Common Cause calls for hearing on vote

Major media outlets, among them the LA Times, the Boston Globe, the New York Times, and the Washington Post, have run articles debunking reports of vote discrepancies. The debunking consisted mainly in sticking the labels “internet” (gasp) and “conspiracy theories” (gasp) on those making the charges. The Washington Post was the most snide; its article referred to “the bloggers and the mortally wounded party loyalists and the spreadsheet-wielding conspiracy theorists.” Salon's and the New York Times' articles were more respectful and more substantive. But the articles tend to miss the point. Most likely the election results won’t be changed by these investigations (although in one North Carolina county 4,500 votes were irretrievably lost and the entire state may have to have a re-do of the election.)

The question remains: is the system broken? If so, how are we going to repair it so we can once again have clear and legitimate election results? Equally important is the question of whether Bush and the Republicans earned the mandate they would like to claim.

Common Cause has put its clout and nonpartisan authority behind calls for investigation. It is organizing a public hearing in Washington DC to ensure that Congress hears about voting problems. They’ve also gathered a report on problems gathered so far and plan to issue a comprehensive report in a few months.

The battle continues in Ohio. As Dennis Kucinich, U.S. representative from Ohio and former presidential candidate, writes, a losing candidate can request a statewide recount only if the margin is within one quarter of one percent (about 16,000 votes). However, a losing candidate or an Ohio citizen who voted for a losing candidate can request a recount of a particular precinct, at a cost of $10 per precinct. Ohio has about 12,000 precincts, so a recount of every precinct would cost about $120,000, not counting legal fees. Blackboxvoting and the National Ballot Integrity Project have created the Help America Recount Fund to collect money for recounts in Ohio, Florida, and elsewhere (where there is anything to recount—with electronic, paperless machines there is nothing to recount).

Dueling statisticians: The New York Times cites the Voting Technology Project, a cooperative effort of Caltech and MIT. Its preliminary study noted that the discrepancies between exit polls and tallied results in favor of Bush were statistically significant nationwide, but not when examined state by state, which the study correctly notes is the appropriate level of analysis, as the presidential election is decided at the state level. The study goes on to demolish the argument that there was significant discrepancy between exit polls and tallies in states that used electronic voting machines. True, such a claim doesn't hold water, but this is a strawman. The serious claims are not about electronic voting machines, but electronic tallying of optically scanned ballots (see my first posting). The MIT/Caltech folks make no mention of this issue... MIT and Caltech are big guns, but UPenn's not so shabby either. A UPenn professor has issued an analysis, showing that the combined chances of the deviation between exit polls and tallies occuring in Florida, Ohio, and Pennsylvania are 250 million to one. Penn professor Steven Freeman used the original exit poll data available on CNN's website until 1 AM Wednesday, November 3. By 6 AM, these data were "recalibrated" to reflect actual tallies, which sounds fishy to me. This might be a source of discrepancies between various studies; I'd tend to trust studies using the "uncalibrated" data (although it is worth checking whether recalibration simply meant a larger sample size, i.e. more voters polled).

Some have raised the issue of high tallies for Bush in certain Florida counties with majority Democratic registered voters. This has been addressed to my satisfaction by, among others, Salon's Farhad Manjoo. This phenomenon is nothing new to this election—Democrats have been voting Republican for president in the South since the 1980s—and it's called Dixiecrats. (What no one has explained is why the heck these people have stayed nominally Democrats. Perhaps they vote Democratic at the state and local level.)

A posting on Freepress.org is urging people to contact Ohio secretary of state Ken Blackwell to urge him to reconsider a new ruling in Cuyahoga County that provisional ballots in yellow packets must be rejected if there is no date of birth on the packet. The Free Press claims that, according to the original Provisional Verification Procedure from Cuyahoga County, date of birth on a provisional ballot is not mandatory. The original procedure required the voter's name, address and a signature that matched the signature in the county's database. The same posting says that once the provisional ballots start being counted this Saturday, anytime there is a question of the legality of one of the ballots it will be voted on by two Democrats and two Republicans. Blackwell breaks any ties, the posting claimed.