Wednesday, June 11, 2014

Forensic Science is Still Bogus

Slate gets to the forensic science mess:
Far from an infallible science, forensics is a decades-long experiment in which undertrained lab workers jettison the scientific method in favor of speedy results that fit prosecutors’ hunches. No one knows exactly how many people have been wrongly imprisoned—or executed—due to flawed forensics. But the number, most experts agree, is horrifyingly high.
Yep.

It has been four years since the NSF issued its damning report on forensic "science," but not much has happened since to change things. Resistance on the part of police and prosecutors to any measure that might undermine convictions is very, very strong, making political headway difficult. But in January the Department of Justice and the National Institute of Standards and Technology announced the creation of the National Commission on Forensic Science, which is supposed to.
work to improve the practice of forensic science by developing guidance concerning the intersections between forensic science and the criminal justice system. The commission also will work to develop policy recommendations for the U.S. Attorney General, including uniform codes for professional responsibility and requirements for formal training and certification.
All fine and good. If you ask me, though, codes of professional responsibility and requirements for certification are bunk; so are things like uniform standards and procedures. What we really need is regular double blind testing of all the labs, with full publication of results.

2 comments:

Shadow said...

Absolutely. Some serious scientific testing whose findings can be repeated by other scientists is badly needed.

We don't even have a nationwide standard for fingerprints. Is it 10 points? 20 points? In some states the minimum number of points required for a match changes within the lab -- it depends on the technician. I've never heard of such nonsense.

And what is this 100% certitude that no two fingerprints can match? Exactly how many points are we talking about? And why don't we have a scale of likelihood: 10 point match = 25% likelihood, 20 point match = 50% likelihood, on so forth. There are 150 points in a good print.

http://www.washingtonpost.com/wp-dyn/articles/A64711-2004May28.html

Unknown said...

Great post. This article is really very interesting and enjoyable. I think its must be helpful and informative for us. Thanks for sharing your nice post .
live football scores