Archive for Saturday, July 18, 2015

Kansas statistician battles government to determine whether vote count is flawed

July 18, 2015


Wichita State University mathematician Beth Clarkson has seen enough odd patterns in some election returns that she thinks it's time to check the accuracy of some Kansas voting machines.

She’s finding out government officials don’t make such testing easy to do.

When Clarkson initially decided to check the accuracy of voting machines, she thought the easy part would be getting the paper records produced by the machines, and the hard part would be conducting the audit. It's turned out to be just the opposite.

“I really did not expect to have a lot of problems getting these (records),” Clarkson said. But Sedgwick County election officials "refused to allow the computer records to be part of a recount. They said that wasn't allowed.”

Instead, Clarkson was told that in order to get the paper recordings of votes, she would have to go to court and fight for them.

Earlier this year, Clarkson filed a lawsuit against the Sedgwick County Election Office and Kris Kobach, Kansas' secretary of state, asking for access to the paper records that voting machines record each time someone votes. The record does not identify the voter.

Clarkson decided an audit was important in part because of national concerns about the voting machines that thousands of Kansans use to cast their votes in elections each year.

Reports of voting irregularities involving the same types of machines have been widespread in other parts of the country for years. And when Clarkson did her own calculations after the November election, she believed she found voting irregularities similar to those in other states such as Ohio.

“I noticed that there are patterns in the data that are suspicious,” said Clarkson, chief statistician for the university’s National Institute for Aviation Research.

Clarkson is in a unique situation.

The voting machines that Sedgwick County uses have a feature that most of the Direct Recording Electronic voting machines, or DREs, in Kansas and around the country do not have: a paper record of the votes known as Real Time Voting Machine Paper Tapes.

Many counties and states opted not to have a paper record to save money, said Pamela Smith, president of Verified Voting Foundation, a nonprofit agency whose mission is to safeguard elections in the digital age.

The DREs were introduced to American voters about 13 years ago to replace “hanging chad” ballots and lever voting machines following the 2000 presidential election showdown and ensuing controversy in Florida involving Al Gore and George W. Bush.

But the DREs have a couple of major flaws, Smith said.

Because no paper records exist in most cases, voters and candidates cannot know whether the machines accurately recorded their votes, Smith said.

That means for a candidate, no recount can ever be done of the votes recorded on those machines, Smith said. And voters can never be sure their votes were recorded correctly.

In addition the voting machine software is proprietary, and even election officials cannot examine it.

“There is a cost for not knowing the results are right in each election,” Smith said. “In our view, it becomes kind of corrosive of voter confidence because over time you can never be sure.”

In addition, post-election audits of the machines cannot be done.

Douglas County does not face those problems. Douglas County is highly aware of the DRE machines and the concerns surrounding them, said Douglas County Clerk Jaime Shew.

In 2006, Douglas County held a number of townhall meetings and brought in vendors to try to determine whether it should move from paper ballots to the electronic machines, Shew said.

The county elected to stay with a system that has paper ballots mostly because there is a paper trail that allows for recounts and audits, Shew said.

Tens of thousands of DRE machines were sold in the early to mid-2000s to election offices around the country. The sales got a boost from Congress when it passed the Help America Vote Act of 2002, which passed along millions of dollars to help states replace their voting equipment.

Johnson County was the first in the country to have a DRE machine in a polling place, said Brian Newby, Johnson County election commissioner, and that was purchased by one of his predecessors before the Help America Vote Act was passed.

Problems with the DREs were apparent early.

By 2004, California decertified its machines because of concerns over security, reliability issues and the inability to have audits.

Johnson County capitalized on that by buying the machines that had been only used once at a cut-rate price and expanded its fleet of machines, Newby said.

Also in 2004, Anita Ramasastry, an associate professor at the University of Washington School of Law in Seattle, wrote about the problems cropping up across the country.

She cited problems in California, Maryland and Georgia and advocated that federal law mandate that paper receipts be available and paper records kept.

“Paper receipts are the obvious answer,” Ramasastry said. “Florida gave recounts a bad name. But there is something much worse than a recount: the utter inability to recount votes, and reconstruct voters' true intent, in light of a serious computer error.”

Newby said his office also recognized the issue and asked Johnson County commissioners to pay to retrofit the machines with printers, but the cost — at $5 million — was too steep. The machines are now aged and the county plans to purchase new ones with some type of paper reporting by 2017.

The problems surfaced in some states because in about half the states, laws require an audit after an election, Smith said.

But Kansas does not have a state law requiring an audit. Machines are checked prior to an election to see if they are operating correctly, several Kansas election officials said.

