Correlation between Bush votes and Touch-screen machines found

I'm normally pretty skeptical of claims of a stolen election. But this study seems pretty interesting:

http://www.computerworld.com/governmenttopics/government/policy/story/0,10801,97614,00.html

The claim is that a statistical analysis showed that there was a significant correlation between "a county's use of electronic voting resulted in a disproportionate increase in votes for President Bush" in Florida.

I think this is worth taking seriously. In an age where companies routinely use statistical analysis to decide how to spend millions of dollars, we can't shrug off massive statistical evidence of vote fraud.