The Register® — Biting the hand that feeds IT

Comments on: 100,000 'erroneous' records on DNA database

PNC taken as gold standard? 

Posted Thursday 17th May 2007 17:56 GMT

I seem to recall that a large number of PNC records are found to be incorrect even after being checked by a supervisor. The PNC is hardly a suitable gold standard for checking data against. How many records have got the wrong data attached to them?

How many records...? 

Posted Friday 18th May 2007 00:30 GMT

> How many records have got the wrong data attached to them?

We'll find out when people start being arrested in Newcastle for crimes committed in Newquay...

which a police officer writes by hand 

Posted Friday 18th May 2007 08:12 GMT

bless.

Yet another... 

Posted Friday 18th May 2007 09:14 GMT

...example of too many organisations being involved in a service where the lives of joe public are concerned and accuracy is a basic pre-requisite. Private firms serving their own interests, poor communication, governmental balls-ups and PC Plod. And who is accountable?

Dubious 

Posted Friday 18th May 2007 10:16 GMT

If you were in court, and the evidence provided by the prosecution was acquired in such a way as to have this high a failure rate, with this many problems, it would surely be deemed inadmissable. Eg - "We've got a partial fingerprint from a piece of glass. It partially matches your little finger. As we've got nobody else to match it against, and have some juicy Labour targets to meet, we'll throw the book at you anyway. If you were on our database to start with, you MUST be evil/a terrorist/a 9-11 plotter/insert convenient gov't label here" - thrown out.

But as soon as the magic Totally Infallible™ DNA records pop out, it's another matter. Never mind the fact that so many records are missing, invalid, incorrectly entered - or that DNA isn't as reliable as it's made out to be to start with!

And what about the remaining errors ? 

Posted Friday 18th May 2007 12:04 GMT

"... load failure rate for DNA samples was wheedled back to its current 1 per cent ..."

So there are still about 1% rejects - or a 1% error rate. So assume we get to 100% of the population DNA tested (which has to be the government traget even if they won't admit it) then that means 600,000 errors MINIMUM !

So we can then stand up in court and ask the prosecution if it's possible that the DNA could be confused with any of the other MILLION+ failed records ?

Bad defense... 

Posted Friday 18th May 2007 20:27 GMT

Merit of having a DNA database at all to begin with aside... Does not sound like the bad quality of the database would help much in court. I mean, just get the DNA tested again. Does your DNA still match the one found in the crime scene? Oh, well...

Cheers

J

Don’t Miss