Windham Election Results Validated

Print More

From the official report

A panoramic view of the room taken from the far corner of the observation area during the hand recount phase of the audit. Screens on the perimeter tables were providing a close-up view of the documents on the hand count tables.


CONCORD — A forensic review of Windham’s election results confirmed the accuracy of the state’s recount of the Rockingham District 7 state representative’s race, according to the report from the Forensic Election Audit Team that Secretary of State William M. Gardner and Attorney General John M. Formella received on July 12.

The integrity of the election came under scrutiny after a recount showed that seven of the eight candidates for the seat picked up about 300 votes each, while the person asking for the recount, Democrat Kristi St. Laurent, lost 99 votes.

The report affirms that, “We found no basis to believe that the miscounts found in Windham indicate a pattern of partisan bias or a failed election,” and that, “All counts agree that Republicans swept Windham’s four seats in the state house of representatives.”

Senate Bill 43, which mandated the audit, requires the Secretary of State’s and Attorney General’s offices to prepare a joint report on the results of the audit and to present it to the New Hampshire Legislature and the Windham Board of Selectmen.

In conducting its audit, the team did not undertake a comprehensive security review of the system, which it said would require additional information and resources, including access to the source code.

“As part of the machine audit, we verified the integrity of the election equipment and inspected the software and configuration for anomalies and irregularities; we did not assess the underlying security properties of the software and hardware, Windham’s procedures for machine and ballot custody (beyond reviewing the documents provided to us), the accuracy of eligibility determinations, and similar matters,” the team stated.

The team’s approach to the audit included performing a hand tally and then confirming the results by scanning the batches with the AccuVote machines. After doing so, the memory cards were forensically examined and compared with a forensic image of the contents of the chips from a reference machine from the Secretary of State.

“Election integrity segments of the memory cards were confirmed to contain identical ballot data, contest data, and candidate data (including the locations of vote targets on the ballots)” and both expected and unexpected differences between the contents of the memory cards were identified and examined, the report states.

Paper ballots were forensically reviewed using a microscope, micrometer, and UV light to determine whether they were genuine printed ballots, and whether any folds were manual or machine-made. It was the folds that contributed to the problems with the machine count.

“To address claims in social media that the selection of the random box was rigged, we asked the audience to select a box of their choice for audit. The audience unanimously picked an additional box for examination,” the report said.

The team concluded that “the large discrepancy between election night totals and both hand counts … can be attributed to unforeseen consequences and misfortune. Harried election officials borrowed a folding machine to send out thousands of absentee ballots more quickly, and votes on roughly 400 ballots were miscounted as a result. … We do not believe town officials could reasonably be expected to have anticipated this problem.

“Even this near-perfect storm was not enough to alter the reported outcome in the State Representative contest (although under different circumstances, it might have). … Nevertheless, the error — along with the smaller discrepancies in the landslide Governor’s contest — expose some vulnerabilities that warrant further attention.”

The Forensic Election Audit Team’s report is posted at

Comments are closed.