What if we had an election without electronic voting machines?

[Stevepol]

25. I appreciate your points. It does seem that Simon ignored a well-done exit poll.

However, the difference between the handcounted 3% of voters and the 97% rest of the state is still highly unlikely. Simon shows that the 97% population of voters was actually MORE likely to vote for Coakley in the previous 2006 election for Attorney General so the 97% population sample was not only equally Democratic but more sympathetic to Coakley than the 3% sample. Here's what he says about the result (p. 235):

"If the handcounted ballots had been distributed randomly throughout the Commonwealth, we would expect the handcount results to fall within 1.0% of the opscan results with better than 99.9999% confidence. The odds of an 8.0% marginal disparity would be beyond astronomical. We have further established that the handcount 'sample' is, for comparison purposes 'better' than random: that is, based on demographics and voting patterns, the handcount voters would be more likely than the opscan voters to vote for Brown. The odds therefore of an 8.0 marginal disparity in the other direction would be, and there is no better way to say this, beyond beyond astronomical. Statisticians never say 'impossible' but that is, for all earthly intents and purposes, what it is."

In the US now, I think we can agree, where the vote is either impossible to verify or almost never verified (even, as is the case in MA apparently, where there is paper to do an audit) it is still essential to democracy that in EVERY ELECTION where electronic voting machines are used, that the vote BE VERIFIED. And about the only way this can be done is through statistical tests, required audits.

Right now, here in Kansas, Beth Clarkson, a statistician at Wichita State, is trying to just gain access to the paper trail in Sedgwick County so that she can rule out machine improprieties in the odd results that came out of the most recent KS elections. She will almost certainly be denied that right by the powers that be. Because we have turned over the counting of votes to voting machines programmed by right-wingers, essentially there's no way to directly verify the vote. If we are going to use these machines, the only recourse we have to VERIFY THE VOTE is statistical tests like those Simon has gone to great pains to do.

Your efforts and those efforts of many others to throw light on our elections is much appreciated. I'm glad to share this forum with you. It's only by efforts like those you are making to come to the truth that the problems we have with the fair counting of votes in the US will ever be resolved.