In the Lucia de Berk case many things went wrong. One of the mean things that was wrong interpreted was the statistics. First the Juliana Children hospital calculated the probability that is was a coincidence that during a work shift of Lucia de Berk there was an incident. – An ‘incident’ in this manner is when someone needed to be resuscitated. – The result of this calculation was a probability of 1 out of 700.000.000. Later Professor Elffers calculated it again and he came to the conclusion that is was 1 out of 342.000.000. This is still a very small probability.
The court (alfa) made the following statement about this evidence: ‘…Out this calculation follows that it is very unlikely that the suspect was there by coincidence when the in the indictment named incidents […] took place. The calculation give therefore a high significance that there is a correlation between the activities of Lucia de Berk and the incidents that took place.’
But this is a misunderstanding about what small probabilities mean. A small probability doesn’t tell us anything about how probable that coincidence is. So it didn’t made sense that the court said: ‘that it is very unlikely that the suspect was there by coincidence’.
This lack of knowledge about statistics on the alfa-part, brings the beta-part (and the alfa-part itself) in trouble.
This is an interesting example of misunderstanding science, because statistics is an important field in science. What makes this particular case so special is that this piece of evidence was first one of the mean reasons Lucia got locked up in jail, but later it completely disappeared from the list of evidence. And maybe even more interesting is that not only the judge interpreted the evidence wrong, but also were the calculations wrongly performed. In this example it is very clear how sensitive scientific evidence is.
(Erasmus University Rotterdam, 12-11-07)