Massive Bostonian Voter Fraud

At least that's what the Fire Andrea Mitchell blog seems to have found while looking at Boston's unofficial voting reports.

There are reports of several places showing more votes cast on election day than actual registered voters. Lets look at Boston, Massachusetts. No one expected that dump of a city, or state to go for Mitt Romney in the election, but the math just doesn’t make much sense ... how does one explain that Boston has 387,142 voters on their roll. Yet, somehow 501,282 ‘cards’ were cast on election day, a 129.48% turnout? Does the chowderhead vote count for like 1 and a half votes now or something? Check out the results yourself. Voter fraud! 129% of registered Boston vote on election day?

So I did. Lo and behold, something sure does seem fishy in the precinct-by-precinct voting reports. Here's a screenshot.

Woah! Nearly every precinct has voter turnout of around 130%. (Except Ward 1, Precinct 15. What's up with them?) That goes on and on for seven pages of this small-fonted report. This old-fashioned ballot-box stuffing is an impressive feat, especially in light of the incompetence required to leave clear evidence of the fact on a public server.

But perhaps (he said sarcastically) another possibility: maybe the data is just mis-coded. If you naïvely assume one card per voter and calculate turnout by cards cast, having a two-card ballot will significantly overstate your voter turnout.

Is there a clue somewhere else in the data that this mis-coding might be what's going on? Put your hands on your hips, it's just a step to the right:

You may have to click to embiggen.

From the top line, the votes for president were counted 855 times out of 1670 cards cast. That's 51.19%.

Glancing over the rest of the wards suggests similar numbers. Drastically fewer total votes for president than there were cards cast. This raises an important question:

If you were stuffing the ballot box, why leave the president blank on half? What are you doing, just adding ballast?

No. I suspect the ballot spans more than one card, and the formula used to calculate voter turnout has not taken that into consideration.

Now you are asking how more than 50% of the cards could count votes for president if it's two cards per ballot? Shouldn't that number always be close to, but less than, 50%?

Sharp thinking, but the answer is simple: some people did not turn in the second card. Boston uses op-scan machines, so this is entirely likely. Most (but not all) voters turned in both cards, but enough turned in only one card to push the presidential votes occasionally over the 50% mark.

Sloppy reporting to be sure, but keep in mind these are unofficial results reported rapidly during an election of high interest.