pbray: (crime)
FBI to review lab work in thousands of cases.

Original Washington Post article here, may require registration to read.

The new review goes back at least as far as 1985. A previous review of FBI Forensic Lab work was done in the mid-90s which uncovered major problems and led to reforms, but obviously they must have reason to believe that the previous review didn't catch all the problems. Nor did their reformed culture stop systemic problems from occurring in later years.

Particularly troubling is that the FBI has apparently known about the problems with hair and fiber evidence for years and failed to act, a pattern that we've seen from them before.

Since 2009 the National Academy of Sciences has been pushing to put the science back into forensic science, and this latest FBI review appears to be at least in part a result of their efforts. Of course you can't help but wonder-- if they get it so wrong at the FBI level, how bad is it in the smaller crime labs around the country?
pbray: (crime)
Interesting article on the BBC website about upcoming changes to the FBI's DNA database.

One thing most people don't realize is that when DNA information is stored or analyzed for comparison purposes, they don't use the entire sequence. Instead they use selected markers, basically a small subset of the total DNA. Which, of course, leads to two potential problems. First, if two databases use different markers, you can't easily compare data between the two systems, a potential problem when trying to share information across jurisdictions.

The second problem is what they refer to as chance matches. The situation where a DNA match is declared based on the subset of DNA examined, when, in fact, a broader comparison would reveal that the DNA is not identical. This is the possibility that troubles forensic scientists, who are still trying to work out how much data needs to be stored in a profile in order to rule out the possibility of chance matches.

And for those who assume well this is the FBI, of course they wouldn't declare themselves certain of a match unless they were, let's not forget Brandon Mayfield. He's the Seattle lawyer whose fingerprints the FBI mistakenly matched to the terrorist bombing in Spain, an analysis which was vigorously defended by the FBI even after Spanish law enforcement declared it to be in error. Eventually Mayfield received a rare public apology from the FBI, but if this had been a strictly domestic case, rather than a high profile case that crossed borders, one wonders if the mistake would ever been caught.
pbray: (crime)
Earlier this year the National Academy of Sciences issued a scathing report on the state of forensic science in this country. It looks like they're firing the next salvo, with an in-depth hearing on the investigative science used in the probe of the 2001 anthrax attacks, specifically those that led the FBI to declare Dr. Bruce Ivins as their sole suspect.
"I think this review is a really good first step," said biological policy expert Cheryl Vos of the Federation of American Scientists. "What I would really like to know is how much did the scientific conclusions drive the investigation, and vice versa. There is a clear intersection between the two." From the USA Today article.

In the vein of seeing what you want to see and finding what you want to find, two recent news stories focused on false evidence linked to scent tracking dogs, including the case of Bill Dillon, sentenced to life in prison based largely on evidence provided by a dog handler.. The article refers to a fake scent tracking dog, but the dog isn't the one who gets on the stand, it's the handler who testifies. Money quote:
Tim McGuire, a dog-tracking expert with Florida’s Volusia County Sheriff’s Department, said it was implausible that a dog could have picked up Dillon’s scent back in 1981 eight days after the murder, and just after a massive hurricane had blown through the area.

The recent issue of Time Magazine also has a story on scent evidence this one focused on a handler in Texas.

November 2014

S M T W T F S
      1
2345678
9101112131415
16171819 202122
23 2425 26272829
30      

Syndicate

RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags