A Deepening Split? The Applicability of Daubert at Class Certification

The defense bar should strive to show the district and appellate courts that the Supreme Court’s dicta strongly favors requiring expert evidence at class certification to be admissible.

In Prantil v. Arkema Inc., 986 F.3d 570 (5th Cir. 2021), the Fifth Circuit joined three other circuits in holding that scientific evidence offered in relation to certification of a class must meet the Daubert standard of reliability. While the Fifth Circuit became the fourth federal court of appeals to endorse such an approach, a circuit split on the issue is unlikely to resolve itself and may be deepening. A look at the development of the circuit split and how other courts have treated the issue may provide defense counsel with insight for opposing class certification in those jurisdictions on either side of the split.

Prantil involved various common law and federal statutory claims against Arkema, Inc. (“Arkema”), which owned a chemical production facility in Texas. Because of a hurricane, the facility dispersed volatile chemicals into the local environment, and local property owners both saw contaminants collect on their properties and experienced physical symptoms. Alleging property damage, adverse health effects, or both, some of these property owners brought a putative class action against Arkema and sought to represent a class consisting of all property owners within a seven-mile radius of the chemical facility.

The plaintiffs moved for class certification and Arkema moved to exclude the plaintiffs’ experts. The district court’s analysis noted that “[w]hether a full Daubert analysis” was required was “unclear,” and found that a “better” expert practice “was not necessary under Daubert at the class certification stage.” The district court ultimately excluded the plaintiffs’ damages expert, allowed the testimony of three other experts, and certified the class.

The Fifth Circuit held the district court’s “hesitation to apply Daubert’s reliability standard with full force” was an abuse of discretion and vacated the certification order. Although the district court conducted a Daubert analysis of experts, the Fifth Circuit explained that the court “was not as searching in its assessment of the expert reports’ reliability as it would have been outside the certification setting.” The Fifth Circuit thus resolved whether Daubert’s “metric of admissibility [should] be the same for certification and trial.” Answering in the affirmative, the court stated “the Daubert hurdle must be cleared when scientific evidence is relevant to the decision to certify.”

The court described the rule requiring Daubert to be applied at the class certification stage “as a natural extension of the Supreme Court’s” precedents. Specifically, in 2011, “the Supreme Court expressed ‘doubt’ that ‘Daubert did not apply to expert testimony at the certification stage of class-action proceedings,’” and two years later “reaffirmed that it is incumbent on plaintiffs to submit ‘evidentiary proof’ of their compliance with Rule 23.” Id. at 575 (quoting Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 354, 131 S. Ct. 2541, 2554 (2011) and Comcast Corp. v. Behrend, 569 U.S. 27, 33–34, 133 S. Ct. 1426, 1432 (2013)).

Because expert testimony that does not satisfy the Daubert analysis cannot constitute evidentiary proof according to the Fifth Circuit, when “an expert’s opinion would not be admissible at trial, it should not pave the way for certifying a proposed class.”

The Prantil court noted that it is the fourth court of appeals to endorse a rule requiring scientific evidence to meet the Daubert standard at class certification. Id. at 575 n.12 (citing In re Blood Reagents Antitrust Litig., 783 F.3d 183, 187 (3d Cir. 2015); Sher v. Raytheon Co., 419 F. App’x 887, 890–91 (11th Cir. 2011) (unpublished); Am. Honda Motor Co. v. Allen, 600 F.3d 813, 816 (7th Cir. 2010)). Despite this agreement, a circuit split exists. The Fifth Circuit has now aligned with the Third, Seventh, and Eleventh Circuits in requiring expert evidence to be admissible for that evidence to support class certification, but on the other hand, the Eighth and Ninth Circuits do not. The Second and Sixth Circuits have also participated in the split in inconclusive ways.

