Princeton Professor: Minnesota Race Is A Statistical Tie

This story was written by Lauren Christensen, The Daily Princetonian

When the recount began last month, Coleman had an edge of just 215 votes out of almost 3 million cast. The narrow margin triggered an automatic recount under state law, which allows representatives from both campaigns to challenge ballots they believe to be invalid. Franken remains behind Coleman, according to The New York Times.

Eric Weitz, a visiting history professor from the University of Minnesota, noted that this race has been extraordinarily bitter by Minnesota standards.

The vote count is very, very close and the process will go on for weeks and weeks, he said in an e-mail.

Though the state canvassing board originally set this Friday as the target date for completing the recount, a recent decision by the Minnesota secretary of state to review absentee ballots may delay the result. Currently, the canvassing board plans to release the results on Dec. 16 after it reviews contested ballots.

For university students from Minnesota, many of who did not travel home to vote, the review of absentee ballots is especially important. Roughly 500 absentee ballots might have been wrongly disqualified, according to the office of the Minnesota secretary of state.

It is imperative that a recount be conducted of these nebulous absentee ballots, and that some explanation be provided of why they havent been counted, and whose ballots they were, Minnesota resident and absentee voter Marian Messing 11 said in an e-mail.

Other members of the Princeton community have taken a more technical interest in the race, which until the victory of Sen. Saxby Chambliss (R-Ga.), confirmed on Monday, could have helped the Democrats reach a filibuster-proof 60 seats in the Senate.

Molecular biology professor Sam Wang has been following the recount on his website, the Princeton Election Consortium, which he created to aggregate and analyze polling data. Since the 2004 U.S. Senate elections, Wang has been look[ing] at state polling data and turn[ing] it into a simple measure anyone would understand, he said.

State polling data is hard[er] to understand than the simple snapshot he presents on his website, he explained.

In a blog post on his website, Wang said that he used binomial distribution statistics to show that the difference in votes is smaller than the standard deviation for two equally probable outcomes in a situation with a similar number of votes cast, implying that, by statistical measures, the Coleman-Franken vote is a tie.

Im just a U.S. citizen with statistical abilities and a blog and spectator to one of the closest Senate races in history, Wang said.

He noted the importance of fair standards being applied to the recount process in Minnesota. The process of declaring a winner involves following the rules in a systematic way [as] a way of ensuring faith in our democracy, he said.

As documented on his blog, prior to the Nov. 4 election, Wang predicted that the race would be too close to call, with Coleman slightly ahead.

Wang said he is satisfied with the accuracy of his predictions.

Basically, the calculation did extremely well in advance, better than any other approach I know of, he said in an e-mail. When I said a tiny edge for Coleman, I didnt think 200 votes, but obviously I was right, he added in an interview.

Wang said that he is obligated to be unbiased in his analysis, noting that the idea s to provide high-quality reporting.

Wang, a registered Democrat, nonetheless acknowledges his political allegiances, saying that it would be weird to do this kind of thing and not care about the outcome.