Quis custodiet ipsos custodes? Does anyone really watch the watchman? In Kansas, the state’s lack of an election post-audit is raising some questions, and a university professor wants to run the numbers on electronic voting machines in and around the state’s largest city.
Like other states across the Union, Kansas began using electronic voting machines following the presidential election of 2000 and the infamous “hanging chad” debacle in Florida. While many Kansas counties use optical scan paper ballots, the two most populous counties in the state, Sedgwick County (home of the state’s largest city, Wichita) and Johnson County (home of some of the most affluent Kansas City suburbs) use electronic voting machines. And while the machines in Sedgwick County print an extensive paper receipt, the machines used in Johnson County do not leave a paper trail.
To some, those paper trails—when they exist—warrant closer inspection to ensure votes are being properly recorded by electronic voting machines. After reading an academic paper that analyzed the 2012 Republican primary and hypothesized that voting machines were being manipulated to cast erroneous votes, Dr. Elizabeth Clarkson, Chief Statistician at the National Institute for Aviation Research at Wichita State University, decided to apply the paper’s statistical reasoning to three later elections: the 2012 presidential election in Ohio, the 2014 gubernatorial election in Wisconsin, and the 2014 U.S. Senate election in Kansas. The hypothesis held. It was likely something strange was going on, but Dr. Clarkson needed better data to confirm her findings. In 2012 and again in 2015, she brought suit in Kansas to gain access to the paper tapes created by Sedgwick County voting machines in order to conduct a more thorough statistical analysis. Though Dr. Clarkson couches the main thrust of her research by saying “statistics don’t prove vote fraud”, she raises a broader, possibly more important point: Kansas doesn’t have a post-election audit statute, so it is difficult to verify votes recorded by electronic voting machines accurately reflect voter intent, even when paper receipts exist.
It’s true. Even though Johnson County was the first jurisdiction in the nation to install Direct Recording Electronic (DRE) voting machines—with installation completed before the Help America Vote Act (HAVA) of 2002 was passed by Congress to help states replace outdated voting equipment—and has had the longest amount of time to discover problems with the machines, Kansas has no election post-audit statute on the books. In fact, except in limited circumstances, it is unlawful for the county board of canvassers or the county election official to unseal ballots in order to examine them, even if they wanted to conduct an election post-audit voluntarily.
That’s not to say Kansas doesn’t have a fairly significant governance scheme for electronic voting equipment. The Secretary of State must initially and continuously certify any voting equipment used in the state, including electronic equipment. All voting machines, including the electronic variety, are subject to a rigorous, public test within five days of an election and are required to pass before being placed into use. The same test is repeated when the ballots are canvassed. Election officials are specifically charged with safeguarding electronic voting machines, and state law defines electronic voting fraud as a non-person felony. But, after the ballots are cast, no audit is required.
Even if Dr. Clarkson’s latest suit does not succeed, as many believe it will not, her curiosity draws attention to an important issue: regardless of the existence of voting machine manipulation, without an election post-audit, there is no way to demonstrate that ballots have been correctly counted or to prove (or disprove) allegations of vote fraud. Indeed, unless ballots, including electronic ballots, can be verified through an election post-audit process, it is difficult to determine who, if anyone, is watching the watchman, especially if the watchman is made of incomprehensible touch screens and circuit boards.