Had Enough of Experts?

Anthony Williams makes the case for a reliability test to admit expert evidence in criminal proceedings.


Forensic science has progressed rapidly throughout the last century. Forensic experts have developed scientific and technological analysis methods which have assisted to secure convictions where crimes may previously have been totally unsolvable, with the true culprits never caught. However, with the prevalence of the “CSI effect” on juries in relation to criminal proceedings reliant on forensic evidence, there is heightened importance on admitting only that expert evidence which is reliable and scientifically valid. Rapid development in forensic science and technology require an equally rapid and robust response from the criminal justice system.

In this article I intend to explore the current law in relation to expert evidence in criminal proceedings. I will then explore the nature of the proposals which were made by the Law Commission of England and Wales (‘the Law Commission’) and the government’s response. I then intend to explore examples of cases in which miscarriages of justice have continued to occur as a result of expert evidence which is unreliable or has questionable probative value. Using these examples, I hope to demonstrate the continuing need for a more robust reliability test and better measures to ensure that expert evidence of limited value is excluded before it reaches a jury at trial.

Expert Evidence – The Current Law

In criminal proceedings, opinion evidence is almost exclusively the realm of expert witnesses. Expert evidence will only be admitted where it is necessary because the expert is able to give an opinion on a matter outside of the knowledge and experience of the tribunal of fact. Otherwise, it is irrelevant and therefore inadmissible (Turner [1975] QB 834).

In terms of the reliability of expert evidence, Thomas LJ identified three principles based on established case law when giving judgment in the case of Reed, Reed & Garmson [2009] EWCA Crim 2698 (at [111-113]):

“First, expert evidence of a scientific nature is not admissible where the scientific basis on which it is advanced is insufficiently reliable for it to be put before the jury… Second, even if the scientific basis is sufficiently reliable, the evidence is not admissible unless it is within the scope of evidence an expert can properly give… Third, unless the admissibility is challenged, the judge will admit that evidence.”

As to the first principle, in Clarke [1995] 2 Cr App R 425, a case which concerned expert evidence regarding facial mapping techniques, the Court observed that there were “no closed categories” of expert evidence which could be placed before a jury, as it would be “entirely wrong to deny to the law of evidence the advantages to be gained from new techniques and new advances in science” (per Steyn LJ at 429-430). This passage was quoted with approval in Luttrell [2004] 2 Cr App R 520, a case featuring lip-reading evidence. In that case the Court of Appeal re-affirmed that English courts were “flexible” in relation to admitting expert evidence.

However, Gilfoyle [2001] 2 Cr App R 57 is an example of a case in which the courts were willing to exclude expert evidence which was not scientifically reliable. This was a case in which the Defendant was accused of murdering his pregnant wife, and the Defendant sought to adduce fresh evidence of an expert’s “psychological autopsy” relating to the mental state of the deceased. The Defendant’s case was that this would demonstrate that his wife had committed suicide. The court declined to admit such evidence on the basis that not only would this have been the expert’s first time providing such evidence, but also that “his reports identify no criteria by reference to which the Court could test the quality of his opinions” (per Rose LJ at [25]).

As to the second principle, whether evidence is within the scope of that which an expert can properly give usually depends on the expert’s area and level of expertise. There are no strict criteria to determine whether someone is an expert in a given case. In Silverlock [1894] 2 QB 766, it was held that a solicitor with ten years of experience studying handwriting was able to give expert opinion evidence as to whether a document was written in the handwriting of the accused or not. By contrast, in Inch (1990) 91 Cr App R 51 it was ruled that opinion evidence as to cause of injury given by a medical orderly with experience of treating superficial injuries should not have been treated as expert evidence, and instead evidence from a properly qualified medical expert should have been sought.

As to the third principle, where the Crown seek to rely upon expert evidence, it is incumbent upon the defence to raise the issue of whether an expert’s evidence is admissible and seek a hearing on the voir dire for that issue to be determined. It is not for the court to order a hearing on the voir dire of its own motion (Francis [2013] EWCA Crim 123). However, once there is an objection to the admissibility of that expert evidence, its admissibility must be proved by the party seeking to rely upon it (Atkins [2009] EWCA Crim 1876).

