With the introduction of DNA analysis three decades ago, criminal investigations and prosecutions gained a powerful tool to link suspects to crimes through biological evidence. This field has also exposed scores of wrongful convictions, and raised serious questions about the forensic science used in building cases.

This week, The Washington Post reported the first results from a sweeping study of the FBI forensic hair comparison unit, finding that 26 of 28 examiners in the unit gave flawed testimony in more than 200 cases during the 1980s and 1990s. Examiners overstated the accuracy of their analysis in ways that aided prosecutors. The National Association of Criminal Defense Lawyers and the Innocence Project are conducting the study with the cooperation of the U.S. Justice Department.

The development is only the latest to shake public faith in what police and prosecutors have often cited as scientific proof. In 2009, the National Academy of Sciences published an exhaustive review of the forensic sciences, concluding that only nuclear DNA analysis has a foundation in research. “Although research has been done in some disciplines,” the report states, “there is a notable dearth of peer-reviewed, published studies establishing the scientific bases and validity of many forensic methods.”

Fields based on matching patterns in fingerprints and hair have unknown error rates. Other methods are believed to be even more dubious, most notably the analysis of bite-mark injuries on victims’ bodies. Still, while the forensic sciences are under scrutiny, unproven practices, both old and new, continue to be used in courtrooms.

We’ve gathered some of the best reporting on questionable forensic science and evidence. Have we missed any? Please let us know in the comments below.

Convicted Defendants Left Uninformed of Forensic Flaws Found by Justice Dept

The Washington Post, April 2012

The Post story that spurred study of the FBI hair matching unit, which details how justice department officials failed to respond to widespread concerns about its fiber analysis. The FBI withheld the information from convicts whose cases included testimony from the troubled lab. “As a result, 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,” Spencer S. Hsu wrote.

Trial By Fire

The New Yorker, September 2009

The story of Cameron Todd Willingham, convicted of setting the fire that destroyed his Texas home and killed his one-year-old twin daughters. Willingham’s conviction rested on investigators’ belief that burn patterns proved the young father had poured a liquid accelerant to propel the blaze. Researchers would prove that theory baseless. The discovery did not benefit Willingham, who had been executed in 2004.

The Bite-Marks Men

Slate, February 2008

Two men in Mississippi were wrongly convicted for the rapes and murders of young girls based on bite-mark analysis by Dr. Michael West, a longtime expert witness in the state’s criminal courts. The entire method of analyzing bite marks is in question.

Weird Science

Texas Monthly, May 2010

Deputy Keith Pikett with the Fort Bend County Sheriff’s Office, outside Houston, connected people to crime scenes through “scent” lineups for his police dogs across Texas. Pikett would expose his dogs to a smell from the scene, and then have them sniff scent samples from the suspect and a series of uninvolved people. Pikett would testify dogs’ “body language” alerted him of alleged matches, helping to indict more than 1,000 suspects using the dubious technique.

Playing With Fire

The Intercept, February 2015

Another case in which fire investigators followed burn patterns to an incorrect conclusion, and a conviction. In 1992, Lorie Lee Lance died of smoke inhalation, crouched on the floor of a utility room as her home in Tennessee burned. Police suspected Lance’s boyfriend, Claude Garrett, had locked her in the room and set the blaze. However, the investigation had found the door was unlocked.

The Hardest Cases: When Children Die, Justice Can Be Elusive

ProPublica, June 2011

The story of Ernie Lopez, convicted in 2003 of sexually assaulting a six-month-old child he had been babysitting. The child, Isis Vas, died from what hospital workers and medical examiners believed were injuries caused by violence. The courts overturned Lopez’s conviction in 2012 after a top pathologist reviewed the case and determined Vas had a severe blood-clotting disorder that caused her to bruise and bleed. Lopez was freed but, to settle the criminal case, accepted a plea deal that prohibits him from proclaiming his innocence.

Shaken-Baby Syndrome Faces New Questions in Court

The New York Times Magazine, February 2011

An examination of the medical science behind shaken-baby syndrome, which has come into question during the past decade, through the case of Trudy Rueda, a caregiver convicted of seriously injuring a baby by shaking in Virginia. “A small but growing number of doctors warn that there can be alternate explanations — infections or bleeding disorders, for example — for the triad of symptoms associated with shaken-baby syndrome,” wrote Emily Bazelon.

Shaken Science

The Washington Post, March 2015

A series examining the child abuse prosecutions attributed to shaken baby syndrome since 2001, including more than 200 cases that were ultimately dismissed, overturned or resulted in acquittal. The vast majority of cases, however, ended with convictions. Many guilty verdicts relied on disputed science that supports the shaken baby diagnosis. “My greatest worry is that I have deprived someone of justice because I have been overtly biased or just mistaken,” Dr. George Nichols, former chief medical examiner of Kentucky, told the Post.

Second Chances Underscore Flaws in Death Investigations

ProPublica, February 2011

The U.S. has a major shortage of qualified forensic pathologists to perform autopsies of suspicious deaths. As a result, doctors who prove themselves unfit for the job continue to rule on causes of death and testify in criminal courts. Dr. Thomas Gill botched autopsies in Indianapolis, Los Angeles, Kansas City and Northern California over two decades. But following each failure, Gill found employers willing to excuse the past.