QUESTION: When is three greater than 37,000?

ANSWER: Now, in the Kansas Secretary of State’s Office, since Kris Kobach has begun using his new power to prosecute voter fraud.

Kobach’s numbers problem is vast.

Recently — and ostensibly to curb this voting-fraud crisis — he approved an administrative rule that will purge about 37,000 “suspense voters” from the state’s voting rolls. These are the voters who did not show proof of American citizenship, such as a birth certificate, when they registered, nor within 90 days afterward. Since a policy’s benefits should exceed its costs, observers are looking for more than 37,000 documented cases of voter fraud to justify this purge.

Kobach has provided us with three.

The people named in these cases — a couple from Johnson County and an individual from Sherman County — are accused of double voting. The charges are that these people each registered and voted twice, in different states.

But wait. The justification for Kobach’s landmark, ironically named, “SAFE Act” was to prevent undocumented immigrants from voting. Kobach’s three prosecutions have nothing to do with that. Double voting could have been identified and referred to federal prosecutors in the pre-Kobach era. Such cases provide no justification for the laws he championed. Catching them is just a matter of good record keeping and interstate cooperation.

At a University of Kansas symposium this past weekend, Kobach said he has evidence that 237 cases of illegal voting occurred between 1997 and 2011. When a questioner labeled these numbers as “petty,” Kobach replied by pointing to close elections that can be swayed by a handful of votes. He indicated that like the first three, these 234 other cases are also about double voting, not undocumented immigration.

One wonders about the time and money invested in tracking down these three malefactors. Kobach wants to prevent voter fraud beforehand, but also prosecute it afterward. Given the small number of cases involved, is this costly belt-and-suspenders approach really necessary?

If all 237 cases he identified are valid, then the number of voting fraud cases is just a tiny fraction of the number of voters Kobach just removed from the rolls: 0.7 percent, to be precise. This is still not a fair comparison because the time periods do not match: Kobach had to go back to 1997 to find these 237 cases, but the 37,000 voters being purged had tried to register just since 2012. Still, at the KU symposium, Kobach was able to keep a straight face while stating that small numbers of voters can sway close elections.

Kobach earns an “A” for acting, but an “F” in math.

These laws, threats, and prosecutions do not increase our faith in the process. Instead, more and more Kansans are wondering how politicians have manipulated the right to vote, to their own benefit. How else can you explain the recent emergence of an unlikely public figure: Wichita statistician Beth Clarkson, whose data analysis has led her to doubt the veracity of this state’s voting machines?

Clarkson concedes that her data analysis is not definitive. She seeks more data from Kobach to test her hypotheses, and his office has not supported her request for a court order. Whatever happens with the voting machines, we already know Clarkson is right about one thing: oftentimes, Kobach’s numbers just don’t add up.

Michael A. Smith is an associate professor in the political science department at Emporia State University and a member of the “Insight Kansas” writing group.