State Court Standards for Expert Testimony

What the state test asks, and how the states divide

Federal Rule of Evidence 702, Daubert and Kumho Tire govern the federal courts and nowhere else. In state court the question turns on the state's own rule and its supreme court's reading of it. Some states ask the Frye standard's question, whether the method is generally accepted in the relevant field; others ask Daubert's, whether the method is reliable; several ask both, and a sizable group apply a test of their own.

Roughly half the states apply Daubert or a rule built on it, a half-dozen keep Frye, and the rest apply hybrid or home-grown tests. At least eight states have adopted the wording of the federal rule as amended December 1, 2023, which requires the proponent to show that it is "more likely than not" that the rule's conditions are met. No survey reliable enough to support a fifty-state table was located, so this page names only the states whose standard was read from the state's own rule, statute or supreme court opinion.

No decision was located, in any court, deciding whether search engine optimization analysis (the examination of how a website is crawled, indexed and ranked by a search engine, and of the traffic that ranking produces) is a generally accepted methodology; what follows about search opinions is analysis built from the rule texts.

The six states that keep a general-acceptance test

Six states apply Frye or a general-acceptance test descended from it.

California calls its test Kelly, after People v. Kelly, 17 Cal.3d 24 (1976); Sargon Enterprises, Inc. v. University of Southern California (Cal. Nov. 26, 2012) confirms in footnote 6 that it "still applies in California courts" to new scientific techniques.

Illinois codified the test in Illinois Rule of Evidence 702, effective January 1, 2011, which places on the proponent of an opinion based on "a new or novel scientific methodology or principle" the burden of showing general acceptance. The Committee Comment states that "Rule 702 confirms that Illinois is a Frye state."

Minnesota is reported to apply its Frye-Mack standard to novel scientific evidence under Goeb v. Tharaldson, 615 N.W.2d 800 (Minn. 2000); only a summary of the opinion was read for this page.

New York applies Frye. In Parker v. Mobil Oil Corp., 2006 NY Slip Op 07391 (N.Y. Oct. 17, 2006), the Court of Appeals recorded that the parties "acknowledge that Frye is the current standard in New York."

Pennsylvania writes the test into Rule 702(c), which requires that "the expert's methodology is generally accepted in the relevant field." Walsh v. BASF Corp. (Pa. July 21, 2020) confirms that this requirement "is commonly referred to as the Frye test."

Washington keeps Frye in criminal cases and has so far assumed it in civil ones. Anderson v. Akzo Nobel Coatings, Inc., 172 Wn.2d 593 (Wash. Sept. 8, 2011), restated the court's "continued adherence to the more stringent Frye test" and, for civil cases, would "assume without deciding that Frye is the appropriate test."

The states that have adopted the December 2023 wording

The December 1, 2023 federal amendment wrote the proponent's burden into the rule text, "if the proponent demonstrates to the court that it is more likely than not that," and rewrote subsection (d) so the opinion must reflect "a reliable application of the principles and methods to the facts of the case." At least eight states have adopted that wording.

Arizona acted first: Supreme Court Order No. R-23-0004, dated August 24, 2023 and effective January 1, 2024, amended Rule 702 "to conform to the changes made to Federal Rule of Evidence 702 that took effect on December 1, 2023." Michigan followed by order of March 27, 2024, effective May 1, 2024. Kentucky adopted the lead-in by Supreme Court Order 2024-19, effective July 1, 2024. Ohio amended Evid.R. 702 effective July 1, 2024, with a staff note stating that the amendment adopts the 2023 federal changes, while keeping its own three-part structure. Louisiana acted by statute, Acts 2024, No. 88, effective August 1, 2024. Oklahoma amended 12 O.S. §2702 by Senate Bill 453, effective September 1, 2025. Wisconsin enacted 2025 Wisconsin Act 92, signed March 6, 2026, to conform its statutory rule "with recent changes made to Federal Rule of Evidence (FRE) 702"; it first applies to actions commenced on or after its effective date. South Dakota amended SDCL 19-19-702 by Supreme Court Rule 26-04, effective March 31, 2026.

Two caveats: the Ohio and Oklahoma texts were read from reproductions, and Wisconsin's effective date was not confirmed, so quote the official source in any filing. Maryland considered the wording in a February 2, 2024 committee report and did not adopt it; Rule 5-702 as amended through November 13, 2024 contains neither "preponderance" nor "more likely than not."

Florida, Maryland and Texas: Daubert by adoption and a test of Texas's own

In In re Amendments to the Florida Evidence Code, No. SC19-107 (Fla. May 23, 2019), the Supreme Court of Florida, receding from its earlier refusal, adopted the Legislature's 2013 Daubert amendments to sections 90.702 and 90.704 "as procedural rules of evidence," effective "immediately upon the release of this opinion." It left the constitutional objections "for a proper case or controversy."

