A prison sentence is in the offing if you can’t show that what happened was an accident. An expert witness takes to the stand and announces that the chances of the event being accidental are 1 in 73 million. Your lawyer takes a sharp intake of breath, and then bumbles his cross-examination. Convinced by the statistic, the jury convicts you.
Who can blame them? Challenging a statistic – particularly one delivered by an expert – can be intimidating. And many people have a pathological fear of numbers: lawyers, jurors and judges are no exception.
Yet courtroom statistics are often misleading, as revealed in the case of Sally Clark – a British solicitor convicted, and eventually cleared, of murdering her eldest two children. In this case, paediatrician Roy Meadow claimed that the chances of two children in the same family dying of sudden infant death syndrome was 1 in 73 million – a statistic that originally came from a UK government-funded inquiry into sudden death in infancy. Although the authors acknowledged neglecting genetic factors that might boost a family’s risk, Meadow didn’t spell this out in court, and no one thought to question him about it.
Advertisement
“If the lawyers and the judges don’t really understand the basics of how to interpret a piece of evidence, jurors have got no hope,” says , director of the risk information management research group at Queen Mary University of London.
And as other forensic disciplines strive to become more quantitative in their analysis of crime scene evidence (see “How faulty crime-scene forensics can make you a murderer“), the potential for misunderstanding can only grow. “The science is becoming more complex, so the interpretation has got to become more complex to keep up with it,” says , chair of the UK Royal Statistical Society working group on statistics and the law. “The numbers that are being produced in court are not answering the questions the courts want.”
95% of reviewed trials involving microscopic hair analysis by FBI experts revealed they had overstated the strength of forensic evidence*
So what to do? One approach is to throw out complex evidence altogether. Another is to educate. The Royal Statistical Society has published a series of guides aimed at teaching lawyers about statistics and probabilistic reasoning. “If the lawyers understood what was going on, they couldn’t misrepresent the numbers because the other side would pick them up on it,” says Aitken. “Many lawyers don’t know the right questions to ask.”
Fenton agrees, which is why he teaches lawyers the basic principles of Bayesian probability – the odds of something happening given the odds of other related events. “Where errors have been made, it’s usually because there have been fairly complex and related bits of evidence,” he says.
But are lawyers and jurors up to the job? Fenton argues that a degree in mathematics isn’t necessary: “We try to think about ways that you could capture the essence of these arguments without getting the lawyers or jurors to actually calculate probability.”
*Source: National Association of Criminal Defense Lawyers
Bad science can put you behind bars – read more of our special report to stay free
This article appeared in print under the headline “Expert opinion”
