June 7, 2009
SD ‘Glitch’ adds 5000 phantom votes
An unexplained ‘glitch’ nearly doubled the reported vote count in a recent local election in Rapid City, SD. Election officials were quick to note that the error was caught and did not change election results.
Comment: election officials always assure the voting public that every thing is OK after a glitch like this is found, metaphorically pat us on the head and tell us to ‘move along … nothing to see here.’ They expect us to ignore the signs that something went wrong that they can’t explain — or how they plan to prevent similar glitchy events in future elections.
Fortunately this ‘glitch’ occurred in a place with paper ballots counted by an optical scanner (ES&S M-650, widely by jurisdictions across the country.
Imagine the case if there had not been paper backup, as has happened elsewhere with touchscreen voting machines. Without the paper there would have been no way to go back and figure out the correct number of ballots cast.
This incident demonstrates the urgent need for paper ballots every where, so that elections can be reviewed, recounted and audited for accuracy. Rep. Rush Holt’s new bill ‘The Voter Confidence and Increased Accessibility Act of 2009’ would move the nation toward paper ballots and post-election audits.
There are a couple of theories as to why the miscount occurred in the recent Rapid City local election. These hypotheses are based on the published account that the ‘glitch’ occurred when the precinct counts were combined in the central tabulator. Data from the M-650 scanners can be uploaded to the central tabulator via zip disk or via local area network.
If zip disks were used it would be possible to upload zip disks more than once, leading to the overcount.
If the scanners were networked to the tabulator then the error could likely have occurred because some batches of ballots were scanned twice.
Either of these would have been possible if there was no reconcilement of the number of ballots cast in the precinct to the number of ballots scanned at the central election office. Newspaper reports suggest there might have been sloppy or non-existent procedures for such reconcilement.
The last possibility would be an error in the tabulating software. This possibility cannot be categorically ruled out based on the information publicly available. However, it is much more likely that the ‘glitch’ was caused by human error in handling the ballots or the zip disks.
It is to be hoped that an investigation is conducted and procedures tightened for future Rapid City election.