“If you’re going to play the game, boy, ya gotta learn to play it right” From “The Gambler”, famously recorded by Kenny Rogers
Multiple articles have been written regarding a ruling in re: Duran v. US Bank National Association, California Court of Appeal, 1st Appellate Dist., (February 6, 2012), which incorrectly claim that statistical sampling cannot be used to prove liability in a class action case. The case is being described with words such as a “game changer” “seismic” and “enormous significance”.
Instead, the case stands for the simple proposition that a trial court needs to obtain competent statistical evidence before creating a sampling and statistics plan on its own. The case does not prevent statistical sampling or representative testimony; one simply needs to follow established statistical rules if one wants to embark on this time-saving path.
The Duran case is an employment class action which challenged the exempt classification of 260 salespersons. The defendant bank was primarily relying on the outside sales exemption. Although a job description for the class members clearly supported the exempt status, the factual issue was whether the actual job implementation comported with the job descriptions.
The opinion is thorough and detailed. This occurs in part because the Appellate Court made its ruling based on the facts of this case rather than broad legal principles. Consequently, the opinion summarizes testimony of numerous fact and expert witnesses. Those wishing to contrast the statistical evidence and results in Duran with the statistical evidence and results in their case will have an easy time of it.
Somewhat surprisingly, this is not a case in which there was widely-contrasting expert testimony. The defendant (appellant) summarized the numerous reasons why the statistics themselves demonstrated that they should not be used. In contrast, at trial, the plaintiff generally said nothing about the statistical failings, other than noting the trial court had already ruled on such matters in the process of constructing the trial management plan.
The case’s underlying statistical issue involves sampling error. Sampling and related inferential statistics provide significant cost savings when compared to attempting to observe an entire population. The samples to be used must be selected randomly, meaning that each item in the population has an equal chance of being selected. Even if the sampling process is entirely random, there remains a possibility that the sample might not be representative of the true state of the population.
For a simple example of sampling error, suppose a large urn contains 300 colored balls – 100 each of red, yellow and blue. In a real situation, we actually would not know that exactly a third of the balls are each of the three colors. Our sampling will be used to estimate the color of the balls in the urn by pulling a sample of balls. Suppose the first ball randomly pulled is red. With this information, we might try to suppose that all of the balls are red. But with only one sample, we cannot be reasonably certain of this result. So we randomly pull a second ball, which luck allows is also red. With two red balls pulled, we are more confident that the urn is full of red balls, but we would still be wrong. The sample size is simply too small for us to reach any reliable conclusion. Inferential statistics instructs us as to how many balls need to be pulled, and how we are to interpret the results that are achieved. Sampling error is the possibility of inaccuracy that is caused by observing only a portion of a population (i.e. a sample) rather than the whole population. Stated otherwise, sampling error is the uncertainty caused by observing a random sample that varies from the characteristics of the whole population. Sampling error is reduced by taking a large enough random sample from the population, although the cost of doing this diminishes the desired cost-savings from using a sample.
In its simplest form, the trial court used too small a sample. The Appellate Court summarized its findings as follows:
“We agree with USB that the trial court here did not follow established statistical procedures … We turn first to the claim that the trial plan violated USB’s right to due process of law. USB contends the court’s strategy of relying solely on the evidence derived from a 21-person sample to determine class-wide liability and restitution violated principles of due process and resulted in a statistically invalid result, as evidenced by the 43.3 percent margin of error in weekly overtime hours. USB also asserts the court infringed on its due process rights when it rejected USB’s efforts to introduce evidence challenging the individual claims of the 239 absent class members. We agree with USB that the trial plan employed here was seriously flawed and the judgment must be reversed….
There was no statistical foundation for the trial court’s initial assumption that 20 out of 260 is a sufficient size for a representative sample by which to extrapolate either liability or damages. Neither party proposed a trial plan based solely on the selection of a representative group of plaintiffs, let alone a group of 20. The court appears to have arrived at this procedure on its own, without reliance on legal precedent or the advice of expert witnesses. In their brief on appeal, plaintiffs state that the trial court “ultimately adopted a trial management plan modeled on Dr. Drogin’s proposal.” While Drogin did propose the use of representative testimony from a randomly selected group of plaintiffs, Drogin did not offer any advice as to the size of the group. Further, Drogin indicated that the group was to be selected only after a survey of the BBO’s [class members’] duties and hours had been conducted.”
The defendant complained that the Court’s sample size was not representative of the entire population. Although agreeing that the defendant in this case was denied the opportunity to present information that contradicted the sample, the Appellate Court continued to approve the use of properly-grounded statistics. The following is representative of the Appellate Court’s analysis:
“USB repeatedly attempted to introduce evidence pertaining to non-RWG [testifying] class members, including those for whom USB had offered specific evidence refuting their potential claims for recovery, such as sworn declarations and/or deposition testimony in which they averred to having spent the majority of their time outside the office. USB notes the judgment awards an average of over $50,000 to each absent class member, notwithstanding that USB offered evidence that potentially could have prevented, at a minimum, approximately one-third of these individuals from receiving any recovery. The four former named plaintiffs who testified at their depositions that they spent more than half their time outside the office (strongly suggesting they were properly classified as exempt), were together awarded over $160,000 in overtime compensation. Yet the trial court excluded their testimony at trial on the ground that it was “irrelevant” because it did not comport with the court’s trial plan. Unfortunately, relevancy was dictated by the court’s trial plan rather than by the trial itself as it unfolded in the courtroom.
While we do not disagree with the proposition that statistical sampling is a tool that may be utilized in appropriate cases, it does not follow that it was proper for the trial court in this case to limit presentation of USB’s affirmative defense solely to the 21 members of the representative group.”
Given the size of both the judgment ($15 million) and the award of counsels’ fees ($18 million) by the trial judge, the Duran ruling will likely be appealed to the California Supreme Court. The Supreme Court should not alter the Appellate ruling.