Daubert and Frye Compared

The two standards ask different questions

The general-acceptance test asks a question about a community. Has the technique “gained general acceptance in the particular field in which it belongs”? Frye v. United States, 293 F. 1013, 1014 (D.C. Cir. 1923). The court surveys the field; the field, in substance, decides. Whether the technique actually works is not the question, and under the stricter state formulations the judge is not permitted to make it the question.

FRE 702 asks a question about a method. Since 1 December 2023 the rule requires that “the proponent demonstrates to the court that it is more likely than not” that the expert’s knowledge will help the trier of fact, that the testimony is based on sufficient facts or data, that it is the product of reliable principles and methods, and that the opinion “reflects a reliable application of the principles and methods to the facts of the case.” The judge decides that, as a preliminary question under FRE 104(a), by a preponderance of the evidence.

Everything else follows from that difference. One inquiry counts heads in a discipline and defers to the result. The other requires a judicial finding about the reliability of a method and of its application. Asking which is stricter imports an assumption — that they are two settings on one dial — that neither standard supports.

How the federal courts stopped applying general acceptance

Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), is a displacement argument built on statutory text, and it runs in four steps. The Federal Rules are a legislative enactment and are construed as a statute. Rule 402 makes relevant evidence admissible unless some other authority provides otherwise. Rule 702 is the specific rule on expert testimony and contains no general-acceptance requirement: “Nothing in the text of this Rule establishes ‘general acceptance’ as an absolute prerequisite to admissibility.” 509 U.S. at 588. And a rigid requirement of that kind “would be at odds with the ‘liberal thrust’ of the Federal Rules.” Ibid.

The conclusion identifies what the Court objected to: “Frye made ‘general acceptance’ the exclusive test for admitting expert scientific testimony. That austere standard, absent from, and incompatible with, the Federal Rules of Evidence, should not be applied in federal trials.” 509 U.S. at 589. The objection was to exclusivity, and general acceptance survived as one consideration among several: “Finally, ‘general acceptance’ can yet have a bearing on the inquiry.” Id. at 594. Saying that the decision rejected general acceptance overstates it.

Two limits are worth carrying. The displacement was federal only, because a 1923 D.C. panel decision never bound any state. And the Court was unanimous on that holding — Chief Justice Rehnquist and Justice Stevens joined it and dissented only from the discussion of reliability that follows. Primary text: the opinion at 509 U.S. 579.

Which is easier to satisfy depends on the kind of method

This is the part that a comparison chart cannot carry, and it is the answer to the question most readers actually have.

• A genuinely novel method. General acceptance is harder, often impossible. A technique published last year cannot be generally accepted however well validated, because acceptance takes time. Under Rule 702 the same technique can be admitted on testing, error rate, peer review and controlling standards, with general acceptance listed last and treated as one factor.
• An old, entrenched, weakly validated method. General acceptance is easier, sometimes much easier. A technique a field has used for fifty years is accepted by definition; whether it has a measured error rate, controlling standards or validation studies is not the question the test asks. Much forensic-identification evidence sits on this axis.
• An experience-based expert. The gap is widest here, and it runs the other way from what most readers expect — see below.

Whether that favors plaintiffs or defendants, or the prosecution or the defense, depends entirely on which side is offering which kind of method in the case at hand. There is no stable answer, and a writer who supplies one is describing a case rather than a standard.

The experience-based expert is where the divergence is largest

Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999), held that the gatekeeping obligation “applies not only to testimony based on ‘scientific’ knowledge, but also to testimony based on ‘technical’ and ‘other specialized’ knowledge.” The operative test for a witness qualified by experience is the intellectual-rigor formulation: the court must make certain that the expert “employs in the courtroom the same level of intellectual rigor that characterizes the practice of an expert in the relevant field.” That witness is fully inside the federal gate.

In most general-acceptance states, the same witness is outside the gate, because the test is triggered only by novel scientific evidence. Illinois says so in its rule text. Pennsylvania’s rule is written more broadly but Grady v. Frito-Lay, Inc., 839 A.2d 1038 (Pa. 2003), confines the analysis to novel science. Minnesota and California are the exceptions, each running a reliability screen that reaches every expert opinion.

