Greg Mankiw directs us to an interesting book review in The Harvard Crimson about university admission preferences for legacies (i.e., children of alumni). It begins:
I’ve never confessed this before—not even to my roommates: I’m a beneficiary of legacy admissions.
That’s an embarrassing fact to acknowledge at Harvard, where “legacies,” the children of alumni, enjoy preferential treatment in the admissions process. Harvard accepts one-third of legacy applicants—more than three times its overall admissions rate. The federal Office for Civil Rights, in a 1990 review of Harvard’s admissions practices, found that legacy preferences allowed applicants with “weaker credentials” to gain acceptance to Harvard.
The comparison presented in the second paragraph is not an appropriate one to make. I think it's a safe bet that applicants who have a parent who went to Harvard have an above average probability of being admitted at every institution to which they apply. Smart people tend to have smart children.
In this particular case, the acceptance rate for legacies should be compared to the acceptance rate for applicants who have at least one parent with an undergraduate degree from an Ivy League (or similarly selective) institution. Has anyone ever seen this comparison made?
In addition, this comparison, like others we've been discussing this week regarding early admission applicants, needs to be made conditional on observable characteristics of the application. At a minimum, this comparison should be made conditional on SAT scores and GPA.
The author's parents are not Harvard alumni--read the whole article to understand his point.