Some eight years or so ago, Sally Clark was convicted of killing her two toddlers, presumably by smothering them. She was then about 30, and a married professional. Sir Roy Meadow an eminent paediatrician was an expert witness. He claimed that the alternative explanation that both had suffered cot death was so unlikely that murder was the only possible explanation. He argued that the frequency of cot death being 1:8000, the probability of two cot deaths is 1:64 million! This is thought to have greatly influenced the jury.
Sir Roy was giving expert opinion in a field in which he had no expertise, namely statistics. The defence council were remiss in not challenging Sir Roy’s competence in statistics and in failing to produce an expert of their own in the field. Both ought to be grounds for appeal but I doubt either are. Sir Roy was struck off by the Medical Council but not for his statistical claim. He was however reinstated, the judge arguing that if an expert witness were held responsible for the accuracy of his evidence, no one would agree to be an expert witness. Sally Clark spent two years in gaol but on appeal the verdict was declared unsafe, although not on the grounds of the inaccuracy or illegitimacy of Sir Roy’s evidence or of the incompetence of the defence.
The assumption behind any probability is that the selection is made at random, in this case that every toddler stands an equal chance of cot death. But that means cot death has no cause and there are no contributing factors, not merely that we know of no cause or propensity but that none exists. All those seeking to understand cot death are wasting their time since cot death strikes randomly. If though there are contributing factors then the probability of cot death in the group to which those factors apply is less than 1:8000. To claim otherwise is to assert that cot death is uniquely causeless. If there are vulnerable families they will include those that have suffered a cot death. One would then expect a relatively large number of second cot deaths, and such is the case.
The situation is similar to that with lung cancer before the link with smoking was known. Then we would have said that the probability of contracting lung cancer was given by the ratio of the number with the disease to the number of adults. But we now know that for smokers the risk was far higher. The risk for smokers was always far higher; the link did not occur because we found out about it. When we thought lung cancer was random we would have been puzzled by a high frequency of two cases of lung cancer in a household. But now we would not be surprised at all since in families where one adult smokes others are likely to smoke as well, and everyone smoker or not will be a “passive smoker”.
If every apparent second cot death is in fact murder then some first cot deaths are likely to be murders also. The notion that “cot deaths” of only children are always what they seem, and that no parent ever murders one child where there are several but always murders the lot, is implausible. So on Sir Roy’s argument there must be a substantial number of child murderers at large; a worrying thought.
But even if a second cot death is very unlikely, the very unlikely does happen – very rarely. Much depends on there being a mechanism. There is small chance of winning the lottery but somebody does win. If someone comes into your local saying “I’ve won the lottery; drinks all round”, you do not ring the police on the grounds that winning is so improbable he must have robbed a bank. No you say, ”Congratulations; mine’s a pint.”
Sally Clark is now dead. I do not know whether the wrongful imprisonment contributed but I would have thought it likely. That apart one should not spend two years in gaol for something you did not do. And Sally Clark is only one of a number of cases where a conviction based in part on Sir Roy’s “expert”, testimony, has been overturned. Of course Sally Clark may have smothered her two toddlers but the law is not concerned with what may have happened.
An appeal as I understand it, requires new evidence. There is nothing new about the incompetence of the defence or Sir Roy’s inexpert testimony. What I have written I suspect would not constitute grounds for appeal but I think it ought to.
Feb 2009
I set out my argument much as here at the time. Perhaps I ought to have tried to get something done then but I didn’t.