Boston Globe

Foolproof Forensics?

by Beth Daley
June 8, 2004


"The premise is interesting that scientific evidence is more reliable than other evidence. . . . It would be nice if it were true," said Simon A. Cole, an assistant professor of criminology, law and society at the University of California at Irvine. "In the cases of wrongful conviction that we know about, scientific evidence is a very significant factor."

Romney's plan, which would be applied only to particular first-degree murders such as killing a police officer or murders involving torture, does not require absolute scientific proof. Rather, it would require a jury to find evidence "reaching a high level of scientific certainty" that will "strongly corroborate the defendant's guilt." While DNA is the most ironclad evidence now available, other categories such as photographs, video and audiotapes, fin gerprints and tool marks may suffice. Multiple layers of review in the plan would ensure "as much as humanly possible" no innocent person be sentenced to death.

"We can't get to zero, but we can get close," said Joseph Hoffmann, a law professor at Indiana University who cochaired the panel that crafted the Romney plan.
Only a minority of murder cases have enough biological evidence to provide DNA, according to defense lawyers and crime experts. This means that the burden of proof could more often fall on far more subjective and much more controversial evidence, such as tire tracks or fingerprints. Though often presented as science by prosecutors and expert witnesses, such evidence is increasingly derided by defense lawyers and academics as an interpretive art.

"[Technicians] are actually told to develop this intuitive sense of certainty when they review fingerprint comparisons that they've obtained a match," said David Faigman, a University of California law professor who wrote "Laboratory of Justice: The Supreme Court's 200-year Struggle to Integrate Science And The Law." He said there are no required standards for fingerprint analysis, and labs often declare a match between two prints based on years of examining fingerprints rather than a clearly spelled-out methodology. "From a scientific standpoint," he said, "that is the voodoo part."
Faigman and other critics argue that science has a long and checkered history in the courtroom. Lawyers once used body characteristics, such the lengths of people's arms or shape of their heads, to prove a defendant's propensity to commit a crime. In 1927, a phrenologist was called into court to "read" a woman accused of murdering her husband; the phrenologist declared that the suspect's chin "tapered like the lower face of a cat," demonstrating "treachery."

As phrenology was being dismissed as quackery, the early 20th century saw the birth of forensic science as a specialized profession, with laboratories and experts who aimed to link suspects definitively to crime scenes. Eventually, handwriting, fingerprints, photographs and blood samples became regularly introduced into evidence, and the belief that "every criminal leaves a trace" became a cornerstone of police investigations.

By the late 1980s, DNA testing had been widely adopted, and today technology is still marching on: A new technique called "brain fingerprinting", a kind of lie detector based on brain signals, was admitted into court in Iowa in 2003 in order to help free a man in prison for murdering a retired police officer. (The man was freed by the Iowa Supreme Court, although the judges did not refer to the technology in their decision.)
If history is any lesson, however, today's certainty is tomorrow's question mark. For example, the rise of DNA testing has revealed enormous failings in the microscopic hair analysis that was considered reliable a generation ago. In 2002, DNA analysis helped free a Montana man who had spent 15 years in prison for rape based in large part on faulty expert analysis of pubic hair found at the crime scene. According to the Innocence Project, which works to free wrongfully convicted people, in 25 of the first 82 DNA exonerations around the country, scientists and prosecutors presented bad or tainted science to convict a defendant.

So concerned was the US Supreme Court about the growing role of science in the courtroom in the 1990s that the court instructed judges to act as gatekeepers for scientific evidence, scrutinizing experts and procedures to be sure scientific techniques were peer-reviewed or tested, with known and acceptable error rates.
That instruction led to the first major court decision questioning fingerprint evidence. Two years ago, a Philadelphia judge ruled an expert could not link fingerprints found at a crime scene to a defendant because the matching technique used by fingerprint experts had never been proven valid. There was no proof, the judge said, that fingerprint analysis had been scientifically tested or its error rates calculated. The judge later reversed his decision after the FBI testified about training, procedure and error rates, but the challenge opened up the floodgates for other defense attorneys protesting fingerprint analysis.

"It has never been demonstrated that fingerprint examiners use a proven methodology," said Lyn Haber, a California forensic researcher.

With DNA analysis, the problem is different. The scientific underpinnings of DNA analysis are well-tested and conceded to be solid even by critics. But the certainty of a DNA match can be overshadowed by the larger question of how the DNA evidence was obtained and handled. In the O.J. Simpson murder case, for instance, defense attorneys cast doubt on DNA results because of sloppy lab work, ultimately suggesting investigators planted the evidence at the scene. And a DNA match to a crime scene, many defense attorneys point out, only proves a suspect was there not that he or she committed a crime.

"The problems with DNA are partly human error, or worse, human corruption,"
said Harvey Silverglate, a Boston civil-rights attorney who fears innocent people may still be convicted under the Romney plan.

Human error is also emerging as a key problem in crime labs,
both in Massachusetts and around the country. Stephan Cowans, who was convicted in 1998 of shooting a police officer in a Roxbury backyard, was freed from prison this year after it was revealed the fingerprint evidence used to convict him did not come from his finger.

A recent article in Champion, a magazine published by the National Association of Criminal Defense Lawyers, noted widespread problems at crime labs across the country, many exacerbated by overwork and small budgets. In Massachusetts, a state report two years ago noted that space in the State Police crime lab in Sudbury was so limited that scientists had to extract evidence from suspect and victim's clothing on alternate days to avoid cross-contamination.

Stung by that report, officials say the state has since gone through a voluntary accreditation by a national board that sets standards for crime labs. But that is only partially true: The Sudbury lab is accredited only in DNA testing and "criminalistics," the analysis of trace evidence, fibers and tool markings. The offices of ballistics and fingerprint analysis are not accreditated; nor is the state's DNA database. State officials say they are attempting to get them accredited and are also seeking a large increase in funds for that lab.

Under the death penalty plan, Romney has pledged to ensure that all labs are operating as flawlessly as possible so there will be no questions about the way evidence is collected or analyzed. If valid questions do arise, prosecutors would not seek the death penalty.

Death penalty opponents agree that if labs were better monitored and funded fully, there would be less suspicion about whether the evidence was tainted or analyzed incorrectly. And the authors of the Massachusetts death penalty proposal are clear in wanting an independent scientific review of the collection, analysis and presentation of evidence, along with other safeguards. But as long as humans are involved in science, either analyzing it or interpreting it, mistakes can happen, others say.

"What we say in forensic science is the more certain the scientist is, the less reliable the scientist is," said James Starrs, a professor of law and forensic sciences at George Washington University. "We all want to be on safe ground, always looking for a magic bullet. But our society can easily be taken in by science, and that is worrisome."


Junk Science
Death Penalty Issues

Truth in Justice