In Sedgwick County, Clarkson, a certified quality engineer with a doctorate in statistics, says her calculations from the November election showed that patterns exist in the voting data to suspect that “some voting systems were being sabotaged, but that doesn't mean that no other explanations are possible for these patterns.”

Clarkson's findings were published in June in StatsLife, the magazine for the Royal Statistical Society and the American Statistical Association.

Clarkson thought getting the data would be the easiest part of her project, but she has learned that obtaining government records in Kansas can be difficult.

In 2013, Clarkson asked Sedgwick County to do a recount but the time to file had expired. So she filed an open records request. Sedgwick County officials refused to release the records and Clarkson filed a lawsuit. But the judge ruled that the paper records were ballots even though they don't identify the voter and so were not subject to the state's open records law.

In November, Clarkson filed for a recount after that election, but even though her request was timely, Sedgwick County officials again refused, saying that only a judge could release the records to her.

So Clarkson again filed a lawsuit in February that asks for a court order giving her access to a certain number of voting records to conduct the audit.

The lawsuit was filed against the state attorney general and amended April 1 to add the Sedgwick election commissioner and Kobach.

But then Clarkson hit another snag.

Instead of paying the sheriff to serve the summons, she mailed it to the Sedgwick county commissioner and Kobach. Under state law, they had 30 days to respond but did not.

A Sedgwick County election official and a Kobach's spokesperson, said they were unaware of the summons.

The Journal-World asked Eric Rucker, assistant secretary of state, whether Kobach had received it.

“I don't know if we did or not,” Rucker said. “We are not going to comment on the status or the nature of this litigation at this time. We certainly have taken the position that during the pendency of any potential litigation that we won't discuss the particulars of the case and that would be the nature of our response.”

Clarkson said she now has paid the Sedgwick County Sheriff's Office to serve the summons this week and is interviewing attorneys for legal assistance going forward.


Richard Heckler 2 years, 8 months ago

In the IEEE Symposium on Security and Privacy 2004, four top computer scientists from the University of California, Johns Hopkins University, and Rice University similarly critiqued Diebold’s voting system:

"We found significant security flaws: voters can trivially cast multiple ballots with no built-in traceability, administrative functions can be performed by regular voters, and the threats posed by insiders such as poll workers, software developers, and janitors is even greater. Based on our analysis of the development environment, including change logs and comments, we believe that an appropriate level of programming discipline for a project such as this was not maintained. In fact, there appears to have been little quality control in the process.

"…The model where individual vendors write proprietary code to run our elections appears to be unreliable, and if we do not change the process of designing our voting systems, we will have no confidence that our election results will reflect the will of the electorate."

Computers are inherently subject to programming error, equipment malfunction, and malicious tampering. If we are to ensure fair and honest elections, and retain voter confidence in our democratic process, we need to ensure that there are no such questions. Therefore, it is crucial that any computerized voting system provide a voter-verifiable paper audit trail and that random audits of electronic votes be conducted on Election Day. Paperless electronic voting machines make it impossible to safeguard the integrity of our vote - thereby threatening the very foundation of our democracy.

Moreover, the seller of the machines, the Diebold Corporation, is a supplier of money to one of the major party candidates, George W. Bush. The CEO and top officers of Diebold are major contributors to the Bush campaign. A corporation with vested political interests should not have control over the votes of the populace.

Richard Heckler 2 years, 8 months ago

Computerized voting devices can be programed to vote wrong which is why these machines should be forever forbidden.

Steve Jacob 2 years, 8 months ago

It is odd Kobach works so hard to keep illegals from voting and does nothing about possible errors with electronic voting. I will even give the benefit that no one is intentionally rigging the machines, but errors happen.

Scott Burkhart 2 years, 8 months ago

Voter I.D. and voting audits are part of a two way street. I am for ensuring that only eligible voters are allowed to vote but I am also, 100% in favor of accurate voting tabulations as well. Corruption on either side of this fundamental right cuts to the core and integrity of a self governing society.

Robert Redmond 2 years, 8 months ago

Its all part of the "vast right wing conspiracy"!

You're being seen by a whole team of psychiatrists aren't you?

Bob Reinsch 2 years, 8 months ago

Kobach's "Interstate Crosscheck" is making any number of eligible voters ineligible for no legal cause. These suspended voters are having their civil rights violated, and their votes are in limbo. He claims the voting records are locked up and sealed, but that further exacerbates the problem. What if a couple thousand of these voters can prove that their rights were violated and their suspended ballots should count? Does the permanent record get unsealed then?

Robert Redmond 2 years, 8 months ago

This is a real non-story, Prof Clarkson's statistical analysis is meaningless. She didn't like the result of the election therefore it must be tainted. Her analysis isn't evidence its conjecture, kinda grasping at straws. She'll do nothing but waste tax money on a foolish endeavor.

