Reply #7: Mecklenburg undervote analysis

[daverts]

I've put a site up that has some info on the Meck recount, plus data and analysis of a significant county-wide undervote that may have potentially affected the Kissell-Hayes results.

Scrutiny of the vote totals has revealed an unexplained undervote: a significantly higher percentage of voters than in the past declined to vote for the most prominent races such as those for US House of Representatives. Never in past years had the Congressional races been "outperformed" by downballot races such as for sheriff, county commissioner etc.... but it did happen in 2006. Mecklenburg County includes parts of three Congressional districts. ALL three districts had a significantly higher undervote in Mecklenburg than they did in other counties, and all three races had a significantly higher undervote than they did past years for the same race.

Many of you will realize that this undervote has implications in NC-08 because of the closeness of the Kissell-Hayes election, but it is also relevant to the undervote in FL-13 (Christine Jennings race) because that election also used the ES&S iVotronic. However, unlike in Florida, the iVotronics used in Mecklenburg do have a paper trail that could provide evidence that the ES&S voting machines are not as infallible as some would suppose.

During a meeting last week with the Mecklenburg Board of Elections, they admitted that there was an unexplained undervote based on the data that is presented on the site. However, they do not believe that it is possible that machine error could have played any role in the undervote, and they did not want to allow scrutiny of precinct-specific machine tapes because they claimed that we could determine individual voter identity from the tapes because they are serial records that could be matched to voter order, a far-fetched defense that presupposes our knowing on which machine each voter voted and memorization of the voter log.

Please visit the site for more details...