I am a bit late in commenting on the en banc 9th Circuit's class certification in Dukes v. Wal Mart, but it is, after all, 137 pages.
The decison is problematic on at least four levels. First, as others have noted, it masks a rather unorthodox view of Title VII. It was necessary, on class certification, to identify common question of law shared by the approximately 1.5 million class members who work or who have worked in 3400 stores located in 43 regions on which the cases of the six named class members are typical. Tall order and, having read all 137 pages, I am still not sure just what the majority thinks that common claim is other than a generalized allegation of gender discrimination that happened in some unspecified way because Wal Mart gives managers too much discretion and yet controls them in a way that warrants trying the claims of 1.5 million women as one. You might imagine that this required substantial dexterity and so it did.
The first problem is the elision of a substantial dispute about what Title VII requires. While courts have recognized that subjective evaluation procedures may enable discriminatory practices, they have made clear that a policy of subjective evaluation is not itself discriminatory. The panel majority recognizes this but, as the dissent points out, it never adequately identifies what additional issue, common to all class members, would make Wal Mart's subjective evaluation procedures unlawful. What it attempted to do was marry the ways in which certain policies might facilitate discrimination with the conclusion of an expert that it did. But that creates further problems.
One of the ways in which the panel majority tried to close this loop is through the use of something called "social framework" testimony. My colleague Paul Secunda and Melissa Hart of Colorado have written a good discussion of such testimony from the perspective of those who favor such testimony. The term refers to social science testimony that attempts to identify "structural" and "hidden" forms of discrimination. "Social framework" testimony is not any one thing and I actually agree with Paul that there may be times when it is admissable.
But that doesn't mean that it is not extremely problematic. I had extensive experience responding to such evidence in systemic racial discrimination cases in an earlier life. The unvarnished truth is that there is often nothing "scientific" about it at all. It sometimes does involves application of some particularized research(that may or may not be specfically related to the case at hand) which, taken alone, may be probative and admissable.
But often the helpful is combined with the obfuscatory. The proferred testimony may contain useful observations about what the ways in which the world might work. It may be that subjectivity may enable discrimination. There might be empirical support for the notion that filling positions through "word of mouth" may tend to result in hiring people of the same ethnic or racial background of existing employees. Under the right circumstances, a trier of fact may benefit from hearing these things.
But information such as this is often combined with other assertions that, while claimed to be "scientific" because they have been "peer reviewed, are, in fact, ideological presuppositions that won't survive fifteen minutes of well planned cross examination. While one might suggest that the latter problem goes to the weight and not the admissability of the opinion, the problem is often confounded by the way in which the testimony is presented. Methodology is often hidden behind an impenetrable mask of supposedly objective "content analysis" - you and I might say that the expert "read" a bunch of things that the lawyer gave her - to reach an ultmate conclusion. While expert testimony is not objectionable for the latter, it is unclear why an expert is in a better position than the trier of fact (perhaps aided by whatever insights might be offered by social science research) to reach that conclusion.
The point is often to avoid having to try the many "facts" that the expert claims to have relied on. It is an obvious - but often obscured - fact that the notion that social scientists have "found" that subjective procedures might facilitate discrimination tells us nothing about whether that has happened in the case at hand. This might be shown through the introduction of admissable evidence about the matter in suit that demonstrates discriminatory treatment. It could be done by proof that these subjective policies were adopted to further, or have been influenced by, policies that run afoul of Title VII's prohibition of disparate treatment or impact,
But, at least in my experience and by description of the testimony proffered in Dukes, an expert narrator tells the trier of fact that he or she has reviewed the evidence and has concluded (trust me) that such discrimination has occurred, offering enough "examples" to make the testimony believable.
The problem with the scheme is that it is unclear that this second step is an application of expertise. To be sure, some of my finest moments as a trial lawyer came in skewering this sort of thing. But, unless it is carefully controlled, it is a bit like peeling back an onion. How many instances of bad analysis are enough to undermine the opinion?
I have not read the testimony of the plaintiffs' expert in Dukes, but it's description - in both the decision and the piece by Secunda and Hart - sounds distressingly familiar. Perhaps it ought to be admitted, yet neither the panel nor the district judge subjected the proferred social framework testimony to a rigorous Daubert analysis The panel majority, ignoring the Daubert challenge that Wal Mart actually made, seemed to think that to do so would be to address the merits at the class certification stage. But. as Professors Secunda and Hart Hart point out, class certification in elephantine employment discrimination is often tantamount to a decision on the merits. Failure to assess whether this evidence really meets the requirements of Daubert would seem to be a prerequsuite to finding that it identifies a common question of which the named plaintiffs' claims are typical.
Nor did the district court or panel majority carefully consider what this social framework testimony, even if credited, could establish. That certain management practices could enable (or even that they have often enabled) discriminatory practices might be useful to the trier of fact (although the latter formulation might be overly prejudicial), but that evidence doesn't tell us that it happened in this case or that it happened in a way that is common to all of 1.5 million plaintiffs. Even allowing an expert to pick out some examples of where it might have occurred doesn't accomplish that .To make that claim, as the defense points out, it is necessary to identify some other common policy or practice (such as an affirmative policy of discrimination or some other policy that might be seen as compelling or encouraging managers to exercise their discretion in discriminatory ways) and, after 137 pages, it is not clear what that might be.
The majority opinion makes vague references to "corporate culture" and "centralization" that somehow left subordinates with too much discretion but yet directed its exercise in some way that was common across 3400 stores and 43 regions. At some point, this becomes hard to distinguish from an assertion that sexual disparity must be the result of discrimination because discrimination is pervasive. But that assumes the matter to be tried.
It is really hard to see what is there other than subjective policies and what the district court and panel decision saw as strong statistical evidence of discrimination.
While the panel seems to have thought the latter resolves the difficulty, it does not. More tomorrow.