18,000 missing votes in FL-13 were largely Democratic

From the Orlando Sentinel:

The group of nearly 18,000 voters that registered no choice in Sarasota's disputed congressional election solidly backed Democratic candidates in all five of Florida's statewide races, an Orlando Sentinel analysis of ballot data shows.

Among these voters, even the weakest Democrat -- agriculture-commissioner candidate Eric Copeland -- outpaced a much-better-known Republican incumbent by 551 votes.

The trend, which continues up the ticket to the race for governor and U.S. Senate, suggests that if votes were truly cast and lost -- as Democrat Christine Jennings maintains -- they were votes that likely cost her the congressional election.

Republican Vern Buchanan's 369-vote victory was certified by state officials Monday. His camp says that, although people may have skipped the race -- intentionally or not -- there is no evidence that votes went missing.

But the results of the Sentinel analysis, two experts said, warrant additional investigation.

"Wow," University of Virginia political analyst Larry Sabato said. "That's very suggestive -- I'd even say strongly suggestive -- that if there had been votes recorded, she [Jennings] would have won that House seat."

David Dill, an electronic-voting expert at Stanford University, put it this way: "It seems to establish with certainty that more Democrats are represented in those undervoted ballots."

The Sentinel reviewed records of 17,846 touch-screen ballots that included no vote in the tightly contested 13th District congressional race to determine whom voters selected in other major races.

In addition to making it look as though Jennings would have won if the votes had been counted, the preponderance of Democratic voters among the missing ballots suggets something else. If the voting machines had simply been suffering glitches, then the votes that were missing should have been essentially randomly distributed among voters. In that case, the breakdown of Republicans v. Democrats with missing votes should have been roughly equal to the breakdown of Republicans v. Democrats not missing votes. Since it wasn't, the Sentinel's results strongly suggest that the problem was a deliberate attempt to throw the election in favor of the Republican.

Unfortunately, the Sentinel article doesn't give the actual figures it came up with in its study. If it had, we could compute the mathematical probability that this was a fraudulent attempt to steal the election.