Posted: January 15, 2009

MN Analysis: Allegations of Double-Counted Duplicates

The Coleman v. Franken Notice of Contest filed last Tuesday alleges that ballots "were counted twice during the Recount due to such Ballots being not marked as 'DUPLICATES' and matched with its 'Original' Ballot in contravention of Minnesota Statutes § 206.86..." (see Notice at 5). Coleman had raised the issue in previous litigation concerning the recount, but the Minnesota Supreme Court decided to wait to address the issue until the evidentiary tools of a post-election legal contest could be invoked. Unfortunately, however, it may turn out that the evidence is so ambiguous that even those substantial tools may not be enough to put Humpty back together again.

First, some background. Legitimate ballots cast by voters are occasionally torn or otherwise damaged in a way that makes them impossible for optical scan voting equipment to read. However, that does not mean the ballots do not count. Instead, Minnesota law requires that at least two poll workers of opposite political parties take a new, undamaged ballot and transfer the discernable votes onto it from the damaged ballot. MSA 206.86. To preserve a record and prevent confusion later on, poll workers are supposed to label the new ballots with the word "duplicate," indicate the precinct associated with the ballot, and place matching serial numbers on both the new ballot and the old, damaged one. The new ballot should then be run through the optical scanner and counted "in lieu of" the damaged or defective ballot card.

The trouble is, what if poll workers did not label the reconstituted ballots "duplicate," did not indicate the precinct associated with the ballot, did not place matching serial numbers on the ballots, or all of the above? To make matters worse, what if, because of mistakes like these, workers in the manual recount that just occurred, without realizing it, counted the discernable votes on both the damaged ballots and the reconstituted ones? That is what Coleman alleges and, if it is true, it means some votes were counted twice. It also may mean that the recordkeeping with respect to duplicates was so poor that it will be very difficult to go back and determine which votes were counted twice. That is bad news for Coleman, because he carries the burden of proving that he was the rightful winner.

About the best Coleman could do is to continue pointing to "excess" ballots in order to suggest that reconstituted votes were double-counted along with their originals. Excess ballots exist when the number of ballots cast in the polling place exceeds the number of voters who signed the poll book prior to voting. Excess ballots indicate one or both of two things: that some legitimate voters were permitted to vote without signing in, or that "extra" ballots somehow made it into the ballot box, as in the case of ballot-box stuffing or the duplicate scenario discussed here. In the present case, Coleman has a better chance of persuading the court that the excess ballots are duplicates if, as he has alleged, the number of excess ballots in each precinct "often" corresponds to "the number of unmatched original damaged ballots counted in the recount" (see p. 3). That would put Franken in the difficult of having to argue that the fact that the numbers match is just a coincidence.

Another wrinkle concerning the excess/duplicate issue is how this issue is supposed to fit together with the issue of what to do with uninitialed ballots (something I discuss in depth here). Briefly, the law requires poll workers to put their initials on the backs of ballots before issuing them to the voter and when, at the end of the day, excess ballots exist, they should be eliminated by discounting the uninitialed ballots until the number of counted votes corresponds to the number of voters who signed in prior to voting. However, what if Coleman persuades the court that the excess exists only because of the double-counting of duplicate ballots? Well, interpreted literally, the law does not care: It says to eliminate the uninitialed ballots first, regardless of whether they are legitimate votes. MSA 204C.09. Officials should not get down to discounting other types of votes to eliminate the excess unless the uninitialed ballots are eliminated first and excess still remains. But the court might flinch from following the strict letter of the law in this case, because it would mean throwing out uninitialed votes that very well might be legitimate, while still double-counting duplicate votes.

Another possible result (unrelated to the excess issue) is this: The court decides that X number of votes were double-counted duplicates, but does not feel comfortable identifying and throwing out individual votes because it cannot be sure which ones were duplicates. In other words, a number of illegal votes have made it into the pile, but they are mixed up with the legal votes and cannot be identified. The Supreme Court of Minnesota has indicated that, in cases like this, in order to "eliminate" the illegal votes it is appropriate to reduce the vote totals of each candidate in proportion to the percentage vote that each of them received in each affected precinct. Berg v. Veit, 136 Minn. 443, 446 (Minn., 1917); Hanson v. Emanuel, 210 Minn. 271, 280 (Minn., 1941). However, the cases approving this procedure are quite old and courts in the state of Washington and other states have rejected the procedure because, they said, it did not fit with modern statistical standards. On this point, we may get a chance to see whether the Minnesota Supreme Court chooses to rely on very old precedent or finds a way to alter the rule. If courts decide to apply the old rule, they will have to be very careful to make sure that it is applied uniformly across all affected precincts, to avoid Bush v. Gore-type issues.

There is more to talk about, but one interesting upshot is that the messiness revealed by the very discovery process that is supposed to assure the public that the true result was reached just might persuade the public of the exact opposite: that whatever result is eventually reached has no more validity than what came before. Therefore, after the lawyers have finished their arguments and the judges retire to their chambers to contemplate the case, I hope they consider, if only for their own amusement, the half-facetious recommendation of elections scholar Michael Pitts has made for close elections like these: Flip a coin. Heads or Tails? A Modest Proposal for Deciding Close Elections, 39 CTLR 739 (Dec., 2006).