DNA law change

**scumdog** · 13th April 2009, 22:22

Originally Posted by Ixion

Not so. Fingerprints are (if a good set be matched) a definative Yes/No. DNA testing is never definative. It's always "this specimen is X% more likely to be form the same person as this one, than one from a random memeber of the population" .The devil is in those percentages. For instance DNA testing will have great difficulty distinguishing between family memebers. Fingerprints have no such difficulty.

So, if Harry's DNA is in the database and the cops find DNA at a crime scene , which , unbeknown to them, comes from Harry's half brother Harvey, then Harry is going to come up as a strong suspect when the database is searched. Once a suspect, the cops will start fitting a case round him. Whereas Harvey's fingerprints will be completely different to Harry's. No risk of the latter being fitted up by fingerprints. Much risk from DNA.

And we do not (yet) require people to contribute their personal fibres to a database. Nor is the driver's licence , our closest thing to a personal ID (as yet) officially recognised as such.

Re the personal identification: I was refering to eye-witnesses, "Yer guv'nor, it were that one armed man wot done it" sort of thing.

And sadly a lot of General Public (tm) can't tell Casius Clay from Santa Claus. ("I dunno, they all look so much alike").

The other stuff I mentioned was just ways of linking somebody to a crime.

**Ixion** · 13th April 2009, 22:31

You nicely state the problem.

If some bystander says "Oh, I think it was that guy Harry done it, looked like him anyway" you receive that information with a degree of caution and scepticism. Well, maybe, Or maybe it was actually Cassius Clay. We'll check it out. And all other cops would be as cautious about assumptions based on such identification - it is notoriously dubious.

But if someone in a white coat says "This DNA sample from Harry is one million times more likely to be that of the perpetrator than one from a random patagonian yak-herder", you (OK, maybe not you, but certainly a very large percentage of cops), will immediately say to themselves " Shit - one million times more likely ! - well, doesn't get any more convincing than that. And its scientific, so it must be right. No problem getting a jury to convict on that basis. OK let's go pick up this Harry guy"

**scumdog** · 13th April 2009, 22:33

Originally Posted by Ixion

You nicely state the problem.

If some bystander says "Oh, I think it was that guy Harry done it, looked like him anyway" you receive that information with a degree of caution and scepticism. Well, maybe, Or maybe it was actually Cassius Clay. We'll check it out. And all other cops would be as cautious about assumptions based on such identification - it is notoriously dubious.

But if someone in a white coat says "This DNA sample from Harry is one million times more likely to be that of the perpetrator than one from a random patagonian yak-herder", you (OK, maybe not you, but certainly a very large percentage of cops), will immediately say to themselves " Shit - one million times more likely ! - well, doesn't get any more convincing than that. And its scientific, so it must be right. No problem getting a jury to convict on that basis. OK let's go pick up this Harry guy"

Still got to have a jury believe the same.

Oh, and put forwards all manner of other compelling evidence type shit too.

**want-a-harley** · 13th April 2009, 23:17

[QUOTE=Ixion;2025856]Not so. Fingerprints are (if a good set be matched) a definative Yes/No. DNA testing is never definative. It's always "this specimen is X% more likely to be form the same person as this one, than one from a random memeber of the population" .The devil is in those percentages. For instance DNA testing will have great difficulty distinguishing between family memebers. Fingerprints have no such difficulty.

QUOTE]

If you get a good DNA sample it can descriminate as much if not more than a fingerprint. If you get good quality single source DNA (such as blood) then you can descriminate even between siblings. The fact you can apply a percentage to DNA profiling is because a larger independent scientific community has examined it unlike fingerprints which is a bit of an old boys club.

For an interesting read on why fingerprints aren't all they cracked up to be read this from the new yorker

http://www.michaelspecter.com/pdf/fingerprint.pdf