Third of four parts
At the State Bureau of Investigation, most forensic scientists are cops. So are their bosses. Their bosses' bosses are prosecutors, the chain stopping at Attorney General Roy Cooper, the state's top lawyer.
This arrangement often forces analysts to become advocates, in lock step with police and prosecutors, shaping evidence needed to deliver a conviction.
Inside the lab, which produces forensic science analysis for police and sheriffs, analysts push past the accepted bounds of science, a News & Observer investigation reveals. The problem is prevalent, a review of lab policy, procedures and questionable decisions in at least a dozen cases shows.
Never miss a local story.
The work of the SBI crime lab has been under fire since February, when Greg Taylor, an innocent man, was freed after judges learned an SBI serologist withheld crucial evidence that proved a stain on Taylor's SUV wasn't blood.
Problems at the lab run deeper than blood. State law puts scientists at the lab on the prosecution's team, instead of assigning them as independent seekers of fact. Analysts sometimes don't run DNA or blood tests that might threaten prosecutors' theories. And they shield themselves from scrutiny, fighting against turning over records and forbidding defense experts from observing their work.
Heather Coyle is a DNA expert who worked in Connecticut's crime lab before taking a professorship at the University of New Haven. She started raising questions about the SBI after being hired by defendants to review the lab's work.
"Science tells us 'when in doubt, let go,' " Coyle said. "The SBI's philosophy is 'when in doubt, convict.' What they do is hardly science."
Such bias is not surprising, given that the lab is controlled by a law enforcement agency; North Carolina is one of 38 states where the lab reports to police or prosecutors. In 2009, the nation's most renowned scientists warned against such a setup in a wide-ranging study that included dozens of top lawyers, professors and judges.
One of their top recommendations: Forensic scientists shouldn't report to law enforcement or prosecutors.
"The best science is conducted in a scientific setting as opposed to a law enforcement setting," said the report from the National Academies, which advises on science, medicine and engineering. "Forensic science serves more than just law enforcement; and when it does serve law enforcement, it must be equally available to law enforcement officers, prosecutors and defendants in the criminal justice system."
In North Carolina, defense attorneys, citing prejudice and sloppiness by SBI analysts, often seek their own costly experts to examine evidence. Many independent experts, like Coyle, find grave mistakes. The cost of that extra review often falls to taxpayers.
The lab, with its primary location in Garner, costs taxpayers $12.7 million a year and employs 181.
Attorney General Cooper, a Democrat, said that the lab is where it needs to be and that policies can safeguard against bias. He said removing the lab from his control would make it more difficult for some lab units to get critical law enforcement data. "You don't want to hobble law enforcement by removing key tools such as technology to prevent them from solving crime," Cooper said.
Meanwhile, juries have sent some defendants to prison with slanted forensic reports. The wrongly convicted, such as Taylor, wonder what they did to deserve prejudice from the state's scientists.
"My hope would have been that the SBI just took a personal vendetta against Greg Taylor and tried to do this to me and nobody else," Taylor said. "Deep down, I kind of felt this wasn't the case. I didn't want it [to happen] to 100 people or 10 people or two people. I didn't want anyone else to go through this."
For every case the SBI works, police deliver a set of clues along with boxes and envelopes of evidence from the crime scene. They offer a story of sorts, a set of assumptions made early in an investigation.
It's a faulty arrangement, according to the National Academies report, one that colors which tests are performed and their results. The dialogue between police and SBI lab analysts continues for months.
"We encourage our analysts to call local law enforcement and say, 'Tell us what you got, what's going on with the investigation,' " said Jerry Richardson, director of the lab.
Analysts are expected to log calls with police and prosecutors. Sometimes those calls don't get noted. Memories of information swapped get foggy.
In 2009, SBI analyst Russell Holley was called to Durham during a pretrial hearing to explain why he chose not to test for blood on several pieces of evidence collected at a convenience market where a clerk had been slain. Holley had not even opened the packages in which Durham police had shipped the blood samples. Without his test to certify the samples were likely blood, analysts wouldn't perform DNA tests to determine whose blood it might be.
Lisa Williams, who represented the defendant in the case, said the blood might have pointed to another assailant and wanted to know the results. She found it curious that Holley chose not to test the evidence.
During a 2009 hearing, Williams asked whether Holley had talked to a Durham police detective, according to a transcript. Holley acknowledged he had talked with Durham police but couldn't remember what they had discussed. Holley did not log the details of those conversations, as required.
"I'll never know if he was advised to not look into that evidence," Williams said. "You can't help but wonder."
