DURHAM — To win a new trial, Durham novelist Michael Peterson and his lawyers are arguing that former SBI agent Duane Deaver tailored his expert witness testimony not to the facts or science of a case, but to what prosecutors wanted him to say.

In court Wednesday, Peterson found support in an unexpected place: the SBI's own files show Deaver's supervisors knew early on that Deaver had a strong pro-prosecution bias.

In 1988, early in his SBI career, Deaver and Michael Budzynski, a fellow scientist at the State Bureau of Investigation crime lab, went through a mock-court training session. The two agents reviewed a set of facts and testified as if they were on the stand and were subjected to the questioning they would receive in court from prosecutors and defense attorneys.

Their supervisor, Mark Nelson, noted one of Budzynski's strengths was "his impartiality as a witness."

In contrast, Nelson noted that Deaver had "a strong bias toward the prosecution."

Deaver was a key expert witness at Peterson's 2003 trial, where the one-time mayoral candidate was convicted of killing his wife, Kathleen Peterson, a Nortel Networks executive who died at the foot of a blood-spattered staircase in the Petersons' Forest Hills mansion in December 2001. As an expert witness at the trial, Deaver was charged with interpreting facts in an unbiased and scientific fashion that favored neither the prosecution nor the defense. Deaver analyzed the bloodstains and said at trial that Kathleen Peterson was murdered and her husband was the culprit.

In court Wednesday, the former director of Connecticut's Forensic Science Laboratory testified that he was troubled by the fact that Deaver's supervisor noted a pro-prosecution bias.

"If an individual has a strong prosecution bias, they can't be objective, they have a horse in the race," said Timothy Palmbach, testifying on Peterson's behalf. "The expert coming into a courtroom shouldn't care about the results, whether it's guilt or innocence."

Lax methodology

Palmbach, a forensic science professor and former state trooper, testified as an expert for Peterson at his 2003 trial. Before his appearance Wednesday, Palmbach reviewed Deaver's seven days of testimony in the Peterson trial, as well as Deaver's notes, reports and experiments.

Palmbach said Deaver's work in the Peterson case failed to meet the basic rules governing high-school science. He said that Deaver didn't document all his work, didn't explain his methodology and did not try to test every possible and competing hypothesis.

Scientists try to disprove a hypothesis to eliminate faulty explanations, and only come to conclusions after trying to eliminate competing explanations, Palmbach said.

Deaver, on the other hand, formed his conclusions in the Peterson case and then designed experiments to back them up, Palmbach said.

For example, Deaver never conducted an experiment to test whether the bloodstain patterns could have come from an accidental fall.

"It's irresponsible that once the alternative theory is out there, an accidental fall, not to have tested that," Palmbach said. "One of the dangers of testing is to conduct one test, get the answer you like, and stop."

Deaver failed to read simple blood spatters correctly and used terminology that had been discredited decades ago, Palmbach said. Deaver's resume and testimony showed he failed to keep up on scientific advances in the field; he attended no national conferences, did not belong to professional organizations and had no outside training since the 1980s. "That's irresponsible. As an expert you always have to expand your knowledge base and expose your work to publication and peer review."

The Peterson hearing, which continues today, is before Superior Court Judge Orlando Hudson, who presided at Peterson's 2003 trial and allowed Deaver to testify as an expert witness. Hudson will have to decide whether he would have allowed Deaver to testify as an expert given the information introduced at this hearing.

Neff: 919-829-4516