Play Live Radio
Next Up:
0:00
0:00
0:00 0:00
Available On Air Stations

Prosecutors Knew Of FBI's Forensics Flaws For Years, The Washington Post Reports

For years, the U.S. Department of Justice has known that flawed forensic work by FBI experts may have led to the convictions of innocent people, but prosecutors rarely told defendants or their attorneys, according to an investigative report in The Washington Post.

The newspaper's series is only the latest by journalists to question some of the most well-known tools of forensic science. NPR, working with PBS Frontline and ProPublica, has done related reporting over the past two years.

That includes a story last month that revealed new evidence in the case of Shirley Ree Smith, a California woman charged with killing her 7-week-old grandson by shaking him violently. The story by NPR, Frontline and ProPublica raised questions about the original medical examiner's report, which was key to convicting her.

Earlier this month, California Gov. Jerry Brown commuted Smith's sentence.

Also this year, a Texas court threw out the conviction in another case NPR questioned. Now Ernie Lopez, who was convicted of raping an infant and charged in her death, faces a new trial.

In today's Washington Post, reporter Spencer S. Hsu says Justice Department officials began reviewing cases after defense attorneys pointed out problems with evidence coming out of FBI labs. But the review was limited.

"As a result," writes Hsu, "hundreds of defendants nationwide remain in prison or on parole for crimes that might merit exoneration, a retrial or a retesting of evidence using DNA because FBI hair and fiber experts may have misidentified them as suspects."

In one case, a Texas man — whose death penalty was based on the FBI's questionable analysis — was executed more than a year after the Justice Department began its review.

The Justice Department, in a statement to NPR, noted that it established a task force in 1996 to let prosecutors know about the investigation of practices at the FBI labs.

"The Task Force undertook an exhaustive effort involving thousands of cases to ensure that defendants' rights to a fair trial were not jeopardized by the performance of a criticized lab examiner," according to the statement. The FBI labs are used by federal, state and local prosecutors. The Post article says that "while many prosecutors made swift and full disclosures, many others did so incompletely, years late or not at all."

The Post series notes that of all the forensic techniques to come out of crime labs, only DNA evidence has been scientifically validated and "able to consistently and accurately link a piece of evidence to a person or single source." Still unproven, but often used, are analyses of fingerprints, hair and fibers, marks on bullets and shell casing, handwriting and the use of polygraphs.

In an interview with All Things Considered airing Tuesday, Hsu tells host Audie Cornish that from the mid-1970s to the mid-1990s, "the FBI lacked written protocols and invoked highly subjective techniques" to conduct hair and fiber analysis.

And Tuesday night, forensic science comes under more scrutiny in a documentary by PBS Frontline, working with ProPublica and the Investigative Reporting Program at the University of California, Berkeley.

Their reporting looks at how flaws in fingerprint analysis led to the false arrest of an Oregon attorney, who found himself on trial for participating in the 2004 terrorist bombings in Spain that killed 191 people. After a judge dismissed the case against Brandon Mayfield, the FBI offered a rare apology.

The Frontline documentary also looks at other dubious uses of forensic evidence, from the testimony of a "smell" expert in the trial of Casey Anthony, who was acquitted of the first-degree murder of her 2-year-old daughter, to a story NPR reported last year on how flawed bite-mark identification was used to convict two innocent men in Mississippi.

Copyright 2021 NPR. To see more, visit https://www.npr.org.

Joseph Shapiro is a NPR News Investigations correspondent.