Wednesday, December 28, 2011

Science in Society

Miscarriage of justice points to fingerprint flaws


The process of fingerprint analysis can sometimes smear judgement <i>(Image: Monkey Business Images/Rex Features)</i>
The process of fingerprint analysis can sometimes smear judgement (Image: Monkey Business Images/Rex Features)

FINGERPRINTS were once the cornerstone of forensic identification. Now a report into a miscarriage of justice has renewed pressure on print examiners to improve their methods, while two new studies reveal the extent of their fallibility. The results could change the fingerprint profession worldwide.
The Fingerprint Inquiry was launched by the Scottish government after detective Shirley McKie was acquitted of perjury. Flawed fingerprint analysis was the only evidence against her. The report, published on 14 December, concludes that human error was to blame and voices serious concerns about how fingerprint analysts report matches. It recommends that they no longer report conclusions with 100 per cent certainty, and develop a process for analysing complex, partial or smudged prints involving at least three independent examiners who fully document their findings.
The recommendations lay bare fundamental problems which have demanded attention for decades, says Jim Fraser, a forensic scientist at the University of Strathclyde in Glasgow, UK. In interviews with 400 fingerprint experts from around the world, he found that 80 per cent believe fingerprint identifications are reliable. And although several recent cases have hinted that fingerprint analysis should be treated with caution, it remains a mainstay of forensic identification.
Following calls to address forensic methods by the US National Academy of Sciences in 2009, a team led by Itiel Dror of University College London has investigated the weaknesses in fingerprint examination. In one study, they looked at whether allowing examiners to see a suspect's print when analysing one taken from a crime scene - common practice in fingerprint labs - would influence their judgement.
They gave 20 analysts from different countries 10 crime-scene prints, five of which were accompanied by a suspect's prints, and asked them to count "minutiae", or distinguishing characteristics, in the crime-scene prints. Seeing a suspect's print usually reduced the number of minutiae found. Examiners also identified different minutiae. Another group, of 10 analysts, differed in their conclusions about 10 prints when shown the same set several months later (Forensic Science International, DOI: 10.1016/j.forsciint.2010.10.013). Dror likens it to a doctor giving two diagnoses - one benign, one fatal - for the same X-ray image.
In a separate study, he looked at how automated fingerprint identification systems (AFIS), which are commonly used to trawl print databases and provide a list of possible matches, could sway judgement. He gave 23 examiners a print and AFIS lists of possible matches - but in some of the lists the position of the actual match on the list was changed.
Examiners are meant to judge all AFIS-generated prints as possible matches, but Dror found they spent longer looking at prints at the top of the list. When the genuine match was lower down, they were more likely to dismiss it and identify a print at the top of the list as the correct match (Journal of Forensic Science, DOI: 10.1111/j.1556-4029.2011.02013.x).
"This is not a fatal blow to fingerprints, but it means we need to take countermeasures to minimise the effects of bias," says Dror. In addition to the recommendations of the Fingerprint Inquiry, he says examiners should always analyse crime-scene prints and document their findings before seeing a suspect's print, and should have no access to other contextual evidence.

http://www.newscientist.com/

No comments:

Post a Comment