The circuit split developed over the past fifteen years. First, in 2005, the Eighth Circuit declined to require a full Daubert analysis at the class certification stage. Blades v. Monsanto Co., 400 F.3d 562 (8th Cir. 2005). That same year, the Fifth Circuit required findings at the class certification stage to be “based on adequate admissible evidence,” but did not specifically mandate a Daubert analysis or even refer to Daubert at all. Unger v. Amedisys Inc., 401 F.3d 316, 319 (5th Cir. 2005). Then, in 2010, a split originated in earnest when the Seventh Circuit specifically held that Daubert applies at class certification. Allen, 600 F.3d at 816. Relying on its own precedents, that court explained “a district court must make the necessary factual and legal inquiries and decide all relevant contested issues prior to certification.” Accordingly, it held that a “district court must perform a full Daubert analysis before certifying [a] class” where an expert is “critical to class certification.”

The following year, relying entirely upon the Seventh Circuit’s opinion, the Eleventh Circuit held that a district court’s “refus[al] to conduct a Daubert-like critique of the proffered experts’s [sic] qualifications” was erroneous and warranted vacatur. Sher, 419 F. App’x. at 890–91. Although the Eleventh Circuit’s opinion in Sher is an unpublished, nonbinding opinion, it continues to be cited by courts within that jurisdiction.

Also in 2011, the Eighth Circuit reaffirmed Blades and “explicitly rejected a request for a full Daubert inquiry at the class certification stage” for the second time. In re Zurn Pex Plumbing Prod. Liab. Litig., 644 F.3d 604, 612 (8th Cir. 2011) (citing Blades, 400 F.3d at 569). As a result, district courts within the Eighth Circuit have the discretion to perform a “focused” or “‘tailored’ Daubert analysis.”

Around the same time, in dicta, the Supreme Court “doubt[ed]” a district court’s conclusion “that Daubert did not apply to expert testimony at the certification stage of class-action proceedings.” Dukes, 564 U.S. at 354, 131 S. Ct. at 2554. Just two years later, in 2013, the Supreme Court certified “the question [of] whether the standards for admissibility of expert evidence set out in Federal Rule of Evidence 702 and Daubert… apply in class certification proceedings” for review. Comcast, 569 U.S. at 39, 133 S. Ct. at 1435 (Ginsburg, J. dissenting) (internal citation omitted). While the Court ultimately did not answer the question, its discussion nonetheless included statements that have been construed as guidance. That guidance boils down to two aspects of the Rule 23 legal landscape. First, the Court reemphasized that Rule 23 “does not set forth a mere pleading standard,” and that instead, “evidentiary proof” must show that the Rule is “in fact” satisfied. Second, the Court “emphasized that it ‘may be necessary for the court to probe behind the pleadings before coming to rest on the certification question.’”

Those statements from the Supreme Court—“doubt,” “evidentiary proof,” and “may be necessary,”—would soon be at the root of the newer iteration of the circuit split.

Following Comcast, the Third Circuit held that Daubert does apply at class certification. In re Blood Reagents Antitrust Litig., 783 F.3d at 187. Noting that the Supreme Court had stated “evidentiary proof” must show that Rule 23 is satisfied “in fact,” the Third Circuit required a Daubert inquiry at the class certification stage. It found further support in the Supreme Court’s dicta indicating “doubt.”

In 2018, in Sali v. Corona Reg’l Med. Ctr., the Ninth Circuit rejected a strict Daubert application requirement at the class certification stage, disagreeing with the Third Circuit and instead claiming to side with the Eighth Circuit. 909 F.3d 996, 1005 (9th Cir. 2018). In the Ninth Circuit, district courts “should evaluate admissibility under the standard set forth in Daubert,” but that evaluation should go to the weight the evidence is given, and mere admissibility must not be dispositive. Like the Eighth Circuit, the Ninth Circuit thus allows district courts a more flexible Daubert analysis at class certification.

Then, as discussed above, in January of 2021, the Fifth Circuit in Prantil joined the Third, Seventh and Eleventh Circuits by requiring a strict application of Daubert at class certification.

