UPDATED JANUARY 14, 2007
Correction of EDA’s previously reported statewide anomaly
On January 10, 2008, analysts at the Election Defense Alliance (EDA) reported that based on the official results on the New Hampshire Secretary of state web site, there was a remarkable relationship between Obama and Clinton votes, when you look at votes tabulated by op-scan v. votes tabulated by hand in a head-to-head contest between the two candidates:
|Category||Votes||Clinton vs. Obama percentage|
|Clinton: statewide optical scan tally||91,717||52.95%|
|Obama: statewide optical scan tally||81,495||47.05%|
|Clinton: statewide hand-count tally||20,889||47.05%|
|Obama: hand count||23,509||52.95%|
EDA has subsequently learned that the list of hand-count voting districts in New Hampshire that it used in its initial analysis on January 10, 2008 was outdated; shortly after that list was downloaded a revised list (1) was published by the New Hampshire Secretary of State with fourteen hand-count precincts converted to Diebold optical scan.
While the actual difference between Obama and Clinton hand count and optical scan margins are not a mirror image of each other to four decimal places as we had initially believed, the undeniable fact is that Obama appears to have carried the hand-counted tally statewide, while Clinton carried the optical scan statewide tally, by almost exactly opposite margins remains a remarkable result.
Whenever the outcome of an election is strongly correlated with the method of voting – given the well-known vulnerabilities of the specific model of Diebold equipment in use – additional investigation is warranted. This is especially urgent when the margin between two candidates for ballots counted by hand conforms to the margin between two candidates reflected in hand-count optical scan vote is so far apart - and the hand count matches the pre-election polling so precisely. Our analysis has continued, and additional findings will be published separately.
Head to Head
Clinton v. Obama Pre-election polling
Head to Head Clinton vs. Obama percentage
|Clinton: statewide optical scan tally||95,843||52.73%|
Real Clear Politics (2)Average
1/5 – 1/7/08:
Clinton 43.9%Obama 56.1%
|Obama: statewide optical scan tally||85,910||47.27%|
|Clinton: statewide hand-count tally||16,767||46.75%|
|Obama: hand count||19,097||53.25%|
2 - See Real Clear Politics for a summary of five public polls from 1/5 to 1/7/08, Obama = 38.3%, Clinton = 30.0%. Head to head percentages were calculated as: Clinton = 30/(30+38.3) and Obama= 38.2/(30+38.3)
ORIGINAL ARTICLE FOLLOWS - see correction above
Analysts at the Election Defense Alliance (EDA) have confirmed that based on the official results on the New Hampshire Secretary of state web site, there is a remarkable relationship between Obama and Clinton votes, when you look at votes tabulated by op-scan v. votes tabulated by hand:
Clinton Optical scan 91,717 52.95%
Obama Optical scan 81,495 47.05%
Clinton Hand-counted 20,889 47.05%
Obama Hand-counted 23,509 52.95%
The percentages appear to be swapped. That seems highly unusual, to say the least.
EDA and others are proceeding with intra and inter-county results and demographic analysis to better understand what this extremely unusual "coincidence" may indicate. The work to understand what really happened in New Hampshire is far from complete.
In the meantime, what are we to make of all this?
On the one hand, everyone has heard of the unanimous verdict of both private and public opinion polls leading up to the New Hampshire primary, showing Obama with about a 10% lead. And a report on Brad Blog today quotes Chris Matthews on "Hardball" who saw a comparable lead for Obama - about 8% - in the media's "unadjusted" New Hampshire exit poll.
On the other hand, it is a fact that the specific models of Diebold op-scan and central tabulators currently in use to count votes in New Hampshire have been proven, by multiple public demonstrations, to be wide-open to insider manipulation through a variety of mechanisms. Some exploits involve computer programs, and others, simple proximity to the central tabulator or precinct scanner.
So there is an undeniable possibility that the optical scan vote in New Hampshire could have been manipulated by insiders at the outsourced companies that run the election there, or by anyone with hand-on access to the voting and tabulating machines.
One recourse might be to recount the paper ballots by hand, but as we saw in Ohio in 2004, in the absence of a secure chain of custody the accuracy of any after-the fact recount of optical scan and hand-count ballots remains problematic - and the clock is ticking.
I've met some of the people who run elections in New Hampshire and I know that they are proud of their state's historic committment to fair elections. When I testified there last year to their State House Subcommitee on Voting Equipment, I advocated New Hampshire adopt the Universal Ballot Sampling hand-count audit protocol, a statistically-robust way of checking the accuracy of optical scan voting and central tabulation systems. If only the state had chosen to implement that simple and universal hand-count audit protocol for the 2008 primary, these lingering questions about the accuracy of the official count could be have been examined independently.
New Hampshire, and most other states still - unbelievably - run their elections with equipment and procedures so open to insider manipulation that they would lead to serious civil or criminal liability if comparable systems were allowed to run, in any bank in America.
If we can't simply return to the best system of voting, hand-counted paper ballots, why can't we at least put in place the capability for an independent citizen-run verification of the accuracy of our computerized voting equipment on election night?
If you have faith that the Diebold equipment is accurate, you are left only with the supposition that many voters in New Hampshire lie to exit pollsters or are secret racists - and the ones who do so also vote on optical scan equipment.
I think New Hampshire is a better state, and we're a better nation than that; and I'd like to know why the people who run our elections think it's acceptable that I can't prove it.