What has had eyebrows raised is a significant discrepancy between the vote counts done by voting machine, and the ones done by hand.
In New Hampshire, 81 per cent of the voting was done in towns and cities that had purchased optical scan machines from the Diebold Election Systems (now called Premiere Election Solutions), a division of Diebold Corp., a company founded by and still linked to wealthy right-wing investors. In those towns, all voting was done on the devices, called Accuvote machines, which read paper ballots completed by voters who use pens or pencils to fill in little ovals next to the candidate of their choice. The ballots are then fed into, read, and tallied by the machines. The other 19 per cent of voting was done in towns that had opted not to use the machine, and to use hand-counted paper ballots instead.
The machine tally was Clinton 39.6 per cent, Obama 36.3 per cent – fairly close to the final outcome. But the hand-counted ballot count broke significantly differently: Clinton 34.9 per cent, Obama 38.6 per cent.
Could something have happened in those machines to shift some votes away from Obama or some of the other candidates in the race, and over to the Clinton total?
If all the votes cast had split the way the hand counts split, Obama would have won New Hampshire by over 10,000 votes, instead of losing to Clinton by about 5500 votes.
"My suspicion is that nothing untoward happened here," says Doug Jones, a professor of computer sciences at the University of Iowa and a member of the board of examiners that approved the use of the same Diebold optical scanning machines in Iowa. "But at the same time, the Diebold machines are vulnerable to viruses that can be spread through the machines by the PCMCIA memory cards, and there are other things that can go wrong too. I’d be much happier if they had a routine random audit procedure in New Hampshire."
A random audit, he says, would involve doing hand counts of some towns’ optical scan ballots, and comparing those results with the results of the machine reading of those same ballots, as recorded election night.
While California does conduct such random audits as a matter of course, most states, including New Hampshire, do not. According to the New Hampshire Secretary of State’s office, any recount of ballots would have to be requested by a candidate, and would have to be paid for by the candidate making the request.
An official in the press office of Obama’s campaign in Chicago, contacted on Wednesday, claimed not to know about the discrepancy between the machine and hand-counted ballots. She said that there was no plan to call for a hand count of machine ballots.
As Prof. Jones notes, requiring a candidate to initiate any hand count makes such hand counts unlikely, since unless the evidence of vote tampering or fraud is overwhelming, such a call would open the candidate to charges of "poor loser."
Kucinich, in making his recount request, resolved that problem.
There is good reason to be suspicious of the results. The counting of the machine totals, in New Hampshire as in all states using the Diebold machines, is handled by a private contract firm, in this case Massachusetts-based LHS Associates, which controls and programs the machines’ memory cards. Several studies have demonstrated the ease with which the memory cards in the Accuvote machines can be hacked, with some testers breaking into the system in minutes.
There are, to be sure, alternative quite innocent possible explanations for the discrepancy between the machine and hand votes for Clinton and Obama. All the state’s larger towns and cities, like Nashua, Concord and Portsmouth, have gone to voting machines. While there are many small communities that have also opted for machines, it is almost exclusively the smaller towns and villages across the state that have stayed with hand counts-most of them in the more rural northern part of the state. So if Obama did better than Clinton in the small towns, and Clinton did better in the large ones, that could be the answer.
But that explanation flies in the face of logic, historic voting patterns, and most of the post election prognosticating.
If it is true that there was "behind the curtain" racism involved in people saying to pollsters that they were for Obama, while privately voting against him, surely it would be more likely that this would happen in the isolated towns of northern New Hampshire where black people are rarely to be seen. Clinton was also said to have fared better among people with lower incomes-again a demographic that is more prominent in the rural parts of the Granite State. Finally, Obama, in New Hampshire as in Iowa, did better among younger voters, and that is the demographic group that is typically in shorter supply in small towns, where job opportunities are limited. Furthermore, in Iowa, it was in the larger municipalities that Obama fared best, not in the rural towns, so how likely is it that his geographic appeal would be reversed in New Hampshire?
David Scanlan, New Hampshire’s deputy secretary of state for elections, whom I contacted Thursday, said that while town election officials are required to do test runs of the Diebold machines in the days before an election, "to make sure that they are reading the ballot markings accurately," and that at that point the machines and the memory cards are sealed until the actual election day, there is no way for his office to independently conduct a post balloting test. The ballot boxes are sealed and the only way they can be opened if for a candidate to request (and pay for) a manual recount, or for a court to order one." Scanlan says that the same is true for the voting machines and the memory cards. While the sealed ballots are retained "for years," however, the memory cards will be back in the hands of the contractor, LHS Associates, in "a few months," to be erased and prepared for use in the general election next November.
Scanlan says that the state legislature is currently considering legislation to provide for routine audits of machines after elections, but that won’t help this election cycle.
Scanlan said that because the machines are freestanding, there is no chance of their being hacked from the outside, but critics note that the hacking can be done in advance to the memory cards, which can pass changes to each other like a virus as each is programmed for a particular election.
Jonathan Simon, an attorney and co-founder of the group Election Defense Alliance, says that the vote discrepancies between machine and hand counts in New Hampshire’s Democratic primary are troubling, and defy easy explanation.
"The trouble is, whenever you have a surprise result in an election, and it runs counter to the polls, the media always say the problem is the polling, not the counting." But he adds, "The thing is, these things always work in one direction-in favor of the more conservative candidate, and that defies the law of quantum mechanics."