When forensic evidence reaches a courtroom, it has to clear two separate bars before a jury ever weighs it. The first is physical: prosecutors must show the item is what they say it is and that nobody tampered with it on the way from the scene to the lab to the exhibit table. That is the chain of custody, a documented record of who handled a piece of evidence, when, and why. A gap in that record gives the defense an opening to argue the sample was switched, contaminated, or degraded. The second bar is legal: a judge has to decide whether the expert testimony built on that evidence is reliable enough to put in front of a jury at all.
For most of the twentieth century, federal courts answered that reliability question with a single test from a 1923 case called Frye, which asked only whether a method was generally accepted in its field. That changed in 1993. In Daubert v. Merrell Dow Pharmaceuticals, the Supreme Court held that the Federal Rules of Evidence had replaced the old general acceptance test, and it put the trial judge in the role of gatekeeper for scientific testimony [1]. General acceptance still counts, but it is no longer the whole inquiry.
Daubert laid out a set of factors a judge can weigh. They include whether a technique can be and has been tested, whether it has been published and peer reviewed, whether it has a known or potential error rate, whether standards exist to control how it is performed, and whether it is widely accepted in the relevant scientific community [1]. None of these is a checklist item that decides the case by itself. The point was to make judges look at the method behind the opinion, not just the credentials of the person offering it.
Those factors now sit alongside Federal Rule of Evidence 702, which governs expert testimony directly. The rule says a qualified expert may give an opinion only if the proponent shows it is more likely than not that the testimony rests on sufficient facts or data, is the product of reliable principles and methods, and reflects a reliable application of those methods to the facts of the case [2]. The rule was amended effective December 1, 2023, to underline that an expert's conclusion has to stay within what the method can actually support [2]. That last clause matters for forensics, because the recurring problem has not been bad science so much as testimony that claims more certainty than the science earns.
How well those standards were holding up came under hard scrutiny in 2009, when the National Research Council of the National Academy of Sciences published Strengthening Forensic Science in the United States: A Path Forward. The report was blunt. It found that, with the exception of nuclear DNA analysis, no forensic method had been rigorously shown to consistently and reliably link evidence to a specific individual or source [3]. It also pointed to wide variation in training, certification, and lab oversight across the country, and it called for an independent federal body to set standards and fund research [3]. For a field that had spent decades describing matches as effectively certain, that was a difficult document to absorb.
Seven years later the President's Council of Advisors on Science and Technology returned to the same ground from a narrower angle. Its 2016 report, Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods, focused on the disciplines that try to match a questioned sample to a known source by comparing features [4]. PCAST drew a distinction between foundational validity, meaning a method has been tested enough to know it works and how often it fails, and validity as applied, meaning an individual analyst actually performed it correctly in a given case [4]. By that measure it treated single-source and simple-mixture DNA and latent fingerprint analysis as on firmer footing, while it found the evidence for bitemark analysis weak enough that it warned against presenting it as reliable [4].
The practical upshot of all this lands on the analyst in the witness chair. A fingerprint examiner or firearms examiner can no longer simply announce a match and call it a certainty. Under the post-Daubert framework and the current text of Rule 702, the testimony is supposed to describe the method, acknowledge its known error rate, and frame the conclusion in terms the evidence can bear [1][2]. The 2009 and 2016 reports gave defense lawyers concrete language to challenge testimony that overreaches, and they gave judges, who already act as gatekeepers, a clearer basis for limiting how far an expert can go.
None of these standards work in a vacuum, and that is worth keeping in mind when a true-crime case turns on a single piece of lab evidence. A clean chain of custody can put a sample in front of a jury, and an expert can be fully qualified, yet the testimony can still be excluded or narrowed if the underlying method has not been validated or if the analyst states a conclusion the science does not back. Daubert decides whether the jury hears the expert. Rule 702 decides how far the expert may go. The two national reports keep pressing the question of whether the method deserves that much confidence in the first place.
Sources
- Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993); Daubert Standard, Legal Information Institute (Cornell Law School)
- Federal Rule of Evidence 702, Testimony by Expert Witnesses (as amended Dec. 1, 2023), Legal Information Institute (Cornell Law School)
- National Research Council, National Academy of Sciences (2009), Strengthening Forensic Science in the United States: A Path Forward, The National Academies Press
- President's Council of Advisors on Science and Technology (2016), Forensic Science in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods (White House archive)