The concerns expressed by Law Commission about the quality of some of the expert evidence admitted before the criminal courts are to be welcomed. Recent miscarriages of justice, particularly in relation to suspected infant homicides, have highlighted the dangers of unreliable expert evidence being placed before the courts.
The proposed partial solution, of a statutory admissibility test for the reliability of expert evidence, may, however, mark a wrong turning in the development of English law. This is for at least five reasons:
(1) There is no clear principle that evidence generally may be excluded because of its unreliability, and it is undesirable as a matter of principle to make ad hoc rules for particular categories of evidence. The discussion in Appendix A suggests the existence of only occasional dicta. In particular, the dictum of Hoffman LJ cited from Vernon v Bosley does not appear to be well formed. We might alternatively consider screening hearings for all evidence which is potentially unreliable, such as eye witness identification evidence, rather than just expert evidence.
(2) The American experience with Daubert suggests that judicial assessment of the reliability of expert evidence, at least on a formal basis, is deeply problematic, both theoretically and practically. Although the Paper suggests that little expert evidence is likely to be excluded under the proposed rules, the US experience is of the generation of an extraordinary volume of satellite litigation, first instance and appellate decisions, and academic commentary. Daubert is one of the few cases for which one can buy a t-shirt (http://www.cafepress.com/daubertontheweb). Of particular relevance here is that there appears a broad academic consensus that Daubert poorly serves the needs of the US criminal justice system. See for example, the Symposium on ‘Daubert, Innocence and the Future of Forensic Science’ in (2007) 43(2) Tulsa L Rev 229-626. The Paper is probably wrong to suggest that the philosophical objections to the Daubert factors are over-stated; rather, the practical problems with a Daubert approach arise in large part from its underlying conceptual difficulties.
(3) The concept of ‘unreliability’ in relation to expert evidence needs to be unpacked, lest we conflate very different causes, such as unreliable experts and uncertainty within disciplines. For example, on p 79 it is suggested that Clark and Cannings are both example of miscarriages of justice arising from unreliable evidence. However, in Clark the unreliability arose from the negligence of one of the experts, while in Cannings the unreliability arose from profound disagreement between reputable experts.
(4) The probative evaluation of evidence is correctly the role of the jury, and a Daubert-style screening by the judge would be to usurp that role. Associated with this, we should be slow to believe that judges are better placed than jurors to evaluate reliability. See, for example, R Munday, ‘Case Management, Similar Fact Evidence in Civil Cases, and a Divided Law of Evidence’ (2006) 10 Int J Evidence and Proof 81-103. On the ability of non-specialists to evaluate expert evidence at trial, see D Dwyer, The Judicial Assessment of Expert Evidence (CUP, 2009) ch 2. At the same time, the role of the judge is changed, to become actively involved in the evaluation of evidence.
(5) There are no clear examples of English criminal cases where a miscarriage of justice would certainly have been avoided if the proposed approach had been in place. If one thinks of a case such as Clark, where much of the evidence of a leading paediatric expert was problematic, defence counsel made little effort at trial to challenge the reliability of the evidence, for example in cross examination. The paper gives some examples of bogus experts, to support a subsidiary suggestion of the accreditation of experts, but these experts would readily have been exposed as bogus by any legal team making simple enquiries.
If we were to proceed with a reliability test for expert evidence, the American experience would appear to suggest that the test for reliability should be as flexible as possible, and left as far as possible to the discretion and common sense of the judge. Much of the Daubert satellite litigation and academic debate arises from an examination of the minutiae of the meanings to be attached to the four Daubert factors. A more effective approach, if we accept judicial screening of evidence, would be to focus on the training of judges and counsel to identify potentially weak expert evidence.
The provisionally proposed wording at para 6.10 is well thought out, and a marked improvement on the wording in Daubert. However, the use of discretionary words such as ‘sound’, ‘properly’ and ‘sufficiently’ may lead to unnecessary satellite litigation.
The proposed wording of the guidelines at paras. 6.26 and 6.35 is similarly well thought out. It may, however, but unduly prescriptive. The proposed distinction here between scientific and experience-based expert evidence may not be as clear cut as statutory guidelines would require. For example, it is unclear why ear-print identification evidence should be tested as scientific expert evidence, while handwriting identification evidence should be treated as experience based.
The proposed use of court-appointed assessors, although perhaps attractive at first blush, is unlikely to succeed. The experience of the English civil courts since 1999 has been of very limited use of assessors. To ask assessors to make decisions on the reliability of expert evidence might be to give them a judicial function that does not sit comfortably with current views in the civil courts on the role of the assessor as an expert witness.