I had exchanged several emails with Professor Hout, taking him to task for not acknowledging that the anomalies revealed by his statistical analysis could be best explained by the strong likelihood that Democrats in Palm Beach, Dade, and Broward counties had inflated Gore's vote totals in 2000 through punched-card ballot fraud, and that the use of electronic voting machines in 2004 had forestalled that method of fraud, leading to an apparent surge in Bush votes over what his model had predicted for those counties. Instead, Hout's publically-reported comments talk about explanatory factors such as software or firmware errors in the electronic voting machines, and he left open the possibility that deliberate fraud (by voting machine manufacturers, perhaps) might have been involved, all of which are substantially less likely than the punched-card vote fraud explanation.
As expected, and as is evident from Hout's dismissal of the academic criticism (in the Wired article), Hout has refused to acknowledge any such possibility. It's evidence of a stunning lack of intellectual honesty, let alone a lack of intellectual curiosity on the part of a "professor", but evidently that's what is to be expected from a "Berkeley professor".