“Low-Income Students and Highly Selective Private Colleges: Searching and Recruiting” (pdf) by Cappy Hill ’76 and Gordon Winston is an interesting read.
The evidence, then, suggests that inadequate attention to geography and the incidence of ACT tests in their search and recruiting activities has contributed to a bias in enrollment against low-income students at highly selective private colleges and universities. Other factors undoubtedly play a role – especially, for instance, widespread
inaccurate information about actual prices at these schools – but these search and recruiting practices do appear to contribute to their relatively meager share of low-income
Perhaps. But the real problem lies in their initial assumption.
So it has been our assumption that these privileged schools should aim to have their student bodies include students from low-income families at least in proportion to their share in the national population of high ability students.
Why would you ever assume that the portion of Williams-worthy students in the richest 5% of US families is the same as the portion in the poorest 5%? Before looking at the data, I would assume the exact opposite. Smart people generally share two attributes: 1) Family income above the bottom 10% and 2) Smart children. Would anyone disagree? (I doubt that Hill and Winston would. At least I hope that they are not infected with the same ridiculous genetic egalitarianism that so hampers the empirical work of, among others, Schapiro and McPherson.)