I personally think Kucinich is right to have called for a recount on principle, but maybe $100k or more is too much money for his campaign and I would have for him to make a fool of himself if the recount finds no irregularities (as in the Nader recount in 2004).

the stormy present:

I have personally got serious reservations about the use of voting machines that don't leave a reliable, verifiable paper trail, and I think their use needs to be discontinued.  I have said this since 2000.

I definitely intend to run a regression on education, income, age, race, gender and town size if I can get my hands on census data. The idea is that after all these effects are taken out there should be no correlation between vote counting method and Clinton/Obama spread.

Examples:  The guy from Election Defense Alliance says he got his data from the NH Secretary of State web site.  He doesn't link to it, but the data is here.

He says Clinton had 91,717 optical-scan votes and 20,889 hand-counted votes, for a total of 112,606 votes.  But the NH Secretary of State says Clinton got a total of 112,610 votes.  He says Obama got 81,495 optical-scan votes and  hand-counted votes for a total of 105,004 votes, but the NH SOS says he got a total of 105,007 votes.

He also doesn't say how he determined which votes were machine-counted and which were hand-counted.  Here's the official list of machine-counted municipalities, fyi.

Again, thanks for this. Scepticism is what we need. I think Georg Polya said the steps in proving something are to convince yourself, then convince a friend, then convince an enemy. Right now we're in the "convince a friend" stage.

We have met the enemy, and he is us — Pogo