The United States Supreme Court resolved a circuit split by holding that class-action wage-and-hour plaintiffs may rely on statistical analysis to calculate average uncompensated work time as a means of establishing liability and damages. In its March 22, 2016 decision in Tyson Foods, Inc. v. Bouaphakeo, the Court upheld a $5.8 million judgment against Tyson Foods in a dispute over whether more than 3,000 current and former workers were owed compensation for time spent donning and doffing protective gear. In the 6–2 decision, written by Justice Kennedy, from which Justices Thomas and Alito dissented, the Court stated that otherwise reliable and admissible statistical evidence would be permissible to "fill an evidentiary gap created by the employer's failure to keep adequate records."
The Supreme Court's decision in Tyson Foods follows its seminal decisions in Wal-Mart Stores, Inc. v. Dukes (2011) and Comcast Corp. v. Behrend (2013), cases that each touched on the use of expert testimony. In Dukes, the Supreme Court decertified a nationwide class of employees alleging gender discrimination claims, holding that plaintiffs seeking to represent a class must "demonstrate that the class members 'have suffered the same injury'" and prove their claims are capable of classwide determination "in one stroke." Because the Wal-Mart plaintiffs did not provide significant proof of a common policy of gender discrimination to which each employee was subjected, they resorted to representative evidence in an attempt to overcome the absence of a common policy. The Supreme Court rejected that approach because individual class members could not rely upon evidence detailing the ways in which other employees were discriminated against to prove discrimination against themselves. Two years later, in Comcast, the Court overturned an order granting class certification because the plaintiffs' statistical model for calculating damages fell "far short of establishing that damages are capable of measurement on a class-wide basis."
In Tyson Foods, the plaintiff-employees of an Iowa pork-processing plant brought both class- and collective-action claims against the company, alleging that they were never paid for time spent "donning and doffing" requisite protective gear. In unsuccessfully opposing class certification before the district court, Tyson argued that the employees spent varying amounts of time donning and doffing gear at the beginning and end of shifts, which precluded class certification because the claims were not sufficiently similar to be resolved on a classwide basis. Specifically, Tyson argued that individual questions would "predominate" over questions relevant to the entire class, defeating certification under Rule 23(b)(3) of the Federal Rules of Civil Procedure. In September 2011, a jury awarded the employees $2.9 million (the later addition of liquidated damages brought the total judgment to $5.8 million), finding that the employees had sufficiently shown that putting on and taking off protective gear was an integral and indispensible part of their work in the factory—donning and doffing was compensable work.
To prove liability and damages before the district court, the plaintiff-employees relied upon a study performed by an industrial relations expert who calculated how long various donning and doffing activities took for 53 representative employees, then used that data to calculate average donning and doffing times per shift by department. The averages were then added to the timesheets of each class member to ascertain who worked more than 40 hours a week and the value of classwide recovery. The Eighth Circuit affirmed the judgment in 2014, permitting the extrapolation of classwide damages from a sample of class members who had varying degrees of injury.
Reliable and Admissible Statistical Evidence Can Support Class Certification if the Same Evidence Could Be Used to Establish Liability or Damages in an Individual Action
Tyson Foods held that class-action plaintiffs may—consistent with Rule 23(b)(3)'s predominance requirement—prove liability and damages with statistical modeling and averaging, which the Court referred to as "representative evidence," if that same evidence could be used in an individual case presenting the same issues. Tyson had argued that there was a great disparity in the time each worker took to get ready based on his responsibilities and that the usage of averages would lead to recovery for individuals who had not worked the requisite hours. According to Tyson, the plaintiffs were using statistical evidence to manufacture predominance for class-certification purposes, while also denying Tyson the opportunity to litigate individual defenses.
The Court rejected this argument, holding that statistical evidence may be used in class actions to the same extent that it could be used in an action by an individual class member. Because the Supreme Court's 1946 decision in Anderson v. Mt. Clemens Pottery Co. allowed individual Fair Labor Standards Act ("FLSA") plaintiffs to rely on evidence like that used in Tyson Foods, the Court held the same evidence could be used to prove liability or damages on a classwide basis. And, because the evidence could be used to establish liability or damages on a classwide basis, it did not give rise to the sort of individualized inquiries forbidden by Rule 23(b)(3)'s predominance requirement.
The Tyson Foods Court insisted it was acting in accord with its five-year-old Wal-Mart decision. It explained that, "[w]hile the experiences of the employees in Wal-Mart bore little relationship to one another, in this case each employee worked in the same facility, did similar work, and was paid under the same policy." Thus, the Supreme Court said, "under these circumstances the experiences of a subset of employees can be probative as to the experiences of all of them." In other words, and in contrast to Wal-Mart, each employee in Tyson Foods could have relied on the statistical evidence to prove Tyson's liability in an individual case. As a result, the same evidence could be used to prove Tyson's liability to the entire class in a class-action case.
The Supreme Court stopped well short of handing down a categorical rule concerning the use of statistical evidence in class actions. Rather, it limited its holding to the facts of the case—"[i]n FLSA actions, inferring the hours an employee has worked from a study … has been permitted by the Court so long as the study is otherwise admissible."
In a concurring opinion, Chief Justice Roberts, joined in part by Justice Alito, expressed skepticism that the verdict would stand: "Article III does not give federal courts the power to order relief to any uninjured plaintiff, class action or not." But certain employees here were not underpaid; they worked fewer than 40 hours and were swept into the case only because statistical evidence was used. These individuals, he concluded, could not be awarded damages.
Justice Thomas, joined by Justice Alito, dissented. He would have held that the district court erred by certifying the class; the variability between the employees was so great that individual questions regarding the hours worked by individual employees would predominate over questions capable of resolution on a classwide basis. Additionally, permitting the use of representative evidence would put employers in the unfavorable position of either tracking all time spent on uncompensated work to avoid litigation or facing the threat of a class action based on statistical averages
Even after Tyson Foods, defendants may challenge the use of statistical analyses by class-action plaintiffs in wage and hour cases, with such efforts coming in the form of a Daubert motion, as opposed to when arguing against class certification. In fact, the ruling in Tyson Foods, which demands a case-by-case analysis of whether such representative and statistical evidence is reliable and admissible, signals defendants to aggressively attack the methodology used by experts and the ultimate reliability of their data. While plaintiffs may argue that the Tyson Foods decision permits them to draw inferences from such evidence, that only is the case if the evidence could be relied upon by a class member to prove his individual claim.
Additionally, the Supreme Court made reference to Federal Rule of Evidence 403 in its discussion of the permissibility of a representative or statistical sample. As Rule 403 permits a district court to exclude relevant evidence for reasons like unfair prejudice, issue confusion, jury confusion, or because it is prejudicial or cumulative, this suggests that employers may be able to leverage the existence of alternative forms of individualized evidence (such as when employees swiped a card to enter and exit through a security gate) to foreclose the use of averages derived from a statistical sample.
Finally, the Tyson Foods Court's embrace of statistical evidence depended heavily on its earlier Mt. Clemens decision, which permitted the use of such evidence in individual cases. Outside the FLSA context, such evidence may be impermissible. And in these contexts, class-action defendants may wish to argue that the need to assess liability or damages on a class-member-by-class-member basis defeats class certification on Rule 23(b)(3) predominance grounds.