By Verified Voting Foundation Limited Scope Investigation Not Conclusive
Verified Voting Foundation concluded after reviewing a leaked copy of a draft GAO test report that the findings were not sufficient to exonerate the voting machines in determining what caused a massive undervote in the Florida District 13 contest of 2006.
"After a lot of investigation, we still don't know what happened," said founder David Dill, a computer science professor from Stanford University who co-authored two papers regarding the problem. "We do know this GAO report cannot be interpreted as a clean bill of health for the machines."
The investigation done by the GAO was limited in scope by agreement with the Congressional Task Force for the Contested Election of the Committee on House Administration. VVF found that the testing was insufficiently ambitious to determine what caused an undervote rate many times higher than in any previous election in that contest, and many times higher than that experienced on other voting systems used in that election.
"This is to point out how hard the problem is, not to criticize the GAO," said Dill. "Had this election been conducted on a voter-verified paper ballot system, as in surrounding counties that form part of District 13, it probably would not have failed. More to the point, it would have been a lot easier to find out what happened."
�The lesson here is that the complexity of computer systems, and the poor quality of evidence trail they produce, can lead to un-resolvable election disputes,� added Dill.
The recently leaked draft report of the GAO study leaves most of the major questions unanswered. Many issues uncovered by activists have not been addressed, and recommendations for further investigation from a study released many months ago have been ignored.
Since as the GAO researchers acknowledge, the interaction between voter and machine plays an important role in the many of the voter complaints from Sarasota, and the hypotheses for what may have gone wrong, those issues should be explored more thoroughly in testing. Other issues such as internal bug data, whether firmware on the ballot cartridges played a role, and even equipment manufacturing quality problems should be investigated before drawing conclusions.