News for hackers. Stuff that matters.
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by Causality1 on Saturday February 02, 2019 @05:11PM (#19066067)
University applications should not contain any data such as race, name, gender, or place of origin that would allow the university to discriminate for or against the applicant based on their genetics.

We've decided as a nation to outlaw discrimination based on race, gender, sexual orientation, or national origin. We should fucking act like it. If the Asian kid down the street who spent five hours a night studying gets into Harvard and I don't, that's the way it should be.

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by wycs on Saturday February 02, 2019 @05:23PM (#19066156)
When this is tried, it does not output ideologically-pleasing student demographics. Presumably because all objective tests of merit are hopelessly biased.

Caltech is such a school. And its population is largely largely Asian and Jewish, both demographics Harvard has a history of discriminating against.

I am sure Caltech’s administration is considered hopelessly reactionary, but their graduates have the highest number of Nobel prizes per capita. But again, what sort of merit does a Nobel prize prove? Look at the demographics of the winners.

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by tzs on Saturday February 02, 2019 @06:04PM (#19066402)
Caltech seems to be considerably less Jewish than Harvard. Caltech is about 3x as Jewish as the general 18-21 population. Harvard is around 14x as Jewish as the general 18-21 population.

I'm getting this from a chart about halfway down the the following very long article [1]. (Note: I have not read that article. I just searched for info on Jewish enrollment at colleges, and found that chart).

[1] https://www.theamericanconservative.com/articles/the-myth-of...

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by wycs on Saturday February 02, 2019 @06:08PM (#19066429)
Harvard does not discriminate against Jews today. They have historically. Now they do the same to Asians. This, one supposes, must be an example of moral progress.
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by tzs on Saturday February 02, 2019 @06:11PM (#19066458)
Oh, I overlooked that you were talking about historical discrimination, not present discrimination.
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by ForHackernews on Saturday February 02, 2019 @05:27PM (#19066182)
Caltech is an unusually focused school. If your only goal is to churn out future STEM PhDs, admitting based on an objective math test is probably not a bad approach.

Harvard's goal, traditionally, has been to [re]produce the ruling class of America, and their admissions criteria are geared accordingly.

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by paulsutter on Saturday February 02, 2019 @06:36PM (#19066611)
Over the coming decades, my bet would be on

- Stanford to produce the ruling class of America, and for

- Harvard to produce the ruling class of Mad Men season 1

(just based on the criteria in the article)

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by trhway on Saturday February 02, 2019 @05:53PM (#19066323)
as a math grad myself i wish the ruling class consisted of Caltech and similar math/physics/other "real" sciences graduates :)
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by wycs on Saturday February 02, 2019 @05:56PM (#19066349)
See Singapore for such a country.
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by zackmorris on Saturday February 02, 2019 @08:32PM (#19067187)
Perhaps college admissions should switch to a two-phase process:

* application score (without race, name, gender, place of origin, etc)

* justice curve (weight score by race, name, gender, place of origin, etc)

The application scoring and justice curves should be carried out by separate impartial groups, with the justice curve preferably applied by a government entity to prevent corruption.

If we truly believe that a student's financial and demographic background have nothing to do with his or her score, then we'd see no statistical evidence of it in achievement tests.

But of course we do. So I think that while your heart is in the right place with respect to fairness, you have overlooked how justice factors in.

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by tomrod on Saturday February 02, 2019 @05:49PM (#19066291)
This point comes up in discussion from time to time. Affirmative action is legal in some domains.[0]

Ultimately, we all come from somewhere, and the interactive, dynamic impact of past will impact an individual's future potential.

It's nice to optimize for a static objective problem, like whom to hire or whom to admit. But people are made up from the cumulative and dynamic series of events that began happening long, long before being born. It behooves a stable society to ensure marginalized groups aren't forever left in the cold. This is often expressed with sayings like "losing privilege feels like discrimination."

Think of it this way: in introductory game theory we learn about the Prisoner's Dilemma, where both players are offered a strictly dominating choice -- and so the outcome is entirely predictable. However, what happens when the game framework is repeated indefinitely, without end? The sum of immediate payoffs can be balanced with future payoffs, and different equilibria ("good policy") can be considered (e.g. tit-for-tat). This is not even accounting for further ex-game interaction, which can open more equilibria options.

[0] https://en.wikipedia.org/wiki/Affirmative_action_in_the_Unit...

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by oh_sigh on Saturday February 02, 2019 @05:54PM (#19066328)
What about in the essays? If I write about overcoming hardship growing up as a black kid with no father in a mostly white neighborhood, should they censor that?
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by betterunix2 on Saturday February 02, 2019 @05:31PM (#19066208)
Would your attitude change if, say, until 11th grade nobody had told you that college prep begins in middle school (or maybe kindergarten, depending on who you ask)? It may shock you but that is actually the situation a lot of students and their parents find themselves in: they find out too late that they were supposed to be preparing for college admissions for the past few years.

(I am not claiming that you have to believe that this justifies affirmative action, but you should know that the situation is not as simple as the media talking points would have us believe.)

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by rohit2412 on Saturday February 02, 2019 @05:48PM (#19066288)
Then instead of admitting students who didn't prepare for reasons of ignorance and lack of guidance, you give better guidance.

Would you award me a fields medal, just because I didn't achieve anything since I wasn't taught and guided well regarding mathematics?

Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by betterunix2 on Saturday February 02, 2019 @06:38PM (#19066625)
I was replying to a comment that said that if Asian students are spending hours every day preparing for college admissions, then they should be admitted. My point was only that the situation is not as simple as "some people work harder than others," not that affirmative action is necessarily the right answer to the problem. Maybe sending out an army of people to work with elementary and middle schools in Black and Latino neighborhoods to explain to students, parents, and teachers what they should be doing to prepare for college is a better approach. Either way we should not pretend that this is a simple problem with a simple answer.
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by whatshisface on Saturday February 02, 2019 @05:50PM (#19066301)
My attitude for that would be that the entire point of holistic criteria is that college prep shouldn't have to begin in kindergarden! The adult success rate of kids that wake up for the first time in college and Tiger Children (TM) isn't all that different, so we don't want to end up in a situation where, for any race, the only way to go to Harvard is to be a Tiger Child (TM). Making it easier for one race to get in to Harvard might help solve the problem for them, but that doesn't help all the other races.
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by turtlecloud on Saturday February 02, 2019 @07:46PM (#19066969)
If these asian kids didn’t work hard what other opportunities are there? Do you think any serious admission committee is gonna look at your avg asian kid and give him the benefit of the doubt for being “holistic and well rounded” when there are 20x more white kids that could take that same spot?
Re:What Statistics Can't Tell Us in the Fight Over Affirmati... by betterunix2 on Saturday February 02, 2019 @06:46PM (#19066666)
Sure, but my point was not that Harvard's admissions process is the best approach; it was that the situation is not as simple as "Asian kids work harder so they deserve admission to the best schools," which was what the person I was replying to had said.