Some important Supreme Court cases are hard to accurately capture in a sound bite, and this is one of them. In a narrow holding, the Supreme Court issued a 6-2 decision in Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. ___ (2016), addressing class claims for overtime compensation certified under Federal Rule of Civil Procedure 23 and the Fair Labor Standards Act of 1938 (FLSA). The Court rejected the employer’s appeal, which sought a blanket rule barring use of statistical proof to establish the elements necessary to certify a class or collective action. Some in the plaintiff’s class action bar will no doubt cheer. However, Justice Kennedy’s opinion for the majority, joined by all but Justice Thomas, provides an analysis revealing that, in the right employment counsel’s hands, this decision can be a relatively powerful weapon in defeating class or collective action certification. Indeed, as we note below, the Tyson decision responds to and is limited by the Mt. Clemens presumption, which notes that in the absence of records it would be presumed that overtime was worked, and the question is how that overtime would be calculated. Although class action lawyers will surely write extensively about the reach of the opinion, at heart, this is a wage and hour case and to understand it properly, one should read it in that context.

The case presented an all-too-familiar fact pattern. The employees brought claims under the FLSA and related state laws which require, in part, that a covered employee who works more than 40 hours a week receive overtime compensation for excess time worked. The plaintiffs were workers in a facility “processing protein” (to be delicate about it, though describing the facility as an abattoir would also be accurate). Given the nature of their work, many workers were required to wear various items of protective gear, some of which can be time consuming to put on as well as to remove. The specific protective gear varied depending on the tasks the workers were required to perform. The employees claimed that if one added the unrecorded donning and doffing time to the hours worked that were recorded, they exceeded the requisite 40 hour workweek and had earned overtime compensation that had not been paid. The primary issue on appeal was whether the plaintiffs’ statistical evidence of the unrecorded donning and doffing time was sufficient to establish a basis for classwide liability. The Court decided that question in the affirmative, but with some significant caveats.

The details matter here. In 2007, Tyson stopped uniformly paying its workers estimated “don and doff” time in favor of compensating only some workers between four and eight minutes for the time Tyson estimated it took the workers to “don and doff” their protective gear. Other employees were not compensated beyond the time spent at their workstations. Notably, “[a]t no point did Tyson record the time each employee spent donning and doffing.” Unhappy with this change, the workers filed suit in the U.S. District Court for the Northern District of Iowa.

Because the employees’ claims were limited to overtime, “each employee had to show he or she worked more than 40 hours a week, inclusive of time spent donning and doffing, in order to recover.” Id. at 5. Plaintiffs relied upon a survey to show the average time it takes workers in two departments to don and doff their protective gear. The plaintiffs’ expert relied on evidence such as employee testimony and 744 videotaped observations of donning and doffing at the plant to analyze how long various donning and doffing activities took for each of the two departments. Using these averages, the plaintiffs’ second expert then estimated the amount of uncompensated overtime work performed by each employee. To no one’s surprise, based upon extrapolations from this survey evidence, the plaintiffs’ experts concluded that nearly all of the employees in the class “had potentially been undercompensated to some degree.”

We pause at the risk of getting into the statistical weeds because, again, the details matter. The sample size was relatively small. The expert used about 53 employees per donning- or doffing-related activity to extrapolate averages for the 3,344-person class. That works out to approximately 1.5 percent. The expert then averaged the amount of time that the sample employees spent per activity, concluding with an estimate that all cut or re-trim department employees spent 18 minutes per day on uncompensated activities (including donning and doffing), while kill department employees averaged 21.25 minutes.

The expert’s data also showed, however, that there were material variances in the amount of time that individual employees spent on the same activities. Cut and re-trim employees took between 0.583 minutes and more than 10 minutes to don pre-shift equipment at their lockers. Post-shift doffing took one employee less than two minutes and another employee over nine minutes. Kill department employees had similar variances. No two employees performed the same activity in the same amount of time, and the plaintiff’s expert observed “a lot of variation within the activity.” That kind of variance resulting in those averages forced the concession that some within the class likely were owed nothing.

Critically, the defendant did not challenge the statistical validity of the plaintiffs’ evidence underDaubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). The defense instead opted to argue to the jury that the variation admitted by the plaintiffs’ expert “made the lawsuit too speculative for class-wide recovery.” That was a strategic decision and, given the state of the law at the law at the time, not an intrinsically poor one. A Daubert challenge may or may not have prevented this evidence from reaching the jury, but it would have preserved the issue for review on appeal. In any case, the expert opinion was admitted and the jury was left to determine what weight it merited. It seems that the jury did not accept the expert’s estimates wholesale — as the $2.9 million damages award was less than half the $6.9 million the experts claimed their statistical analysis showed. 

