A divided panel of the Fourth Circuit ruled that a South Carolina district court judge improperly applied Wal-Mart Stores, Inc. v. Dukes in decertifying disparate treatment and disparate impact claims challenging promotion decisions under federal discrimination statutes on behalf of black employees at a single plant. In 2009, prior toDukes, the Fourth Circuit had previously ordered the district court to certify both disparate treatment and impact promotion classes comprised of approximately 150 class members as well as a separate hostile environment class because the plaintiffs provided sufficient statistical, direct, and anecdotal evidence of discrimination to meet Rule 23 (a) and (b) certification requirements. The direct and anecdotal evidence included specific examples of racial slurs and epithets along with liberal displays of the Confederate flag and nooses in the workplace that established that racial discrimination permeated the workplace and transcended the multiple departments within the plant.

In 2012, the district court found that the intervening Dukes decision warranted decertification of the promotion classes because 1) the analytical rigor now mandated by the Supreme Court’s 23(a)(2) analysis for plaintiffs’ statistical evidence was lacking; 2) Dukes placed the burden on the plaintiffs to provide “significant proof” of a general policy of discrimination tied directly to the promotion claims; and 3) a policy of discretionary promotion decisions could not sustain a class unless the plaintiffs could show that decision makers exercised the discretion in the same way. The district court went on to find that even if the plaintiffs still met the (a)(2) commonality standard after Dukes, decentralized promotion decision making in the individual plant departments precluded a finding that common issues predominated under (b)(3).

Interestingly, the district court did not decertify the hostile environment class, finding that the workers sufficiently demonstrated a general policy of discrimination through evidence of an environment that was hostile to black workers and a finding that management ignored or acquiesced to the harassment.

In overturning the district court’s decertification order, the Fourth Circuit initially chastised the district court for engaging in a separate (b)(3) predominance analysis, finding that re-evaluation of the Fourth Circuit’s 2009 mandate was not compelled by Dukes because the Supreme Court only addressed the (a)(2) commonality requirement. The Fourth Circuit then went on to examine the evidence again under (a)(2), and spent a considerable portion of its opinion explaining why the plaintiffs’ evidence was sound and statistically significant, including making specific distinctions between the six-department plant site at issue there versus the proposed nationwide class of 3,400 stores that troubled the Dukes court.

The Fourth Circuit also seemed hard pressed to divorce the overwhelming evidence supporting the hostile environment claim with the promotion claims. The court specifically found that when the workers’ statistical evidence was combined with their anecdotal evidence, the “glue” of commonality that Dukes demanded was met, especially where, according to the court, plaintiffs provided, “substantial evidence of unadulterated, odious racism throughout the Nucor plant, including affirmative actions by supervisors and a widespread attitude of permissiveness of racial hostility.” That was enough for the Fourth Circuit to find that there was a “common answer” to the question of why black workers were disadvantaged at the plant.

In a sharply worded, detailed and lengthy opinion, the dissent challenged the foundation and significance of the plaintiffs’ statistical evidence, disputing that the plaintiffs met the “rigorous analysis” of Dukes. The dissent asserted that because discretion for promotion decisions had been granted to each individual department, Dukesmandated that the plaintiffs were required to prove department-by-department statistical disparities, rather than purported aggregated plant-wide disparities. The dissent also challenged the quality and quantity of plaintiffs’ anecdotal evidence suggesting that the hostile environment evidence that supported the majority’s findings was not tied to the specific promotion decisions at issue, was temporally unrelated to the majority of the claims, and arose mainly in only one of the six departments in the plant. Finally, the dissent disputed that the district court’s Rule 23(b)(3) predominance analysis was inappropriate.

While decentralized decision making cases will still pose challenges for plaintiffs in broad-based employment discrimination class actions post-Dukes, the Brown decision is not surprising given that the overwhelming evidence of a racially hostile environment that swayed the court in 2009 appeared to still play a role in the Brownmajority finding that the Dukes (a)(2) commonality standard was met. With a fairly small class size limited to a single plant site, the Brown case did not suffer from the same commonality disconnect that a nationwide class of 3,400 stores and hundreds of thousands of potential decision makers had for the court in Dukes. Setting aside the dispute over the validity of the plaintiffs’ statistical analysis, in cases without evidence of the “glue” of discrimination found in Brown, purported statistical evidence of discrimination alone will unlikely carry the day.

Brown v. Nucor Corporation, 2015 WL 2167646 (4th Circuit, May 11, 2015).