have easy access to the programming
2.) Edison began building in a bias towards Clinton in an effort to stop getting the primaries wrong, which could be rationalized by having to account for such things as the purported enthusiasm gap, and
3.) Edison’s adjustment to polls upweighting women, and/or African Americans and/or older voters backfired in Oklahoma where the elections were not rigged, so the margin of error disparity was reversed.
New York and Arizona both received widespread press for other perceived irregularities which seemed to point to voter suppression and purging tactics. Although James Carter believes election fraud exists in the United States, he saw no evidence of it in either New York or Arizona:
The voter purge in NY happened in a heavily pro-Hillary area, and there is no way that some vote rigger would have known who to purge. People purged who were planning to vote Bernie were the ones we heard from, though. Has anyone actually tried to figure out who the rest of those voters were planning to vote for? In Arizona … Hillary’s campaign signed on to the DNC’s lawsuit against the county before Bernie’s did. Again there’s no evidence that it was an anti-Bernie conspiracy…. There is evidence of huge problems in our election system that should be discussed more regularly and that need to be fixed. It is not fraud, however.
Brad Friedman also does not see evidence of fraud in either New York or Arizona — and he is quick to say, “And I’m no fan of Hillary.”
He cites a study by New York Public Radio (WNYC) analyzing 122,454 people purged from the Brooklyn voter rolls. The study found that a disproportionate percentage were Latino. Friedman reasons since Latinos tended to favor Clinton, the purge probably hurt her more than Sanders. (Median age purged was 53).
California, on the other hand, is receiving a lot of scrutiny for other reasons. Bev Harris thinks the difference between early and late counted votes warrants investigation, “It doesn’t really pass the sniff test.”
Harris feels that the early count does not match the enthusiasm by Sanders voters. It was initially reported that Clinton had 63 percent of the vote in California. Harris felt that this was a staggering number, and over time, we have seen a trend reversal where late votes went 75 percent for Sanders, which she thinks signals a red flag. She also notes a similar pattern in Oregon, which was 100 percent vote by mail.
Harris hypothesizes there is a simple way this could be done. Vote by mail ballots go down a conveyer belt where a camera photographs the barcodes and signatures for voter authentication. If the signature does not match the one on file, the ballot gets diverted so it can be checked by humans later. It would be easy to have a database that flags Clinton vs. Sanders ballots. A higher level of scrutiny could be set for the signature verification on Sanders ballots so they are disproportionately diverted. When ballots are delayed from being included in original count, there is a greater chance they’ll be tampered with, lost or discarded.
Harris postulates early/late vote differences were part of a pattern to prematurely present Clinton as the winner. She says it has been documented for decades that there is a huge pressure to get candidates to concede early so nobody cares about final vote tally. The Associated Press declared Clinton the winner of the Democratic primary 24 hours before voting even began on June 7, and said 100 percent precincts had reported — despite millions of uncounted ballots. (The AP says 100 percent precincts have reported when any votes have been registered. It is more accurate to say 100 percent precincts have “checked in”). So on election night, viewers were led to believe vote count was complete, when in fact it was not.
Reporter Greg Palast has also been analyzing early and late votes in California. He says Bernie Sanders actually won the state. Palast said that most of the late mailed-in ballots were No Party Preference (NPP) and estimates that 3 to 1 went for Sanders. This assumption is based on previous polls showing Sanders at about 3 to 1 amongst NPP voters. Palast says most provisional ballots were NPP voters, so even conservatively estimating half the votes went for Sanders, the number of late votes would have put him over the edge.
Some feel Palast was premature in making such a strong statement, which was based on certain theoretical assumptions. However, Palast’s initial statements were part of an interview rather than an official article, and he did say there were more details to come.
The Black Vote and Absentee/Early Ballot Discrepancies
Richard Hayes Phillips is conducting a nationwide analysis of voting results. So far, he is seeing a noteworthy trend with disparities between absentee votes compared with poll votes.
Michigan is a prime example. In Detroit, Clinton won 69.1 percent to 30.0 percent at the polls, but 86.9 percent to 11.2 percent with absentee ballots. In neighboring Oakland County, Clinton lost 51.9 percent to 47.4 percent at the polls, but won among absentee voters 67.9 percent to 25.5 percent. Absentee ballots made up 21.05 percent of total ballots in Detroit and 19.18 percent in Oakland. Phillips feels these kinds of large differences between absentee ballots and poll results warrant audits.
Phillips is also seeing questionable trends with African American votes. The presidential preference of African Americans was reported in 21 states — 10 of which were in the South. Clinton reportedly received 80 percent to 91 percent of the Black vote in Southern states, whereas she got 67 percent to 75 percent of the Black vote elsewhere.
African American votes appear to have been adjusted upward in the final exit polls in 15 of 19 states. Phillips thinks this trend raises questions about whether Sanders really got less than half the African American vote in the South, or whether vote counts may have been altered.
This cursory summary is by no means comprehensive. It does not even begin to address caucus states such as Iowa, Wyoming and Nevada, as just a few examples. Nevertheless, the various election integrity experts raise a lot of legitimate questions. It leaves voters wondering, “If there’s so much smoke, could there be fire?” The WikiLeaks emails showed that at the very least, the election was rigged in favor of Hillary Clinton. But what we don’t know is if there was actual fraud. We may never know the answer, but the question will remain an indelible stain on the Democratic Party.