How trustworthy are electronic voting systems in the US?

I’ve recently written an article for the Royal Statistical Society on the trustworthiness of US voting systems. Here is an excerpt:

My statistical analysis shows patterns indicative of vote manipulation in machines. The manipulation is relatively small, compared with the inherent variability of election results, but it is consistent. These results form a pattern that goes across the nation and back a number of election cycles. I’ve downloaded data and verified the results from several states for myself. Furthermore, the manipulation is not limited to a single powerful operator. My assessment is that the data reveals multiple (at least two) agents working independently to successfully alter voting results.

You’ll find the article in its entirety, here:
http://www.statslife.org.uk/significance/politics/2288-how-trustworthy-are-electronic-voting-systems-in-the-us

Interesting Question: Couldn’t other factors explain the patterns in the data?

The answer to this question is “Yes, it’s possible”. I’ve seen several claims that people can explain the patterns in the data that I find concerning with voter registration of party and/or % turnout by party and/or other demographic variables. I haven’t yet done much analysis into these factors, primarily because I haven’t had access to precinct level data. I’m working on that and may have an acceptable database of those statistics ready soon.

At the county level, for the data I’ve looked at, such factors only explain a portion of the trend. Further, the pattern in the cum sum % analysis is not what would be expected by demographic type factors. We should see a leveling out after a rise due to those factors, not the continuous increase those models frequently show. If you check the Wisconsin graph, broken down by voting machine type, the pattern I would expect to see exhibited for that type of issue is shown by the purple line as opposed to the red lines (increasing) or the green line (paper ballots only).

2014 Wisc. Gov Race

However, even if all these patterns (which are consistent across different states and elections while showing similarities by voting machine type) are explainable in such a manner. At this point, I consider the fact that our voting machines in Sedgwick County have never been audited or checked for accuracy after an election to be equally concerning. Without such an audit, we do not know whether the votes have been counted accurately. IMO, that’s a serious problem and my lawsuit at least illuminates that issue even if I never get access to the paper audit trail.

In the news – Monroe Evening News

Statistician is suing Kansas for release of voting records Kansas seeks to block release of voting machine paper tapes

The top election official in Kansas has asked a Sedgwick County judge to block the release of voting machine tapes sought by a Wichita mathematician who is researching statistical anomalies favoring Republicans in counts coming from large precincts in the November 2014 general election.

In the news – Esquire (blog)

Why Kansas Keeps Killing Itself

Clarkson has been at this for a while now. Her attempts to pry open the voting records for Sedgwick County have barely made a ripple nationally but, in the world of voter-suppression, this is a big deal I guess it would have been to the advantage of Kobach, who is the dark genius behind nationwide anti-immigrant and nationwide voter-suppression laws, if the Kansas courts actually had shut down.  Clarkson is speculating not only that votes were suppressed, but also that there was some severe monkey-mischief played with the votes that were counted. This is an entirely different – and more serious – thing.