Since the Supreme Court certified the question of whether a district court must undertake a Daubert analysis at the class certification stage and then did not reach the merits of that question in 2013, the Third, Fifth, and Ninth Circuits have issued divergent, binding opinions on the issue. In addition, the question has been raised before the Second Circuit, which did not reach the question, and the Sixth Circuit, which addressed it in a nonbinding opinion. See In re U.S. Foodservice Inc. Pricing Litig., 729 F.3d 108, 129 (2d Cir. 2013) (holding the question need not be resolved and affirming class certification where district court “opinion makes clear” that “it considered the admissibility of the expert testimony,”; In re Carpenter Co., No. 14-0302, 2014 WL 12809636, at *3 (6th Cir. Sept. 29, 2014) (unpublished) (noting that the Supreme Court “has suggested that such [a Daubert] analysis may be required in some circumstances” at the class certification stage, and that in the case before it, the district court did not abuse its discretion where it “only applied Daubert to critical witnesses,”). Both those courts held that lower courts’ respective Daubert analyses—one of which occurred without a hearing or explicit findings, and one of which pertained only to select experts—did not constitute an abuse of discretion. The frequency with which this issue has arisen, the divergent opinions, and the two nonbinding opinions related to the issue, all indicate the split will probably continue to develop.

Indeed, the Ninth Circuit’s Sali opinion and related proceedings show the split may be worsening. After a panel of judges on that court issued the 2018 opinion departing from the Third Circuit and renewing the split, five circuit judges dissented from denial of rehearing en banc. The dissent explains that the panel decision both “puts [their] court on the wrong side of a lopsided circuit split,” and “defies clear Supreme Court guidance.” Sali v. Corona Reg’l Med. Ctr., 907 F.3d 1185, 1189 (9th Cir. 2018) (Bea, J., dissenting from order denying rehearing).

Adding to the confusion, the Ninth Circuit and the Eighth Circuit do not appear to agree on a single minority position. While in the Eighth Circuit, courts are empowered to “conduct[] a focused Daubert analysis which scrutinize[s] the reliability of the expert testimony,” courts in the Ninth Circuit “may consider whether the [] proof is, or will likely lead to, admissible evidence,” and “should evaluate admissibility under” Daubert. While the standards of these courts are both a far cry from a categorical rule requiring expert evidence to meet a strict Daubert analysis, they are not exactly the same. Moreover, what constitutes a focused Daubert analysis in the Eighth Circuit, and the minimum requirements of a district court to meet the Ninth Circuit’s standard, are both unclear.

Another indicator that the split is not resolving is the Sixth Circuit’s recent opinion in Lyngaas v. Curaden AG, 992 F.3d 412 (6th Cir. 2021). Lyngaas, which issued after Prantil, held “that [] ‘evidentiary proof’ need not amount to admissible evidence, at least with respect to nonexpert evidence.” While the Sixth Circuit’s opinion did not choose a side of the split as to expert evidence, it agreed with “the Eighth and Ninth Circuits” as to the admissibility of evidence required for class certification generally and did not limit its holding to nonexpert evidence. Yet it did not distinguish between the two approaches of those courts. It is also notable that, in declining to limit its holding to nonexpert evidence, the court did not refer to either the Supreme Court’s dicta doubting Daubert’s application to such evidence, or its own previous unpublished, nonbinding opinion that found a district court did not abuse its discretion in applying Daubert to only “critical witnesses.” See In re Carpenter Co., No. 14-0302, 2014 WL 12809636, at *3 (6th Cir. Sept. 29, 2014).

