Reply #76: Re Franklin

[Febble]

When I first looked at this, my estimate was that around 18,000 votes were lost, of which about two thirds would have been for Kerry, 40% for Bush, and the remainder for independents, making an 8000 dent in Bush's margin.

This was assuming that the turnout differential observed in precincts in which the supply of machines per active voter was below the median would have been the same as that in precincts above the median.

However, on further investigation, I found that turnout in 2000 figures appeared to have been used to allocate machines (or similar figures) which meant that it is not legitimate to extrapolate observed turnout figures from the adequately-supplied precincts to the potential turnout in the inadequately supplied precincts. So I think the 8000 is probably a generous maximum, although I am prepared to be persuaded otherwise.

What is absolutely clear is that Democratic precincts in Franklin county were systematically provided with fewer machines per active voter, although whether this was due to conspiracy, incompetence or bad luck is up to you to say. I'd put it somewhere between the first two, but I'm not in a position to judge.

My paper is still posted here:

http://uscountvotes.org/ucvAnalysis/OH/FranklinCountyRe...

despite the fact that the site host thinks I am an incompetent liar.