From Frye to Daubert: How Courts Navigate Scientific Evidence

Much has been written advising laypeople to navigate conflicting medical opinions and identify the golden truth amid the pervasive muck and mire. These questions also vex judges when evidence is presented in litigation seeking damages from allegedly related injuries. Although judges do not decide which competing position is correct, they serve as evidentiary “gatekeepers,” responsible for determining the soundness of the science presented and barring junk science from the courtroom.

Once assessed, all sound evidence from both sides is presented to the jury for deliberation, essentially based on which witness they found most credible. Whether the decision is made by a judge or a jury, specific, albeit different, strategies are used to make these assessments.

For 70 years, from 1923 to 1993, judges could only admit novel scientific claims if the consensus of the relevant scientific community supported it – known as the Frye standard. This approach significantly limited the introduction of cutting-edge science or emerging critiques by younger researchers, effectively silencing newcomers who challenged existing dogma.

Recognizing that Frye inhibited the admissibility of innovative scientific perspectives, the Federal Rules of Evidence broadened the standard in 1970, allowing all evidence based on sound science while relying on cross-examination to expose unreliable experts. The Supreme Court clarified this position in the landmark 1993 decision Daubert v. Merrell Dow, which has since been adopted in all federal courts and all but six states.

Daubert and subsequent cases refined the standards for distinguishing between sound and junk science. The standard requires that scientific evidence meet specific theoretical parameters before a jury can consider it: reliability, relevance, and “the fit” of the evidence to the facts of the case. This implicitly requires compliance with the scientific method in the research being presented.

Later cases expanded Daubert’s scope, requiring that expert opinions be tethered to methodology and research, preventing experts from merely pontificating based on personal assessment. The General Electric Co. v. Joiner case shifted focus from the soundness of methodology to the legitimacy of the expert’s conclusion.

Scientific reliability is typically measured by statistical significance, while validity may be satisfied if the research adheres to methodological norms of the particular field and yields objective results. Daubert suggests additional tests judges may use to evaluate claims:

• Whether the research was peer-reviewed and published
• Whether the theory was tested experimentally with objective results
• Whether an acceptable error rate was calculated
• The degree of scientific consensus (the Frye index)

The landscape has changed dramatically since Daubert was decided in 1993. The number of journals has mushroomed, along with instances of fraudulent research. Many publications are now financed by organizations supporting the claims being litigated, weakening the reliability of peer review as a quality control mechanism.

To address these concerns, courts have developed practical remedies, including identifying retracted studies through databases like Retraction Watch, examining financial and ideological conflicts of interest of publishers and authors, evaluating a journal’s impact factor, and assessing the editorial board’s credentials and potential partisan ties.

When implementing the Daubert standard, the 9th Circuit Court noted that expert opinion based on methodology that “diverges significantly from the procedures accepted by recognized authorities in the field” cannot be considered reliable. Of particular concern is the independence of the expert’s research. The court favors testimony based on “legitimate, preexisting research unrelated to the litigation” rather than opinions developed expressly for testifying in court.

Another red flag for courts is “cherry-picking” – when experts selectively cite only studies supporting their position while ignoring contradictory research without explaining their selection criteria. Numerous judges have rejected such approaches, including in recent high-profile cases involving acetaminophen-autism claims and Zantac litigation.

Courts also typically reject unpublished reanalyses of epidemiological studies that haven’t undergone peer review, particularly when they contradict the “massive weight” of original published studies that have been fully scrutinized by the scientific community.

The publication process itself is subject to manipulation. While reputable journals require declarations of conflicts of interest, some less scrupulous tactics include citing impressive-sounding articles that don’t actually support the underlying premise, referencing AI-generated “hallucinated” sources, or presenting letters to the editor (which aren’t peer-reviewed) as authoritative research articles.

Even prestigious publications like The Lancet and the New England Journal of Medicine have occasionally published erroneous articles, highlighting the importance of replication. At the same time, concerns have emerged about valid research being removed from the internet when outcomes conflict with political positions.

When evaluating scientific evidence – whether in court or for personal knowledge – the final question remains crucial: Who is presenting the information, who is objecting to it, and what interests do they have in the outcome?