There is a valuable reminder to be drawn from "Group vows to sue state over voter data" (Aug. 12), which reported on Minnesota Voters Alliance's (MVA's) recent lawsuit demanding public election data be released by the Minnesota secretary of state.

The article confirms that when it comes to issues of voter fraud, there is often an alignment between the mainstream press and liberal groups such as the Brennan Center, which has a long-standing commitment to demonstrating the nonexistence of illegal voting.

After an otherwise professional accounting of the issues and events, the last paragraph of the article sailed off into an effort at "balancing," using a misleading statistic. It reported that "one 2014 study" — which, after our inquiry, the newspaper confirmed as a Brennan Center study — examined 1 billion votes between 2000 and 2014 and found only 31 incidents of documented fraud.

Our fundamental problem with the statistic cited is that it was ill-defined and, it turns out, has almost nothing to do with the issue at hand, which is how much illegal voting is occurring in Minnesota.

The illegalities identified in the study were limited to impersonating a legal voter. That is a small part of the issue we are addressing. Even if accurate, the study did not seek to identify all types of voting illegalities, which was the inappropriate impression given to the reader.

The author of the study simply reported that he found very little that he could "substantiate," to which we say "So what?" Anyone can look up at the sky and not see many stars, but that should hardly be reported as proof that astronomer Carl Sagan was wrong when he exclaimed there are "billions and billions" of them. Not finding something isn't proof of its nonexistence — unless the methodology is extremely tight and the researcher is unquestionably unbiased.

And how could the methodology of this study be relevant to what the MVA is looking at when court records show that known voting crimes in Minnesota occur at a rate at least 1,000 times the frequency cited in the study?

Court records show there were 369 voting convictions in Minnesota between 2008 and 2014. That is more than 10 times the number of violations the article said the study found, in a shorter time period and among more than 100 times fewer ballots cast (slightly fewer than 10 million). So in Minnesota, the rate of voting convictions is more than 1,000 times the rate of "fraud" cited in the article.

Keep in mind, also, that these are 369 convictions. Convictions represent a small fraction of the actual amount of ineligible voting, because not all ineligible voters are identified and, in Minnesota, statutes require the state to show that the person "knowingly" voted while ineligible. What that means, as a practical matter, is that a person is convicted only if they can be detected and they admit they are guilty.

MVA's own study, fully documented and submitted to the Minnesota Supreme Court in a 2016 petition, found 1,670 provable instances of ineligible voting by felons alone during the four general elections of 2008 through 2014, more than four times the number of convictions during the same time period. Further, the amount of additional ineligible voting — for example, by noncitizens — has not been able to be fully investigated. Data being hidden by the secretary of state might help get a start on quantifying its magnitude.

The Star Tribune should have just reported the facts and should not have attacked the MVA's claims by injecting an unrelated study and implying that it might apply to Minnesota and to claims the MVA makes about the types of ineligible voting that occur here.

Andrew Cilek is executive director of the Minnesota Voters Alliance.