The Inns of Court College of Advocacy (ICCA) and Royal Statistical Society (RSS) Guidebook for Advocates on the Use of Statistical Evidence in Court

This important and detailed guidebook is designed to identify the main pitfalls that advocates are likely to meet when handling statistics and so help us to avoid the miscarriages of justice that can occur – and have occurred – because of misunderstandings, and inappropriate use, of statistics and probability in court.

Taking the simple, famous but salutary example of a mother who finds herself prosecuted for the murder of her two children.

On the results of a relevant study, the chances of a child dying as a result of Sudden Infant Death Syndrome (SIDS) in a similar (non-smoking, professional etc.) family is 1 in 8453. The prosecution proceeds, relying on expert evidence from a paediatrician, on the basis that the probability of two children dying from SIDS is 1 in 8453 x 8453 = 1 in 73 million. The mother is then convicted on that statistically flawed basis because counsel, solicitors and the trial judge failed to recognise it as statistically flawed.

The example teaches an important lesson about the importance of independence: It is only where two outcomes are independent (and not related, in this example by other genetic or environmental factors) that one can multiply them to get to the probability of both occurring.

Another important example addressed in the guidebook is the so-called prosecution fallacy, which arises most obviously in circumstances of a DNA match. Taking and analysing two statements in turn:

1. “The probability of a match if the semen came from another person is one in a billion” This is unobjectionable. In statistical terms, it answers the question: “Assuming the defendant is innocent, what would be the probability of finding this match?”; and
2. “The probability that the semen came from another person is one in a billion”. This is objectionable. In statistical terms, it answers the question, “Assuming the match has been made with the defendant, what is the probability of his innocence?”, which is a question which of course no expert should be asked or should purport to answer.

Section 2 of the guidebook “Key Concepts in Statistics and Probability” will be a helpful place to dip into whenever statistics and probability arise in a case. Section 3 “Into Practice” provides a useful list of Guidelines for experts, condensing the various rules of court and good practice on this subject. It concludes with a series of case studies, including one in the criminal sphere, to assist further with the practical application of these important principles.

The overall message is clear: If we are to meet our professional responsibilities, we need to get to grips with statistics and probability. We need to understand at least enough to be able to challenge and probe the underlying statistical bases of expert opinion, rather than hide in the more familiar territory of experts’ respective credibility or experience.

The guidebook itself is available to download on the ICCA and RSS websites.