4 ms·

sjburt said they are less accessible to poor students, which I was defending as true.

• sokoloff 10y ago
  Is accessibility best measured by: P ( student_is_poor | student_attends_ivy ) P ( applicant_is_admitted_to_ivy | applicant_is_poor ) I take the view that the latter is the better measure, but I can see people arguing that the former has some relevance.
  • esrauch 10y ago
    The point is that both are low, I would accept the second metric and I don't think it shows what you seen to claim it does. If you take all the applications submitted, income would be positively correlated with acceptance even though it is "needs blind". Some of that is implicit, but some of it is likely explicit (admit kids who played polo over kids who worked at McDonald's if everything else is equal)
    • sokoloff 10y ago
      I agree that both are low. MIT's acceptance rate is a little under 1 in 12, and several Ivy League colleges have an even lower acceptance rate. I further agree that there is a positive correlation between income and acceptance. I further agree that there is a positive correlation between income and application. What I don't see is any productive way to 100% eradicate those correlations. I'm only familiar with the admissions policies as practiced by MIT in the late 80s and early 90s. There was explicit bias for diversity (which is good; no one wants an utterly homogenous population), but I also think it's perfectly reasonable to deny admission to a very selective university to a student who is utterly undistinguished in all aspects of their high-school career. Admissions to selective schools ought not become a totally random lottery. I think that removing financial gates for those students who are accepted (need blind admissions) is mandatory and effective. Going beyond that is fraught with peril I think.