957 F.3d 979
9th Cir.2020Background
- Plaintiff Phyllis Grodzitsky, as class representative, sued Honda alleging design defect in window regulators of 2003–2008 Honda Pilots that caused windows to fall and become inoperable, asserting CLRA and UCL claims.
- Plaintiff’s expert, engineer Glenn Akhavein, opined that regulators failed from vibration-induced metal fatigue at the ferrule–carrier interface and that Honda’s tests were inadequate because they used a single frequency rather than a range.
- Akhavein testified regulators "should last the life of the vehicle," relied on a small, non‑statistical sample (26 regulators, ~12 failed), performed limited vibrational testing, cited no industry standards or peer‑reviewed literature, and offered no specific redesign prescription.
- District court excluded Akhavein’s testimony under Daubert for lack of reliable methodology (no objective durability standard, small/unknown sample provenance, no durational correlation of vibration testing, no supporting studies) and found that exclusion fatal to class certification for failure to show Rule 23 commonality.
- Ninth Circuit majority affirmed exclusion and denial of class certification, holding Akhavein’s opinions were methodologically unreliable and insufficient to show a common defect across the putative class; a dissent argued the court should have admitted non‑flawed, observational portions of his testimony.
Issues
| Issue | Plaintiff's Argument | Defendant's Argument | Held |
|---|---|---|---|
| Admissibility of expert under Daubert | Akhavein’s engineering analysis of fatigue and vibration is reliable and relevant to common defect | Akhavein’s opinions lack scientific method, objective standards, and adequate testing | Excluded: expert unreliable (methodology and standards deficient) |
| Adequacy of sample & proof of common defect | Small sample and forensic observations suffice to identify common design defect | Sample too small/biased; no statistical or industry data to generalize to >400k parts | Excluded: sampling and causal inferences insufficient for classwide defect opinion |
| Validity of Honda’s vibration testing | Honda’s single‑frequency tests miss real‑world frequency range; Akhavein’s testing shows vulnerability | Akhavein performed no durational testing or correlation to actual failures | Excluded: Akhavein’s vibration testing did not reliably show causation or durability impact |
| Effect of excluding expert on Rule 23 commonality | Even if parts of testimony excluded, remaining observations could generate common answers for class certification | Without the expert’s opinion plaintiffs cannot show commonality; remaining evidence is individualized | Affirmed denial of class certification: exclusion left only highly individualized evidence |
Key Cases Cited
- Daubert v. Merrell Dow Pharms., 509 U.S. 579 (1993) (trial‑court gatekeeping standard for expert admissibility)
- Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999) (Daubert applies to engineering and non‑scientific expert testimony)
- Wendell v. GlaxoSmithKline LLC, 858 F.3d 1227 (9th Cir. 2017) (expert reliability focuses on principles and methodology)
- Sali v. Corona Reg'l Med. Ctr., 909 F.3d 996 (9th Cir. 2018) (apply Daubert in class certification context)
- Ellis v. Costco Wholesale Corp., 657 F.3d 970 (9th Cir. 2011) (Daubert reliability—not persuasiveness—controls expert admissibility at class stage)
- Reed v. Lieurance, 863 F.3d 1196 (9th Cir. 2017) (district court may not exclude an expert’s entire testimony when only portions are unreliable)
- Wal‑Mart Stores, Inc. v. Dukes, 564 U.S. 338 (2011) (commonality requires common answers apt to drive litigation)
- Wolin v. Jaguar Land Rover N. Am., LLC, 617 F.3d 1168 (9th Cir. 2010) (distinguishing existence of defect vs commonality inquiry)