Chances of Wisconsin Presidential Vote Results shown to be 1 in 850, and Worse for Other States

The Columbia Free Press has published images showing the statistical likelihood of variance from the exit poll results that the reported vote counts represent.  In the case of Wisconsin (Figure 1), Trump received 44.3% of the exit poll share, but won the state with 47.9%.  The likelihood of such departure is 1 in 850, which is much less than 1% chance of happening, actually at 0.1% chance.

Figure 1.  The likelihood of the Wisconsin vote result relative to exit poll result.

Pennsylvania is another state slate for a recount request by Jill Stein.  This time, the Clinton vote share is shown, as underperforming the exit poll result of 50.5% with only 47.6% of the vote count (Figure 2).  The likelihood of this result is 1 in 60 elections, or 1.6%.

Figure 2.  The likelihood of the Pennsylvania vote result relative to exit poll result.

North Carolina is another state with odd results.  There, Trump received 46.5% of the exit poll share, but received 50.5% of the vote count, flipping the state in his favour (Figure 3). The likelihood of such an outcome is shown as 1 in 19,000, or 0.005%.

Figure 3.  The likelihood of the North Carolina vote result relative to exit poll result.

Results for other states and even senatorial races are shown at their article linked below.

Bob Fitrakis & Harvey Wasserman | Why the U.S. State Department Would Not Certify Trump’s Election as Legitimate | Columbia Free Press

Exit Polls in Other Countries Are Used to Verify Election Results, But Not in the U.S.

Fritrakis and Wasserman provide some background on exit polling, and how these are used to verify election results in other countries.  It is the best type of polling conducted because it’s not polls about people and how they are likely to vote (and may change their minds, opt not t0 vote, or are prevented to vote by inadequate IDs or the Cross-Check program), but exit polls are conducted on actual voters immediately after they have voted. Accuracy of polls is usually even less than 1%, and were tight with voting results for the Republican primaries, for instance.

In the lead-up to November 8, pre-election polls strongly indicated a Clinton victory. Post-Election exit polls showed her winning as well, most critically in the swing states whose Electoral College votes could give her the presidency.

Exit polls are the accepted international standard for indications of election fraud and vote tampering. 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.

It adds that “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.”

In countries like Germany and Switzerland, which use hand-counted paper ballots, exit polls are accurate to a margin error of less than 1%.

Here the 2016 exit polls were paid for by a major corporate media consortium, as has been standard practice for years. Here they are designed to reflect the actual vote count within a 2% margin of error nationally.

But in the US, if exit polls don’t agree with official vote counts, they are regularly “adjusted” to conform to official results, no matter how implausible. This makes fraudulent elections appear legitimate.

Bob Fitrakis and Harvey Wasserman | Did the GOP Strip & Flip the 2016 Selection? | Columbus Free Press

Clinton Camp On Board with Recount– Wisconsin to Proceed Perhaps by Next Week

This is a surprising new development, and a good sign.  NPR reports that the Clinton campaign supports the recount.

Hillary Clinton’s campaign said Saturday it will participate in the recount efforts in Wisconsin spearheaded by Green Party presidential candidate Jill Stein. If Stein also pursues recounts in Pennsylvania and Michigan, as she has pledged, the Clinton campaign will participate in those efforts, as well.

The recount in Wisconsin could begin as early as next week.

In a post on Medium, Clinton campaign lawyer Marc Erik Elias wrote that “regardless of the potential to change the outcome in any of the states, we feel it is important, on principle, to ensure our campaign is legally represented” in any recount proceedings. It was the first time anyone from the Clinton camp had publicly weighed in on the potential recounts.

From Clinton lawyer, Marc Erik Elias, in his post about this at Medium.  Citing the clear instances of Russian (and/or other) hacking interference in the politics of this campaign, he writes that they have been quietly assessing the situation:

For all these reasons, we have quietly taken a number of steps in the last two weeks to rule in or out any possibility of outside interference in the vote tally in these critical battleground states.

First, since the day after the election we have had lawyers and data scientists and analysts combing over the results to spot anomalies that would suggest a hacked result. These have included analysts both from within the campaign and outside, with backgrounds in politics, technology and academia.

