Scientist argues fingerprint evidence can never be 100 per cent accurate

A leading lecturer in psychology at the University of Southampton is casting doubt on the commonly-held belief that fingerprint analysis is infallible.

Dr Itiel Dror has worked with more than seventy fingerprint experts from the UK, USA, Netherlands, Australia and Israel over the last two years to test his theory that mistakes can occur because of the way the human brain processes information.

The issue hit the headlines when an American Muslim from Oregon was wrongly identified as one of the Madrid bombers after his prints appeared to match those taken at the crime scene.

"The mind is not a camera," says Dr Dror. "It is a dynamic machine which can distort what it sees, not a passive recorder of visual information. Perception is far from perfection."

He argues that forensic scientists, along with everyone else, cannot avoid the risks of subjective bias. Frequently they have to use their skill and expertise to decide whether a fragment of a fingerprint left at a crime scene matches a precise impression given by a suspect in a police station.

Research by Dr Dror and his students at Southampton Dave Charlton and Ailsa Peron has tested whether experts can be affected by outside influences. Five examiners were tested with examples of prints that they had previously identified as positive matches in court five years earlier. In the new external context only one of the experts agreed with his/her previous decision. Three contradicted their previous judgement and now decided that the prints are a no-match and the fourth wasn't sure. The findings will be presented at the Biometrics 2005 conference on 20 October.