Law Commission Report

In 2011 the Law Commission published its report entitled Expert Evidence in Criminal Proceedings in England and Wales following concerns that “expert opinion evidence was being admitted in criminal proceedings too readily, with insufficient scrutiny”.  

The Law Commission highlighted examples of miscarriages of justice resulting from “unreliable” expert evidence. Those examples included the convictions of Sally Clark and Angela Cannings. Both were convicted of murdering their own infant children based on a wide range of expert medical evidence which was later discredited or cast into doubt.

In Clark’s case this included expert opinion evidence given by Dr Alan Roy Williams, a Home Office accredited pathologist who had conducted post-mortems on both of her deceased children and gave evidence that injuries were non-accidental. It later transpired that Dr Williams had failed to disclose microbiological reports he had considered, which the court opined was “likely to mislead others, who may work on the same case and who will be denied the opportunity of considering the material… in reaching their own properly informed conclusions” (Clark [2003] EWCA Crim 2010, per Kay LJ at [138-171]).

Both cases also included evidence from Professor Sir Roy Meadow which purported to exclude the possibility of two children within the same family dying of Sudden Infant Death Syndrome (‘SIDS’), implying that two such deaths must therefore have been murder.

The Law Commission also referred to the case of Mark Dallagher who was convicted based on expert evidence in relation to ear print identification. This conviction was subsequently cast into doubt when scientific opinions were later published diminishing the forensic value of ear print evidence (Dallagher [2002] EWCA Crim 1903).

Proposals for Reform

The Law Commission made a range of proposals which included the introduction by primary legislation of a statutory reliability test which would provide that “expert opinion is admissible only if it is sufficiently reliable to be admitted”. Expert opinion evidence would be deemed sufficiently reliable if, under a two-part test, the opinion is “soundly based”, and “the strength of the opinion is warranted having regard to the grounds on which it is based”.

The Law Commission also considered that any legislation should give “high-order” examples of reasons why expert opinion evidence would not be deemed reliable. These included where the opinion was based on hypotheses which had not been subjected to sufficient scrutiny or had failed to stand up to scrutiny, or where the opinion relies on an inference or conclusion which has not been properly reached.

It recommended in addition that the common law as discussed above be codified, so that expert evidence is only admitted where an expert’s evidence is likely to be necessary and the expert is qualified “on the balance of probabilities” to give such evidence.

The Law Commission did recommend that there should be a power to disapply the reliability test so that it was not applied dogmatically, but that such a power should be limited to prevent the test from becoming a “nominal barrier” to the admission of unreliable evidence.

Government Response

The government published its formal response on 21 November 2013. In summary, the government declined to introduce the full reforms proposed by the Law Commission. The government recognised that the proposal would “enhance the existing common law rules, and potentially increase the likelihood that unreliable expert evidence would be excluded from criminal trials”. However, the government cited cost concerns with regards to their implementation and concluded that any savings promised by the Law Commission were not guaranteed.

Instead the government determined to refer many of the recommendations to the Judicial College and the Criminal Procedure Rule Committee to consider whether amendments would be needed to the Criminal Procedure Rules (‘the CrimPR’) and the Crown Court Bench Book to reflect the concerns raised in the report.

Current position

Amendments were made to what was then CrimPR 33 to set out in greater detail the overriding duty owed to the court by experts and the required content of an expert’s report, to include details of their qualifications, any person who had conducted tests for the purposes of the report, discussion of any range of opinion and reasons for their own, “such information as the court may need to decide whether the expert’s opinion is sufficiently reliable to be admissible as evidence”.

In 2015, further reforms were brought about to the CrimPR, resulting in the new Part 19 relating to Expert Evidence. This largely mirrors the provisions within CrimPR 33 regarding the expert’s overriding duty to the Court and the contents of expert reports. However, it remains the case that the defendant must take issue with the admissibility of expert evidence by seeking a hearing on the voir dire.

