Dukes V. Wal-Mart: Analyzing Round Two

As noted in my previous posting, the Ninth Circuit Court of Appeals has agreed to review the District Court’s decision to certify the class in Dukes v. Wal-Mart sex discrimination lawsuit. In this posting, I will review two selected issues in Wal-Mart’s petition for permission to appeal the class certification, as well as some of the responses to these two issues from the plaintiffs to the petition. I will focus only on the issues of particular interest to me as an expert witness, and will let others (maybe George?) attend to additional relevant issues in this appeal and opposition to the appeal.

The role of subjective decision-making.

Wal-Mart’s arguments. A key point in the plaintiffs’ case was that highly subjective decision-making processes in pay and promotion decisions were susceptible to discrimination. Indeed, Wal-Mart’s petition to appeal the class certification asserts that “District Courts within the Ninth Circuit are divided” with regard to such decision-making processes and offers another case, where certification of the class was denied. In that decision, the high degree of subjectivity was viewed by the court as a criticism, but not an actual policy that would make it amenable to meeting the notion of commonality. Thus, Wal-Mart’s petition for review suggests that the “divergence of opinion” regarding the role of highly subjective decision-making in the courts means that the district court’s opinion is worthy of review.

Plaintiffs’ response. The plaintiffs response is that this district court has argued elsewhere that local decision-making does not defeat an argument of commonality. Moreover, the plaintiffs point out that Wal-Mart has a strong degree of centralization, which in turn creates a uniformity of policies and practices. The subjectivity is merely one of part of the argument, according to the plaintiffs. The common practice needed for class certification is a combination of uniformity of practices combined with subjectivity of decision-making.

The prevalence of statistical evidence

In last month’s analysis I provided of the judge’s decision to certify, I pointed out that the “level of analysis” issue is critical in this case.

Specifically, what I said then was:

“different results may be found, depending on the level of the analysis. That is, in the present case, the plaintiffs analyzed pay differences at the regional level (there were 41 regions) and found that women on average earned less, even taking into account various factors (e.g., seniority, performance, etc. see below “Missing Variables”). The defendants analyzed pay differences at the store sub-unit level (of which there were about 7,500) and found that there were very few instances of women being paid less. Without going into the explanation, the judge ruled that the plaintiffs’ expert’s analysis was sufficient to pass initial muster.”

Wal-Mart’s arguments. Turning back to Wal-Mart’s petition, it is not surprising that Wal-Mart criticized the judge’s ruling in this regard. Their argument is that other district courts have noted that focusing on aggregate numbers (e.g., regional level) masks differences from district to district and from supervisor to supervisor, which would preclude a finding of commonality.

Plaintiffs’ response. The plaintiffs cite a different case in the ninth circuit, which they assert addressed aggregated statistical data. I found another case, in a different area of the country, where the district court did accept (for class certification purposes) aggregated analyses (e.g., Warren v. Xerox, 2004) by the plaintiffs. The plaintiffs also contend that in a class certification situation, the court’s role is not to make decisions as to which expert’s evidence is correct; rather, that the purpose is to determine whether the plaintiff’s evidence is “sufficient.” Finally, the plaintiffs noted that Wal-Mart actually analyzed pay differences at the store sub-unit, which they contend resulted in very small sample sizes, making it difficult to find statistical significance (I could not locate the typical sample sizes here, but if true, this could make it difficult to find statistical significance).

In my opinion, these two issues (i.e., subjectivity of decisions and aggregation of statistics) have been critical turning points in court decisions involving class certification.

In sum, the Ninth Circuit Court of Appeals has agreed to review the Wal-Mart sex discrimination class certification decision. Stay tuned for the next round. It will be interesting to see whether the class certification holds up or not!

Sphere: Related Content