Sally Clark

Sally Clark 1964-2007

Sally Clark

More about the statistics and the appeal

One part of Sally's appeal focussed on the appalling `one in 73 million' statistic quote by Roy Meadow. The defence argued

  1. that this figure was wrong, since it was based on an absurd squaring of the probability of one death; and
  2. that there was cause for concern that the jury might have been misled by the `prosecutor's fallacy', which, in a nutshell, runs like this: `natural double infant deaths are very rare (hitting one family in 100,000, say); so if you find a family with two deaths, the odds are 100,000 to 1 that the deaths were unnatural'; and
  3. that Roy Meadow himself founded his opinion that the deaths were unnatural in part on the statistical evidence in breach of established legal guidelines.
The prosecution countered these two assertions in a remarkable way, which Roy Meadow (the `expert witness' who quoted the statistic) repeated in the January 2002 issue of the British Medical Journal [BMJ 324 (7328): 41]. In a nutshell, they said `everyone agrees the statistic was irrelevant, so its role in the case was irrelevant'.

In more detail, the prosecution's tack was to say "the statistic concerns the probability of two SIDS deaths (a particular type of unexplained natural death); the prosecution don't believe the deaths were SIDS deaths; and the defence also agree that the two deaths were not both SIDS deaths (because the defence suggest that Christopher's death may have been, as originally recorded by the doctor, caused by lung infection, which would mean it was not SIDS; and Harry's death may have been caused by vaccination - he died four hours after a vaccination!); so since neither prosecution nor defence thinks a double-SIDS is the explanation, the probability of double-SIDS is completely irrelevant to the case. The jury should have ignored this evidence."

[This argument by the prosecution raises the question `if the prosecution views this evidence as irrelevant, why did they not only introduce it, but also continue quoting it right through to the summing up of the trial?]

Nevertheless, the appeal judges bought the prosecution's argument. They concluded that the prosecutor's fallacy had not been committed. They accepted that inaccurate statistical arguments had been made, saying they were `not impressed' with mistakes made by the prosecution. One of their main reasons for concluding that the 73-million-to-one figure had not misled the jury went as follows:

[We] are not persuaded that ... the Crown ... submitted to the jury that the odds against the appellant being innocent were 73 million to one against. That submission would in our judgment have been obviously fallacious, and had it been made, we would have expected Mr Bevan for the defence to have objected, the judge to have upheld the objection, and the 1 in 73 million figure would have gone as an unnecessary distraction. That there was no such application suggests the lack of impact of [the 1 in 73 million figure quoted in the summing-up].
The judges found comfort in the idea that, while inaccurate, the 1 in 73 million figure was only ever taken by the court as a `ballpark figure'. This attitude is criticised in the Royal Statistical Society's press release of 23 October 2001.

So, in a nutshell, the appeal judges said `the prosecution can't possibly have used statistics wrongly because the correct way to use statistics is so obvious that the defence would surely have objected'.

The appeal court judges concluded that "In the context of the trial as a whole, the point on statistics was of minimal significance and there is no possibility of the jury having been misled so as to reach verdicts that they might not otherwise have reached."

David MacKay Mon 7/1/02

Site last modified Wed Oct 12 09:58:06 BST 2011