How Is “Error” Defined in Forensic Science?
by Moderator on Jun.25, 2010, under Topics for discussion
How is “error” defined in forensic science? What are the various types of error? What does “error rate” mean? If an investigator knows the exact error rate for a technique, how can that information be used in interpretation of evidence and testimony?
December 27th, 2010 on 7:01 am
So far, the “error” has only been defined as the examiner making a wrong call on the hypotheses. The real error of forensic science is however that the examiner would make such a call at all.
As he is not in a position to know about the posterior probability of the hypotheses (let alone taking decisions on them!), such error rates only have some meaning in comparing the outcome of proficiency tests between individual examiners.
If it turns out to be impossible to have an examiner interpret his observations properly (e.g. due to lack of scientific level), one might regard the “utterance of the expert” as the evidence. Under such (regrettable) circumstances false positives and negatives etc could be used to obtain a likelihood ratio for the (yes / no) utterance of the expert.
However, I would hope that such an approach would be reserved for scent dogs, who have the excuse of a rather limited vocabulary (bark / no bark).