# Shimoda Law

Statistical Analysis In Class Actions - A Mathematician's Dream

#### Written by By Galen T. Shimoda and Erika R. C. Sembrano

Any lawsuit will involve numbers.  Employment law, especially wage and hour law, revolves around numbers, such as an employee’s salary, unpaid overtime wages, unpaid minimum wages, penalties, hours worked, and much more.  In individual lawsuits, or even lawsuits with a small number of plaintiffs, numbers are not daunting.  It is generally not difficult to calculate total hours worked, total unpaid hours, and thus total unpaid wages.  Arriving at an almost-100% accurate damages amount is therefore entirely possible in individual actions.  But, class action lawsuits can make damages calculations more difficult.  Class actions can be comprised of anywhere from thirty to literally thousands of class members.  One can easily imagine how difficult damages calculation can become without a systematic method in place.

The current answer to this conundrum is “representative evidence,” a statistical sampling of a select number of class members and then projecting those results to cover the entire class.  Courts and attorneys have seen the advantages and disadvantages of using statistical analysis in support of damages calculations.  An advantage is having the answer sooner rather than later; a disadvantage is the reliability of the relatively small sample to be used to cover the entire class.

Not too long ago, in 2014, the California Supreme Court in Duran v. U.S. Bank Nat’l Assoc. criticized the use of statistical samples and representative testimony, essentially stating that not enough was done by the lower courts to determine whether a sample size was statistically appropriate and whether there was a reasonable margin of error in arriving at the resulting numbers.  The sampling errors, the Court noted, resulted in an unreliable conclusion.  Interestingly, in Duran, statistical sampling was used to provide both liability and damages in a class action context.  The Court did note, however, that using statistical sampling to provide damages is “less controversial.”  In the end, the statistical sampling used in the case was simply too flawed but using statistical sampling to prove liability seemed to make it even harder for the plaintiffs to prevail.

Then, this year, the U.S. Supreme Court finally addressed representative evidence and did not denounce it!  In Tyson Foods, Inc. v. Bouaphakeo, the plaintiffs filed a class action and collective action in federal court.  In general, the plaintiffs claimed that their employer failed to pay them for all hours worked.  In support of their request for damages, the plaintiffs testified, introduced videos of unpaid hours worked, and brought in an expert to testify as to the average time spent that was unpaid.  Another expert even analyzed the employer’s time records, pay records, and the unpaid amounts owed.  The jury eventually found in favor of the plaintiffs and awarded \$2.9 million of the \$6.7 million initially requested.  The U.S. Supreme Court affirmed the trial court’s ruling and the appellate court’s affirmation of the trial court’s ruling.  The U.S. Supreme Court allowed the use of representative evidence emphasizing that whether it is admissible depends on whether the evidence is reliable.  In Tyson, the representative evidence helped reach a damages amount due to the employer’s failure to keep accurate time records.

Tyson ultimately provides more insight as to how and when to use representative evidence in our favor so that a Court and/or jury will consider it seriously.  Specifically, whether a party may properly use representative evidence will depend on the purpose of the evidence and on the claim’s elements themselves.  In Tyson, using the representative evidence was acceptable because of the lack of time records and because of the reliability of the representative evidence to demonstrate class-wide damages thereto.  However, Duran’s concerns still prevail regarding whether representative evidence can (and should) be used in determining liability.  In this latter scenario, parties should strive to have the representative evidence (and statistical sampling) as reliable as they can get it.

Nevertheless, parties should keep the option of representative evidence open, especially in cases where representative evidence will help “fill in the gaps,” like the Tyson employer’s gap in records filled in by representative evidence.  The first step is whether the case is proper for the use of representative evidence.  Once that issue is overcome, then the second step is to demonstrate the reliability of the representative analysis, and one would do well, if in an unpaid wages case like Tyson, to have more than one expert, credible testimony, and other forms of reliable evidence to support the claims for damages.  One should also consider the flaws found in Duran and consider overcompensating when trying to obtain a valid statistical sampling.  Unfortunately, the Tyson court was reluctant in establishing a bright-line rule to follow to ensure the use of representative evidence.  Yet, Tyson presents a small but sure step for representative evidence in class action lawsuits.

If you think you may have potential claims relating to your employment or believe you have claims appropriate for a class action, contact our office to have your claims evaluated.

Back to Articles