Problems with current approach – novel science

One problem with the current approach relates to the question of what can be regarded as scientifically reliable and how that is measured. Currently it is the role of the Forensic Science Regulator to ensure that forensic science services in the United Kingdom meet the appropriate standards of scientific quality. However, as described above the courts are hesitant to deprive the criminal justice system of new methods of detecting crime. The Law Commission has even described the courts’ historical approach to the admission of expert evidence as “laissez-faire”.

The difficulty with allowing expert evidence to be admitted based on scientific techniques which are still developing is that flaws may not be identified until it is too late to prevent a miscarriage of justice, as demonstrated by the case of Dallagher above.

One developing area in forensic science is the use of proprietary software to analyse mixed DNA profiles and provide a likelihood ratio of a match with the Defendant’s DNA. One such example is TrueAllele’s probabilistic genotyping software. This type of software purports to be able to effectively separate individual DNA profiles from a mixed sample with multiple contributors and provide a likelihood ratio as to how much more likely it is that those partial profiles match the DNA profile of a Defendant than not. This type of software has been in use in the UK for many years, and has recently made regular appearances in Crown Court cases involving low-level mixed DNA profiles.

Such technology has been the subject of controversy overseas. In 2015 in Queensland, Australia an error was found in the coding of STRmix (another brand of probabilistic genotyping software) which affected the calculation of likelihood ratios. Out of 4500 cases in which STRmix had been used, it was found that likelihood ratios had been affected in 60. In 24 of those cases, new witness statements from the experts who gave evidence were required because the likelihood ratios were lower than those originally given. In 2016 during the trial of Oral Hillary in the US for murder, both TrueAllele and STRmix were used to analyse the same set of DNA samples, but gave different results. One exonerated Hillary and the other implicated him. Hillary was acquitted.

This growing field and its pitfalls have seen little scrutiny in the courts of England and Wales. The case of Tsekiri [2017] EWCA Crim 40 involved a robbery trial in which the only evidence connecting the Defendant to the crime scene was a mixed DNA profile found on the handle of a car door. Davies LJ ruled that even in the absence of any other evidence connecting a defendant to a victim, crime scene or object, “the fact that DNA was on an article left at the scene of a crime can be sufficient without more to raise a case to answer where the match probability is 1:1 billion or similar” (per Davies LJ at [14]). However, the Court did not concern itself with the question of how the mixed DNA profile was analysed, or how that match probability ratio was calculated. Commentators have observed this represented a missed opportunity to critically evaluate the limitations of this type of expert opinion evidence.

Problems with the current approach – experts self-regulating

A second issue with the current approach is the reliance on experts to conduct themselves in accordance with their duties to the court. In effect, experts police their own evidence in the absence of further scrutiny by the courts or the Forensic Science Regulator. CrimPR 19.2 sets out the expert’s duty inter alia to “help the court to achieve the overriding objective” by giving evidence which is objective, unbiased, and within their area of expertise. This appears to be clear and unambiguous, but there are numerous examples of cases in which experts have purported to give opinions in fields outside their areas of expertise.

In the field of medicine, one example is that of Professor Sir Roy Meadow mentioned above. Meadow subsequently became the subject of General Medical Council (‘GMC’) fitness to practise procedures and was erased from the List of Registered Medical Practitioners (‘LRMP’). Although it was held on appeal that his conduct did not constitute serious professional misconduct and the erasure decision was reversed, the Court of Appeal expressed the view that he was “undoubtedly guilty of some professional misconduct”. It found that he had expressed an opinion firstly based on “statistical figures of uncertain source and scientific validity”, and then on figures which “he misunderstood and, by implication and the use of an inappropriate analogy, misapplied” (Meadow v General Medical Council [2006] EWCA Civ 1390, per Sir Anthony Clarke MR at [251]).

