Forensic Follies: FBI Analysts Send Man To Jail For 28 Years By Confusing Dog Hair And Human Hair
link
I long ago concluded that forensic analysis is about a half-step above voodoo.
Give a lab ten samples and ask whether any of them belong to the accused--or whatever. If you can't pick the damn sample out of a lineup, then you don't know what you're doing. And if you can't afford to do it that way, then you can't afford to use these methods.
Confirmation bias alone makes this shit near enough to witchcraft so far as I can tell.
I long ago concluded that forensic analysis is about a half-step above voodoo.
Give a lab ten samples and ask whether any of them belong to the accused--or whatever. If you can't pick the damn sample out of a lineup, then you don't know what you're doing. And if you can't afford to do it that way, then you can't afford to use these methods.
Confirmation bias alone makes this shit near enough to witchcraft so far as I can tell.
posted by Winston Smith at
11:18 AM
4 Comments:
There are many forms of forensic analysis that are quite accurate. The instrumental methods used to test for drugs, alcohol and DNA, for example. I've always been a bit suspicious of hair analysis since it's so qualitative.
Oh yes, what the FBI Analysts did is inexcusable and I hope they are severely punished to the maximum extent of the law.
Dammit, I knew as I posted that that I was overgeneralizing and talking out my ass, and that, for all I knew, some of that stuff was sound...
I should also point out that even if a technique such as hair analysis were 99.9% accurate, if the technician makes shit up the accuracy of a technique is meaningless. What they did, as I understand it, was take a technique that's qualitative at best and make shit up.
Post a Comment
Subscribe to Post Comments [Atom]
<< Home