Evan Tager is a Partner with Mayer Brown LLP in the firm’s Washington, DC office and is the WLF Legal Pulse’s Featured Expert Contributor on Judicial Gatekeeping of Expert Evidence. Jonathan S. Klein, an Associate, with the firm. Along with five other Mayer Brown attorneys, they co-authored a January 2020 WLF Monograph, Admissibility of Expert Testimony: Manageable Guidance for Judicial Gatekeeping.
Plaintiffs often must use expert testimony to support their motions for class certification, especially in attempting to show commonality under Federal Rule of Civil Procedure 23(a)(2) and predominance under Rule 23(b)(3) (the WLF Monograph noted above discussed this issue at length). A recent Ninth Circuit decision that affirms a district court’s exclusion of expert testimony, and consequently affirms the denial of class certification, underscores the unreliability of an expert opinion that purports to extrapolate from a miniscule sample to a nationwide class that comprises tens or hundreds of thousands of consumers.
In Grodzitsky v. American Honda Motor Co., 957 F.3d 979 (9th Cir. 2020), plaintiff Phyllis Grodzitsky sued Honda on behalf of a putative class of owners and lessors of 2003 through 2008 Honda Pilots. The plaintiff claimed that Honda designed the “window regulators” in Honda vehicles defectively, which supposedly prevented people from rolling the windows up and down. A vehicle’s windows experience force or friction when the vehicle moves, and the plaintiff alleged that Honda’s testing (which purportedly tested for vibrations at a single frequency) failed to account for the different types of vibrations that a vehicle experiences in the real world. According to the plaintiff, the vibrations from the vehicles’ movement allegedly caused the “window regulator” on Honda Pilots to fail prematurely. After surviving a motion to dismiss, the plaintiff moved for class certification.
Attempting to meet Rule 23(a)(2)’s commonality requirement, the plaintiff proffered expert testimony from Glenn Akhavein, a putative engineering expert who opined on the cause of the window regulator’s failure and the purported inadequacies in Honda’s testing. The expert opined principally that Honda’s design team failed to account for the stress that the window regulators suffer from constant and varying vibrations from a vehicle’s movement. When asked in his deposition how long a window regulator should last if properly designed, Akhavein testified that the regulator “shouldn’t fail ever” but conceded that he could not articulate any alternative design that would avoid the problem alleged by the plaintiff.
Akhavein also admitted that his testing consisted of “just a real world driving around test” and that he never identified any industry standard governing the durability or lifespan of a window regulator. In addition to forming an opinion based on the so‑called “real world driving around test,” Akhavein inspected just 26 regulators and found that twelve had failed.
The district court rejected Akhavein’s testimony as unreliable under Daubert for several reasons. First, the district court observed that “without some objective basis to indicate how long the regulators should last, alleging that they are defective as soon as they are installed is simply circular—the window regulators are defective because they do not last as long as they should.”
Second, the district court rejected as unreliable Akhavein’s opinion on the specific cause of the defect (a purportedly defective “ferrule interface”). The court emphasized the inadequacy of Akhavein’s sample; he extrapolated from twelve failed regulators to a conclusion about the 441,600 regulators covered by the proposed class definition. Although the plaintiff denied that statistical significance mattered to her claims, the district court noted that Akhavein “framed his opinion in terms of probabilities” even though he lacked “an adequate methodology supporting his assessment that it was ‘more likely than not” that a regulator failed specifically because of the ferrule‑interface defect.
Third, scientific jargon aside, the “common defect” purportedly identified by Akhavein boiled down to the assertion that none of the twelve “defective” regulators lasted “the entire duration the vehicle is on the road.” But the district court had already rejected that assertion as circular.
After excluding the expert’s opinion in its entirety, the district court denied the plaintiff’s motion for class certification because “without Akhavein’s opinion, the plaintiffs were unable to demonstrate the requisite commonality.” The district court observed that “all [the plaintiffs] have is a series of window regulators that may or may not have broken before they were supposed to, and these breakages may or may not have been caused by a common defect which may or may not exist.”
Finding no abuse of discretion, the Ninth Circuit affirmed the district court’s exclusion of the expert testimony and the resulting denial of class certification. On the contrary, in explaining that the district court properly applied the “rigorous” gate‑keeping mandated by Daubert, the Ninth Circuit noted that Akhavein’s purported “vibration testing” undisputedly “provided ‘no direct correlation’ to the durability of the regulators.” In other words, even if the regulators experienced vibrations different from those in Honda’s testing, Akhavein failed to identify any correlation (much less causation) between real‑world vibrations and the failure of the window regulators.
The opinions of both the district court and the Ninth Circuit deserve commendation for faithfully applying the rigorous gate‑keeping requirements of Daubert and its progeny. In addition to highlighting the unreliability of an expert opinion that applies a subjective or unreasonable standard (“the regulator … shouldn’t ever fail”), the district court in Grodzitsky properly excluded the expert opinion on a basis that many courts shy away from: the impropriety of extrapolating from a handful of occurrences to a massive, nationwide class without identifying any objective and reliable methodology that would permit a class‑wide resolution of the key question in the case. Grodzitsky stands as a shining example of the importance of carefully scrutinizing the fit between the expert’s sample size and the class size to which the plaintiff wants to extrapolate.