Judge skeptical of claims that Rossi got more votes

WENATCHEE – The judge presiding over the trial to decide who won the 2004 governor’s election expressed some skepticism Thursday about studies that show Republican Dino Rossi really got more valid votes than winner Christine Gregoire.

But Chelan County Superior Court Judge John Bridges refused Thursday to keep those studies out of the case, saying he wants “to make sure as much evidence as possible is available” for the state Supreme Court, which everyone expects will get the case after he issues his ruling as early as next week.

Whether he’ll use the studies to determine that the vote totals should change enough to overturn Gregoire’s 129-vote victory margin, remains to be seen.

“I have some concerns, based on what I heard today, but (both sides) should go ahead and put on their cases,” he said.

Bridges’ ruling came after Democrats and Republicans spent much of the day playing a courtroom version of “my expert is better than your expert” when it comes to analyzing and manipulating data on voters who cast ballots they weren’t legally entitled to cast.

A pair of experts for the state Republican Party looked at some information on felons who voted even though they didn’t have their rights restored, plus a smattering of other allegedly illegal votes, and concluded that Rossi actually beat Gregoire.

That would bolster the case for the Republican Party, which wants Bridges to order a new election for governor.

Jonathan Katz, a professor of political science at California Institute of Technology, explained his method for dividing some 1,183 “invalid” voters among Gregoire, Rossi, Libertarian Ruth Bennett and other categories like write-in candidates and ballots that weren’t marked in that race. Under his formula, Gregoire would lose between 250 and 300 more votes than Rossi, and the Republican would wind up ahead.

Anthony Gill, a professor of political science at University of Washington, said he did a different type of analysis with the same data supplied by the GOP, and concluded that it “may have resulted” in Rossi winning by 62 votes.

Both Katz and Gill said they couldn’t vouch personally for the data, which continues to change as both parties discover new examples of possibly illegal votes and cross suspected illegal votes from old lists.

Democrats contend the list of possible illegal voters the Republicans gave their experts is heavily weighted toward King County, which had the highest margin of victory for Gregoire, and ignores much of the pro-Rossi areas of the state.

When Democratic attorney David Berman asked if the data in the report was skewed, Gill said he had no way of knowing but agreed “the results that you get will depend on the data.”

Even before Katz could discuss his report, he was grilled by Berman on whether he used proper scientific methods.

The analysis assumes that the illegal voters chose their candidates in the same proportions as legal voters, Berman noted. “You can’t say for certain how any illegal voters voted in the governor’s race, or even if they voted?” he asked.

“Given the secret ballot, no,” Katz replied. “This is the best method, given the limitations of the data. The method is always subject to the data.”

Katz insisted it was possible to use his formula to determine who would have won if any number of illegal votes were subtracted from the statewide totals.

In a special hearing to challenge the validity of the Republicans’ science, Democrats called statisticians of their own to say the GOP analysis wasn’t an acceptable way to determine if illegal votes changed the outcome.

Christopher Adolph, a professor of political science and statistics at University of Washington, said the Republicans’ studies are “badly flawed” because they make a broad assumption that felons would vote the same way non-felons vote in a particular precinct or neighborhood.

That would be like assuming felons who illegally registered to vote matched legal voters on another feature, such as gender, Adolph said. But in a demonstration precinct he checked, male and female voters were about equally divided, but illegally registered felons were about 75 percent male.

“It’s really possible for those people to have voted every which way,” he said.

But GOP attorney Mark Braden pointed out that Adolph was an untenured professor who has never been an expert witness in a trial, while Katz is “one of the most eminent political scientists in the country.”

Finding a way to subtract hundreds of illegal votes from the 2004 vote totals and determine who really won the governor’s election might be a “hopeless case,” Adolph said. But it might be possible to get valid survey data that would allow a statistician to reach that conclusion.

How would one conduct a survey of people who may have committed a crime by casting an illegal ballot and determine how they vote, Braden asked.

“I would suggest simply asking them. I think it would be worth trying,” said Adolph, who later conceded he had never actually conducted a survey.

Mark Hancock, another expert called by the Democrats, said the data used by the GOP’s experts don’t meet scientific standards and “if science says you cannot rely on that (data), science shouldn’t be invoked.”

Hancock is a statistician and social scientist at the University of Washington. Under questioning from Braden, he said he wasn’t a political scientist, and has never appeared as an expert witness in an elections case. But Berman quickly followed up with questions that pointed out he was a social scientist, and an expert in sampling and other types of statistical analysis.

At one point, Braden and Hancock went back and forth over a scientific conclusion on who won the election could be reached, if the court decided which votes to remove. The GOP eventually may ask Bridges to come to that conclusion on illegal ballots, and attempt to plug them into their experts’ formulas.

Hancock argued that wouldn’t necessarily be good science. No matter what way Braden asked the question, Hancock wouldn’t budge.

Finally, Bridges broke in: “We have a lawyer talking, and a scientist answering, and that’s a problem.”