Second, we have had numerous meetings and calls with various outside experts to hear their concerns and to discuss and review their data and findings. As a part of this, we have also shared out data and findings with them. Most of those discussions have remained private, while at least one has unfortunately been the subject of leaks.

Third, we have attempted to systematically catalogue and investigate every theory that has been presented to us within our ability to do so.

Fourth, we have examined the laws and practices as they pertain to recounts, contests and audits.

Fifth, and most importantly, we have monitored and staffed the post-election canvasses — where voting machine tapes are compared to poll-books, provisional ballots are resolved, and all of the math is double checked from election night. During that process, we have seen Secretary Clinton’s vote total grow, so that, today, her national popular vote lead now exceeds more than 2 million votes.

In the coming days, we will continue to perform our due diligence and actively follow all further activities that are to occur prior to the certification of any election results. For instance, Wisconsin and Pennsylvania conduct post-election audits using a sampling of precincts. Michigan and many other states still do not. This is unfortunate; it is our strong belief that, in addition to an election canvass, every state should do this basic audit to ensure accuracy and public confidence in the election.

Laura Wagner | Clinton Campaign Says It Will Participate in Recount Efforts | NPR

Marc Erik Elias | Listening and Responding To Calls for an Audit and Recount | Medium

Four Preliminary Pollsters Show a Closer Match to Exit Polls Than Vote Count

Figure 1.  Differences from exit polls, comparing a combined average from polls against the vote count (for states in which the margin of error of the vote count exceeds the margin of error of the exit polls).

Many have identified discrepancies between the 2016 presidential vote counts with the exit polls.  The preliminary polls are also good to check.  The punditry have generally dismissed the pollsters as just being wrong this year (even while they placed a great deal of trust in them right up until the day of voting).  One way to check the polls, though, is to evaluate the final polls against the exit polls.  In the chart above (Figure 1), I compare the differences of five polls against the vote count for the differences from the exit polls.

Earlier, I had posted a table comparing the final poll forecasts of FiveThirtyEight against the Exit Polls and the vote counts. This chart indicates the average of four poll-tracking sites:  TalkingPointsMemo, Huffington Post, RealClearPolitics, as well as FiveThirtyEight. I include states for which the vote counts exceed the margin of error by greater than 1% (I also exclude New Jersey, as an outlier); so, all other states (of the 28 with exit polls) exhibit closer proximity between vote counts and exit poll results.

The pattern here is that the polls average difference of 2.48%, which is an acceptable margin of error from the exit polls.   The vote count in these states, however, exceeds the exit polls by nearly 6% (+5.95).

FiveThirtyEight exhibited the closest forecast-to-exit-poll differences at 2.04.  The other sites were close, with Huffington Post averaging 2.48, TalkingPointsMemo averaging 2.58, and RealClearPolitics averaging 3.13 points difference from the exit polls.

Figure 2.  Differences from exit polls, comparing individual polling sites against the vote count (for states in which the margin of error of the vote count exceeds the margin of error of the exit polls).

Jill Stein is focusing attention on Wisconsin, Michigan, and Pennsylvania, which have closer margins of victory.  Other irregularities beyond differences from exit polls are figuring into those states, in addition to the unexpected flipping of those states for Trump.

Here, these comparisons indicate substantial discrepancies of the vote count to both the exit polls and the preliminary polling in Iowa, South Carolina, Maine, Ohio, and Missouri, each of which was contentious in this election; we could also include Georgia, Indiana, and New Hampshire (Utah is an outlier here given the prominence of McMullen’s candidacy*). The difference with Missouri is quite stark, with 10.7 points difference from the exit polls; all of the pollsters were within ± 2.08 points of the exit polls for the state.

Although Stein is focusing audits on those three states of Wisconsin, Pennsylvania, and Michigan, the indications here are that irregularities appear to be occuring in numerous states, especially among those in contention.

*If we exclude Utah from the charts, the average difference of the pollsters rises slightly to 3.3 points from the exit polls, while the vote count average difference is slightly higher at 6.1 points.  So, the vote count discrepancies are nearly double that of the pollsters.

