The New York Times this morning reports a doozy of an election-error story: In 80 voting districts around the city, unofficial results that were put out on Super Tuesday showed Barack Obama winning, unbelievably, no votes at all.
New York, which uses ancient lever-style mechanical voting machines that produce no paper record of each vote cast, has a complex process of reviewing Election Night results for errors. That review, says the Times, is turning up hundreds of Obama votes that had previously gone unrecorded.
The upshot is this: Results in the city may turn out to be closer -- maybe much closer -- than we'd previously suspected, possibly helping Obama gain additional delegates on Hillary Clinton's home turf.
"The history of New York elections has been punctuated by episodes of confusion, incompetence and even occasional corruption," says the Times, but many who commented for the article, including reps from both the Obama and Clinton campaigns, say there probably wasn't corruption this time.
But why unofficial results gave Obama zero votes in many areas -- and why even Clinton, in some districts, also got zero votes -- seems to be a complete, fishy mystery. Speaking to the Times, local officials offer unhelpful explanations like "has to be a mistake," "probably the result of human error," and "clerical error of some sort."
What kind of mistake, human error, or clerical snafu? Nobody ventures a theory.
Neither do we know how officials are mining up additional Obama votes in their official review. In Harlem's 70th Assembly district, for instance, Election Night results gave 141 votes to Clinton and zero to Obama (which made no sense, considering Obama's appeal to African-American voters).
A review since then has put the tally at 118 votes for Clinton to 116 votes for Obama. A Clinton Sweep, in other words, turned into a dead tie.
The most innocent explanation is that something got fouled up in the reporting of results on Election Night.
As this Times graphic shows, on Election Night inspectors record the results from each voting machine in the city. Each inspector turns his results over to a police officer, who takes the results to a precinct house and records the numbers into a computer. The computer disseminates these numbers to the press. At any step in this process -- recording the votes from the machine, passing the votes to an officer, entering the votes into a computer -- there could have been a clerical error.
Importantly, though, these numbers are not official; they have no bearing on who wins an election, or, in this case, how delegates are awarded to the candidates.
Official results are determined through a "recanvass," a process in which a Republican and a Democratic official go to each machine and collect official tallies. These tallies are then entered into a computer, which adds up the binding results.
It's a system, in other words, of which Rube Goldberg would be proud. And it does little to inspire confidence in election results.
Of course, for years now, we've known that voting on systems that produce no human-countable record of the ballot -- whether those machines are electronic touch-screen systems, or whether they're mechanical lever machines -- is a very bad idea. The shifting numbers in New York's election prove that.