It's a crime when science gets it wrong
Houston Chronicle By Rick Casey
September 18, 2005
HERE'S a test I would like you to take.
I would also like the staff of Houston Police Department's crime lab, the district attorney and his staff to take it.
An analysis was done of 86 criminal convictions that DNA evidence later found to be wrong. Please rank the factors most often found to have contributed to the wrongful convictions:
•Incompetent defense
lawyers.
•Police misconduct.
•Eyewitness errors.
•False testimony by forensic scientists.
•Prosecutorial misconduct.
•False confessions.
•Errors in scientific testing.
•False testimony by lay witnesses.
•Dishonest informants.
•Police misconduct.
•Eyewitness errors.
•False testimony by forensic scientists.
•Prosecutorial misconduct.
•False confessions.
•Errors in scientific testing.
•False testimony by lay witnesses.
•Dishonest informants.
If you ranked false confessions last, you are right. But did you guess 17 percent of the cases, nearly one in five, featured false confessions? That percentage tied with false testimony by lay witnesses.
If you ranked eyewitness errors first, you were also right. Erroneous eyewitness accounts showed up in a stunning 71 percent of these cases.
The most important finding
Prosecutors will be glad to know that their brothers and sisters engaged in misconduct in fewer of these cases (28 percent) than police officers (44 percent).
I was most surprised by two findings. One is that incompetent defense attorneys figured in just 19 percent.
The other is the most important finding of this study.
It is that fully 63 percent of the cases featured errors of forensic science. What's more, in 27 percent of the cases forensic scientists gave false or misleading testimony.
(Percentages exceed 100 percent because more than one factor was found in many cases.)
Time after time, juries believed experts in ballistics, fingerprint identification, hair analysis and other forensic disciplines who were wrong, lying or both.
And the vast majority of the cases weren't handled by the Houston police crime lab. The problem is broader than that.
The analysis was presented last month in an article in the respected journal Science. The authors are Jonathan Koehler, a behavioral scientist specializing in legal issues at the University of Texas McCombs School of Business, and Michael Saks, a professor of law and psychology at Arizona State University College of Law's Center for the Study of Law, Science, & Technology.
'The world's top expert'
It's an optimistic article, sort of.
DNA analysis, the authors say, has provided a scientific standard for forensics. It was developed through a scientific process. It was critiqued by a wide variety of scientists. Standards were developed for conducting DNA tests and for reporting them.
"In DNA, we say, Here are the chances of a match, and here is the frequency with which we make an error," Koehler told me Friday.
By contrast, he said, a ballistics expert might say, in effect, "It's a match. This bullet came through this gun. I know because I'm the world's top expert."
Yet, say the authors, the studies haven't been done that show statistically how singular are the identifying marks of fingerprints, bullets, hair or other evidence.
What's more, "there has been remarkably little research on the accuracy of traditional forensic sciences."
These "sciences" became science largely by saying they were.
Forensic scientists often argue that the science is infallible. Any problems are due to "human error." But as Koehler points out, where the error occurs matters little to the jury, or to the person wrongly imprisoned.
Koehler and Saks propose subjecting crime labs to "blind testing," in which technicians are tested with samples that appear to be part of their regular work. The tests should be conducted by an independent agency, and the results should be made public, as with mainstream science.
"The FBI does their own testing and says, 'We've done our testing and we're perfect, and you can't look at our data because we're the FBI,' " said Koehler.
One of the reasons forensic science is so fallible, the authors write, is that 96 percent of the positions are held by persons with bachelor's degrees or less. By contrast, in "normal science, academically gifted students receive four or more years of doctoral training where much of the socialization into the culture takes place. This culture emphasizes methodological rigor, openness, and cautious interpretation of data."
Forensic science, by contrast, is "an adversarial, crime-fighting culture" where there often is pressure to produce findings that support police and prosecutors.
Koehler thinks we're headed in the right direction, but we have a long way to go.
"Some people twist our words into saying we believe no forensic science should come into court," he said. "I don't believe that. I just believe that we're far better off with forensic science that is real science."
A science that doesn't rank right behind eyewitness errors in sending innocent people to prison.