León Cosgrove

Class ‘Commonality’ and Statistical Evidence in the Dukes and Tyson Opinions

By: León Cosgrove LLC

scotus11On March 22, 2016 in Tyson Foods v. Bouaphakeo, the U.S. Supreme Court upheld a $5.8 million judgment against Tyson Foods, Inc., in a “donning and doffing” wage & hour class action by more than 3,000 workers at a pork-processing plant in Storm Lake, Iowa. Tyson had hoped the Court would rely on its 2011 Wal-Mart Stores v. Dukes decision blocking the use of statistical evidence in class certification to reverse certification of the Storm Lake class in a lower court ruling.

However, the Supreme Court found the use by workers at a single location of statistical inferences about donning and doffing times, in the absence of any Tyson company records, to be significantly different from the Wal-Mart workers’ reliance on statistics to define their purported 1.6 million-member nationwide class. The High Court’s Tyson opinion also relied on a 70-year-old Supreme Court decision allowing workers to use statistical evidence in compensation claims when their employer didn’t keep adequate records of their hours.

In the following Q-and-A, León Cosgrove Partner Alec Schultz discusses the pivotal issue of class commonality under FRCP 23(a), and why the Supreme Court’s pro-class decision in Tyson is actually harmonious with its anti-class decision in Dukes.

Q: What was the high court’s opinion, in brief, on the use of statistical evidence to mount a purported nationwide class in Dukes v. Wal-Mart?

A: The crux of the certification issue in Dukes was the issue of commonality under FRCP 23(a).  To support a finding that the requisite commonality existed, the class provided statistical evidence regarding back pay and promotion disparities between men and women, which the Ninth Circuit found to be adequate evidence of commonality. The Supreme Court disagreed, noting that a commonality finding requires all plaintiffs to suffer the same injury, which must be more specific than a general violation of the same statute (which can be violated in unique ways unamenable to class certification).

The plaintiff class had attempted to use statistical evidence to demonstrate that Wal-Mart’s policy of allowing local supervisors discretion over employment decisions resulted in a company-wide practice of discrimination. However, the court found that the statistical evidence, comprised of dividing the country into regions and using comparative analysis to show the percentage of women promoted at Wal-Mart versus the number of women in the regional employment pool, failed on two grounds.

First, the court noted that such a broad framework could not provide evidence of the exact, same type of discrimination on a store-by-store basis, and thus could not demonstrate a common/uniform practice. Second, the court noted that even if the statistical evidence was adequate, simply showing an overall result of disparity is not enough. The plaintiffs had to show a “specific employment practice,” as opposed to a policy of general discretion that could lead to a disparate result.

Q: How did SCOTUS rely, or not rely, on its Dukes precedent in deciding Tyson?

 A: The Court found its pro-class decision in Tyson to be harmonious with its anti-class decision in Dukes because of the key difference that all employees in Tyson worked in the same plant, under the same company policy. In Dukes, the class comprised employees from thousands of different stores across the country.

As a result of all the Tyson class working in the same plant, each plaintiff could have used the same statistical evidence in an individual suit, which was not the case in Dukes.

Q: What will be the immediate impact of Tyson on state and nationwide class actions?  

A: Tyson should not be read as condoning the carte blanche use of statistical evidence to meet the commonality or predominance requirements of Rule 23. The Court was specific in saying that the use of statistical evidence must be analyzed on a case-by-case basis, with the aim of ascertaining whether it can be applied to each class member uniformly.

The likely result of the opinion will be district and appellate courts being more open to the consideration of statistical evidence in specific contexts (i.e. workers in one plant, buyers of one specific product), but the Dukes opinion still stands as an obstacle to the use of statistical evidence on a nationwide scale.

Q: Are there other related issues waiting to be decided that would bear on statistically-created classes, and where are they in terms of being decided by the courts?

A: In Tyson, the Court raised a concern about the class damage award potentially leading to a recovery for uninjured class members, calling it a question of “great importance.”  However, the Court also noted that this specific issue wasn’t properly before it in the case. The conservative block of the Court will likely remain focused on the issue of Article III injury in fact in future class certification cases. Plaintiffs would be wise to seek a bifurcated trial on liability and damages, which could help avoid the issue.


AlexSchultz_BW_webAlec Schultz is a partner at León Cosgrove who specializes in the trial and appeal of complex commercial and antitrust disputes, employment matters and privacy law actions on behalf of plaintiffs and defendants.