Criminal defense in Austin Texas.

DNA Testing

Posted By on January 5, 2010

DNA is often considered definitive proof in criminal cases.

In “Calculated Risk” Gerd Gigerenzer reminds us it may not be as reliable as people think it is.

The expert witness testifies that there are about 10 million men who could have been the perpetrator.  The probability of a randomly selected man having a DNA profile, that is identical with the trace recovered from the crime scen is approximately 0.0001 percent.  If a man has this DNA profile, it is practically certain that a DNA analysis shows a match.  If a man does not have this DNA profile, current DNA technology leads to a reported match with a probability of only 0.001 percent.

A match between the DNA of the defendant and the traces on the victim has been reported.

Question 1.  What is the probability that the reported match is a true match, that is, that the person actually has this DNA profile?

Question 2.  What is the probability that the person is the source of the trace?

Question 3.  Please render your verdict for this case: Guilty or not guilty?

You might want to think about it for a minute.

According to Gigerenzer, the answer to the first question is that the probability of a “true match” is 9%, or 1/11.  In other words, the chances are 91% that the defendant’s DNA doesn’t match the perpetrator’s – that the lab made a mistake.

Gigerenzer says the chances that the defendant is the “source of the trace” – in other words, that it was his DNA – is even lower: 0.9%, or less than one in a hundred (1/110, to be precise).

Of course, all that depends on the defendant having been selected at random, from a pool of 10 million men.  That’s not usually how it happens.  If there is other evidence that the defendant is guilty – in other words, some other reason the defendant was tested, other than being one among 10 million men – Gigerenzer’s calculations won’t work.

Still, his calculations show how dangerous it is to rely on DNA alone.

More importantly, Gegerenzer discusses the possibility of lab error, about which surprisingly little is known.

The laboratories that do the work are surprisingly hostile to outsiders. Independent testing of the results of their work is rare.  The FBI has resisted efforts to discover the results of their own internal tests.

Gigerenzer says that one of the few outside studies of laboratory error, two out of three labs reported false positives (matches that weren’t there), when sent a batch of 50 samples.  In a subsequent test, one year later, one out of the three reported a false match.

Experts say the error rate is between 1 in 100 and 1 in 200.

All of which makes DNA expert-witness testimony about the accuracy of DNA tests meaningless.  When an expert testifies that the chances of a random match are 1 in a million, or whatever, he’s talking about the chances excluding the possibility of human error.

Human error, however, can’t be excluded.


Leave a Reply

Security Code:

%d bloggers like this: