RAWLINSON, Circuit Judge

:

In this design defect case, Appellant Phyllis Grodzitsky (Grodzitsky), the class representative for a proposed class of purchasers and lessees of 2003–2008 Honda Pilot vehicles, appeals the district court’s order excluding her expert’s opinion and denying class certification. Grodzitsky alleged that window regulators inside Honda Pilot vehicles were defectively designed because the regulators failed to properly support the side windows, rendering the windows inoperable. Plaintiff’s expert Glenn Akhavein (Akhavein) opined that Honda window regulators were not sufficiently durable when exposed to vibrations at certain frequencies. We affirm the district court’s order excluding Grodzitsky’s expert and denying class certification.

I. BACKGROUND

In her third amended class action complaint, Grodzitsky alleged that the window regulators installed by Honda were defective because they caused windows to fall into the doorframes, which increased the likelihood of injuries or accidents. Based on the alleged defect, Grodzitsky alleged causes of action for: (1) violations of California’s Consumer Legal Remedies Act, Cal. Civ. Code § 1750 et seq . ; and (2) violations of California’s Unfair Competition Law, Cal. Bus. & Prof. Code § 17200 et seq .

Grodzitsky initially sought certification of a class of "[a]ll persons in the United States who purchased or leased [one of seven Honda models, for model years 2000–2011, including the Honda Pilot] with the Window Regulator," as well as various subclasses based on the residencies of the vehicle owners. During the course of the litigation, Grodzitsky narrowed the proposed class to include only individuals who leased or owned 2003–2008 Honda Pilots.

In her renewed motion for class certification, Grodzitsky described the asserted design defect as window regulators in Honda Pilot vehicles that were "insufficiently strong and insufficiently durable to withstand the forces required to perform [their] intended function." In support of her motion, Grodzitsky relied in part on Akhavein’s expert opinion. In his report, Akhavein, an engineer, explained that "[a] window regulator, including a Honda Pilot regulator, has a primary purpose of moving the window glass from where it is to where the user wants [it] to go and stay there." Akhavein conveyed that static loading, which occurs "when the load or force on an object is constant," and dynamic loading, involving changes in force on an object, may impact the efficacy of a window regulator. Akhavein opined that, "[b]ased on [his] comprehensive review of the failed Honda Pilot regulators, all appear to have failed at the ferrule-carrier interface, that is the portion of the carrier that supports the cable ferrule."

According to Akhavein, dynamic loading affected the window regulators due to the force exerted when the vehicle was in motion. Akhavein stated that "[a] significant design mistake made by Honda, and missed or ignored by [Honda’s expert], is just how quickly a high number of cycles is applied to the carrier due to the dynamic vibrational loading that occurs in the Vehicle Moving state." In other words, the internal mechanisms of the window regulators were ultimately unable to withstand the vibrations caused by the vehicle’s movement. Akhavein opined that "Honda did not adequately design the Window Regulators to be strong and durable enough to withstand this high-cycle dynamic loading." Akhavein also explained that "[f]atigue, in this case caused by vibration exposure, is important as it has been estimated that 90% of all mechanical service failures can be attributed to fatigue."

Akhavein determined that Honda’s testing of its window regulators was deficient. Akhavein faulted Honda for only testing dynamic loading at a single, constant frequency, as opposed to subjecting the window regulators to a range of vibrational frequencies. Akhavein stated that "[t]his is crucial because a mechanical object can function perfectly when vibrated at one frequency but will tear itself apart if vibrated at or near its natural frequency." Akhavein concluded that "[t]he Pilot Window Regulators were not adequately designed to [be] strong and durable enough to withstand the continual affects [sic] that vibration induced metal fatigue had upon the metal portion of the carrier to ferrule interface which is what supports the window."

During Akhavein’s deposition, Honda asked several questions concerning his standard for the effective performance of window regulators. In response, Akhavein stated that window regulators "shouldn’t fail ever," and agreed that, if someone owned a vehicle for thirty years, the "window regulator assembly should work the same way it worked when it was brand new ... [i]f the car is still running." Akhavein further asserted that the window regulator "should work for the life of the car." Akhavein articulated that "[t]here’s no universal reason that a window regulator should fail." Akhavein explained that the window regulators had a common defect because they were not "durable enough," and that the standard for durability was that a window regulator "should last the life of the vehicle." Akhavein acknowledged that there was no industry standard establishing that a window regulator should last for the life of the vehicle. Akhavein stated that his testing protocol was "just a real world driving around test," and that he did not identify "a common solution" for resolving the defect in the window regulators. Akhavein did not have an opinion concerning the proper method that Honda should have utilized in designing its window regulators. Akhavein also acknowledged that the vibrational testing he conducted indicated the manner in which the window regulators responded to various frequencies, but there was "no direct correlation" with how long the part should last or whether it was "durable enough to perform its function[.]"

Honda filed a motion to exclude Akhavein’s expert opinion, which the district court granted. The district court determined that Akhavein’s opinion was deficient under Daubert v. Merrell Dow Pharms., Inc. , 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), reasoning that Akhavein’s opinion was premised on "half-baked, warmed-over conclusions" that Honda window regulators were defective "because they do not last the life of the vehicle, which [Akhavein] define[d] as the entire duration the vehicle is on the road." The district court excluded that opinion because:

(1) Akhavein provide[d] no industry standards for window regulator replacement rates; (2) Akhavein relie[d] on no peer-reviewed literature relating to window regulator durability; (3) Akhavein provide[d] no information on average window regulator replacement rates; and ( [4] ) Akhavein provide[d] no consumer studies to substantiate his claim that consumers expect a window regulator to last forever.

The district court observed that Grodzitsky attempted to recast Akhavein’s opinion by asserting that the window regulators were "defective from the moment they are installed , and that Akhavein’s claims that they should last forever [were] therefore irrelevant." The district court rejected Grodzitsky’s attempt because "[w]ithout 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, therefore they are defective."

The district court was also unpersuaded by Akhavein’s opinion that the internal cable or ferrule interfaces for the regulators were defective. The district court opined that the methodology employed by Akhavein to reach this opinion was deficient because:

(1) he examined an extremely small sample size of window regulators; (2) he does not know who made the regulators he examined; and (3) of the failed window regulators he inspected, he does not know if the failures were attributable to Plaintiffs’ theory of liability.

The district court emphasized that:

Mr. Akhavein inspected a total of 26 regulators over the course of his investigation. Of these, three were sent to him by Plaintiffs on the instruction of Plaintiffs’ counsel, eight were inspected in Plaintiffs’ vehicles, ten came from junkyards Mr. Akhavein visited, and three regulators came from a test vehicle. Not all of these regulators had failed when Mr. Akhavein inspected them. In fact, only roughly a dozen had failed. ... This means Mr. Akhavein has no information whatsoever regarding why the other failed regulators in class vehicles

failed. He is essentially extrapolating from the 12 failed regulators he inspected to draw conclusions about the 441,600 regulators installed in class vehicles. This is insufficient.

The district court articulated that, although Grodzitsky maintained that statistical significance was irrelevant in design defect actions, Akhavein framed his opinion in terms of probabilities without an adequate methodology supporting his assessment that it was "more likely than not that a failed Honda regulator ... failed because of this design defect."

The district court similarly rejected Akhavein’s opinion that Honda’s vibration testing was defective because it was conducted at only one frequency. The district court determined that Akhavein’s vibration testing did not support his opinion because "he conducted no durational testing to confirm that these additional frequencies cause the regulators to fail when they are exposed to those frequencies over time." The district court opined that Akhavein failed to demonstrate that his testing would have identified the defect relative to the tests performed by Honda.

Finally, the district court held that "[t]he exclusion of Mr. Akhavein’s testimony [was] fatal to Plaintiffs’ motion for class certification," because "without this testimony Plaintiffs [were] unable to meet the requirements of Rule 23." The district court reasoned that, without Akhavein’s opinion, the plaintiffs were unable to demonstrate the requisite commonality because "[a]ll they 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."

We granted Grodzitsky’s petition for permission to appeal the district court’s order.

II. STANDARDS OF REVIEW

We review the district court’s exclusion of Akhavein’s expert opinion and the resulting denial of class certification for an abuse of discretion. See Crowley v. Epicept Corp. , 883 F.3d 739, 752 (9th Cir. 2018) (expert opinion); see also Sandoval v. Cnty. of Sonoma , 912 F.3d 509, 515 (9th Cir. 2018) (denial of class certification).

"A class certification order is an abuse of discretion if the district court applied an incorrect legal rule or if its application of the correct legal rule was based on a factual finding that was illogical, implausible, or without support in inferences that may be drawn from the facts in the record. ..." Sandoval , 912 F.3d at 515 (citation and internal quotation marks omitted).

III. DISCUSSION

Grodzitsky contends that the district court erred in excluding Akhavein’s expert opinion because his opinion was reliable, relevant, and comported with the admissibility standards established in Daubert .

The district court did not abuse its discretion in excluding Akhavein’s opinion. "[I]n evaluating challenged expert testimony in support of class certification, a district court should evaluate admissibility under the standard set forth in Daubert ...." Sali v. Corona Reg’l Med. Ctr. , 909 F.3d 996, 1006 (9th Cir. 2018), as amended (citation omitted). Under Daubert , "the district court judge must ensure that all admitted expert testimony is both relevant and reliable." Wendell v. GlaxoSmithKline LLC , 858 F.3d 1227, 1232 (9th Cir. 2017) (citation omitted). "Scientific evidence is reliable if the principles and methodology used by an expert are grounded in the methods of science." Id. (citation and internal quotation marks omitted). "The focus of the district court’s analysis must be solely on principles and methodology, not on the conclusions that they generate," and "the court’s task is to analyze not what the experts say, but what basis they have for saying it." Id. (citations, alteration, and internal quotation marks omitted). In conducting this analysis, the district court may consider "whether the theory or technique employed by the expert is generally accepted in the scientific community; whether it’s been subjected to peer review and publication; whether it can be and has been tested; and whether the known or potential rate of error is acceptable." Id. (citation omitted).

The district court properly excluded Akhavein’s opinion under Daubert . Although Grodzitsky vigorously attempts to disassociate her design defect claim from the standard utilized by Akhavein during his deposition, Grodzitsky’s recasting of Akhavein’s opinion is unavailing. In his deposition, Akhavein faulted Honda’s design for its window regulators because the window regulators "shouldn’t fail ever." This was not a passing observation by Akhavein. Instead, he stated that "a window regulator assembly should work the same way it worked when it was brand new," the window regulator "should work for the life of the car," there was "no universal reason that a window regulator should fail," if an owner operated the vehicle for thirty years, the "window regulator assembly should work the same way it worked when it was brand new," the window regulators had a common defect because they were not "durable enough," and the applicable standard for durability was that the regulators "should last the life of the vehicle." Although Grodzitsky maintains that Akhavein’s opinion was premised on his determination that the regulators were defective because they lacked the requisite strength and durability to withstand force produced during the vehicle’s operation, this assertion elides the life-of-the-vehicle defect design theory permeating Akhavein’s deposition testimony.

Moreover, the methodological flaws in Akhavein’s opinion were not limited to his overly expansive standard for a design defect. Indeed, Akhavein admitted that he failed to identify a common solution to the defect, but relied on "just a real world driving around test." Additionally, Akhavein explained in his rebuttal report that he examined only twenty-six Honda Pilot window regulators. Akhavein stated that:

Of the 26 regulators that [he had] seen, ... [f]our of them were sent to [him] directly by the named Plaintiffs ... 9 of them were on Plaintiffs’ Vehicles whose inspection pictures [he] reviewed, 10 of them are Regulators [he] removed from Class Vehicles [he] found at random in Florida salvage yards, and 4 are Regulators from the test vehicles used by Honda’s expert[.]

In addition to utilizing this small sample size to prove a common defect in over 400,000 window regulators for class vehicles, Akhavein acknowledged that there was "no direct correlation" between his vibrational testing and failure of the window regulators. Akhavein further conceded that he did not conduct a comparison with window regulators from other manufacturers, and did not review any industry data concerning replacement rates for window regulators. Finally, Akhavein confirmed that he did not "have an opinion on what [Honda] should have done" in designing a proper window regulator.

The dissent fully acknowledges that the district court properly excluded Akhavein’s opinion that the regulators should have lasted the life of the vehicle, the crux of Akhavein’s defect theory, and that Akhavein entirely failed to cite industry standards, peer-reviewed literature, or even "test a statistically significant number of regulators to opine on the probabilities that any given Honda Pilot regulator failed because of the alleged defect." Dissenting Opinion , p. 988. Nevertheless, the dissent unsuccessfully attempts to salvage Akhavein’s deeply flawed opinion by artificially separating Akhavein’s fatally flawed opinion into a series of opinions, including some based on "first-hand observations and testing." Id. at p. 988. However, as the district court noted, Akhavein articulated no scientific basis for his observations. And the only testing he performed was not designed to identify any defects, let alone a common defect. See, e.g. , Cates v. Whirlpool Corp. , No. 15-CV-5980, 2017 WL 1862640 at *12 (N.D. Ill. May 9, 2017) (excluding an expert opinion in the class certification context based on the expert’s "merely point[ing] to an undifferentiated mass of potential problems"). In other words, identification of general "potential problems" in the absence of the identification of a specific defect present in all the regulators did not constitute "an opinion that fits Plaintiffs’ class-certification argument." Id. (citation omitted). See Ellis v. Costco Wholesale Corp. , 657 F.3d 970, 982 (9th Cir. 2011) (explaining that " Daubert does not require a court to admit or to exclude evidence based on its persuasiveness; rather it requires a court to admit or exclude evidence based on its scientific reliability and relevance. Thus, an expert’s inference or assertion must be derived by the scientific method to be admissible.") (citations and internal quotation marks omitted). The dissent’s reliance on Akhavein’s generalized opinion, while ignoring its acknowledged flaws, does not comport with our standards for assessing expert opinions. See id.

According to the dissent, if Akhavein could cobble together some form of generalized opinion, even one riddled with scientific and methodological flaws, the district court would abuse its discretion in not accepting that opinion. See Dissenting Opinion , pp. 988-89. Relying on this premise, the dissent contends that the district court should reconsider its denial of class certification based on "the admissible portions of Akhavein’s testimony." Id. at p. 990. However, we have recognized that a district court errs when it "limit[s] its analysis of whether there was commonality to a determination of whether Plaintiffs’ evidence on that point was admissible." Ellis , 657 F.3d at 982. Instead, the district court must engage in a "rigorous analysis" of commonality, rather than "merely conclud[ing] that, because ... evidence was admissible, a finding of commonality was appropriate." Id. at 984.

The district court conducted the requisite "rigorous analysis" in determining that Grodzitsky was unable to demonstrate commonality. Contrary to the dissent’s contention, Akhavein did not proffer an opinion based on "general product engineering principles" completely untethered from the scientific and methodological flaws identified by the district court. Dissenting Opinion , pp. 988-89. In his own testimony, Akhavein stated that the regulators had a "common defect" because they were not "durable enough" to "last the life of the vehicle," conceded that he was unable to identify a "common solution" to the purported defects in the regulators, and confirmed that he had no opinion concerning the proper manufacturing method that should have been utilized by Honda. Finally, his vibration testing provided "no direct correlation" to the durability of the regulators. Due to these concessions, Akhavein did not and could not provide a reliable opinion demonstrating a common defect for over 400,000 regulators. See Ellis , 657 F.3d at 982–84 ; see also Wendell , 858 F.3d at 1232 (reminding us that the focus is "solely" on "principles and methodology, not on the conclusions they generate") (citation omitted). We "analyze not what the experts say, but what basis they have for saying it." Wendell , 858 F.3d at 1232 (citation omitted). The district court in this case properly performed its gatekeeping function in assessing Akhavein’s proffered opinion and concluding that the opinion was not predicated on reliable scientific methodology. See id.

The dissent’s reliance on McKendall v. Crown Control Corp. , 122 F.3d 803 (9th Cir. 1997) is misplaced. See Dissenting Opinion , pp. 988-89. In that case, which was a products liability action "for injuries incurred when a sofa fell on [the plaintiff] while he was operating [a] forklift," we held that the expert’s "testimony, based on his engineering experience and his having investigated hundreds of fork lift cases over the past thirty years, that a safety device is feasible, [was] both facially helpful and relevant and seemingly reliable." McKendall , 122 F.3d at 804, 807 (citation and internal quotation marks omitted). We did not hold that an engineer’s generalized opinion suffering from severe methodological and scientific flaws was otherwise reliable. Moreover, we have since recognized that McKendall’s holding that "the Daubert factors are relevant only to testimony bearing on scientific knowledge and did not apply to an expert testifying on how a product ought to have been designed" has been overruled. See White v. Ford Motor Co. , 312 F.3d 998, 1007 (9th Cir. 2002). Thus, McKendall provides little assistance in salvaging Akhavein’s opinion.

Neither do United States v. Laurienti , 611 F.3d 530 (9th Cir. 2010) and United States v. Finley , 301 F.3d 1000 (9th Cir. 2002) compel a finding that the district court abused its discretion in this case. See Dissenting Opinion , p. 989. In Laurienti , a criminal prosecution for securities fraud, we concluded that the district court erred in excluding testimony concerning National Association of Securities Dealers rules and the expert’s "general knowledge of the industry as a whole." 611 F.3d at 548. In Finley , we held that the district court erred in excluding the entirety of an expert’s testimony in a criminal case as a discovery sanction under Federal Rule of Criminal Procedure 16(d)(2). See 301 F.3d at 1018. Because these cases did not address a deeply flawed and unsupported expert opinion offered in the context of the rigorous analysis applied to assess commonality for a putative class action, they are unpersuasive on this point. Finally, our opinion in Wolin v. Jaguar Land Rover N. Am., LLC , 617 F.3d 1168 (9th Cir. 2010), lends no assistance to the dissent’s position. In that case, the existence of a defect was undisputed. See id. at 1170–71 (focusing on whether the defect was common rather than on the existence of a defect).

IV. CONCLUSION

The district court did not abuse its discretion in excluding Akhavein’s expert opinion under Daubert . The district court properly held that Akhavein’s opinion was unreliable due to Akhavein’s failure to utilize a workable standard supporting his design defect theory; the lack of supporting studies or testing to demonstrate a common design defect; and deficiencies in Akhavein’s methodology.

In the absence of the report, the plaintiffs failed to demonstrate commonality, as the remaining evidence consisted solely of highly individualized complaints.

AFFIRMED.

MURGUIA, Circuit Judge, dissenting:

Although I agree that the district court acted within its discretion in excluding certain parts of Akhavein’s expert testimony, the district court abused its discretion when it excluded the testimony in its entirety, the admissible portions of which may have critically supported Plaintiffs’ motion for class certification. I would therefore reverse the district court’s order excluding Akhavein’s expert testimony.

To be sure, for many of the reasons identified by the majority, the district court did not abuse its discretion in excluding portions of Akhavein’s testimony. The properly excluded portions included Akhavein’s opinion that regulators should last the life of the vehicle, Akhavein’s views on the probabilities that any given failure is due to the alleged defect, and Akhavein’s opinion that the alleged defect in fact caused the regulators he examined to fail. In rendering these opinions, Akhavein failed to, among other things, cite any industry standards for normal or expected regulator replacement rates, cite any peer-reviewed literature regarding regulator durability, and test a statistically significant number of regulators to opine on the probabilities that any given Honda Pilot regulator failed because of the alleged defect.

The district court, however, abused its discretion when it failed to separate these opinions from the remainder of Akhavein’s report, choosing instead to exclude Akhavein’s testimony in its entirety. See Reed v. Lieurance , 863 F.3d 1196, 1208–09 (9th Cir. 2017) ("While the district court may have had a proper basis to exclude portions of the expert report in its discretion, ... the district court abused its discretion in excluding the entirety of [the expert’s] testimony[.]"). There is no dispute that Akhavein—an engineer with twenty-five years of education and training—is qualified to opine on product design and forensic engineering. See City of Pomona v. SQM N. Am. Corp. , 750 F.3d 1036, 1044 (9th Cir. 2014) ("[Expert testimony] is reliable if the knowledge underlying it has a reliable basis in the knowledge and experience of the relevant discipline." (quoting Primiano v. Cook , 598 F.3d 558, 565 (9th Cir. 2010) )). Nevertheless, the district court went too far when it excluded even those of Akhavein’s opinions that were rooted in first-hand observations and testing—including his opinion that the regulators were designed in a way that made them susceptible to failure as well as his basic opinion regarding the vibrational frequencies that the regulators would be exposed to with normal vehicle use. These opinions—"conclusion[s] from a set of observations based on extensive and specialized experience"—have a reliable basis and would be helpful to a jury tasked with assessing whether the regulators are defective. Kumho Tire Co. v. Carmichael , 526 U.S. 137, 156, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999) ; see Sali v. Corona Reg’l Med. Ctr. , 909 F.3d 996, 1006 (9th Cir. 2018) ("[I]n evaluating challenged expert testimony in support of class certification, a district court should evaluate admissibility under the standard set forth in Daubert ."); City of Pomona , 750 F.3d at 1043–44 ("Under Daubert and its progeny ... ‘[t]he district court is not tasked with deciding whether the expert is right or wrong, just whether his testimony has substance such that it would be helpful to a jury.’ " (quoting Alaska Rent-A-Car, Inc. v. Avis Budget Group, Inc. , 738 F.3d 960, 969–70 (9th Cir. 2013) )).