In contrast, a finding of serious professional misconduct was made against Dr Williams for his failure in Clark’s case to disclose microbiological reports which he had considered in reaching his conclusions (amongst other clinical failings in carrying out post-mortems generally), which was upheld by the High Court on appeal (Williams v General Medical Council [2007] EWHC 2603). Although Professor Meadow and Dr Williams are examples of experts whose flawed evidence was detected and eventually dealt with, it is notable that this was not until after the Crown Court proceedings had taken place, too late to prevent an initial miscarriage of justice. It has been said that the consequences in Clark’s case were fatal.

A more recent example of a witness who purported to give expert evidence but was found to be acting outside their competence and expertise is Dr Waney Squier. Dr Squier was a neuropathologist who was significantly involved in cases in which certain injuries suffered by babies were alleged to be indicative of Shaken Baby Syndrome (‘SBS’) rather than accidental injury. However, she disputed the medical consensus regarding those injuries and the existence of SBS generally.

In 2016 Dr Squier was the subject of Medical Practitioners Tribunal (‘MPT’) proceedings, which found her fitness to practise to be impaired as a result of her involvement as an expert witness in six cases and she was erased from LRMP. Although she was cleared of dishonesty on appeal, Mitting J ruled that the MPT had been entitled to find that Dr Squier’s fitness to practise was impaired on the basis she had failed “to work within the limits of her competence, to be objective and unbiased and to pay due regard to the views of other experts” (Squier v GMC [2016] EWHC 2739 (Admin), per Mitting LJ at paragraph 82). Mitting J imposed conditions on Dr Squier’s registration including a prohibition on acting as an expert witness in any court for 3 years. Dr Squier’s case is another example of the flaws inherent in depending on independent experts to regulate themselves when purporting to give expert opinion evidence – even where the expert is a member of an otherwise highly-regulated profession.

Problems with the current approach – reliance on challenge

A further concern with the current approach is that, as the burden is upon the accused’s legal team to raise the question of whether expert opinion evidence is admissible in a given case, there may be a question as to whether they are sufficiently experienced in the subject matter of an expert’s report to challenge it.

It has been suggested that in light of the laissez-faire approach to the admission of expert evidence, better training in this field would enable more robust challenges to be made to expert evidence before it is admitted. This approach still leaves the onus wholly on the shoulders of Counsel to take issue with any expert evidence. Respectfully, this relies either on lawyers having enough understanding of the basis upon which the expert has given their report, or incurring expense by instructing further experts to assist in interpreting the original expert evidence.


It is disappointing that this government has failed to take up the Law Commission’s recommendation of implementing a statutory reliability test despite its clear benefits. It is also disappointing that the Court of Appeal has failed to take the opportunity in recent authorities to address this growing area of concern. A relatively simple legislative amendment would add a level of emphasis to the requirement for scientific evidence to be reviewed and robustly scrutinised before it is relied upon by the Crown to secure a conviction.

It is submitted that such legislation could and should go further – preliminary hearings in cases which rely solely or heavily upon expert evidence could be made mandatory in order to prevent such evidence from reaching a jury unnecessarily. It would no longer be incumbent upon the Defence to take the initiative to have unreliable evidence excluded, and the burden would be properly placed upon the Crown to demonstrate that such evidence is reliable before it reaches a jury. The outcome of this would be that cases built upon unreliable scientific evidence could be formally concluded before they reach trial, which would safeguard the public purse, prevent miscarriages of justice, and uphold the integrity of forensic science in the UK.


Other articles by Tony Williams include:

The Criminal Procedure (Amendment No.2) Rules 2018 – What’s Changing?

Tony Williams is able to give advice on any cases where a lay client is being prosecuted with expert evidence being deployed.

Central Chambers has a criminal team with experience in defending a variety of serious criminal matters involving complex scientific/forensic evidence.

If you would like to instruct Tony Williams or another member of the criminal team, please do not hesitate in contacting the criminal clerks by clicking here or by calling on 0161 236 1133