Hair Analysis Problems Shine Light on Forensic Science

The Washington Post reports that the FBI has now admitted that virtually every bit of testimony offered by their “experts” on microscopic hair fiber analysis during criminal trials was based on bad science. Hundreds of cases are now under review by the DOJ.

The Justice Department and FBI have formally acknowledged that nearly every examiner in an elite FBI forensic unit gave flawed testimony in almost all trials in which they offered evidence against criminal defendants over more than a two-decade period before 2000.

Of 28 examiners with the FBI Laboratory’s microscopic hair comparison unit, 26 overstated forensic matches in ways that favored prosecutors in more than 95 percent of the 268 trials reviewed so far, according to the National Association of Criminal Defense Lawyers (NACDL) and the Innocence Project, which are assisting the government with the country’s largest post-conviction review of questioned forensic evidence.

The cases include those of 32 defendants sentenced to death. Of those, 14 have been executed or died in prison, the groups said under an agreement with the government to release results after the review of the first 200 convictions.

The FBI errors alone do not mean there was not other evidence of a convict’s guilt. Defendants and federal and state prosecutors in 46 states and the District are being notified to determine whether there are grounds for appeals. Four defendants were previously exonerated.

The admissions mark a watershed in one of the country’s largest forensic scandals, highlighting the failure of the nation’s courts for decades to keep bogus scientific information from juries, legal analysts said. The question now, they said, is how state authorities and the courts will respond to findings that confirm long-suspected problems with subjective, pattern-based forensic techniques — like hair and bite-mark comparisons — that have contributed to wrongful convictions in more than one-quarter of 329 DNA-exoneration cases since 1989.

The show CSI and its roughly 1400 spinoffs have probably done a lot of damage here, leading the public to think that forensic science is this undeniably true and accurate way of catching criminals. But the truth is that many types of forensic science are based on almost no science at all and a whole lot of innocent people are in prison or dead because of it.

It’s time for Congress to bring together real scientists to examine every type of forensic science used in criminal trials and pass a law with the necessary and appropriate safeguards to ensure that all techniques used are evidence-based and reliable. But I’m not holding my breath for that to happen. There’s no money in it.

"You mean the sanctions that most of the Republicans signed too. You do know who's ..."

Gorka Lies About Clinton and Uranium ..."
"God lies in the first few pages of the bible.Shortly after he created two people ..."

Christian Right Still Oblivious to Their ..."
"It's possibly criminal in trump's case. But I guess that doesn't really matter to people ..."

Gorka Lies About Clinton and Uranium ..."
"Given the current climate of outing sexual harassment from decades ago, I don't think I'm ..."

Gorka Lies About Clinton and Uranium ..."

Browse Our Archives

Follow Us!


What Are Your Thoughts?leave a comment
  • JustaTech

    I’m curious about those 2 analysts who *didn’t* overstate the matches. Were they new and inexperienced and not allowed to testify at trial? Were they old and stuck in their ways? Honest and marginalized?

    And what happened in 2000 to make the hair data better (or less used)?

  • sugarfrosted

    Even finger print analysis is flawed to hell. One person from Oregon was almost convicted for the Madrid Train Bombing due to it. He had never even been to Madrid. Multiple experts said they matched. He was only exonerated when they caught the real guy whose finger prints also matched.

  • colnago80

    I believe that I noted in a comment on the subject of forensic testing of hair samples some time ago that, during the O. J. Simpson trial, the FBI expert was limited in what he could testify to by the judge. He was not allowed to testify that hair samples found at the crime scene were a match to hair samples taken from Simpson but could go no farther then testifying that the crime scene samples were from an African American and were consistent with the samples taken from Simpson. He was forced to testify under cross examination that the comparison of the two samples were in no way definitive, unlike fingerprints and DNA. Fingerprints are definitive provided that a sufficient number of points of identification are found.

  • colnago80

    Re sugarfrosted @ #2

    As I understand it, federal statutes require 9 points of identification for declaring a match. California, for instance, requires 12. Obviously, 9 points isn’t sufficient, although it probably is sufficient for exclusion. Much like 1st generation PCR is only useful for exclusion.

  • eric

    It’s time for Congress to bring together real scientists to examine every type of forensic science used in criminal trials and pass a law with the necessary and appropriate safeguards to ensure that all techniques used are evidence-based and reliable.

    This may solve some of the problems but not all of them. Some if the issue surrounding misinterpretation of forensic evidence has more to do with the legal side; how prosecutors and defense attorneys talk about evidence to try and support (or undermine) a showing of ‘beyond a reasonable doubt,’ and how juries understand both legal and scientific aspects of evidence. You are never going to get a scientist on the witness stand saying that she is 90% confident that the X matches the accused’s X within +/-5%, until the lawyers themselves decide that is the sort of argument they want to put in front of the jury. Even if its true, even if its the scientifically best description of what the evidence says, if the prosecutor thinks this will mean nothing to the jury, they are going to have the witness re-parse results in simpler language that is probably less accurate and almost certainly rhetorically designed to convey high confidence. Defense attorneys will of course do the same, only in support of a good outcome for their client.

