Expert Liability, Bias and Misconduct

Not every wrong forensic conclusion represents the same kind of failure. Courts and regulators distinguish three points on a spectrum, and the legal and professional consequences attached to each point differ substantially.

Honest error is a conclusion that falls outside the correct result but within the range of conclusions that a competent practitioner could reasonably reach given the available methods and data. Every measurement has uncertainty; every pattern-recognition discipline involves judgment calls. A fingerprint examiner who concludes a partial mark is inconclusive when another examiner would have concluded a match has made a judgment call, not a breach of duty. Provided the examiner applied the correct methodology and disclosed the basis for the opinion, honest error creates no civil or criminal liability in most jurisdictions, though it may be challenged on cross-examination or countered by a rival expert.

Negligence is a failure to meet the standard of care expected of a competent practitioner in the field. It requires a duty of care, a breach of that duty, and resulting harm. In forensic contexts, negligence can take the form of using a method that is below current professional standards, failing to disclose a known limitation of the technique, or conducting the analysis in conditions that a careful practitioner would have rejected as unsuitable. The standard against which breach is assessed is typically that of the ordinary skilled practitioner in the relevant sub-discipline, not the standard of the most meticulous specialist in the country.

Misconduct involves deliberate or reckless wrongdoing: fabricating results, falsifying records, suppressing exculpatory data, or knowingly testifying to conclusions the expert knows to be unsupported. This is not a failure of method; it is dishonesty. It can constitute perjury or a statutory equivalent, as well as professional misconduct leading to loss of registration. Large-scale misconduct by a single forensic scientist can affect hundreds of convictions, triggering mass case reviews and generating civil claims from wrongfully convicted individuals.

Common-law courts historically extended immunity from civil suit to statements made in the course of judicial proceedings, on the ground that participants in litigation must be free to speak candidly without the chilling effect of potential liability. This immunity protected witnesses, advocates, and judges alike. For expert witnesses, the doctrine meant that a forensic scientist who gave negligent or even dishonest evidence could not be sued in tort by the party who retained them or by the party harmed by the evidence.

England and Wales reconsidered this position in Jones v Kaney [2011] UKSC 13, in which the UK Supreme Court held by a five-to-two majority that expert witnesses retained by a party are no longer immune from negligence claims. The court reasoned that the policy justification for immunity, protecting candid participation in litigation, did not apply in the same way to paid party experts as it did to compellable witnesses, and that immunity from suit was inconsistent with the enforceable duty owed by the expert to the court. This was a significant change: an expert in England and Wales who provides a report that falls below the standard of a competent practitioner and causes the client loss can now be sued in negligence.

In the United States, the position varies by state, but the dominant approach grants experts a qualified quasi-judicial immunity for testimony given in court. This immunity generally does not extend to out-of-court work such as laboratory analysis, report preparation, or work done for a party before proceedings commence. A forensic scientist whose laboratory analysis was negligently conducted may therefore face civil liability even if the courtroom testimony itself is protected. Some jurisdictions also allow professional licensing boards to sanction an expert for false testimony independently of civil liability.

In India, the Bharatiya Sakshya Adhiniyam 2023 (BSA) governs how expert opinion evidence is received in court, but the liability framework for experts who give false evidence sits primarily in the criminal law. Under the Bharatiya Nyaya Sanhita 2023 (BNS), fabricating evidence or giving false evidence in a judicial proceeding is an offence. Civil negligence suits against expert witnesses remain uncommon in Indian practice, partly because the tradition of court-appointed experts (rather than party-retained experts) diffuses liability and partly because the procedural infrastructure for such claims is less developed than in common-law systems with a long history of party-expert retention.

Cognitive bias describes any systematic pattern of deviation from rational judgment. In forensic science, the concern is not that examiners are dishonest: it is that even honest, competent, experienced professionals reach different conclusions depending on what irrelevant information they were exposed to before conducting an analysis. This has been demonstrated empirically across multiple forensic disciplines including fingerprint examination, bite-mark analysis, bloodstain pattern interpretation, and document examination.

Confirmation bias is the tendency to search for and weight evidence that is consistent with a pre-existing hypothesis while underweighting or ignoring evidence that contradicts it. In a forensic context, an examiner told that a suspect has confessed may interpret an ambiguous fingerprint impression as matching the suspect when the same impression, presented blind, would have been assessed as inconclusive. The confirmation is not conscious: the examiner genuinely believes the conclusion is objective. A study by Dror et al. (2006) presented the same fingerprint evidence to experienced fingerprint examiners in two conditions: in one condition they were told the suspect had confessed; in the other, that the suspect had an alibi. A significant proportion of examiners reached different conclusions in the two conditions.

Contextual bias is the broader category: being influenced by any case context that is logically irrelevant to the scientific question. An examiner who knows the victim was a child, or that the accused has a prior record for the same offence, or that the lead investigator is convinced of guilt, has been exposed to contextual information. Research shows this information affects conclusions even when examiners explicitly try to set it aside. The National Commission on Forensic Science in the United States and the UK Forensic Science Regulator have both identified contextual bias as one of the primary threats to the reliability of forensic evidence.

Expectation bias and allegiance effect are related phenomena. Expectation bias occurs when an analyst interprets ambiguous findings in line with what they expect the result to be, based on training or prior experience with similar cases. Allegiance effect occurs when an expert retained by one party unconsciously skews conclusions in favour of that party's case, even without any conscious intention to do so. Studies comparing independent court-appointed experts with party-retained experts have found systematic differences in conclusion rates that are difficult to attribute purely to the selection of better cases.

Because cognitive bias operates below conscious awareness, telling examiners to be objective is not sufficient. Effective mitigation requires structural changes to how casework is organised and information is managed within forensic laboratories.

Sequential unmasking, developed by Itiel Dror and colleagues, is a workflow in which case information is released to the analyst in stages, from the least to the most potentially biasing, with the analyst completing and recording their assessment of each evidence type before receiving the next piece of context. A DNA analyst, for example, would complete profile interpretation before learning whether the profile matches a specific named suspect. The principle is to preserve the scientific integrity of each analytical step by ensuring it is not contaminated by information that belongs to a later step.

Linear flow procedures require that once an examiner has formed and recorded a conclusion, they do not revisit it simply because they have subsequently received additional case information. If new scientific data requires re-examination, the re-examination is documented as a separate event with its own reasoning, not presented as a continuous revision of the original conclusion. This creates an audit trail that makes it possible to identify whether a change of opinion followed new data or new case context.

Blind verification requires a second examiner to verify the first examiner's conclusion without knowledge of what that conclusion was. Standard verification, where the second examiner is told the first examiner's result and asked to check it, is susceptible to confirmation bias in the verifier. Blind verification has been shown to catch more errors but requires more resource and is not yet universal in most forensic laboratory systems.

Forensic misconduct is often difficult to detect at the time it occurs, precisely because it involves a scientist with apparent expertise presenting results that other participants in the proceedings cannot independently verify. Cross-examination can expose inconsistencies but is limited if the fabrication is technically credible. Detection most commonly occurs through one of four routes: whistleblowers, internal audits, post-conviction retesting, or the discovery of inconsistencies across cases.

In the United States, the most consequential forensic misconduct scandals of recent decades were uncovered through combinations of these routes. The FBI forensic hair-analysis scandal, in which examiners testified for decades to microscopic hair comparison evidence using statistically unsupported claims, was exposed through a review initiated after a National Research Council report (2009) criticised the scientific foundations of the technique. The Department of Justice and the FBI subsequently reviewed over 2,500 cases and found that examiners had given erroneous testimony in at least 90% of the cases reviewed. The Dookhan scandal in Massachusetts involved a drug laboratory analyst who fabricated test results and forged co-workers' signatures; it was exposed through an internal audit and led to the dismissal of over 21,000 drug convictions.

In England and Wales, the Criminal Cases Review Commission (CCRC) refers cases back to the Court of Appeal when it believes there is a real possibility the conviction would not be upheld. Forensic science issues, including both fresh scientific analysis and newly discovered evidence of expert error or misconduct, are among the most common grounds for CCRC referrals. Once a referral is accepted by the Court of Appeal, fresh forensic evidence can be admitted under the Criminal Justice Act 2003.

In India, the Bharatiya Nagarik Suraksha Sanhita 2023 (BNSS) replaced the Code of Criminal Procedure and governs the procedural framework within which forensic evidence is obtained and presented. There is no equivalent of the CCRC in India, but the High Courts and the Supreme Court of India have jurisdiction to review convictions through revision petitions and special leave petitions. The National Forensic Sciences University (NFSU) and state forensic science laboratories operate under departmental oversight, and internal audit mechanisms vary significantly across states.

The legal rules on expert witnesses in most jurisdictions converge on a core principle: the expert's overriding duty is to the court or tribunal, not to the party that retained them. This principle is codified in different forms across jurisdictions. In England and Wales, CPR Practice Direction 35 (applicable to civil proceedings) and the Criminal Procedure Rules require the expert to state that they understand and have complied with their duty to the court. The Ikarian Reefer guidelines, articulated by Cresswell J in 1993, remain influential: they require the expert to provide independent assistance, to state facts or assumptions on which the opinion is based, to distinguish those from matters within the expert's own knowledge, and to make clear when a question falls outside the expert's expertise.

In the United States, Federal Rule of Evidence 702 requires that expert testimony be the product of sufficient facts or data, reliable principles and methods, and that the expert has reliably applied those principles and methods to the facts. Rule 26 of the Federal Rules of Civil Procedure also requires disclosure of the expert's report, qualifications, prior testimony, and compensation, creating a transparency framework that enables opposing parties to identify potential bias from financial relationships. The Daubert standard itself, requiring the court to act as gatekeeper, partly serves the function of filtering out testimony that would not withstand scrutiny as to its scientific basis.

The duty of disclosure covers more than just the expert's qualifications. It includes disclosure of adverse findings: data or results that cut against the party's position. An expert who selects only the findings that support the retaining party and suppresses contrary results has breached their duty of impartiality, even if every individual statement made is technically accurate. Courts in England and in Australia have sanctioned experts for selective reporting and in some cases excluded their evidence entirely. The obligation to disclose applies to the entire evidentiary base, not only the parts that support the conclusion.