UC Berkeley Research Team: "Someone Must Investigate and Explain These Anomalies"

NOV 16 - 30, 2004

VOL. 2 ISSUE 22

UC BERKELEY RESEARCH TEAM: “Someone Must Investigate and Explain These Anomalies” With Florida E-Voting That Gave Between 130,000 and 250,000 Extra Votes To President Bush

Problem Only In Largely Democratic E-Voting Counties, “99.9 Percent” Certain Error Of E-Vote Tally Responsible

by Betsy R. Vasquez

NOVEMBER 18, 2004 – A University of California at Berkeley research team told a press conference today that a study they have done of the e-voting in Florida shows there is a statistical anomaly in 3 heavily Democratic leaning counties which gave President Bush between 130,000 and 250,000 more votes than he should have received.

“Someone must investigate and explain these anomalies,” said Professor Michael Hout, a Sociology professor and head of the UC Berkeley research team that issued the report.

At issue are Broward, Palm Beach, and Miami Dade counties, listed “in order of magnitude” of the discrepancy.

The report indicates that, “In Broward County alone, President Bush appears to have received approximately 72,000 excess votes.”

Asked if there was any chance people in these counties simply decided to vote for President Bush in these excess numbers, Hout said there is “less than 1 in 1000” chance.

What happened in these counties, said Holt, “was out of pattern with what occurred in the other 64 counties in Florida.”

Even in other counties that used e-voting, but in which Bush support was stronger, there were not statistical anomalies found like the ones in the three large, heavily Democratic-leaning counties cited.

Hout described himself as a skeptic, and noted he has been acting to explain the “Dixiecrat” issue that caused a number of counties in Florida which have a higher number of Democrats to vote more heavily for Bush. He noted that these counties did not present as anomalies in this study.

But in Broward, Palm Beach, and Miami Dade, the results of the analysis showed so dramatically a problem that, “that’s why I’m not a skeptic now,” he said.

“Something went awry with the voting in Florida,” he said plainly.

What exactly the issue is Hout and his team left for others, repeatedly responding to questions on the topic by saying, “I am not an engineer or a programmer.” He said the errant vote counts could be due to embedded software glitches, smudged screens, or, in one case, to a reported count in which at some point the tallying reversed and started counting downwards. But he was clear to say that this list was not definitive as this was not his area of specialty.

“All I know is that the smoke alarm’s gone off, it’s up to the Fire Department now,” to figure out exactly what occurred, by which he meant election officials in the three mentioned counties.

Due to the “significant departure” from what vote count should have occurred in these counties, President Bush received a total of 1,157,435 votes while he should have received between 900,000 and 1,020,000.

The team said they studied both Florida and Ohio, but only spoke about the Florida results at this press conference. There is a possibility that they could also look into e-voting in other states, but, as one of the researchers noted, “We are busy grad students,” and so that might not happen immediately.

The purpose of this study was deal with the issue that there were not statistics evaluating potential e-voting problems. “Just rumors in the blogosphere,” said Hout. And so the team sought to both “shoot down false claims,” which Hout said they have done, as in the case of the Dixiecrat Florida counties, as well as confirm actual problems, such as the ones found in Broward, Palm Beach, and Miami Dade.

“We are not a political action committee, not a lobby, we are just a bunch of researchers who have something we think is important,” said Hout.

Exit polling data was not used for this project, as they would present a whole new group of statistical issues into the mix.

Again Hout repeated, “The smoke alarm is beeping, people need to see if there’s a fire.”

“We can be 99.9% sure that these effects are not attributable to chance,” the report says. “Compared to counties with paper ballots, counties with electronic voting machines were significantly more likely to show increases in support for President Bush between 2000 and 2004. This effect cannot be explained by differences between counties in income, number of voters, change in voter turnout, or size Hispanic/Latino population.”

“The data show with 99.0% certainty that a county’s use of electronic voting is associated with a disproportionate increase in votes for President Bush,” according to the report.