Maryland followed on August 28, 2020 in Rochkind v. Stevenson, 471 Md. 1, a 5–3 decision that overruled the state's Frye-Reed line and adopted "a single standard by which courts evaluate all expert testimony: Daubert." The court listed ten factors, five from Daubert and five from later federal cases, including whether the expert "has adequately accounted for obvious alternative explanations," whether the expert is "being as careful as he [or she] would be in his [or her] regular professional work," and "whether the field of expertise claimed by the expert is known to reach reliable results" for the type of opinion offered. It also adopted the Joiner analytical-gap rule. For a search opinion the alternative-explanations factor is the whole difficulty: a Google algorithm update (a change to the ranking system that Google rolls out on its own schedule) can coincide with the conduct complained of.

Texas applies its own test, built in E.I. du Pont de Nemours & Co. v. Robinson, 923 S.W.2d 549 (Tex. 1995), and extended to all expert testimony in Gammill v. Jack Williams Chevrolet, Inc., 972 S.W.2d 713 (Tex. 1998), as restated in Cooper Tire & Rubber Co. v. Mendez, 204 S.W.3d 797 (Tex. 2006), which is the source of the quotations here. Alongside factors familiar from Daubert, Robinson adds two with no federal counterpart: "the extent to which the technique relies upon the subjective interpretation of the expert" and "the non-judicial uses which have been made of the theory or technique." Gammill adopted the Joiner rule against opinion "connected to existing data only by the ipse dixit of the expert." The first bears directly on search work: a rank-tracking export (a third-party tool's record of where a page appeared in search results for a chosen keyword, location and device when the tool checked) is a tool output, and the inference that a drop in it was caused by a redirect error is the analyst's.

New York and California: the second layer beneath the test

In New York, Parker v. Mobil Oil states that "[t]he Frye inquiry is separate and distinct from the admissibility question applied to all evidence," meaning whether a proper foundation exists. The experts in Parker failed foundation, not Frye, because their comparison to refinery workers did not state the exposure level or how the plaintiff's exceeded it. An expert who uses accepted tools but never quantifies the thing the opinion claims to measure, such as the visits lost to a specific defect over a specific period, can fail at that second layer without ever triggering the first.

In California, Sargon construed Evidence Code sections 801 and 802 so that "the trial court acts as a gatekeeper to exclude speculative or irrelevant expert opinion," including opinion based on matter on which an expert may not reasonably rely or on reasons unsupported by the material relied on. The limits are stated with equal force: the gatekeeping role "does not involve choosing between competing expert opinions," and the court "must not weigh an opinion's probative value or substitute its own opinion for the expert's opinion." Sargon affirmed the exclusion of a lost-profits expert's market-share projection as speculative, which is the posture a search damages opinion most often occupies in California: a section 801 question about whether a projection of lost traffic rests on more than assumption. Sargon cited Daubert and Kumho Tire but left Kelly in place, so the sources do not support calling California a Daubert state.

What general acceptance means for a field with no licensing board

The objection a search expert hears in a Frye state is that there are no peer-reviewed studies of SEO and no body that licenses its practitioners, so nothing in the field can be generally accepted. The rule texts answer most of that objection, as analysis rather than holding.

The Frye states define the trigger narrowly. Illinois applies the burden only to "a new or novel scientific methodology or principle." Washington's Anderson states that "[t]he Frye test is only implicated where the opinion offered is based upon novel science"; application of accepted science "is a matter of weight." California applies Kelly only to new scientific techniques. Pennsylvania's Walsh confines the judge to method: "trial courts may not question the merits of the expert's scientific theories, techniques or conclusions"; "[t]hose are questions for the jury to decide."

The methods a search expert uses are measurement and comparison, not novel theory. Reading server logs (the web server's own per-request record of what was requested, by whom, and when), exporting Google Search Console data (Google's own report of the queries, impressions and clicks it recorded for a verified site), pulling Google Analytics reports (the site owner's own record of visits, sessions and conversions), running a crawl (an automated fetch of every reachable page that records status codes, redirects and directives such as noindex), and comparing traffic before and after a dated event are techniques used daily outside litigation, on the platform's or the party's own records. On the Illinois, Washington and California formulations the gate opens only if the expert relies on something novel. In my experience the thing that qualifies is a proprietary model that claims to predict what a page would have ranked; that is the opinion that invites a Frye motion, and it is the one I decline to build.

In the Daubert states the factors do the same work. Maryland asks whether the field "is known to reach reliable results" for this type of opinion; Texas asks about non-judicial uses and subjective interpretation. The points to brief are that the tools are used daily outside litigation, that the comparisons can be re-run from the exported data, and that the opinion is connected to that data by more than "the ipse dixit of the expert." The last point decides most motions, under any test.

Frequently Asked Questions

Published 2026-11-04