Statistics and the law
Posted: Mon Apr 19, 2010 10:04 am
Lucia de Berk, a Dutch nurse who served more than six years of a life sentence for murdering seven patients under her care, has been fully exonerated. The country's attorney general personally apologised for her ordeal.
This article describes the case:
http://www.guardian.co.uk/commentisfree ... nurse-case
The Lucia de Berk case, like the Sally Clark case, is not just an example of bad statistics, it is also an example of statistical evidence being misunderstood by criminal justice professionals and juries.
These cases, like the Shirley McKie case, relied on expert testimony as the only evidence that a crime occurred as well as indentifying the perpetrator. If my experience over the last couple of years is anything to go by, criminal justice professionals have no idea how risky this is - or even why it is risky.
If fingerprint evidence is one day going to be presented in probabilistic or likelihood terms, somebody is going to have to find a way to explain to the people who use it how it should be interpreted (and how to account for the fact that experts can be wrong). This is most important when there is very little other evidence against the accused.
This article describes the case:
http://www.guardian.co.uk/commentisfree ... nurse-case
The Lucia de Berk case, like the Sally Clark case, is not just an example of bad statistics, it is also an example of statistical evidence being misunderstood by criminal justice professionals and juries.
These cases, like the Shirley McKie case, relied on expert testimony as the only evidence that a crime occurred as well as indentifying the perpetrator. If my experience over the last couple of years is anything to go by, criminal justice professionals have no idea how risky this is - or even why it is risky.
If fingerprint evidence is one day going to be presented in probabilistic or likelihood terms, somebody is going to have to find a way to explain to the people who use it how it should be interpreted (and how to account for the fact that experts can be wrong). This is most important when there is very little other evidence against the accused.