In yesterday's and other postings, I praised election officials in Minnesota, and said that election officials nationwide can learn a lot from how Minnesota conducts elections, including but not limited to audit and recount. Today I'd like to point out some improvements to the MN recipe, starting from the root cause of the need for improvements: lack of public confidence in voter registration systems. Across the country, we've seen many voters doubt whether their upcoming election day experience would be successful, because of concern over errors in voter rolls, or polling snafus in interpreting them. Based on earlier horror stories from Ohio and elsewhere, it's an undertandable concern. In many states, it's difficult even to check on one's registration status. The result of these concerns is a desire by many voters to voter early, in the theory that if there are problems with voting, they can be diagnosed and fixed early, in time for election day; but if you wait until election day to find probems, you likely won't be able to vote. I think that this concern will persist for some time, despite creditable work by state election officials to improve the accuracy and accountability of their voter registration systems and procedures.

In MN in 2008, the only form of early voting was vote-by-mail, and the huge jump of vote by mail participation is interpreted as being in part a desire by many Minnesotans to vote early for precisely the above reasons. But that huge jump on vote by mail ballots also set the stage for the Coleman-Franken dispute, and also illustrates the need for improving the recipe. Here's the problem: every county in MN has its own rules for its own election officials to perform the process of examining a vote by mail envelope, and determining whether the ballot inside should be counted. Further, each county has its own practices for logging these activities and decisions. And of course there is risk of human error as well, magnfied by the large number of VBM envelopes to examine.

Reviewing the VBM envelope processing was a key part of the MN recount procedure, as well as the controversy following it. Part of the purpose of the review was to find those cases where an election official had made an error in following their county's rules, and had mistakening excluded a VBM envelope that actually was just fine, such that the ballot inside should have been counted. When these cases had all ben reviewed and resolved, the next bit of fuel for fire came from cases where in one county a VBM envelope was excluded and in another county a VBM envelope was included even though the two envelopes had very similar imperfections. These cases were correct, but confusing because of the variation between counties of the rules for VBM envelope review.

That brings us to two improvements on the current MN recipe, that some MN election officials would like to put in place: early voting to cut down on the number of VBM ballots and hence the number of cases where an election official has to decice whether a ballot counts or not; and state-wide uniform VBM envelope processing rules, so that all such decisions are made in the same way, and all review of those decisions (esp. in a re-count) is performed uniformly. To those two procedural improvements, I would add a third, technological improvement: apply to VBM envelope processing the same kind of accountability and transparency measures that we're starting to see in some state voter registration systems. What are those measures? Stay tuned. ;-)