It was a busy week for the Supreme Court. Among the many headlines it made were allowing same-sex marriages to move ahead in California, extending federal benefits to same-sex couples, and declawing the Voting Rights Act. Getting slightly shorter shrift—partly because the decision came earlier, partly because the court's ruling was harder to understand, partly because it was simply less consequential—was another decision, handed down 7–1, on affirmative action.
In a nutshell, the Supreme Court decided not to rule itself but rather to send the case back to a lower court to decide whether the University of Texas at Austin's use of race as a factor in admissions is acceptable. The Supreme Court included instructions to the lower court to take a harder look at the admissions policy to ensure that there was definitely, positively no other way to ensure a diverse student body other than looking at race. In other words, the court made affirmative action harder, but it acknowledged that it is sometimes necessary (and thus constitutional).
I am personally not sure where I stand on affirmative action, but I think I think that the Supreme Court did the best they could with an extremely sticky issue. Don't get me wrong—I certainly understand the gripes that many have had with the decision, namely that the court "punted" their ruling and passed it off as not their problem. But if I were a Supreme Court justice, and I were wrestling with my position on affirmative action, I would do the same thing: leave the option open but let someone else decide how to apply it.
That's because, while I believe there must be some accounting for students' different backgrounds in the admissions process, I have no idea of the best way to do so. That's for someone smarter than me—and smarter even than the nine Supreme Court justices—to decide. Someone like a sabermetrician. Let me explain.
In baseball, like in academic achievement, players do not always play under the same conditions. As a result, statistics that purport to be measured on the same scale—earned run average for pitchers, batting average, home runs, etc. for hitters—don't always tell us who is the better player. For example, how do you compare Jorge De La Rosa's 3.09 ERA so far this year, with half of his games coming at hitter-friendly Coors Field, with Hyun-Jin Ryu's 2.83 for the Dodgers, who play in a much better pitcher's park?
Baseball's answer has been adjusted ERA (better known as ERA+), or adjusted OPS (OPS+) for hitters. These statistics put ERA and OPS (on-base percentage plus slugging percentage) on a universal scale by weighting them for the ballpark (hitter-friendly or pitcher-friendly) that players have achieved them in. A league-average ERA+ or OPS+ is 100, and the number basically equates to a percentage—if you have an OPS+ of 90, you have an OPS that's 10% worse than the league average; if you have an ERA+ of 150, you have an ERA that's 50% better. Jorge De La Rosa's ERA+ is 145; Hyun-Jin Ryu's is 129. These more accurate measures of a player's performance therefore show us that, if De La Rosa and Ryu had pitched every game under identical conditions—much like a controlled experiment, in scientific-method terminology—De La Rosa would have given up fewer runs (despite what has actually occurred in real life).
Affirmative action should take the form of a similar universal statistic—let's call it adjusted grade point average, or GPA+. Colleges should seek to ascertain how all students would have performed relative to each other if they had all been playing on a level field—and then take the X number of students with the highest scores on that universal scale. Where OPS+ is adjusted according to the difficulty of the ballpark, GPA+ can be adjusted according to the quality of the high school where the child was educated, the difficulty of the economic circumstances that the child had to overcome, and other obstacles that might cause a student's true ability not to be reflected in their raw GPA.
Again, I don't know what all those factors would be, much less how to weight them properly. Would race be included? It's hard to see how race can affect or understate a student's ability (rather, race is all too often a proxy for the economic conditions that truly do that), but contributing to the diversity of the student body does carry value, no? For that reason, maybe the University of Texas at Austin would prefer to use a statistic like VORS (Value Over Replacement Student) than GPA+.
What statistic to use, and how to calculate it, would certainly be controversial; I have no illusions about that. But the first step is to get smart statisticians creating these education SABR-equivalents, and then we can have a debate about them (much as Fangraphs and Baseball-Reference have a healthy debate over how to calculate WAR). Great, unbiased data minds—ones whose loyalty is to truth in numbers, not ideology–are out there. It's time to get them on the job for something bigger than baseball.