Jill Stein Announces Fundraising Effort for a Recounts in 3 States That Flipped

Jill Stein calls for a recount in Wisconsin, Michigan, and Pennsylvania, each of which have close margins of victory for Trump.  As I write, an hour ago, the webpage for fundraising efforts was at $640,000 already, having announced today.  In posting this now, it’s already over a million dollars, which is enough to do one of the recounts.  The cost of all three is $2.5 milliion.

Jill Stein, the Green party’s presidential candidate, is prepared to request recounts of the election result in several key battleground states, her campaign said on Wednesday.

Stein launched an online fundraising page seeking donations toward a $2.5m fund she said was needed to request reviews of the results in Michigan, Pennsylvania and Wisconsin.

Stein said she was acting due to “compelling evidence of voting anomalies” and that data analysis had indicated “significant discrepancies in vote totals” that were released by state authorities.

“These concerns need to be investigated before the 2016 presidential election is certified,” she said in a statement. “We deserve elections we can trust.”

Stein’s move came amid growing calls for recounts or audits of the election results by groups of academics and activists concerned that foreign hackers may have interfered with election systems. The concerned groups have been urging Hillary Clinton, the defeated Democratic nominee, to join their cause.

Jon Swaine | Jill Stein Hopes to Request Election Recounts in Battleground States | The Guardian

Advocate to Clinton Campaign Calls for Paper Recount to Test Against Computer-Based Results

J. Alex Halderman has been advising the Clinton campaign to audit votes in three states.  At the Medium, he has published an explanation about why these should be audited, explaining scenarios of hacking the results.  Again, the time for that to happen is in the next few days, lest it just be research published years after.

The only way to know whether a cyberattack changed the result is to closely examine the available physical evidence — paper ballots and voting equipment in critical states like Wisconsin, Michigan, and Pennsylvania. Unfortunately, nobody is ever going to examine that evidence unless candidates in those states act now, in the next several days, to petition for recounts.

…

America’s voting machines have serious cybersecurity problems. That isn’t news. It’s been documented beyond any doubt over the last decade in numerous peer-reviewed papers and state-sponsored studies by me and by other computer security experts. We’ve been pointing out for years that voting machines are computers, and they have reprogrammable software, so if attackers can modify that software by infecting the machines with malware, they can cause the machines to give any answer whatsoever. I’ve demonstrated this in the laboratory with real voting machines — in just a few seconds, anyone can install vote-stealing malware on those machines that silently alters the electronic records of every vote.

J. Alex Halderman | Want to Know if the Election was Hacked? Look at the Ballots | Medium

NY Times Reports that Clinton’s Popular Vote Keeps Rising, Now Over 2 Million

With her popular vote continuing to rise, the Times (Nov. 23) notes that people are calling for recounts or audits of computer-based machines that appear to produce results trending in Trump’s favour.  The difference between paper ballot percentages does appear at odds with other voting methods, and these trends also match the pre-election polls and the exit poll trends. 

With new votes tallied from New Jersey, Illinois, Maryland and California, Mrs. Clinton’s popular vote lead reached 2,017,563 overnight, prompting new calls for an audit of voting machines in battleground states.

Vanessa Friedman et al. | Hillary Clinton’s Lead Expands | New York Times

BradCast Focuses on Whether the Presidential Election Be Recounted in Wisconsin, Michigan, and Pennsylvania

The Brad Blog podcast (Nov. 22) focuses on the exit poll disparity, including an interview with Steven Rosenfeld of Alternet.

On today’s BradCast: Will the 2016 Presidential election be publicly hand-counted (what some call a ‘recount’)? Should it be? Experts, citing anomalies in the reported results and other concerns, are beginning to say ‘yes’. But action would need to be taken — and a lot of money raised — by one or more of the Presidential candidates quickly in order for that to happen. [Audio link to full show follows below.]

I haven’t been the only one asking questions about the reliability of the reported results of the November 8, 2016 election and whether voters should have confidence in the computer-tallied results. As we’ve been reporting since Election Day, there are a lot of folks looking at the numbers and asking questions about what actually did — or didn’t — happen. A number of world class computer science, voting system and election integrity experts are beginning to urge for a public count of the election in a number of key states.

