Election audit sufficient?

By RICHARD VALENTY Colorado Daily Monday, November 14, 2005

Boulder County ran a post-election audit procedure Monday, satisfying Secretary of State election rules but leaving local voting activists asking for more.

County staff pulled 100 ballots at random from a batch of 250 ballots that were processed on one of the County's eight scanners during the tabulation process.

The Secretary of State's office determined prior to the audit that the County would do a hand test of results among the 100 ballots from two races; the Town of Erie Issue 2A and the Longmont City Council Ward 2 contest, and compare the hand test with the County's voting system tabulation.

A form delivered from Secretary of State Gigi Dennis to County Clerk Linda Salas directed the County to audit results from Referendum C or D if no Erie or Longmont races were counted on the ed scanner.
 

The 100 ballots included 44 with Ward 2 on the ballot and 15 with Erie 2A, while others were from other cities, including Boulder, and were run through the scanners but not hand-counted. The hand-count results matched the machine tabulations perfectly.

Citizens attending the audit were required to be official representatives of candidates or issue committees. Boulder resident Ralph Shnelvar attended on behalf of the anti-Referendum C committee ?Vote No, It's Your Dough,? and had several issues with the scope of the audit.

Secretary of State election rule 11.5.4.4 says election officials shall 1 percent ?but not more than 100 ballots? counted on the specific audited device.

Shnelvar said such an audit would not be statistically significant enough to provide citizens with confidence of election accuracy.

Roughly 86,000 County ballots were cast in 2005, although the counting was distributed among the eight scanners.

?There's just no way of knowing that these machines did the right thing,? said Shnelvar during the audit.

Josh Liss, County elections coordinator, said the County intended to follow the letter of the Secretary of State rules during Monday's audit, but said he would work with local voting activist Neal McBurnett, who attended the audit, to try to craft new audit-rule language in efforts to satisfy a broader segment of concerned citizens.

?Let's all work together to change the rule,? said Liss to the four citizens in attendance.

Citizen Mary Eberle, who also attended a 600-ballot audit after the March 2005 City of Boulder election, expressed concern that the audit was not truly ?random? since County staff and not citizens ed the 100 ballots.

Shnelvar said he remains concerned about the accuracy of Ref. C results. The County's election system had the capacity to ?read? a fold in a ballot as a voter choice, and many, if not all, 2005 County mail ballots had a fold running through the ?yes? choice on Ref. C.

County election staff caught the ?fold? situation during a pre-election Logic and Accuracy Test, and staff devised a strategy to visually inspect each ballot before scanning.

The fold would, in theory, only affect undervoted races since the County system would ?flag? races with overvotes for visual resolution, and County staff separated ballots with undervoted ?fold? races into distinct batches for processing.

The County will perform canvass board post-election operations Tuesday, and Liss said the County would only legally be allowed to conduct another specific examination to deal with Ref. C results if it received a written request explaining specific reasons for the test prior to the 9 a.m. commencement of canvassing.

Shnelvar said he and McBurnett would work on creating the written request after the audit, but neither had a draft prepared at roughly 3 p.m. Monday.