Was the Buchanan vote much larger than expected in Palm Beach County?

I began my own analysis of the impacts of the Palm Beach ballot as an intellectual exercise on the evening of Wednesday, November 8th, and I completed it by Thursday afternoon, November 9th. I then posted it on the web at http://socrates.berkeley.edu/ ucdtpums/). Many others also posted their analysis in the next few days. Among the numerous excellent pieces of research, I rely upon the paper by Jonathan Wand, Kenneth Shotts, Jasjeet Sekhon, Walter Mebane, and Michael Herron which is available on the web at http://elections.fas.harvard.edu/. My analysis relies upon standard and widely used statistical techniques. It also relies upon my experience as a Ph.D. in Political Science and Economics, as a teacher of courses in statistical methods at the University of California, Berkeley, and as a widely published author in statistical methods and electoral research. The papers mentioned above and many others present overwhelming evidence that the percentage of the vote for Patrick Buchanan in Palm Beach County is extraordinarily atypical compared to all other Florida Counties and compared to other reporting districts around the country once we control for state variations in Buchanan support and for the different sizes of reporting units. These analyses prove with a very high degree of scientific certainty that the Palm Beach County Buchanan vote of about .8% was exceptionally large - between 2.6 to six standard deviations away from the expected value of around .19 Thus, the likelihood that the observed Buchanan vote is produced by the same factors that were at work in the rest of Florida's counties is certainly no greater than one out of 200 and possibly as small as one in a billion.