Are Crime Labs really unbiased?

This article from the whyfiles shows more evidence that crime laboratories feel pressure to "find" the guilty party. Can we really trust DNA evidence to be accurate when there is pressure to have the forensic evidence match?
http://whyfiles.org/133crime_lab/5.html
Don't give that girl a gunBehind the debate over the proper role of crime labs is this question: Are labs tools of the prosecution, or of justice? Certainly, crime labs are touted as impartial, scientific means of finding the truth. But the truth is that most exist in a prosecutorial milieu -- as branches of the department of justice or the state police.
Geurts, of the Wisconsin lab, acknowledges that he works for the Department of Justice, but says, "Once you become an advocate, you lose your status as a scientist. We try to avoid that, but in general we work for the prosecution, and there's a small chance an individual analyst could become more prosecution oriented. In general, we get equal satisfaction from eliminating someone from suspicion, even though police usually respond, 'How can that be? We know this is the guy.'"
Evidence expert Jennifer Mnookin sees merit in the proposal -- raised in Oklahoma among other places -- to make labs independent of the prosecution. Although it would not solve everything, she adds, "Closely tying the forensic scientist to the prosecutorial apparatus may increase the likelihood of some of these problems. People may feel that to get promoted, they need to provide the evidence that makes a conviction; they may turn into cops in lab coats."
Mnookin describes the challenge thusly: "How do you create an incentive structure that does not encourage scientists and technicians to err on the side of the prosecution?"
Somebody bad stole de wedding bell We've heard other suggestions for improving labs:
Faigman suggests creating a "firewall" to reduce contact with police." Examinations should be as blind as possible, he says, "so the officer does not walk in and say, 'We found the guy who has these prints with the stereo equipment, and we want you to match them.'" The expectation of guilt, he says, "will intrude on the analysis."
Confirm the theory with basic scientific techniques -- and confirm the examiner's practice with blind tests. The Wisconsin State Crime Lab, says Geurts, tests employee proficiency at least annually. When drug examiners test proficiency samples, he says, they think they are working on actual cases.
Use peer review. Geurts says that having more than one examiner handle subjective analyses, such as handwriting, can reduce errors.
Subterranean homesick blues A final means to make sure that crime labs do their work honestly and intelligently is accreditation from the American Society of Crime Lab Directors. To date, 196 U.S. crime labs have attained certification, that's about half of the full-function scientific crime labs in the United States. That's the estimate of Robert Conley, director of the American Society of Crime Lab Directors Accreditation Board -- a separate group established by the Society. Accreditation, he says, requires "100 percent compliance" to a long list of standards affecting the "quality of the work product."
Covered areas include:
Minimum education for various disciplines -- typically at least a bachelor's degree -- with a concentration in science.
Annual proficiency testing for all employees.
Written procedures covering, according to Conley, "literally everything the lab does." Procedures must be validated in the lab, so "the method works in their lab, with their equipment and their personnel."
The standards are tough enough, he says, that almost all labs fail the first inspection. They then have a year to get up to snuff.
Could certification have helped at the Oklahoma City police lab, whose misconduct apparently doomed Jeffrey Pierce to spend 15 years in prison for another person's crime? Perhaps, Conley says, noting that the lab was not accredited.
"Without being specific or concluding about what happened in Oklahoma, because I only read about it in the papers, our process requires external review of all procedures used in the lab. If we were looking at microscopic analysis of hair, certainly the kind of results we have been hearing about would not be acceptable scientific practice."
However, accreditation does not address the issue of independence of the prosecution, or the overall question raised by the Supreme Court -- whether the forensic sciences are actually science.
While DNA profiling has been proven to be science after years of hard work and intense argument, that remains the exception. As Professor Faigman says, "Courts are wondering, scratching their heads, about why the other forensic sciences don't look like DNA profiling."