Footwear Database analysis by Experts using Bayesian statistics (R v T [2010] EWCA Crim 2439)

Published: 15 November 2010

The Court of Appeal has handed down a redacted ruling in the case of R v T [2010] EWCA Crim 2439. The basis for the judgment is the reliance on an insufficiently large database, the Footwear Database, and the failure to disclose any statistical analysis used in coming to an expert opinion.

The ruling only applies to evidence based on the Footmark Database. The principles should not, at this stage, be applied to other database analysis.

Footmark comparison evidence is still admissible and experts can still give subjective, evaluative opinion based on their examinations.

For current cases involving footmark comparison evidence, where experts are using the Bayesian statistical approach to support their findings, they should be asked to provide a further statement clarifying any statistical basis they have used. Where forensic science providers are using this approach, they should provide a summary of their data and interpretive methods.

Past convictions are safe unless and until set aside by an appellate court.

Please note the following specific points which support a narrow interpretation of this case, in particular pending the re-trial which is expected to take place early 2011:

  1. The ruling confined itself "solely" to the Footwear Database placed before the Court (see paragraphs 16 iii) and iv), and 32).
  1. This was in fact the Forensic Science Service Ltd. Footwear Database; other such Footwear / mark databases do exist and the reliability of those was not considered by the Court, although the statistical approach to such data was.
  1. The judgement concluded (in paragraph 86) "We are satisfied that in the area of footwear evidence, no attempt can be realistically made in the generality of cases to use a formula to calculate the probabilities."
  1. Paragraph 46 contains a reference to the submission that the Bayes statistical approach should not be used in databases other than DNA; it is important to note that the Court did not rule on this submission and went on to say (paragraph 91) "It is not necessary for us to consider the matter further and consider how likelihood ratios and Bayes theorem should be used where there is a sufficient database" (emphasis added).
  1. Please also note paragraph 90 refers to databases (other than DNA databases) where the statistics are reliable; this is a clear reference to databases existing which do / may have a suitably reliable set of data. It should be remembered that since R v Doheny was heard in 1997, many other reliable forms of databases have been established.
  1. Very clear guidance was given to the issue at the heart of this case in paragraph 70 indicating that the Case Management guidance given by the Court of Appeal in R v Reed and Reed [2009] EWCA Crim 2698at paragraphs 111 – 113 should always be adopted in determining admissibility (as opposed to assessments of the evidence, which should be left to the jury). These paragraphs prescribe the steps that should be taken by the prosecution team in applying Criminal Procedure Rules 2010, rule 33.
  1. R v T application is limited to expert evidence in footwear mark cases and clearly requires effective case management to be used in ensuring that if an expert has used a statistical formula to assist them in reaching their conclusion, they must specifically refer to the details of the formula and the impact its use has had on their conclusion. In short, experts are required to be transparent about the statistical data and methods used in coming to their conclusions.
  1. Footwear comparison evidence remains admissible and experts can still give subjective, evaluative opinion based on their experience and examinations.

Karen Squibb-Williams

Strategic Policy Adviser

CPS Headquarters