Using social science to assess sex discrimination cases

A lawsuit's overall odds of success sound useful. But what do those odds tell me about my particular case? That is the problem we will need to understand here.

First, we get a reason to welcome social science into legal research. Traditional research sometimes concentrates on a few unrepresentative cases or on obscure issues that do not help people with actual legal problems. Notice the word “justifiably”: the author agrees with that criticism. Zirkel and Schoenfeld have done something useful by applying social science tools to employment discrimination against women in higher education. We should not read the coming criticism as a rejection of their entire project.

Then comes “However.” Their enthusiasm for one particular technique, outcomes analysis, is misguided. What is that technique? The next paragraph tells us: counting how many plaintiffs win and how many lose. We do not need to know the technical term in advance. We need to understand what the researchers are doing and why the author objects.

One problem is that plaintiffs who feel certain they are right and will win are not persuaded by discouraging statistics. More importantly, an overall success rate lumps together cases with very different evidence. Imagine that 30 percent of plaintiffs win. That made-up number still would not tell me whether my case is hopeless or unusually strong. Do I have a written admission of discrimination? Did the employer seriously violate its procedures? Those details could make my situation very different from the average. Tell me why people won or lost. Just counting them does not do enough.

The last paragraph gives us two ways to investigate those reasons. Policy capturing sounds complicated, and the sentence describing it is long. Slow it down: researchers read each court opinion, identify factors discussed there, and analyze whether those factors help predict the outcome. For example, how were employees evaluated? Were the evaluators trained? We do not need to memorize the entire list. The point is to connect features of a case with its result.

The other approach reads complete trial transcripts and examines matters such as the alleged misconduct, the remedy, and the evidence that helped the plaintiff prevail. Here is a distinction worth keeping straight: policy capturing uses the court's opinion; the other technique uses the full transcript. Both study litigated cases. The difference is not that one looks at events before a lawsuit and the other looks at events afterward. A full transcript can include relevant details the judge did not discuss in the opinion.

Neither approach promises certainty. The final sentence even acknowledges a limitation: those studies cover a particular period. Still, they give plaintiffs and defendants a more useful basis for assessing a case. The author supports social science that explains outcomes and criticizes a technique that merely totals them.