3 ms·

While I tend to agree, I hope we don't get lazy with the law and prosecute on DNA alone. DNA matching is not 100% accurate, and juries aren't the best at apprec

by madrox 6y ago

While I tend to agree, I hope we don't get lazy with the law and prosecute on DNA alone. DNA matching is not 100% accurate, and juries aren't the best at appreciating the nuance of statistics. When I was a stats major, Peter Donnelly's TED Talk was required viewing on the matter: https://www.ted.com/talks/peter_donnelly_how_juries_are_fooled_by_statistics?language=en https://www.ted.com/talks/peter_donnelly_how_juries_are_fool...

sillysaurusx 6y ago
What got me was, if it's not 100% accurate, how do you explain their ability to hunt down the golden state killer using his family tree? They narrowed the possibilities to 1,000 or so, then checked all 1,000. Straightforward police work. Then a final DNA match confirmed their hypothesis that they'd found him, which turned out to be true. I don't know. I'm vividly aware that if the matching isn't accurate, innocents will go to prison. But the truth seems less simple than that.
- dxdm 6y ago
  I'm assuming the "police work" you mention would yield more material to prosecute on, so they won't gi by "DNA alone", which GP was concerned about.
- Spivak 6y ago
  It's the same thing with IAFIS. If you think you're going to trial and winning with just a fingerprint match in the database you're gonna be sorely disappointed. These tools are better understood as "lead generation" than any kind of proof.
542354234235 6y ago
It is also a problem when the DNA isn’t obvious that it is connected to the perpetrator of the crime. Semen found inside a victim is very different than DNA found inside a room where a crime occurred, which is different than DNA found at a crime scene in a public or semi-public place. If you can just collect up a bunch of DNA and run it against everyone, with no initial justification of probable cause, you can get innocent people that are guilty of no more than leaving their DNA somewhere and not having an alibi being victims themselves of lazy police investigations.

by madrox 6y ago

sillysaurusx 6y ago

What got me was, if it's not 100% accurate, how do you explain their ability to hunt down the golden state killer using his family tree? They narrowed the possibilities to 1,000 or so, then checked all 1,000. Straightforward police work. Then a final DNA match confirmed their hypothesis that they'd found him, which turned out to be true. I don't know. I'm vividly aware that if the matching isn't accurate, innocents will go to prison. But the truth seems less simple than that.

dxdm 6y ago
I'm assuming the "police work" you mention would yield more material to prosecute on, so they won't gi by "DNA alone", which GP was concerned about.
Spivak 6y ago
It's the same thing with IAFIS. If you think you're going to trial and winning with just a fingerprint match in the database you're gonna be sorely disappointed. These tools are better understood as "lead generation" than any kind of proof.

542354234235 6y ago

It is also a problem when the DNA isn’t obvious that it is connected to the perpetrator of the crime. Semen found inside a victim is very different than DNA found inside a room where a crime occurred, which is different than DNA found at a crime scene in a public or semi-public place. If you can just collect up a bunch of DNA and run it against everyone, with no initial justification of probable cause, you can get innocent people that are guilty of no more than leaving their DNA somewhere and not having an alibi being victims themselves of lazy police investigations.