I’m asking for funds to support academic research on the accuracy of machine counted election results.

In Kansas, the only way to get access to the physical ballots cast in an election is through an official recount. I know. I’ve sued in the past and lost because in Kansas, that is not a right granted to citizens.

Candidates can request recounts if they are willing to pay the costs if it does not result in a reversal. Citizens have the right to request recounts for ballot questions, such as the proposed constitutional amendment on the August primary ballot, with the same caveat. However, the individual making the request can specify what counties and precincts to recount. For this effort, the entire state is not required, just a representative sample.

In August, I was able to obtain the needed financial support to conduct two recounts, one of the Amendment question and one for House District 84 with the cooperation of candidate Ford Carr. After the recount, I had data from 49 precincts in the constitutional Amendment recount ($35,000), but less than a dozen in the District 84 race ($10,000).

Neither of those races showed statistical flags indicating further scrutiny was appropriate. The outcome changes showed no signs of manipulation of the results. This sample now constitutes a valid measurement of the expected outcome of a well-conducted hand count of the physical ballots. There were ten reported ballot changes in six precincts, an average of one change in 2,500 ballots with no observable pattern. This is well below any reasonable level of concern regarding the accuracy of the outcome. We can reject the hypothesis of deliberate manipulation favoring one outcome.

I want to run recounts in the Nov. election. I have to persuade candidates to assist me by requesting a recount and asking for the precincts I specify. If you are acquainted with candidates in the state wide Kansas races, I’d appreciate your support in suggesting to them that they support my research.

I also need to fund the costs. If I can get cooperation from more than one candidate, which I am hoping for, the costs double. I’m told it’s a requirement of the official state protocol that each recount be conducted separately. Also, I expect the costs to be higher with the general election rather than the primary because of the increase in number of ballots cast. Counting ballots is an activity with little reduction in cost for scaling up. That’s why statisticians spent a lot of time trying to minimize the required sample size. I am estimating costs to run from $150,000 to $500,000, depending on how many races show statistical flags and how many candidates are willing to support my research. I’ll be concentrating on state-wide races for this effort.