Do our Minnesota Diebold machines count correctly?

Today I was able to observe an audit of one of our Diebold Accuvote vote counts. Basically 4 precincts are randomly selected for a county to be have a hand count that is matched against the machine count. I did not observe the "random" selection process, other people did. I watched a one precinct audit of 3 races: U.S. Representative, Senate, and Gubernatorial races. The difference between what the audit showed that the machine counted and people counted was at the highest, 0.2%, off by 4 in a count of 1920. Some of these problems can be attributed to jams (which are usually written up in an incidence report) and some could be because the pen used to mark the ballot was too light in color. Some differences do not have an explanation.

All of our ballots are paper, which is then counted optically by machine. In this way, we always have the original ballot for a recount. We also can continue voting during a power outage or if the counting machine goes on strike. We also can have many, many vote marking spots and just one counting machine, which speeds up our process. I think we have the cleanest most efficient voting processes in the US. Now with this audit, we can have some confidence in the machine count.

Above all, the most reassuring aspect of the audit was the attitudes of everyone involved in the audit. Everyone wanted the counts to match exactly and recounted several times to ensure that there really was difference. One comment was that close was good only in horseshoes, not in counting votes. I am proud of our voting process and the Minnesotans who are working hard to keep it the best. To learn more about election integrity, check out the Election Integrity website.