3 ms·

It's scary because while DNA is more accurate, it's not 100%. Nobody wants to be the one that is thrown under the bus by a bad DNA result that is given more wei

by pyre 13y ago

It's scary because while DNA is more accurate, it's not 100%. Nobody wants to be the one that is thrown under the bus by a bad DNA result that is given more weight than it should because, "well it's DNA, it's a fact."

rayiner 13y ago
It's true that people think DNA evidence is more accurate than it is.[1] That said, juries think all "scientific" forensic evidence is more accurate than it is. The difference is that DNA evidence is actually pretty accurate, while things like hair analysis are wildly inaccurate. And eyewitness testimony tends to be insanely persuasive, even though its also insanely inaccurate. [1] Generally, juries are instructed about random-match probabilities, but aren't usually allowed to be told about lab error rates, which totally dominate the random-match error rates (1-2% error in typical scenarios).
- venomsnake 13y ago
  Why does then the defenders not attack the eye witness testimony on these grounds - that people and memory are unreliable, so what this guy remembers even if honestly believe so is not to be trusted? They could also throw some Bayesian math in the mix.
  - jlgreco 13y ago
    They can attack the eye witness testimony on those grounds. Of course if the witness is a cop (community hero!) or god forbid a girl scout delivering cookies, then legitimately attacking the credibility of eye-witness accounts can put you in a very unsympathetic light before the jury. An attack on the poor recall abilities of homo-sapiens becomes warped into an attack on the character of a plainly honorable witness. Public understanding of even the basics of Bayesian thinking is abysmal.
  - gknoy 13y ago
    The judge probably would not allow them to do that, and even if so, that means that pretty much any defensive testimony would likely face similar hurdles.
- rmc 13y ago
  As in: "The actual accuracy of DNA is closer to the percieved accurancy, in the public's mind, of DNA"
ajross 13y ago
This is true, but almost exactly the same thing could have been said about fingerprints 60 years ago. Bad fingerprint analysis certainly has convicted innocent people. DNA is much (!) more accurate than a fingerprint, so even given that criticism it still represents progress.

by pyre 13y ago

rayiner 13y ago

It's true that people think DNA evidence is more accurate than it is.[1] That said, juries think all "scientific" forensic evidence is more accurate than it is. The difference is that DNA evidence is actually pretty accurate, while things like hair analysis are wildly inaccurate. And eyewitness testimony tends to be insanely persuasive, even though its also insanely inaccurate. [1] Generally, juries are instructed about random-match probabilities, but aren't usually allowed to be told about lab error rates, which totally dominate the random-match error rates (1-2% error in typical scenarios).

venomsnake 13y ago
Why does then the defenders not attack the eye witness testimony on these grounds - that people and memory are unreliable, so what this guy remembers even if honestly believe so is not to be trusted? They could also throw some Bayesian math in the mix.
- jlgreco 13y ago
  They can attack the eye witness testimony on those grounds. Of course if the witness is a cop (community hero!) or god forbid a girl scout delivering cookies, then legitimately attacking the credibility of eye-witness accounts can put you in a very unsympathetic light before the jury. An attack on the poor recall abilities of homo-sapiens becomes warped into an attack on the character of a plainly honorable witness. Public understanding of even the basics of Bayesian thinking is abysmal.
- gknoy 13y ago
  The judge probably would not allow them to do that, and even if so, that means that pretty much any defensive testimony would likely face similar hurdles.
rmc 13y ago
As in: "The actual accuracy of DNA is closer to the percieved accurancy, in the public's mind, of DNA"

ajross 13y ago

This is true, but almost exactly the same thing could have been said about fingerprints 60 years ago. Bad fingerprint analysis certainly has convicted innocent people. DNA is much (!) more accurate than a fingerprint, so even given that criticism it still represents progress.