That produces symmetrical errors. In a general-acceptance state, filing a general-acceptance motion against an experience-based expert invites the answer that the test does not apply. In a Rule 702 jurisdiction, reciting the reliability factors and observing that none fits the witness is a losing argument after Kumho Tire, which held the factors “neither necessarily nor exclusively appl[y] to all experts or in every case.” In both forums the productive challenge is to the rigor of the reasoning and to the gap between what was relied on and what is being concluded.

Scope, decider, and what happens to a misapplied method

Three structural differences do the work that the labels are usually asked to do.

Scope. The general-acceptance gate reaches novel scientific evidence in most formulations. Rule 702 reaches all expert testimony, scientific or not, and its reliability requirements apply to each opinion offered.

Who decides. Under general acceptance the relevant field effectively decides and the judge surveys it. Under Rule 702 the judge decides, by a preponderance, under FRE 104(a) — and in doing so is not bound by the rules of evidence except those on privilege, so the report and the deposition transcript are available to the court even though they are hearsay.

Misapplication of an accepted method. Classically this was a matter of weight in a general-acceptance jurisdiction: the gate asks about the method, not about this expert’s execution of it. Federally it is an admissibility question under Rule 702(d), and the 2023 amendment rewrote that subsection to say the opinion must “reflect a reliable application” of the method. The states applying general acceptance have converged on the same problem from the other direction, each adding a second screen — Washington’s ER 702, New York’s foundation inquiry under Parker v. Mobil Oil Corp., 7 N.Y.3d 434 (2006), Minnesota’s foundational reliability requirement, and California’s Sargon gatekeeping duty.

On appeal, a federal ruling either way is reviewed for abuse of discretion under General Electric Co. v. Joiner, 522 U.S. 136 (1997), and that deference extends to how the court chose to test reliability. State practice on review varies.

What the empirical work says about whether the choice matters

Edward K. Cheng and Albert H. Yoon, “Does Frye or Daubert Matter? A Study of Scientific Admissibility Standards,” 91 Va. L. Rev. 471 (2005), approached the question indirectly and cleverly. Rather than coding opinions, the authors used removal rates from state to federal court as a proxy for how much litigants themselves believed the difference was worth. If defendants valued the federal standard, they should remove more often out of general-acceptance states. The finding was that “a state’s choice between Frye and Daubert does not matter in tort cases,” and the authors suggest the 1993 decision’s real effect was to raise judicial awareness of unreliable science rather than to change the test being applied.

The limits belong with the finding. It is one study, on tort cases, using an indirect measure, and its data predate the moves by Florida, Maryland, New Jersey and the District of Columbia. It is evidence, not proof. What it supports is a modest proposition that practitioners tend to reach independently: the label on the forum’s standard is frequently less important to an outcome than the standard of appellate review, the timing and availability of a pretrial hearing, and whether the state’s summary-judgment practice reaches the sufficiency of expert proof. Published study: Cheng and Yoon, 91 Va. L. Rev. 471.

The binary is a convenience, not a description

Three facts break the two-column framing.

There is a third group. Utah requires only “a threshold showing” that the underlying principles or methods are reliable, based on sufficient facts or data, and reliably applied — and then makes general acceptance a sufficient rather than a necessary way to make that showing. Nevada under Higgs v. State, 222 P.3d 648 (Nev. 2010), Virginia by statute and decision, South Carolina, Tennessee and others each apply their own articulated standards.

Some states apply different standards to different cases. Georgia used one standard in civil cases and another in criminal cases until 1 July 2022. New Jersey aligned its civil standard in 2018 and extended a comparable standard to criminal and quasi-criminal matters in 2023. Alabama and Missouri carry statutory carve-outs.

Some states run both screens at once. Minnesota’s Rule 702 is the clearest three-sentence proof that the categories are not exclusive: every expert opinion “must have foundational reliability,” and novel scientific theory additionally requires general acceptance. California does the same thing through Kelly and Sargon.

The practical consequence for anyone briefing an admissibility question is that the forum’s current rule text and its own high court’s decisions are the authority. A chart — including any list on a reference site — is a starting point for research, not a citation. Rule text: Minn. R. Evid. 702.

Frequently Asked Questions

Published 2026-08-12