THE STUNNING admission by federal law enforcement officials of flawed testimony by forensic experts in hundreds of criminal cases has focused attention on the dubious use of hair analysis. Add this to a list of factors that has played a role in securing the conviction of people who later proved their innocence, including mistaken eyewitness identifications, bad informants and police or prosecutorial misconduct. Forensic technologies must be reassessed, but other parts of the system also need scrutiny if the chances of wrongful convictions are to be reduced.

The Justice Department and FBI, as The Post’s Spencer S. Hsu reported, have acknowledged widespread instances of scientifically indefensible testimony involving microscopic hair comparison by members of an elite FBI forensic unit. A review of cases over two decades before 2000 concluded that nearly every examiner — 26 of 28 — gave flawed testimony in ways that favored the prosecution. Of the cases, 32 defendants were sentenced to death. Fourteen of those have been executed or have died in prison.

The FBI errors do not necessarily mean defendants were innocent. But the confirmation of long-held suspicions about hair analysis, the inability of courts to keep junk science out of courtrooms (and hair analysis is not the only suspect technique) and the questionable reliability of other evidence provide reason to worry about how many innocent people are being convicted.

Consider, for example, the three defendants from the Washington area — Kirk L. Odom, Santae A. Tribble and Donald E. Gates — whose exonerations in separate cases were pivotal in prompting the hair analysis review. Examiners had testified in each case that hair found at the crime scenes matched that of the suspects, analysis later undermined by DNA results that exonerated the three men.

In each case, improper forensics combined with other failings to produce wrongful convictions. Mr. Odom, who spent 22½ years in prison for a rape he didn’t commit, was mistakenly identified by the victim after a questionable police lineup. For the other two, testimony from police informants with something to gain was a factor. Mr. Tribble spent 27 years and 10 months in prison for a murder he didn’t commit. Mr. Gates spent 28 years for a rape and murder he didn’t commit. In other cases, confessions came from susceptible people, or information that would have been helpful to the defense was withheld by the government.

With the exception of the rare bad apple, police and prosecutors don’t set out to convict the wrong person, but even the most well-meaning people can be affected by contextual or cognitive bias, particularly in a system with crushing caseloads and intense pressure. Better protections are needed. Steps in the right direction are improvements by D.C. police in procedures for eyewitness identification and the creation of a conviction integrity unit by the U.S. Attorney’s Office in the District.

But more needs to be done, including, as we’ve argued before, requiring the prosecution to share information with the defense more thoroughly and consistently. We hope the Justice Department undertakes an analysis not only of the breakdown in hair forensics but also of other vulnerabilities in the system.