USDOJ should embrace, not fight science-based forensics

Now that Eric Holder, President Obama's appointee to lead the Department of Justice, has been confirmed as US Attorney General, one of his first orders of business should be removing all obstacles to the release of a new National Academy of Sciences report analyzing the scientific basis (or lack thereof) for techniques used by forensic crime lab analysts. The FBI has been fighting the report's release, the New York Times reported this week, but Holder should insist they embrace it.

Justice has nothing to fear from real science, but too much so-called forensic "science" amounts to little more than folklore and wives' tales, particularly disciplines requiring comparative judgments by individuals as opposed to science-based analysis. Notes Rick Casey in the Houston Chronicle:

Ironically, our awareness of the ineptitude of traditional police “science” is largely the result of the greatest advance in forensic science, DNA matching.

One of the studies considered by the academy panel examined the trial transcripts of 137 cases in which DNA showed that a suspect had been wrongly convicted.

According to Peter Neufeld, co-director of the Innocence Project, and University of Virginia Law professor Brandon Garrett, who conducted the study, in 60 percent of the cases false or misleading analysis was presented regarding such things as hair, blood, fingerprints, fiber or soil.

The study is scheduled to be published next month in the Virginia Law Review, but its findings in an amicus brief filed with the U.S. Supreme Court were described by Justice Stephen Breyer as being “filled with horror stories” of erroneous conclusions by police labs.

Such misjudgments can even occur in disciplines long considered gold-standard evidence, like fingerprint matching. According to the Baltimore Sun:

When British researchers asked five crime lab examiners to evaluate a series of fingerprints, they were told one pair had been mistakenly matched to a terrorism suspect. The experts reached conflicting results. Only one judged the prints identical. The fingerprint examiners later learned that the samples were prints they each had previously reviewed and found to be the same. The study by Itiel E. Dror and two colleagues underscores what some defense attorneys in Maryland and elsewhere have argued - forensic experts can be influenced, and not in justice's favor.

That example shows how researcher bias can influence results, rendering analysts subject to swaying their findings, perhaps subconsciously, to match the expectations of police and prosecutors.

One of the most telling instances of real science debunking "junk" forensics used to send thousands of people to prison arises in arson cases, where more than a decade after old investigative techniques were overturned by actual scientific experimentation, some arson investigators (who are typically cops or firefighters, not scientists) are still using flawed analyses and courts are still using that testimony to convict people.

An excellent piece this week by Sue Russell at MillerMcCune.com gives a good backgrounder on the controversy over invalid arson forensics and details specific, debunked "arson indicators" that have long been relied upon by the courts with no scientific basis at all. According to Russell:

several key players say that finally divesting the shrinking faction of holdouts of their misguided faith in arson indicators like "crazed glass," "alligator blisters" and "concrete spalling" (see sidebar) remains a work in progress.

"I think it certainly is still a problem; I don't think it's as widespread as it used to be," said Bobby Schaal, a 22-year veteran of the ATF speaking only in his capacity as incoming first vice president of the International Association of Arson Investigators. But investigators, lawyers and scientists say that old investigation errors are still being made, and old myths are still being heard in courtrooms.

Common themes emerge when discussing reforms. One hot topic is the routine dependence on negative corpus evidence —simply put, investigators rule out electrical faults and exploding coffee pots, for example, rather than rule in evidence of how a fire did in fact start. So rather than a more accurate description of "cause undetermined," fires are often called arsons based on investigation by exclusion.

To veteran investigator Patrick Kennedy, that practice is unethical and immoral. "I don't know what it is, so it must be arson?" he said. "That is a pretty poor reason to kill somebody."

Kennedy's reference is to a Texas case - that of Cameron Todd Willingham, who was executed in 2004 based on now-debunked arson forensics. The state's Forensic Science Commission is re-investigating that case to determine if bogus arson science led to execution of an innocent man.

I'd expect a similar reaction from crime lab analysts to the National Academy of Sciences report as we saw among arson investigators to scientific advancements in their field - defensiveness, anger, and denial. But AG Eric Holder could move forward the process a lot faster by embracing the NAS report and insisting that FBI analysts change the way they do business to accord with science-based methods instead of subjective judgments. New best practices would be implemented much more rapidly by state and local crime labs if DOJ would lead the way instead of attempting to thwart actual, science-based forensics.

See related Grits posts:

• Arson cases may be next venue for innocence claims
• Many arson convictions based on invalid science
• College developing screening processes to vet old arson cases for junk science
• Are crime lab reports 'testimonial' documents? SCOTUS will decide
• CSI-Houston would be dark comedy

MORE: See commentary from Radley Balko at The Agitator.