Are these statistics enough to establish a firm-wide pattern of discrimination? They might be. A plaintiff can prove a "pattern and practice" of employment discrimination by showing an unexplained statistical disparity between the composition of the employer's workforce and the composition of the qualified labor pool from which the employer draws its workforce. A nondiscriminating employer would wind up with a workforce that reflects the demographics of the relevant labor market, absent some unusual circumstances. Of course, that's just where the analysis starts: If fewer women are qualified for or interested in the jobs in question, the imbalance wouldn't suggest discrimination. Sophisticated statistical methods can take account of the other factors that might contribute to disparities, like differences in qualifications, interest, or availability.
Once such legitimate factors have been taken into account, it's reasonable to conclude that any remaining disparity is the result of discrimination. And here Wal-Mart's size works in the plaintiffs' favor. Unexplained disparities in small workforces may be attributable to chance. But in larger workforces like Wal-Mart's, random accidents will cancel each other out, making statistical evidence more reliable.
In the classic pattern-and-practice case, the plaintiff doesn't have to point to a corporate policy that joins all of the individual cases of discrimination—it's enough to show, using statistics, that such a pattern exists. The employer as an entity is responsible for the decisions of its managers: Even if each manager exercises his or her own independent judgment, he still acts on behalf of the employer whenever he hires, fires, promotes, or disciplines an employee. So, if a plaintiff can show—using a combination of statistics and a sample of representative cases—that an employer discriminated repeatedly, it is considered sufficient to establish liability. That's what the Dukes plaintiffs hope to show at trial.
But while statistics might tell us that a lot of women have been discriminated against, they can't tell us which ones. A lot of unexplained decisions that harm women suggests a pattern of discrimination. But any single unexplained decision that goes against an individual woman might be due to chance. And it's not practical for the court to hear evidence about 1.5 million cases. As a result, if the case goes ahead and the plaintiffs win, the trial plan calls for any damages to be determined for the entire class and then apportioned to the individual women according to a mathematical formula. As some of the judges who opposed class certification at earlier stages of the litigation complained, this guarantees that women with strong claims will effectively have to share their damages with those who have weak claims. For individuals, the class action offers only rough justice.
Still, rough justice is better than no justice at all. Most of the women in the Dukes class action would never bring an individual lawsuit. Proving individual discrimination is hard. Even if the statistics prove that Wal-Mart discriminated against a lot of women, very few would be able to prove that they were one of them. And even for those who could prove it, the damages they would be entitled to often aren't worth suing over.
More important, civil rights have always been as much about social justice as individual justice. And social justice requires that employers who discriminate be held accountable so they won't keep discriminating. The deterrent effect of a large class-action lawsuit may be the only thing that will encourage employers to root out discrimination among their managers: Tellingly, Wal-Mart has made an impressive effort to treat women more equitably in the 10 years since the case was filed. That's why large pattern-and-practice civil rights cases have been one of the most important types of civil rights litigation since the passage of the Civil Rights Act of 1964. If large civil rights class actions like the one in Dukes can't get off the ground, then for the most part the government will be the only entity able to sue for firm-wide patterns of discrimination.
Ultimately what's at stake in Dukes v. Wal-Mart is whether class-action lawsuits will continue to be a way to address pervasive discrimination, or whether America's battle against prejudice will have to be fought on a case-by-case basis. This has been the central question in many of the most important civil rights disputes of the last 30 years. And in almost every case since the early 1980s, the Supreme Court has come down on the side of individualism against social justice. That statistic doesn't look so good for the women suing Wal-Mart—or for equal opportunity generally.