The police brush the crime scene for fingerprints. They crunch the numbers. One name tops the list as a match for the physical evidence: yours.
But what if you weren鈥檛 even there? In the absence of a cast-iron alibi, your protestations aren鈥檛 likely to do you much good. The idea put forward in the 鈥渂ible of forensic science鈥�, Paul L. Kirk鈥檚 1953 manual Crime Investigation, still holds sway today: 鈥淧hysical evidence cannot be wrong.鈥�
Advertisement
But the human interpretation of that evidence can be very wrong indeed. According to New York鈥檚 Innocence Project, an organisation devoted to exonerating the wrongfully convicted, has contributed to half the convictions overturned thanks to DNA evidence in the US since 1989.
Some techniques still used in the prosecution of capital offences, such as the interpretation of bite marks, are based on junk science. 鈥淪tudies have shown that experts can鈥檛 tell whether a bite mark is even human, let alone who left the mark,鈥� says , head of the Forensic Science Center of Excellence at Iowa State University in Ames.
In other areas, such as the analysis of shoe prints and tool marks, matches can be identified more accurately, but what that means is still unclear. 鈥淭rying to compute probability without really understanding what the frequency of different types of shoes and sizes in the population is 鈥� it鈥檚 practically impossible,鈥� says Carriquiry. Efforts are therefore under way on both sides of the Atlantic to create large footwear reference databases, allowing for more realistic calculations of the chances of getting a match.
But even when such databases exist, as with fingerprints 鈥� and even some types of DNA evidence 鈥� the definition of what constitutes a 鈥渕atch鈥� can lie in the eye of the beholder. Which is bad news for whoever the examiner thinks of as the prime suspect. Consequently, computer algorithms are now being developed to aid human fingerprint analysts in determining the best match.
46% of overturned convictions examined by the US聽Innocence Project involved dodgy forensics*
Itiel Dror, a cognitive neuroscientist at University College London, believes things could be further improved by reversing how matches are obtained. He is campaigning for evidence to be examined in isolation from the suspect, and only compared directly once a forensic profile has been built. 鈥淭his way the evidence is driving the comparison, rather than the suspect,鈥� says Dror. Although the need for this safeguard has been accepted by regulators in both the US and UK, it has yet to filter down to all labs carrying out forensic work. 鈥淢any forensic examiners and crime scene investigators still do not know about bias and how to minimise it,鈥� Dror says.
And these biases are just as likely to undermine any new techniques, no matter how powerful they might be. 鈥淚t鈥檚 not only specific methods that we need to develop, but general principles,鈥� Carriquiry says.
*Out of a total of 342 convictions overturned in the US since 1989, according to the US Innocence Project. Multiple factors were involved in many of these cases.
Bad science can put you behind bars 鈥� read more of our special report to stay free
This article appeared in print under the headline 鈥淎t the scene of the crime鈥�
