4 ms·

Here's a thought experiment. You are a college admission counselor. You have room to accept 50 people. You get 500 Asian American applicants with 4.0 GPAs, 1

Here's a thought experiment. You are a college admission counselor. You have room to accept 50 people. You get 500 Asian American applicants with 4.0 GPAs, 1500 SATs, and IB degrees. You get 500 White applicants with the name numbers. You get 1 Black applicant with the same numbers. Do you throw all 1001 applicants in a lottery and pick 50. Or are you more likely to accept the one Black applicant outright. Put the remaining 1000 in the lottery and pick 49?

bane 12y ago
Well right, that's the heart of the matter isn't it? "Diversity" in a pure form means that the people accepted should be accepted in proportional numbers to their racial percentages in the general population. If you agree with this or not isn't terribly important, it's the goal that's been set. Ideally there'd be some set of criteria that would allow this naturally happen. Some mix of SAT score, GPA, etc. that when summed up would give you x% white, y% black, z% hispanic and so on. But there isn't. Some ethnic groups, for whatever set of complex reasons you would like to use, under represent on these measurable criteria. But basically it boils down to this, there aren't enough Black and Hispanic students, scoring at high enough levels, AND applying to these schools, to supply a proportional representation in the student body (about 13% and 17% respectively). "Diversity" programs operate under the assumption that there's some kind of systemic cultural, social or environmental pressure on these groups that keeps them underrepresented. What those are, seem to be complex, and for whatever reason don't appear to impact White, Asian (and a couple other identifiable groups) in the same way. Poverty, racism, culture, whatever, seem to be chronic issues for underrepresented groups while proportionally or overrepresented groups seem to either be immune to these problems or able to work through them. The idea is that, if you get enough of these underrepresented minority applicants into prestigious schools, get them good educations, start getting them well integrated into upper socio-economic echelons, it'll have a "viral" like effect on other people in their same cultural ethnic group and eventually (over generations) they'll naturally start to represent at the "right" ratios. On the flip side, there's well documented reasons why some ethnic groups (notably Asians, but other so called "model minorities" seem to show the same patterns) seem to excel and over-represent. It seems to mostly come from a high cultural value for education, combined with constant social and parental pressure to get educated and into prestigious jobs combined with an intense work ethic. It seems to be a repeated pattern that crosses all racial, genetic, socio-economic, geographic and other boundaries. Work hard, work towards a high goal and focus on that goal, and you'll make it out of being poor and make it into being reasonably rich. So in your example, a pure "diversity" sample would overrepresent the White students to ensure the acceptances are as close to the natural population diversity as possible. The black student would get in, and for the remaining 49 seats, about 38 would got to White applicants, and 11 would go to Asian applicants. The problem is obvious, setting "diversity goals" to arbitrarily say that proportional attendance is an appropriate goal is entirely arbitrary. Why is that virtuous? Because it seems fair? Why is race or ethnicity used and why not gender or economic quintile or hair color or height? Why is the concept or proportionality the right model? What's wrong with a student body that's 50% Asian? Shouldn't hard work get rewarded instead of genetics? And that's the billion dollar question.
- kenjackson 12y ago
  Shouldn't hard work get rewarded instead of genetics? But even in the model you suggest, you're not rewarding "hard work", you're rewarding test scores. Not the same thing. As my old professor would say, "Don't tell me you studied hard. I don't care how hard you studied." In many regards, all these metrics are arbitrary. I wonder if there is a utilitarian perspective on college admission metrics.
  - bane 12y ago
    Yeah, you are absolutely correct. But nobody ever seems to have come up with a better way to sort out regular snowflakes from special ones, especially in a scalable way.

bane 12y ago

Well right, that's the heart of the matter isn't it? "Diversity" in a pure form means that the people accepted should be accepted in proportional numbers to their racial percentages in the general population. If you agree with this or not isn't terribly important, it's the goal that's been set. Ideally there'd be some set of criteria that would allow this naturally happen. Some mix of SAT score, GPA, etc. that when summed up would give you x% white, y% black, z% hispanic and so on. But there isn't. Some ethnic groups, for whatever set of complex reasons you would like to use, under represent on these measurable criteria. But basically it boils down to this, there aren't enough Black and Hispanic students, scoring at high enough levels, AND applying to these schools, to supply a proportional representation in the student body (about 13% and 17% respectively). "Diversity" programs operate under the assumption that there's some kind of systemic cultural, social or environmental pressure on these groups that keeps them underrepresented. What those are, seem to be complex, and for whatever reason don't appear to impact White, Asian (and a couple other identifiable groups) in the same way. Poverty, racism, culture, whatever, seem to be chronic issues for underrepresented groups while proportionally or overrepresented groups seem to either be immune to these problems or able to work through them. The idea is that, if you get enough of these underrepresented minority applicants into prestigious schools, get them good educations, start getting them well integrated into upper socio-economic echelons, it'll have a "viral" like effect on other people in their same cultural ethnic group and eventually (over generations) they'll naturally start to represent at the "right" ratios. On the flip side, there's well documented reasons why some ethnic groups (notably Asians, but other so called "model minorities" seem to show the same patterns) seem to excel and over-represent. It seems to mostly come from a high cultural value for education, combined with constant social and parental pressure to get educated and into prestigious jobs combined with an intense work ethic. It seems to be a repeated pattern that crosses all racial, genetic, socio-economic, geographic and other boundaries. Work hard, work towards a high goal and focus on that goal, and you'll make it out of being poor and make it into being reasonably rich. So in your example, a pure "diversity" sample would overrepresent the White students to ensure the acceptances are as close to the natural population diversity as possible. The black student would get in, and for the remaining 49 seats, about 38 would got to White applicants, and 11 would go to Asian applicants. The problem is obvious, setting "diversity goals" to arbitrarily say that proportional attendance is an appropriate goal is entirely arbitrary. Why is that virtuous? Because it seems fair? Why is race or ethnicity used and why not gender or economic quintile or hair color or height? Why is the concept or proportionality the right model? What's wrong with a student body that's 50% Asian? Shouldn't hard work get rewarded instead of genetics? And that's the billion dollar question.

kenjackson 12y ago
Shouldn't hard work get rewarded instead of genetics? But even in the model you suggest, you're not rewarding "hard work", you're rewarding test scores. Not the same thing. As my old professor would say, "Don't tell me you studied hard. I don't care how hard you studied." In many regards, all these metrics are arbitrary. I wonder if there is a utilitarian perspective on college admission metrics.
- bane 12y ago
  Yeah, you are absolutely correct. But nobody ever seems to have come up with a better way to sort out regular snowflakes from special ones, especially in a scalable way.