A jury later convicted the defendant, Keith Kidwell, of first-degree murder.
Cooper defended communication between law enforcement and analysts, saying that he expects them to stay in contact to better solve crimes.
Analysts are told, however, that they must let prosecutors know before speaking with defense attorneys.
SBI examiners are delighted when they can deliver what prosecutors seek.
Durham prosecutors turned to the SBI to try to replicate the blood patterns found in the stairwell where Kathleen Peterson was found dead in 2001. Back at the lab, analysts tried to figure how Michael Peterson had pushed his wife and beaten her.
Finally, analysts nailed it, finding an experiment that helped them see a blood pattern similar to that on Peterson's shorts. The two analysts high-fived, a film of the experiment shows.
It's important that prosecutors think highly of analysts. They depend on prosecutors to provide favorable feedback on their courtroom testimony as part of a certification requirement for the lab.
ASCLD-LAB, an independent group that works to ensure proper protocols and practices at crime labs, requires that analysts who testify in court be observed once a year, either by supervisors, fellow agents or by court officials.
This requirement does not cover field service examiners, including bloodstain pattern analysts such as Duane Deaver and Gerald Thomas, whose tests in a Davie County murder trial drew ridicule. No supervisor is required to monitor their court testimony, and problems have gone unaddressed for years.
The majority of the time, examiners turn to prosecutors to provide feedback on their testimony.
In hundreds of feedback forms reviewed by The N&O, prosecutors offered glowing responses. They checked boxes certifying the analyst spoke clearly and remained calm. The prosecutors also affirmed that the analyst's testimony was reasonable and within the bounds of accepted science.
Ann Kirby, a former Johnston County prosecutor who now works in the Craven County District Attorney's Office, offered her praise of drug analyst Lisa Edwards in 2003.
"If Lisa Edwards gets any better on the witness stand," Kirby wrote, "the Johnston County defense bar is going to try and have her banned from the county!"
Demonizing the defense
To SBI analysts, defense attorneys are often the bad guys. Training manuals and directives paint defense attorneys as tricksters who are driven to let criminals go free.
"Tell the D.A. in advance of any weaknesses in the case so that the trial of the case can be planned to minimize the weaknesses' impact," says a 2007 manual used to teach analysts how to testify in court.
The N&O showed Cooper the training manual on July 29.
"There's some language in this document that should not be there, and there should be a thorough review of it," he said.
Robin Pendergraft, the former SBI director, said she had never seen the manual when The N&O asked about it. She said she found it interesting.
Cooper replaced her shortly after her interview. On Wednesday, new SBI Director Greg McLeod suspended use of the manual.
Until recent years, analysts were discouraged from speaking with defense attorneys about a case before trial. Now, analysts are advised to speak with defense attorneys only after notifying the prosecutor, so that he or she may choose to be present for the conversation.
Greenville lawyer Ernest "Buddy" Conner experienced the alignment with prosecutors in 2004 when he called to question an analyst about a lab report he suspected was faulty.
Conner, a veteran defense attorney, was certain his client, Leslie Lincoln, had nothing to do with her mother's killing. He was sure the SBI lab report would confirm that. Instead, it locked on Lincoln as the suspect. A bloodstain found near her mother's bed bore Lincoln's DNA profile. It was a match so sure that statistically, it could belong to no one but Lincoln.
Conner couldn't believe the results and called the lab to ask the analyst to double-check her work.
Conner said he was told that the lab wouldn't rerun the test at his request. Only if a prosecutor or judge demanded it would the lab repeat the test, Conner said he was told.
Conner said he was so flabbergasted by the resistance that when he went to court, he asked the judge to order that the evidence be sent to an independent lab for a retest.
Only then did the SBI analyze the tests further.
When it did, analysts found an egregious error: The analyst had swapped the victim's sample with Lincoln's. The DNA at the crime scene was her mother's, not hers.
The analyst, Brenda Bissette, was removed from lab duties, records show. She soon resigned.
Lincoln, who was acquitted, has filed a lawsuit against Bissette and local police and prosecutors. Bissette declined to comment.
"I will never understand why they didn't jump at the request to make sure they were right," Conner said. "What would it have cost them?"
Trouble with DNA
DNA is an exact science. Every person has a unique profile that belongs only to him or her.
But at crime scenes, perfect samples can be hard to find. In sexual assaults, the victim's profile has mixed with the rapist's. Often, the evidence is degraded, and scientists can only get a good look at a fraction of the 16 unique identifiers that make up one's DNA profile.
Crime labs adopt their own protocol on what constitutes a match between a suspect and a mixed or partial sample.