    But I’m not holding my breath for that to happen. There’s no money in it.

    This is true, but again only half the battle. No prosecutor wants to know if some technique has a less than 99.9999% accuracy. If the accuracy is in doubt, best not to quantify it at all; best that nobody (not even the scientists performing the analysis) even knows the accuracy, so the defense can’t bring it into the case. And since we can bet that the accuracy of pretty much any sound forensic scientific technique isn’t going to be that high, there is no incentive for federal, state, or local police and justice departments to quantify the errors in these techniques. Its not just ‘the public doesn’t want to fund it,’ its just as much ‘the state doesn’t want to know the answer.’

  • sugarfrosted

    @3 Perhaps this is true, but in that case, they had 16 points of comparison, which at that point was the gold standard of comparison.

    The scary fact was that this was viewed as proof of his guilt contrary to everything else, which raised reasonable doubt.

  • http://en.uncyclopedia.co/wiki/User:Modusoperandi Modusoperandi

    sugarfrosted “Even finger print analysis is flawed to hell. One person from Oregon was almost convicted for the Madrid Train Bombing due to it. He had never even been to Madrid. Multiple experts said they matched. He was only exonerated when they caught the real guy whose finger prints also matched.”

    To be fair, there was no way the prosecution could’ve known he had an evil twin.

  • colnago80

    Re sugarfrosted @ #6

    According to an FBI internal investigation, apparently the FBI examiners really only had 7 points of identification, the others were misidentified. The bottom line is that the examiners were careless in identifying points of identification. Clearly, there is some subjectivity in fingerprint identification, much like 1st generation PCR DQ-Alpha analysis.

    http://goo.gl/VF3uyD

  • tbp1

    I was on jury duty a few years ago. During voir dire for a criminal case lawyers for both sides were diligent in explaining that we shouldn’t expect CSI-style razzle-dazzle or 100% certainty, but that instead we would be called upon to use our best judgment in weighing evidence. I was rather impressed by that. As it happens, I didn’t end up on the actual jury, so I don’t know what happened. Except for the fact that it looked like it was going to be a long trial, I was actually a little sad I wasn’t chosen.

  • http://www.thelosersleague.com theschwa

    Well, MOST of the examiners might be using flawed science. But when I need an expert, I can always rely on Prof. Vidal Sassoon to bring the real McCoy!

  • JustaTech

    This comes back to a problem that comes up in the general understanding of statistical tests. You really can’t say “well, the p-value is less than 0.05, so X is better than Y!” All you can say (and be mathematically correct) is “The p-value is less than 0.05, so we reject the null hypothesis that X and Y are the same.”

    The problem is that no one talks like that, even in most of the sciences, let alone when speaking to the lay public.

    i can only imagine what I would be like on a jury: “Proof is for math; show me evidence.”

  • http://artk.typepad.com ArtK

    I’ve recently been binge-watching the CSI:XXX series. Not only is the science presented badly, but so is basic logic. “We found this dirt on the floor of the car and it must have been brought in by the killer.” Well, no. Unless you can prove that the car was steam cleaned right before the killer got in, that dirt could have come in at any time. It would take them years to analyze the trace evidence to be found in my car. The repeated “it can only mean…” when there are half a dozen viable alternate explanations drives me nuts.

  • EnlightenmentLiberal

    The show CSI and its roughly 1400 spinoffs have probably done a lot of damage here, leading the public to think that forensic science is this undeniably true and accurate way of catching criminals.

    Hey hey now. Sometimes it is.

    When it’s with DNA analysis or something equally reputable.

    And when the chain of custody of evidence has not been compromised.

    And when the cops are not planting evidence.

    And when the CSI guys are not faking evidence.

    And when the CSI guys are not lying under oath for the prosecutors.

    And when the CSI guys, who IIRC often don’t even have a relevant college degree, don’t make honest mistakes when they don’t know what the hell they are doing.

    Yea. Anyway.

    Even finger print analysis is flawed to hell.

    IIRC, only DNA analysis techniques have had actual good peer reviewed work behind it to show that it works. I bet finger print analysis is generally solid with an honest analyst, but the false positive rate also must be significantly higher than the miniscule rate that is modern DNA analysis. I would still love to know if there are actual papers that put a number on the false positive rate.

  • slc1

    Re Enlightenment Liberal @ #13

    I recall mention of a study the FBI supposedly conducted a decade or so ago in which they accessed their AFIS database which, at that time, had some 40 million individuals and found a number of entries linking to more then one individual with a 9 point of identification criterion. I did not see an actual article describing this, it was mentioned in a comment on another website.

  • slc1

    In the unlikely event that someone may be curious as to why I have logged in under my slc1 identity, I have been unable to log in under my colnago80 identity.

    To follow up with my comment @ #14, the number of false positives will be heavily dependent on the criterion as to the number of points of identification in common used to declare a match.