Lyngaas is remarkable because, like Prantil, it relied upon the analytical principles set out in Comcast but came to a different conclusion. The Sixth Circuit noted that, while a “party seeking to maintain a class action must ‘satisfy through evidentiary proof at least one of the provisions of Rule 23[],” binding precedent has also “suggest[ed] that admissible evidence is not always required.” Although the “analysis sometimes requires the trial court to ‘probe behind the pleadings,’ at other times, ‘the issues are plain enough from the pleadings.’” The court’s focus on the suggestion that admissible evidence is not always required may indicate it would require a less stringent Daubertanalysis at the class certification stage, aligning it on that more specific issue with the Ninth and Eighth Circuits. But see In re Carpenter Co., 2014 WL 12809636, at *3 (unpublished) (noting that the Supreme Court “has suggested that such [a Daubert] analysis may be required in some circumstances” at the class certification stage, and that in the case before it, the district court did not abuse its discretion where it “only applied Daubert to critical witnesses.”).

Thus, although this circuit split has existed for many years, the courts do not appear to be converging on a rule. This split is therefore unlikely to resolve itself. While the Fifth Circuit in Prantil has added further weight to the majority position, the Ninth Circuit’s denial of en banc rehearing cemented that court’s decision reviving the split.

At the root of the disagreement is Supreme Court dicta, susceptible to multiple interpretations and capable of being afforded different weight. While the Supreme Court has suggested that admissibility is only sometimes required, lower courts are left with no analytical tools to determine when it is. Just as the Supreme Court did in 2013, it should grant certiorari and this time answer this important question: whether evidence—all evidence, or expert evidence at least—must be determined to be admissible before it may support class certification.

Until the Supreme Court steps in, practitioners defending putative class actions will benefit from an awareness of this overall situation, and will be better capable of applying the following lessons learned from the cases outlined above.

First, defense practitioners must preserve claims of error as to the admissibility of evidence at the class certification stage. Several jurisdictions have no binding precedent on this matter, and it seems likely that the Supreme Court will take the issue up at some point. Thus, practitioners must be vigilant in raising all objections to the admissibility of evidence used to support class certification. A failure to preserve such a claim resulted in forfeiture of the objection in Comcast, which is why the Supreme Court did not answer the certified question in 2013.

Second, defense practitioners should consider factual findings regarding admissibility to be paramount. The Second Circuit, for example, was able to avoid answering whether or when a Daubert analysis is a necessary aspect of a class certification decision because the objecting party waived a Daubert hearing, and the district court’s “opinion ma[de] clear” that the court “did make the requisite findings.” In re U.S. Foodservice Inc. Pricing Litig., 729 F.3d at 130. The defense bar should seek to avail itself of every protection available by requesting hearings when they are necessary and seeking to ensure that appropriate findings of fact issue.

Third and finally, in jurisdictions where there is no binding precedent, practitioners should be fully aware of state of this circuit split. The heavy weight of appellate authority—the Third, Fifth, Seventh, and Eleventh Circuits—explicitly require expert evidence to satisfy Daubertin order to support class certification. While the Eighth and Ninth Circuits do not agree with the majority, those courts actually maintain distinct standards and do not agree amongst themselves. And although the Sixth Circuit indicated some agreement with those courts, the Lyngaas opinion may not ultimately apply to expert evidence, and thus, the Sixth Circuit’s unpublished opinion—finding no abuse of discretion where a court applied Daubert—may still be persuasive authority, illustrative of that appellate court’s position on the issue.

In sum, although Prantil characterizes itself as a natural extension of binding precedent, there is a significant circuit split on the issue, and the Supreme Court should resolve the inconsistency. The defense bar, meanwhile, should strive to show the district and appellate courts that the Supreme Court’s dicta strongly favors requiring expert evidence at class certification to be admissible, that most courts that have considered the issue agree, and that the courts that disagree do not maintain a consistent position among themselves.

Click here to read the article on DRI's For the Defense. 

About the Author: Matt Gibbons is an associate at Shipman & Goodwin LLP. He authored the article as an associate at Ellis & Winters LLP in Greensboro, North Carolina. Matt’s practice focuses on complex commercial litigation. Prior to pursuing his law degree, Matt was a former Infantry Squad Leader in the United States Marine Corps.