At the SBI's crime lab, the threshold is low.
"We are not here to establish guidelines on just how many is enough," said Mike Budzynski, a veteran analyst who is in charge of the DNA section. "There are no minimum standards as far as I'm concerned."
According to an administrative order, seeing a suspect's profile at one of 16 unique identifiers in the mixture is enough.
Such standards are low, and far below other labs' practices, said Coyle, the DNA expert from Connecticut.
"It is highly irresponsible to make any conclusions with this level of evidence," Coyle said. "It's highly circumstantial. The benefit of the doubt should always be given to the defendant."
Most labs, Coyle said, won't consider anything a viable match unless three or four of the identifiers are consistent between the mixed crime scene sample and the suspect's profile.
Budzynski said it's up to prosecutors to determine how much weight to give the DNA evidence they offer.
The only thing that linked Johnny Maness to the rape and slaying of Ruthie Morgan, an elderly hog farmer in Moore County, was a match to one of 16 locations on a mixed DNA profile extracted from her body.
As DNA goes, these are extremely low odds. It means that one in 25 white people in North Carolina would have had just as much likelihood of being linked to the semen mixture found in Morgan. Later, further analysis with a new statistical formula set the odds to one in 98 white people; a different sample showed odds of one in 30,000.
Maness' lawyer said her expert disputed that later analysis.
Maness swore that he was innocent but pleaded guilty to second-degree murder in 2009 to avoid the death penalty. Police had found Maness' DNA on a cigarette butt in the victim's truck, which was not linked to her murder. Maness had worked odd jobs for Morgan and admitted to having sat and smoked in her truck previously.
Before taking the deal, Kim Stevens, Maness' attorney, asked the district attorney to request the SBI to run the male portion of the mixture profile in the state's DNA database. She wanted to see if analysts could find another suspect who matched with higher certainty than Maness.
Peter Strickland, assistant district attorney in Moore County, said he asked the SBI to run the test. He doesn't recall getting a response.
Amanda Thompson, the analyst in the case, said that she asked Strickland for a written request from the defense attorney before she performed further work. Before that happened, Thompson said that Maness agreed to plead guilty and that she never reviewed the case further.
Stevens, an assistant capital defender in Winston-Salem who represented Maness, said she never knew she needed to write another letter. She is befuddled by any hesitation or resistance to perform extra analysis.
"I don't know why they wouldn't want to be absolutely sure," Stevens said. "They were willing to take a man's life over this."
Defense fights back
Over the past decade, defense attorneys have grown increasingly wary of reports generated by the SBI. They hire their own experts to double-check evidence they don't have the expertise to understand.
The SBI has met this trend with resistance, according to training manuals, directives and court testimony.
A 2008 memo issued to all analysts advises that canceled versions of lab reports are not considered part of the packet of material turned over to defense attorneys. That means that any changes in the reports before they are sent to prosecutors and defense teams are not disclosed - unless they are specifically requested.
The lab has also resisted turning over other material that defense attorneys say is crucial to helping represent their clients. Examiners don't like to be challenged. And their bosses don't like them being questioned or observed.
They are wary of defense experts. Included in analysts' training manual is a memo prepared for prosecutors across the state by Michael Parker, now district attorney in Stanly County.
In the memo, Parker warns against "defense whores." Parker goes on to promise prosecutors that the SBI will vet the defense expert, finding background information on him for prosecutors.
The SBI's attorney sometimes goes to court to fight orders compelling that information be provided to defendants.
John Watters, a lawyer at the state Department of Justice who has represented the SBI since 1993, fights requests for information not specifically listed in the discovery law that ensures that defense attorneys have access to investigative reports.
Cooper said he will ask his new SBI director to review the practice and will likely assign more lawyers to assist the SBI with such matters.
Watters is particularly resistant to defense experts observing tests in the crime lab. The philosophy was put into policy in 2009, banning any observers from the lab.
Jerry Richardson, head of the lab, defended the policy, saying analysts needed to work in solitude so they won't be swayed by observers.
But SBI officials even resist observation when the test would consume all of the evidence, leaving none for defense tests.
In May 2009, Watters argued at a hearing that the defense's scientists could contaminate the lab. He also insisted that the lab isn't equipped to handle a rush of outside scientists observing their case work.
"I'm telling you that the thing that concerns us most is the precedent this would set, and the potential for harm," Watters said. "I think if you look at the discovery statute, since their inception, we've never been a testing laboratory for the defendant. We are the state's laboratory."
News researcher Brooke Cain and database editor David Raynor contributed to this report.