The Supreme Court Endorses Statistical Sampling to Prove Liability

Last week, in Tyson Foods v. Bouaphakeo, — S. Ct. —-, 2016 WL 1092414, the U.S. Supreme Court affirmed a district court decision certifying a class of workers who sued Tyson Foods under the Fair Labor Standards Act (“FLSA”) and Iowa state labor law for short-changing them the time it took to change in and out of their protective gear, thus unlawfully depriving them of overtime pay.  To establish that the “donning” and “doffing” of protective gear pushed them over the 40-hour overtime threshold, the workers used sampling evidence.  An expert videotaped a representative number of employees donning and doffing the gear, calculated the average time that the process lasted, and applied that number on a class-wide basis.

Tyson appealed the order granting certification, contending that the time it took employees to change in and out of protective gear was an individual question that substantially predominated the class-wide issues, and that the use of sampling evidence “assum[ed] away the very differences that make the case inappropriate for classwide resolution.” Tyson sought a ruling not only that the use of sampling evidence was inappropriate in this case, but also that it was necessarily an improper means of establishing liability in a class action.

The Court, in a 6-2 opinion, rejected Tyson’s arguments. “A representative or statistical sample,” it concluded, “is a means to establish or defend against liability. Its permissibility turns not on the form a proceeding takes—be it a class or individual action—but on the degree to which the evidence is reliable in proving or disproving the elements of the relevant cause of action. Indeed, “[i]n many cases, a representative sample is the only practicable means to collect and present relevant data establishing a defendant’s liability.”  This, the Court held, was one such case.

While Boutaphakeo involves a class action and labor law, its language – the Court’s most definitive statement on the propriety of statistical samples as a means of proving liability – suggests broad applicability. In the False Claims Act (“FCA”) context, courts often permit statistical sampling to prove damages once liability is established, but cases permitting such evidence to prove liability exclusively based on representative evidence are far fewer in number.  Perhaps most notably, in United States ex rel. Martin v. Life Care Centers of America, 2014 WL 4816006 (E.D. Tenn. Sept. 29, 2014), the government alleged that skilled nursing facilities submitted false claims to Medicare for medically unnecessary services. The government sought to establish its case by identifying a small sample of these claims and extrapolating to the more than 154,000 claims alleged to be at issue.  The court permitted the sampling over Life Care Centers’ objection, concluding that “limiting FCA enforcement to individual claim-by-claim review would open the door to more fraudulent activity because the deterrent effect of the threat of prosecution would be circumscribed.”

Thus, in Boutaphakeo, the Supreme Court seemingly endorsed the proposition set forth in Martin: where it is impractical to “collect and present” all of the false claims at issue, presenting a “representative sample” may suffice in establishing a defendant’s liability.