Regina versus Denis John Adams

A rape trial in the United Kingdom in 1996. The woman described her attacker as in his 20s. A suspect, Denis Adams, was arrested and an identity parade was arranged. The woman did not pick him out, and on being asked if he fitted her description she said No. She had described a man in his 20s and when asked how old Adams looked, she replied about 40. Adams was 37. However, his DNA profile fitted that of evidence left at the scene. The prosection forensic scientist calculated the DNA match probability as 1 in 200 million. Adams had an alibi for the night in question, as his girlfriend said he had spent the night with her. The DNA was the only incriminating evidence heard by the jury, as all the other evidence pointed towards innocence. The defence argued that the match probability figure was incorrect, and that a figure of 1 in 20 million, or perhaps even 1 in 2 million, was more appropriate. The issue of how the jury should resolve the conflicting evidence was addressed by the defence by a formal mathematical method. The jury was instructed in the use of Bayes's theorem by Professor Peter Donnelly of Oxford University. The judge told the jury they could use Bayes's theorem if they wished. Adams was convicted, but the case went to appeal. The Appeal Court judges noted that the original trial judge did not direct the jury as to what to do if they did not wish to use Bayes' theorem and ordered a retrial. The defence lawyers again wanted to instruct the new jury in the use of Bayes's theorem (though Prof. Donnelly had doubts about the practicality of the approach [1]). The judge asked that the statistical experts from both sides work together to produce a workable method of implementing Bayes' theorem for use in a courtroom, should the jury wish to use it. A questionnaire was produced which asked a series of questions such as "If he were the attacker, what's the chance that she would say he looked nothing like the attacker?" Also, "If he wasn't the attacker what's the chance that she would say he looked nothing like the attacker?" These questions were intended to allow the Bayes factors of the various pieces of evidence to be assessed. The questionnaires had boxes where jurors could put their assessments and a formula to enable them to produce the overall odds of guilt or innocence. Again, Adams was convicted and again an appeal was made. The appeal was unsuccessful and the Appeal Court ruling was highly critical of the appropriateness of Bayes's theorem in the courtroom.

Points to note

The only evidence against Adams was the DNA evidence. His age was substantially different from that reported by the victim, the victim did not identify him and he had an alibi which was never disproved. The 1 in 200 million match probability calculation did not allow for the fact that the perpetrator might be a close relative of the defendant - an important point, since the defendant had a half-brother in his 20s whose DNA was never tested.

The Court of Appeal subsequently produced guidelines for the way match probabilities should be explained to jurors. Judges should say something along the lines of the following. "Suppose the match probability is 1 in 20 million. That means that in Britain (population about 60 million) there will be on average about 2 or 3 people, and certainly no more than 6 or 7, whose DNA matches that found at the crime scene, in addition to the accused. Now your job, as a member of the jury, is to decide on the basis of the other evidence, whether or not you are satisfied that it is the person trial who is guilty, rather than one one of the few other people with matching DNA. We don't know anything about the other matching people. They are likely to be distributed all across the country and may have been nowhere near the crime scene at the time of the crime. Others may be ruled out as being the wrong sex or the wrong age group."

References

[1] Donnelly, P. Appealing Statistics, Significance, March 2005, Volume 2, Issue 1