Is forensic science “under attack?”

Deputy Attorney General Rod Rosenstein said so in a speech at the National Symposium on Forensic Science back in August, saying that forensic science is not only quantitative and automated—that it can involve human assumptions, and subjective expertise honed through years of observation and study.

But a new paper by the immediate past president of the American Academy of Forensic Sciences in Forensic Science International asserts that challenging “routine” forensic methods should be done constantly, to get at the right answers. Challenges are not “attacks,” it adds.

“Regularly facing Frye and Daubert motions is not ‘being under attack;’ it is necessary for the mission of the justice system,” writes Betty Layne DesPortes, a Virginia-based criminal defense attorney.

Rosenstein’s notable speech was made on Aug. 7 at the Symposium meeting. Speaking to forensic experts in the crowd, he cited some of the “observation, comparison, interpretation and judgment” aspects of evidence gathering.

At one point, Rosenstein said, bluntly: “Some critics would like to see forensic evidence excluded from state federal courtrooms.”

DesPortes, whose year as AAFS president ended in February, drafted her letter and submitted it to the journal about a week later.

She balks at the characterization of the forensic science critics “demonizing” disciplines and experts, just because they want to question how reliable it is.

The fundamental misunderstanding, she writes, is over the use of the words “scientific” and “admissible”—and how they pertain to the Federal Rule of Evidence 702. The Rule admits some non-scientific expert testimony—so conflating admissibility with scientific validity is counterintuitive, DesPortes writes.

“DAG Rosenstein’s assertion that pattern evidence methods are ‘scientific methods’ by citing the federal rule on the admission of expert testimony is like asserting cupcakes are healthy because the Centers for Disease Control (recommends) a variety of food groups for a balanced diet,” she asserts. “We ‘critics’ just do not believe that something is reliable simply because law enforcement has used it for years or because a judge has admitted it in the past.”

Rosenstein’s prime example in his speech of improving the reliability of pattern evidence came from building a massive databases of automotive paints, including automated and quantitative methods of determined color and chemical analysis by make and model and year, using chemical methods and standardized instrumentation and protocols. The deputy AG held that up as paradigm of reliability and forensic improvement. DesPortes posits that the same can be done for other pattern evidence methods.

“We seek the scientific testing of the method to prove the assurance of its reliability,” she writes. “We do not want to exclude science from the courtroom; we want more science in the courtroom.”

This debate, of course, has been raging for almost a decade, from police departments and crime laboratories, to courtrooms. The watershed 2009 National Academies of Science Report entitled “Strengthening Forensic Science in the United States: a Path Forward” sparked a reevaluation of a wide swath of forensic disciplines, especially hair analysis, bite marks and certain impression evidence—but also even long-accepted fingerprint “matching” and the relatively-new interpretation of complex DNA mixtures.

The contention of Rosenstein and some forensic experts: sciences such as fingerprints—which are based on sometimes-subjective “calls” made by a trained expert—should not be thrown out just because there is not hard quantitative data to accompany courtroom testimony. Indeed, Rosenstein and the Department of Justice have rolled out a series of 13 “uniform language” guidelines to attempt to put all examiners on the same standardized playing field—a group including fingerprint testimony. Latent fingerprint identification is one of the disciplines that has proven to be consistently accurate, despite some criticisms that it is not quantitative. A critical paper published in 2005 identified 22 fingerprint misattributions internationally over the first century of the use of forensic fingerprint comparisons. Ten of those resulted in convictions that were later overturned. None of those convictions were reached after 2000.