In the fights over affirmative action, many people voice very strong opinions, either for and against, and are not shy about stating them. Supporters say categorically that considerations of race in university admissions are fully justified, either because of our egregious history of discrimination in this country or the critical benefits of being educated in a racially diverse environment. Opponents are equally vociferous, contending that racial preferences undermine the entire premise of the civil rights movement: that individuals should be judged on their merits, not their skin color.
It strikes me that the debate over legacy preferences in college admissions has a very different feel than the one over affirmative action. Opponents (including me and the other contributors to a 2010 volume, Affirmative Action for the Rich) state their case forcefully, calling preferences based on lineage un-American and illegal. Supporters, on the other hand, are usually fairly hesitant, even slightly apologetic in their advocacy.
Case in point is Russell K. Nieli, a senior preceptor in Princeton University’s James Madison Program in American Ideals and Institutions. Nieli recently weighed in with “A Reluctant Vote for Legacies,” at the Manhattan Institute’s Minding the Campus blog. He argues that legacy preferences are form of “defensible corruption.”
One way to interpret the difference in tone is that supporters of legacy preferences are more pragmatic, reasonable, and less hysterical and ideological than opponents. Another way to interpret this, though, is that there really isn’t much of principled argument to be made on behalf of legacy preferences; that unlike racial preference, legacy preference doesn’t present a “hard case,” with strong arguments on both sides.
Consider Nieli’s two central points in defense of legacy preference.
First Nieli suggests that legacy preferences are a necessary evil because they probably do increase alumni giving. He doesn’t cite a single study but rather “common sense.” He acknowledges that the most comprehensive study of legacy preferences finds the existence of preferences per se doesn’t increase alumni giving, and he doesn’t criticize its methodology. Nor does he grapple with the evidence suggesting that alumni giving has not significantly declined at institutions that have dropped legacy preference policies. As Michael Dannenberg of the Obama Administration has noted, universities are institutions that are supposed to be dedicated to research. So where is the evidence that alumni giving increases as a result of legacy preferences? Is having a hunch really enough to justify discrimination based on ancestry?
Nieli’s second defense is that legacies preferences are smaller than preferences for under-represented minorities, so they involve less of a compromise with merit. There is some evidence from William Bowen, Martin Kurzweil and Eugene Tobin to back this up: they find that controlling for test scores, legacy preferences increase one’s chances of admissions by 19.7 percentage points, while being an under-represented minority increases one’s chance of admissions by 27.7 percentage points. But as I pointed out in a recent blog post, new evidence from Harvard University researcher Michael Hurwitz shows that legacy preferences are larger than previously thought. Looking at Fall 2007 applicants to 30 elite schools, Hurwitz concludes that after better controlling for variables than previous researchers did, legacy preferences of all kinds increase one’s chances of admissions by 23.3 percentage points. More importantly, “primary legacy” candidates (sons and daughters, as opposed to siblings, nephews, nieces, or grandchildren) see a 45.1 percentage point increase in the chances of admission. (Bowen’s study of an earlier cohort of applicants to 19 selective schools counted legacies as applicants of the children or grandchildren of alumni and did not distinguish between the two; and Bowen’s study controlled for SAT’s but not other factors such as teacher recommendations and extracurricular activities.)
More importantly, from a meritocratic perspective, the simple comparison of legacy and affirmative action weights ignores the larger picture: that one set of preferences (for under-represented minorities) is going to students who, on average, are more economically disadvantaged than the general applicant pool, while the other set of preferences (for legacies) is going to a students, who, on average, are more advantaged than the general pool. When honoring merit, it matters a great deal whether a preference is being provided to students who are likely to have more potential in the long run than their test scores suggest, given obstacles they’ve overcome, or less potential, given advantages they’ve enjoyed. Indeed, I’ve argued for sharpening the link between affirmative action and disadvantage by calling for preferences based on economic disadvantage per se.
In short, I’m convinced by Nieli that legacy preferences are corrupt, but not that they’re defensible.
