In the last couple of days I've seen a lot of discussion of the "undervote" count in the Minnesota race for US Senate. Franken fans are excited about the recount, and seem to be building up a lot of enthusiasm for what they see as a likely Franken win.
I would love to see Al Franken become our Senator, but recent reports are overly optimistic. In particular, I think that this AP article tends to misread the situation here on the ground in Minnesota.
By way of background, it's worth having another look at the graphs I made in this diary. The information presented there is the foundation for my reasoning today.
Here's the logic that I think is faulty: There were 25,000 people in Minnesota who voted for President but not Senate! And most of those came from counties won by Obama! And optical scan machines can overlook some marks on ballots! So a hand recount will probably show a Franken win!
That may very well be correct. But there are a lot of reasons why it may not be.
First, the Franken-Coleman race was considered very bitter by Minnesota standards, and the advertising and phone calls turned off a lot of people. A lot of voters hated both candidates, resulting in a huge third-party turnout, and, undoubtedly, a lot of people who intentionally cast no vote for Senate. There's no reason to believe that an undervote of 25,000, or about 1%, shows anything but a lot of disgusted voters.
Second, the fact that more undervotes came from Obama-friendly counties doesn't shock me: Obama won Minnesota by 10 points. As I showed earlier, it would make more sense to look at whether the undervotes occur in Franken territory rather than Obama territory. And county-level analysis is too crude (Hennepin county includes a lot of urban voters, but also a lot of suburbs): it's important to look at raw precinct data, as I did last time. Even then, you get a pretty mixed picture: any new Franken votes would probably be balanced out by new Coleman votes (the net result is about +1 Franken for every 15 newly counted votes). There would have to be an awful lot of undervotes that get turned into counted votes in order to be able to tip this race. It's hard to say with any confidence which way things will go, because the "accidental" undervotes are swimming in a sea of "intentional" undervotes.
Third, at least here in Hennepin county, home to Minneapolis and a huge chunk of the Franken vote, votes are tallied on election day using ES&S optical scan machines. Voters themselves insert their ballot into the machine. And the machines are programmed to alert voters to an undervote: If there is no vote recorded for a contested race, the machine will beep and reject the ballot. Update: If the machine rejects an undervote, the normal course of action is that the voter takes their ballot back and fixes it (or exchanges it for a fresh ballot).
To summarize, there's a perfectly good explanation for all the undervotes.
And it's pretty hard to get an accidental undervote recorded. And newly counted undervotes are not likely to break for Franken by a huge margin. Is any of this reason to give up on the recount? Of course not. As far as I'm concerned the recount is a coin toss, just as likely to give the seat to Coleman as to Franken, but that's still a lot better chance than just surrendering and letting Norm Coleman return to his seat for another six years.