Amy Varoli Elliott 2 years, 8 months ago

What have the vast statistics that you run told you? What form of statistics did you use? What was your margin of error? Unless you can provide with your results I will just have to assume your a cranky old man who doesn't like when people don't see things your way.

Tom Weiss 2 years, 8 months ago

She did not say she disliked the results of the election. She is simply trying to find out whether the machines have produced an erroneous tally regardless of the direction of the bias.

Dorothy Hoyt-Reed 2 years, 8 months ago

So you would trust a computer after all the hacks that have happened? You would trust politicians not to cheat if they could? Why don't you want a paper trail? Why wouldn't you want an audit? What are you afraid of. What is Kobach afraid of?

And you do realize that they will charge her for putting all these records together, so tax dollars wouldn't be used for that, but tax dollars are being used to keep her from getting the records. So with whom should you be upset? Clarkson or the pols who are spending a lot of money to keep her from seeing the records. Why are they fighting this? What do they have to hide?

And here is another brand new today poster who has poor punctuation and grammar, similar to some other recent posters who no longer post here.

Cheryl Nelsen 2 years, 8 months ago

Did you read beyond the first couple of paragraphs? Kobach and his crew have filed lawsuits out the wazoo, so why would one more lawsuit matter? This is a very important story, not only to Kansas. She is not grasping at straws. She is not interested only in what is happening in Kansas, so I don't think this is about not liking an election result.

Dorothy Hoyt-Reed 2 years, 8 months ago

He has never been one to read everything, and if he does his mind is so clouded with his extremist positions he would never truly comprehend it. In his mind, you never question authority unless you perceive it to be just left of Senator McCarthy. To him, Kobach is pure and above criticism. He would follow him anywhere and do anything Kobach and crew would order of him. He is loyal to this radicalism that has swept this nation, making rich people idols of the poor and making winning by any means ok. And all scientists are liars too. Science deals with evidence of what is true, and can change if evidence proves something wrong. It goes against the authoritarian mindset held by these radicals.

Art War 2 years, 8 months ago

Diebold machines are designed to hack. America needs to return to paper ballots and inked thumbs before elections can be fairly held. The ability to swing a few percent here and a few there means a 15% spread could be erased. Judges are empowered to defend the publics rights and liberties, why would they stonewall an investigation unless someone told them too? Because they are bought and paid for that's why.

Paul Beyer 2 years, 8 months ago

Hi Kevin, welcome back. Your mindless jokes are welcome.

Steve King 2 years, 8 months ago

Ok Robert. She had a PhD in statistics. And you've a PhD in what? Your professional and educational background allows you to challenge her opionion on emotions? I think it's your emotions that taint your logic.

David Carson 2 years, 8 months ago

Steve, Kevin is back again. Notice how he joined just to comment on this article?

Tom Weiss 2 years, 8 months ago

Kobach has established a website where people can post information about suspected voter fraud. This would seem to be a clear case of suspected fraud - clearer than anything Kobach has come up with so far. If it is fraud, and Clarkson has not concluded anything about that (although the reluctance of Kobach and the Sedgwick county commission offer some hints) we don't know who might have committed such fraud or how. So everyone should post this example on Kobach's website.!

Paul R Getto 2 years, 8 months ago

If they had computers when Stalin was alive, he would have said: "It's not who voted who counts. He who owns the software controls the counting." Bring back paper ballots. The media can wait a few hows to speculate on who won.

Ralph Reed 2 years, 8 months ago

Putting myself at hazard of receiving another poison pen letter via the USPS ...

I'd like to welcome Robert Redmond to the LJW forum. He/she/it just joined today and has already taken up the mantle of a banned member we all know and love. The first post includes a disparaging comment about a previous poster to this article and the second post is a withering personal attack against Prof Clarkson.

We've missed these types of posts and look forwards to the humor you will bring to this otherwise stodgy forum.

Greg Cooper 2 years, 8 months ago

Note that "ART WAR" has also joined today. Could it be that he "AT WAR" with the Journal World over an unfair disappearance?

As for poison pen letters, the guy you refer to has contacted my employer to let him know that he will "probably" not do business with us because of my "hatefull" comments toward him.

I ask again, Nick, is there not a way to track your members? I, for one, am sick to death of this guy and his sneaky, snarky ways.

Paul Beyer 2 years, 8 months ago

Mr Reed, for laughs check "Robert's" post on his Facebook site prior to latest name change.