With approximately 100,000 total votes (out of more than 13 million cast), reportedly separating Clinton and Trump in the states of Wisconsin, Michigan and Pennsylvania alone, I can confirm that a number of those experts believe it would be a worthwhile exercise to file for hand-counts to make sure the results are correct. (Remember, just 50,000 votes recorded for Clinton instead of Trump across those three states would mean she, not Trump, becomes the next President of the United States.)

Today we look at just a few of the anomalies that have been widely cited today by some of the nation’s top election experts, such as more votes reportedly cast in the Presidential election than the “Total Votes Cast” in a number of Wisconsin counties — at least according to the counties’ initially reported results.

Brad Friedman with Steven Rosenfeld | Will the Presidential Election Be ‘Recounted’ in WI, PA, MI? Should it Be? | BradBlog

Clinton Campaign Urged to Challenge Election Results in Three States

NY Mag reports that analysts and lawyers have recommended election recounts in Wisconsin, Michigan, and Pennsylvania.  I wonder why they didn’t also include North Carolina.  The piece notes that official challenges have to be filed in those states by this Friday for Wisconsin, Monday for Pennsylvania, and Wednesday for Michigan.

Hillary Clinton is being urged by a group of prominent computer scientists and election lawyers to call for a recount in three swing states won by Donald Trump, New York has learned. The group, which includes voting-rights attorney John Bonifaz and J. Alex Halderman, the director of the University of Michigan Center for Computer Security and Society, believes they’ve found persuasive evidence that results in Wisconsin, Michigan, and Pennsylvania may have been manipulated or hacked. The group is so far not speaking on the record about their findings and is focused on lobbying the Clinton team in private.

Last Thursday, the activists held a conference call with Clinton campaign chairman John Podesta and campaign general counsel Marc Elias to make their case, according to a source briefed on the call. The academics presented findings showing that in Wisconsin, Clinton received 7 percent fewer votes in counties that relied on electronic-voting machines compared with counties that used optical scanners and paper ballots. Based on this statistical analysis, Clinton may have been denied as many as 30,000 votes; she lost Wisconsin by 27,000. While it’s important to note the group has not found proof of hacking or manipulation, they are arguing to the campaign that the suspicious pattern merits an independent review — especially in light of the fact that the Obama White House has accused the Russian government of hacking the Democratic National Committee.

Gabriel Sherman | Experts Urge Clinton Campaign to Challenge Election Results in 3 Swing States | New York Magazine

Brief summary posted at The Stranger‘s blog, Slog:

Jen Graves | Computer Scientists and Election Lawyers Are Privately Lobbying Hillary Clinton to Challenge the Vote | The Stranger – Slog

“Not Even ‘Dewey Defeats Truman’ was this big” of a Poll Failure. Or was it?

Dewey Defeats Truman

Clinton Defeats Trump

When the headlines showed “Dewey defeats Truman,” the pollsters had it wrong in 1948. But, maybe they were not so off the mark here in 2016?  David Moore questions the exit polls as they compare to the vote count, and highlights how this is different from 1948.

Prior to the election, Natalie Jackson, Senior Polling Editor at HuffPollster, wrote that “A Donald Trump win would signal the biggest poll failure ever.”

Then she added in the past tense, as though she had some crystal ball or the disaster had already occurred (her piece was published two and a half weeks before the election): “Not even ‘Dewey beats Truman’ was this big.”

Well, yes and no.

In 1948, it was the national polling that predicted Gov. Thomas E. Dewey of New York to beat President Harry S Truman. Besides, the Republicans had swept the Democrats out of office in the 1946 congressional mid-terms, so Dewey’s victory was seen as a natural conclusion to a political shift in the country. Truman’s surprise victory was indeed a big polling failure.

As it turns out in 2016, however, the national polling figures were not all that bad. They predicted Hillary Clinton to win by about four percentage points. In fact, it appears she may win the popular vote by close to two percentage points.

Moreover, Moore also highlights the discrepancies in U.S. Senate elections as well.  He provides tables and screenshots of the exit poll results.

David W. Moore | Why the Exit Polls Sow Doubt About the Vote Count | iMediaEthics