The majority argues that these opinions are—like the others—"riddled with scientific and methodological flaws." But the majority fails to convincingly explain why Akhavein’s observations about the design of Honda Pilot regulators and the normal-use vibrational frequencies that the regulators would be exposed to—which, again, are rooted in his twenty-five years of education and training in the field of engineering— are "fatally flawed." While the majority contends that these observations had "no scientific basis," it notably fails to explain with any level of specificity how it comes to that conclusion. Rather, the majority comes to its conclusion by indiscriminately lumping these opinions together with the above-mentioned defective opinions, seemingly concluding that a few bad apples must mean that the whole barrel is rotten. But contrary to the majority’s contention, there is nothing "artificial" about declining to exclude an expert’s testimony wholesale simply because some of his opinions are flawed.

The majority unpersuasively cites as an example of Akhavein’s "fatally flawed" opinion Akhavein’s examination of "only twenty-six Honda Pilot window regulators" "to prove a common defect in over 400,000 window regulators for class vehicles[.]" But as the district court explained, Akhavein’s failure to test a statistically significant number of regulators posed a problem only to Akhavein’s opinion regarding the probability that a given regulator failed because of the alleged defect. That is, Akhavein does not need to test a statistically significant number of regulators to give his general observations and conclusions about the design of Honda Pilot regulators. Additionally, the majority cites Akhavein’s acknowledgment of the lack of direct correlation between his vibrational testing and failure of the regulators. But this does not undermine Akhavein’s opinion that—as a matter of general product engineering principles—products should be designed with the vibrational frequencies they are expected to experience in mind, and that Honda failed to take this into account. See McKendall v. Crown Control Corp. , 122 F.3d 803, 807 (9th Cir. 1997) ("[The expert] reached his expert conclusions by drawing upon general engineering principles and his twenty-two years of experience as an automotive engineer." (quoting Compton v. Subaru of Am., Inc. , 82 F.3d 1513, 1519 (10th Cir. 1996) )).

While Akhavein’s admissible opinions may not alone definitively establish that the regulators are defective , Plaintiffs do not need to demonstrate that they will prevail on the merits to satisfy commonality; they need only show that a classwide proceeding would "generate common answers apt to drive the resolution of the litigation." Torres v. Mercer Canyons Inc. , 835 F.3d 1125, 1133 (9th Cir. 2016) (emphasis omitted) (quoting Wal-Mart Stores, Inc. v. Dukes , 564 U.S. 338, 350, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011) ); Wolin v. Jaguar Land Rover N. Am., LLC , 617 F.3d 1168, 1172 (9th Cir. 2010) (finding that plaintiffs "easily satisf[ied] the commonality requirement" because each class member’s claim involved "the same alleged defect, covered by the same warranty, and found in vehicles of the same make and model").

For these reasons, the district court abused its discretion when it excluded the entirety of Akhavein’s expert testimony, an error that cannot be construed as harmless. United States v. Laurienti , 611 F.3d 530, 548 (9th Cir. 2010) (finding district court’s erroneous exclusion of expert testimony harmless because, "[u]nlike in many cases, where the district court prohibits all testimony by a proffered expert, the district court here permitted testimony by [the expert] on a wide range of topics and sustained objections only to a limited set of questions"); United States v. Finley , 301 F.3d 1000, 1018 (9th Cir. 2002) (holding that the exclusion of expert testimony was not harmless because it was "the only evidence of [defendant’s] diagnosed mental disorder, and the [district] court excluded the entire testimony").

Finally, because the district court explicitly denied class certification on the basis that "[w]ithout [Akhavein’s] testimony, Plaintiffs have no way of demonstrating the commonality required by Rule 23," I would also reverse the district court’s denial of class certification and remand so that the district court could reconsider Plaintiffs’ motion in light of the admissible portions of Akhavein’s testimony.

Therefore, with respect, I dissent.