Devin Wilson 2 years, 8 months ago

Wow. I assumed that our voting machines had a paper trail for accountability. This is a HUGE problem. We should be able to verify EACH election. I was ridiculed by my Senator Smithw when I offered suggestions for making voting easier, like a smartphone app. You can automate a screen capture that emails results to the voter + the Secretary of State's office. That is way more accountable than our current electronic voting machines.

Dorothy Hoyt-Reed 2 years, 8 months ago

A small piece of paper printed out like a receipt and put into a secure box should be enough. And they should be counted every time, and if they don't balance with the machine, the paper is wins. What's hard about that, and it's saves paper.

Richard Crank 2 years, 8 months ago

Even if it's unprovable, I will always believe Brownback is illegitimate.

Glenn Reed 2 years, 8 months ago

The statistical oddity is reason enough to give the ballots another look. There's no good reason to push against it.

If Kobach is being honest and the system is robust, then the records are going to be accurate and we have a fun statistical oddity to study. More knowledge of how things work is always good.

If Kobach is being honest and the system has failed, then we can identify where the failures are and get to work fixing them before the next election. Identifying problems and fixing them is important.

Really, the pushback against examining the available evidence is reason enough to cast doubt on how much Kobach cares about violations regarding voting laws.

Glenn Reed 2 years, 8 months ago

One more thing, the software that runs our voting machines should be open-source. If it can be examined by anyone, we'd be less likely to run into these kinds of issues.

Levi Bowles 2 years, 8 months ago

I'm no fan of the election results, but I don't think that Clarkson's analysis is indicative of election fraud. I detailed my reasons for that on my website a couple months ago. That said I definitely think that she should have access to source code and the election machines. Here is the link to my analysis it also links out to 2 additional analyses I did.

Jim Riley 2 years, 8 months ago

Clarkson's latest analysis is here:

She segregates the results by voting machine type, which has a secondary effect of segregating them by county.

Since centrally-counted optical scan machines were 59% for Robers can we conclude potential fraud? No!

Voting machines are expensive, and it may be difficult to justify in a rural township with 48 voters. So instead you use a central scanner in the county seat. Clarkson's graph understates the number of votes counted on such equipment since it only includes precincts with 500 votes cast, which will only happen in larger towns (each ward in a town may have its own precinct). Rural area => Central Optical Scan, Rural area => Republican. No surprise here. If Clarkson had included all precincts the Roberts percentage would likely be over 60%.

Optical scan precinct ES&S. This shows Roberts under 40%. 29 counties used this type of equipment, but Clarkson says there were only 44 precincts with more than 500 votes cast. Douglas County (Lawrence/KU) was by far the largest county to use ES&S precinct scanners. Though Douglas was only one of 29 counties using this equipment it represented about 1/4 of the population. When you screen based on votes cast in precinct you are going to select a larger share of precincts in more populous counties. What is remarkable is that the Roberts percentage is so high. He only received 29% of the vote in Douglas County. This suggests that there were enough non-Douglas precincts to bring the average up.

Now let's look at DRE Premier. Its use was dominated by Johnson County, which is a suburban county west of Kansas City, Missouri and south of Kansas City, Kansas. About 190,000 votes were cast in 2014, but Clarkson only shows 130,000 votes. So she discarded about 1/3 of the votes because they were from smaller precincts. Overall this increased the Roberts vote from 49% to 52%. That is, votes cast in smaller precincts were more non-Republican than in the county as a whole.

Johnson County gets more Republican as you head west and south. If Clarkson were to arrange precincts by how southwesterly they were, there would be an upward trend just like in her charts (likely a sharper increase). But if you were to plot cumulative average votes cast vs. southwest, there would also be an upward trend. Precincts tend to be larger in newer subdivisions, turnout is higher, and voters more Republican.

About half of the DRE ES&S votes were cast in Sedgwick County (Wichita), but they were used in 5 of the largest 9 counties. Butler, in particular, has some very large precincts. If Clarkson were to create choropeth maps based on votes cast in a precinct, and the Republican percentage of the vote, the source of her erroneous conclusions would be obvious.

Cait McKnelly 2 years, 8 months ago

Sam Brownback won by less than 3% (not even outside the margin of error. ) It was confirmed earlier this year that over 5,000 votes from Saline county were not even counted last November until AFTER the state wide election results were officially posted.
Election fraud exists, just not VOTER fraud. It exists in gerrymandering, election manipulation and targeted voter suppression. I am still quite thoroughly convinced that Anonymous blocked Karl Rove from hi jacking the 2012 presidential election. Kobach has his fingerprints all over this kind of activity.

Evan Ravitz 2 years, 7 months ago

Dan Rather did his greatest investigative work about the Florida 2000 problems. Here seven people at the plant that printed the ballots that caused problems said they were forced to do things wrong.

Commenting has been disabled for this item.