http://www.expertclick.com/NewsReleaseWire/default.cfm?Action=ReleaseDetail&ID=9474By Gigi Wasz
Gazette news magazine, Chicago
A University of Illinois at Chicago (UIC) professor recently issued a report calling into question the results of the 2004 presidential election.
In the 30-plus page report, Ron Baiman, PhD, of UIC's Institute of Government and Public Affairs, along with eleven other colleagues from other prestigious universities, applied quantitative data to explain the discrepancy between exit poll projections and votes actually recorded in the November election and to understand the analysis given by Edison Media Research & Mitofsky International (E/M), the pollster of record for the national election.
E/M's poll projections predicted a win for Democrat John Kerry by 3%, however, when votes were tallied, Republican President George W. Bush was given the win by 2.5%-the largest discrepancy in the poll's history. In its post-election analysis and report, E/M discredited its own poll projections, claiming the official vote was not corrupted and that "Kerry voters were more amenable to completing the poll questionnaire than Bush voters."
"Using their (E/M's) data tables, we demonstrated that their hypothesis of outspoken Kerry supporters is implausible," explained Baiman. "If the polls were faulty because Bush voters were shy in the presence of Kerry voters and less likely to cooperate with pollsters, the polls should be the most accurate in the precincts where Bush voters were in the overwhelming majority and where exit poll participation was also at its maximum. What we find is just the opposite. In fact, the mean exit poll discrepancy was dramatically higher in Bush strongholds than in Kerry strongholds."
Baiman and his research team then set out to explain the discrepancy between the national exit poll and the official vote. They initially targeted random chance or exit poll errors. Said Baiman, "There have been several methods to estimate the probability that the national exit polls would be as different as they were from the national popular vote by random chance. These estimates range from one in 1,240 to one in 16.5 million. No matter how one calculates it, the discrepancy cannot be attributed to chance."