University researchers challenge Bush win in Florida
'Something went awry with electronic voting in Florida,'
says the lead researcher
By Dan Verton
NOVEMBER 18, 2004 (COMPUTERWORLD) - Researchers at the University of California, Berkeley, said today that they have uncovered statistical irregularities associated with electronic voting machines in three Florida counties that may have given President George W. Bush 130,000 or more excess votes. The researchers are now calling on state and federal authorities to look into the problems.
The study, "The Effect of Electronic Voting Machines on Change in Support for Bush in the 2004 Florida Elections," was conducted by doctoral students and faculty from the university's sociology department and led by sociology professor Michael Hout.
Hout is an expert on statistical methods at the Berkeley Survey Research Center and a member of the National Academy of Sciences.
According to the study, counties with electronic voting machines were significantly more likely to show increases in support for Bush between 2000 and 2004 compared to counties with paper ballots or optical scan equipment. This change cannot be explained by differences between counties in income, number of voters, change in voter turnout, or size of the Hispanic/Latino population, said Hout.
In Broward County, for example, Bush appears to have received approximately 72,000 excess votes, Hout said, adding that the research team is 99.9% sure that these effects are not attributable to chance. The other two counties that experienced unexplained statistical discrepancies in the vote are Miami-Dade and Palm Beach. The three counties revealed the most significant irregularities and were the most heavily Democratic counties in the state. Smaller counties that showed strong support for Bush didn't produce any statistical anomalies, Hout said.
"For the sake of all future elections involving e-voting systems, someone must investigate and explain the statistical anomaly we found in Florida," Hout said at a news conference today.
The researchers said they used a widely accepted method of study known as Multiple-Regression Analysis. It is a statistical technique widely used in the social and physical sciences to distinguish the individual effects of many variables, which in this case included number of voters, median income, Hispanic population, change in voter turnout between 2000 and 2004, support for President Bush in the 2000 election and support for Republican candidate Bob Dole in 1996.
"No matter how many factors and variables we took into consideration, the significant correlation in the votes for President Bush and electronic voting cannot be explained," said Hout. "The study shows that a county's use of electronic voting resulted in a disproportionate increase in votes for President Bush. There is just a trivial probability of evidence like this appearing in a population where the true difference is zero -- less than one in a thousand chances."
Hout, who describes himself as a nontechnical statistical researcher who has long been a skeptic of the criticisms levied against electronic voting machines, said he's "always taken a show-me approach to the theories of problems [with e-voting systems]." But when he saw the results of this study, "that's when I converted from skeptic [to believer]. I have concluded that something went awry with electronic voting in Florida."
The researchers also studied electronic voting results in Ohio, which Bush also won, but found no problems there, said Hout.
In an effort to explain what might account for the statistical irregularities related to counties that used touch-screen e-voting systems instead of optical scanning systems, Hout said there could be embedded software glitches or other potential hardware problems as were reported on election day in the press.
"We have no political ax to grind," said Hout. "We're interesting in leaving no vote behind."