RALEIGH, N.C. (AP) -- The accreditation group that certifies North Carolina's top law enforcement agency says the agency should have made it clear in crime lab reports from the early 1990s when tests of evidence produced conflicting results, appearing to contradict the agency's position that they were scientifically acceptable.

"We have always expected that reports would be clear and revealing of anything that should be revealed," Ralph Keaton, executive director of the accreditation group, said Tuesday.

The director of the State Bureau of Investigation, Robin Pendergraft, has said agents were told to use certain language in the lab reports submitted to courts, and she defended its use at the time. She reiterated her stance Tuesday.

"Back in the '90s, they complied -- as I understand it and I don't know, I haven't gone and researched it -- with the statutory requirements," she said. " ... The language was acceptable in the scientific community, as has been explained to me."

The issue arose in the case of Greg Taylor, who served more than 16 years behind bars for the murder of a prostitute in Raleigh before he was exonerated two weeks ago in a groundbreaking innocence hearing.

During the hearing, SBI Agent Duane Deaver testified that agents were told to write in lab reports that evidence gave chemical indications for the presence of blood when the first test for blood came back positive. He said agents were told to use that language even when a follow-up test came back negative.

The information about a negative follow-up test would be part of an agent's bench notes, but wasn't included in the lab reports routinely provided to courts, Deaver said.

Although standards have changed since the society began accrediting crime labs in 1982, the type of language used by SBI was never acceptable, said Keaton, executive director of the American Society of Crime Lab Directors/Laboratory Accreditation Board, which certifies SBI's lab and hundreds of others.

"There should have been a qualification in the report that you might have gotten screening tests that were positive, but the report should have clarified that confirmatory tests were negative," he said.

One reason is that the average lay person, including a juror, "doesn't necessarily understand the significance of any test," he said.

Keaton emphasized that the society has not communicated with the SBI, and he based his comments on what he had read about the case.

Pendergraft said she didn't know whether the directive for the wording in lab reports was in writing or was part of SBI culture at the time. She also said she didn't know when it ended, although she said that question is part of a review.

The society, which accredits more than 350 labs, first accredited the SBI lab in 1988 and has reaccredited it every five years since. The society would have examined bench notes and lab reports as part of accreditation, but didn't find the differences between the two, Keaton said.

He declined to say whether the society would re-inspect the SBI lab. Such a decision wouldn't be made public, he said. He didn't immediately return a phone call to answer questions about whether the society would re-examine its accreditation procedures because it didn't catch the differences between SBI lab reports and bench notes.

Keaton's comments are not the first time the SBI's choice of words has been questioned.

In a federal court order issued last year in the case of George Goode, a federal judge said Deaver's testimony was faulty because he led a 1993 jury to believe that he found blood, when he had only conducted a preliminary test that indicated the possibility it was present.

"A reasonable consideration of the record demonstrates the State, through Agent Deaver, presented misleading evidence about the testing done on petitioner's boots being conclusive for the presence of blood," U.S. District Judge Malcolm Howard wrote in his order, which vacated Goode's death sentence because of ineffective counsel at trial.

___

On the Net:

http://www.ncdoj.gov/About-DOJ/State-Bureau-of-Investigation/Crime-Lab.aspx

http://www.ascld-lab.org/dual/indexdual.html