Although the Court found that Tyson “present[ed] no occasion for adoption of broad and categorical rules governing the use of representative and statistical evidence in class actions,” the Court’s opinion is by no means an outright boon for the plaintiffs’ bar. Rather, the Court emphasized that its decision was specific to the unique facts at issue and that “[w]hether or when statistical evidence. . .can be used to establish class-wide liability will depend on the purpose for which the evidence is being introduced and on ‘the elements of the underlying cause of action.’” Slip Op. at 2 (quoting Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011)). The Court found that it was not appropriate to “establish general rules governing the use of statistical evidence, or so-called representative evidence, in all class-action cases.” Id. at 10.

The Court specifically pointed to the similarities between the workers in this action as a basis for affirming the judgment, explaining “One way for respondents to show, then, that the sample relied upon here is a permissible method of proving class-wide liability is by showing that each class member could have relied on that sample to establish liability if he or she had brought an individual action.” Id. at 11. Therefore, it remains critical that the defense still be allowed to present defenses and there must be an underlying policy or practice giving rise to the potential liability.

Additionally, the Court ruled that the evidence considered must lend itself to statistical sampling. Wide variations in class member experiences — as is often the case in, for example, exempt status or independent contractor misclassification cases — will be critical in challenging whether such statistical evidence may be relied upon absent evidence specific to each class members’ experience. Such a defense will likely rely heavily on the sophisticated use of expert testimony to discredit improper statistical sampling.

The reason plaintiffs relied upon the statistical evidence in the first place was because the employer had not kept records of the time each employee spent donning and doffing their protective gear. Mirroring the Court of Appeals’ reliance on Anderson v. Mt. Clemens Pottery Co., 328 U.S. 680 (1946), the Court explained that when an employer violates a statutory duty to keep proper records, an employee will have carried his or her burden if “’he proves that he has in fact performed work for which he was improperly compensated and if he produces sufficient evidence to show the amount and extent of that work as a matter of just and reasonable inference’” Id. at 12 (quoting Mt. Clemens, 328 U.S. at 687). The burden then shifts to the employer to present “‘evidence of the precise amount of work performed’” or “’evidence to negative the reasonableness of the inference to be drawn from the employee’s evidence.’”Id. (quoting Mt. Clemens, 328 U.S. at 687–88). Applying Mt. Clemens to the facts in Tyson, the Court noted that the plaintiffs relied on their experts’ statistical evidence to fill an evidentiary gap created by the employer’s failure to keep adequate time records. As a result, the Court found that had the employees proceeded with individual lawsuits, reliance on the representative evidence would have been permissible, which is a key point. The case does not discuss use of statistical evidence in any other context.

This outcome is a stark reminder of an employer’s burden to maintain proper records. Had such records been maintained for each individual, the plaintiffs’ reliance on a representative sampling would not have been an issue. Notably, in his dissent, Justice Thomas aptly notes that whether the time itself is compensable — and therefore gives rise to a duty to maintain records — may also be at issue in the litigation. Put simply, if the time was not compensable the employer had no obligation to keep track of it. Therefore, employer defendants should carefully consider the impact that a lack of records on any given claim will have and should fashion other defenses accordingly.

The Court did not decide an issue that had been closely watched in the case; namely, whether a class could be certified that included persons that had not been harmed. That issue was sent back to the district court to sort out on remand in devising some way to distribute the damages award. The Court — in both the majority and dissent, and especially in the concurring opinion of Chief Justice Roberts — highlights that whether this can properly be carried out remains an open question, and subject to challenge by the defense on remand. Because the jury awarded a lump sum so much less than the experts claimed their evidence showed, the trial court’s challenge will be sorting out (a) precisely who among the class are eligible to recover, and (b) how much to award each individual. That may prove to be a daunting task on a class wide basis for the trial court on remand, and the experts’ statistical evidence will be of little assistance.

In sum, the Court’s decision was not a sweeping endorsement of statistical evidence. It also is not an across-the-board prohibition. Instead, it serves as a reminder of the strategic importance of documenting employee activity in a pre-litigation labor and employment environment, as well as careful and individualized assessments of newly filed and existing class and collective cases to shape the analysis of experts to ensure that impermissible statistical analysis is effectively challenged and limited.