The Supreme Court Charts a Narrow Course in the Use of Statistical Evidence at Class Certification
The United States Supreme Court recently ruled in Tyson Foods, Inc. v. Bouaphakeo, No. 14-1146, --- S. Ct. ---, 2016 WL 1092414 (U.S. Mar. 22, 2016), as to when a plaintiff may use statistical sampling in seeking to certify a class. The decision was narrowly tailored to the specific facts and cause of action at issue in the matter. Thus, the Court declined to adopt a categorical rule, cautioning that the admissibility of such evidence must be made on a case-by-case basis.
In Bouaphakeo, the plaintiffs––employees at a Tyson Foods (“Tyson”) plant––asserted a “donning and doffing” claim for overtime wages. They claimed that they were entitled to compensation, including overtime compensation, for the time spent putting on and taking off protective gear. Specifically, the plaintiffs asserted that they were only being paid for the time worked on the actual production assembly line, not for the preparation time. The plaintiffs brought a class action for state-wage claims and a collective action for Fair Labor Standards Act (“FLSA”)-based claims. At the class certification stage, the plaintiffs faced an evidentiary problem in attempting to establish on a classwide basis how much time each putative class member spent putting on and taking off protective gear, in part because different employees wore different types of gear.
To overcome the evidentiary hurdle, the plaintiffs relied on a study performed by an industrial-relations expert, who conducted videotaped observations and examined how much time a sample of workers took to put on protective gear before a shift and how long they took to remove it after a shift. The expert averaged the time spent on these activities by the sample population. The averages were then added to the time sheets of each employee to determine which class members worked over 40 hours (and thus were entitled to “overtime” pay) and to calculate the value of classwide recovery on that basis. Tyson argued that the varying amounts of time it actually took employees to put on and take off different protective gear made reliance on the expert’s sampling method improper and that the method’s use would lead to recovery for individuals who had not, in fact, worked more than the requisite 40 hours and thus were not entitled to recover damages. At the same time, Tyson did not bring a Daubert challenge to the expert opinion, which the Supreme Court would later consider in rejecting Tyson’s position.
In certifying plaintiffs’ proposed class, the District Court held that common questions, such as whether putting on and taking off protective gear were compensable under the FLSA, were amenable to classwide resolution even if not all employees wore the same gear. After a class trial, the jury found in favor of the class of employees and awarded $5.8 million in unpaid wages together with liquidated damages to the class as a whole. On appeal, the Eighth Circuit affirmed the judgment and the award.
The Supreme Court’s Decision
The Supreme Court affirmed in a 6–2 decision, with Justice Kennedy delivering the opinion of the Court. The Court stated, “[w]hether and when statistical evidence can be used to establish classwide liability will depend on the purpose for which the evidence is being introduced and on the elements of the underlying cause of action.”  The Court held that statistical or representative evidence is permissible where “each class member could have relied on that sample to establish liability if he or she had brought an individual action.”  The Court emphasized that representative evidence was appropriate in this case “to fill the evidentiary gap” created by Tyson’s failure to keep adequate time records as required under the FLSA. The Court ruled that in a circumstance where required records are absent, an employee may produce “evidence to show the amount and extent of work as a matter of just and reasonable inference.” 
In deciding the matter, the Court distinguished Wal-Mart Stores, Inc. v. Dukes.  There, the Court had rejected the use of representative evidence in a nationwide employment discrimination class action. In particular, the Dukes plaintiffs sought to rely on deposition testimony of a statistical sample of the class in an attempt to overcome the fact that the putative class members were not subject to a common policy of employment discrimination. The Court, however, rejected this proposed “trial by formula,” determining that because the individual class members were not similarly situated, none of them could have prevailed in an individual suit by relying on depositions detailing the ways in which other employees were discriminated against by their particular managers. In contrast, the Bouaphakeo court stated that the employees worked at a single facility, did similar work, were paid under the same policy, and could have introduced the expert’s study in a series of individual lawsuits.
The Court did not accept Tyson’s argument that reliance on statistical evidence would strip it of the ability to defend the suit. In doing so, the Court pointed out that a defendant has the ability to challenge the statistical study under Daubert or otherwise establish that the study was not representative, but that Tyson had not pursued those avenues.
The Court did not reach the second issue presented in the case, namely whether class members who did not work more than 40 hours per week and were thus ineligible for overtime pay could recover based upon the statistical evidence, despite lack of actual harm. Instead, the Court ruled that the issue was not ripe because the $5.8 million award had not yet been distributed to class members.
The Chief Justice separately concurred (joined in part by Justice Alito) to emphasize that the Court’s ruling does not recognize a relaxed standard of proof under the FLSA.  In addition, the Chief Justice expressed that the jury verdict could not stand if the statistical evidence did not establish a right to recover damages on behalf of each class member.  In particular, the Chief Justice stated, “[i]f there is no way to ensure that the jury’s damages award goes only to injured class members, that award cannot stand. This issue should be considered by the district court in the first instance.” 
In a dissent joined by Justice Alito, Justice Thomas went further, opining that the District Court erred in certifying a class where the statistical evidence was insufficient to establish common proof of liability (i.e., uncompensated overtime worked) as to each class member. 
Implications of the Decision
Because the circumstances under which the use of statistical evidence will vary from case to case, the Court declined to adopt a categorical rule governing the use of such evidence in class actions. Rather, Bouaphakeo mandates that in assessing whether a class may be certified, a trial court must review the admissibility of statistical or representative evidence based on the specific circumstances of the case at hand, the purpose for which the plaintiff seeks to introduce the evidence, and the nature of the underlying cause of action.
Important to the Bouaphakeo decision is the unique scenario the FLSA creates. Employers are specifically obligated under the FLSA to maintain time records. In Bouaphakeo, however, the employer did not have time records of how long the workers spent donning and doffing protective gear. The Court thus relied on the absence of the requisite time records in finding permissible the plaintiffs’ use of statistical evidence to support their claims. While the Bouaphakeo decision permits the use of statistical evidence to establish classwide liability in that narrow circumstance, such evidence may not be appropriate in many other legal and factual settings. Nonetheless, business entities should carefully examine the Court’s guidance in Bouaphakeo, including the following admonitions:
- To maintain adequate records (especially time records for FLSA compliance) to avoid creating “evidentiary gaps” that might entitle a plaintiff to use representative evidence;
- To scrutinize and challenge the methodologies deployed when statistical evidence is proffered, including bringing appropriate Daubert challenges; and
- To consider options such as bifurcating the liability and damages stages of litigation to mitigate potential problems with the inclusion of uninjured class members.
 Bouaphakeo, 2016 WL 1092414, *8 (internal quotation omitted).
 Id. (quoting Anderson v. Mt. Clemens Pottery Co., 328 U.S. 680, 687 (1946)).
 564 U.S. 338 (2011).
 Bouaphakeo, 2016 WL 1092414 at *13 (Roberts, C.J., concurring).
 Id. at *14–15 (Roberts, C.J., concurring).
 Id. (Roberts, C.J., concurring).
 Id. at *16–17 (Thomas, J., dissenting).