During the 2009 municipal elections the Auditor’s Office expanded its experiment with the electronic poll book (e-poll book) to a total of five precincts.  Our objective is to evaluate how much e-poll books will increase accuracy and efficiency in the polling place, and save costs in after election processing of voter history.  We plan to use this information to decide if the increases in accuracy, efficiency and cost savings offset the investment cost in e-poll books.

The e-poll book continues to demonstrate increased accuracy and some increase in efficiencies in the polling place, especially for special case voters, when compared to the current system.  The experience reported by precinct election officials was generally very good.  The amount of time it took to process the average voter was a little faster with the e-poll books than traditional paper methods, and promises to only improve as poll workers gain more experience and confidence.  Also, e-poll books were more accurate in recording who voted than with the paper roster form.  Finally, the e-poll books made processing special case voters, such as address or name changes, provisional voting and election day registration much easier for poll workers, and therefore faster for the voters involved.

The cost savings for post-election processing promises to be very significant.  In the five precincts using the e-poll books we processed history for 1080 voter records in 26 minutes.  A comparable number of records in three precincts processed by hand took 182 minutes.  By this measure processing voter history was seven times faster with the e-poll book than with hand processing.  Processing measured by precinct was 11.7 times faster with the e-poll book than with hand processing.

Further, with e-poll books it takes a little more than five minute to process a precinct regardless of the number of voters.  In contrast, hand processing is measured in units of average voters.  Average voter units include the time needed to identify and correct errors in the printed election register.  Therefore, if e-poll books were deployed to all 63 precincts in Scott County processing those precincts would take approximately five and one half hours.  Hand processing records in a heavy turnout election such as the 47,582 voters in the 2008 general election would take almost seventeen days.  Experience shows that election roster errors greatly increase in heavy turnout elections, so that the time to hand process a heavy turnout election would be even longer than seventeen days.  In fact the 2008 voter update was not completed until the end of December that year.  Thus, there is potential for substantial cost savings in utilizing e-poll books, especially in precincts with significant numbers of voters.