Hacker News Reader

Comment by Retra | Hacker News Reader

We all have to make decisions with imperfect information. That can be assumed outright with little contention.

Yes, AA is fundamentally flawed, but that is because it is a biasing correction to a pre-existing and difficult to manage bias. You might be tempted to think "if the goal is to eliminate bias, you shouldn't introduce more bias" but the fact is, a counter bias can mitigate another. No, it won't work in a utopia, but we don't live in a utopia. And maybe it can't even help create one. But it doesn't actually prevent any progress.

> "What is worrying is that the truly great minority leaders/thinking/doers/etc.."

Well, don't worry about the truly great leaders, thinkers, and doers. They are truly great, so they will overcome. It is results that make one great, not rewards.

View on HN

Perhaps leveling the playing field at the primary school level would do more good than introducing yet more racism higher up. As it is, the way we fund our primary schools perpetuates downward spirals and we are still talking about Affirmative Action as if it HASNT failed in almost all its predictions from decades ago. At some point, programs have to achieve measurable results or be stopped or reformed, not perpetuated for no good reason than "feeling good" that we are "doing something".

http://www.pbs.org/wnet/wherewestand/reports/finance/how-do-...

The solution to racism is to institutionalize treating people as individuals and providing them with equal opportunity early on.

There is a pretty easy way to eliminate bias in medical school admissions: blind applications.

This is widely cited as a success for orchestral auditions (which resulted in women being hired at a higher rate) and a demonstration that bias exists, yet strangely the solution is opposed by most activists. It's almost as if those same activists are afraid a blind application process might reveal admissions are biased the other way for medical schools.

The problem is not a matter personal/selection bias, but of historical or cultural bias.

I.e., if some communities are given poorer educations, they will naturally perform worse when tested against it. The solution is not to exclude them from opportunities that value education, because then they will be incentivized to avoid furthering their educations. Additionally, a small amount of fear can cause a lot of damage.

If people have an impression that they can't get accepted because of their race, they tend to avoid even trying, which results in undervalued education, and more unqualified minorities. AA in this sense provides a measure of reassurance. (Even if people are racist, you may still get a job in spite of that, so you might as well still put in the effort.) Which is better than fostering whole minority culture that has given up on success. Because that tends to lead to crime, selfishness, and hopelessness.

Secondly, the goal isn't simply to hire the top performers. That's short-sighted. The goal is to produce a better performing society in the long run. To do that, your orchestra will have to hire worse performers for now.

...historical or cultural bias.

What you describe is not bias, it's some groups being actually less qualified.

Further, you don't even get the incentives right. If the message sent is "we let anyone qualified in without even looking at who you are", that provides incentives for anyone to acquire education.

If people have an impression that they can't get accepted because of their race, they tend to avoid even trying,

Given that Asians can't get accepted because of their race, yet keep trying, that doesn't appear to be true.

If your theory is true, at what point do you expect Asians to avoid even trying? When that point occurs and Asians don't avoid trying, will you change your views?

>What you describe is not bias, it's some groups being actually less qualified.

They are less qualified due to historical and cultural bias, not due to their race or gender.

As to the rest of your post: I disagree. Maybe it's time for you to bring out some reliable studies, because I am really not going to take your word for it.

What, specifically, do you disagree with or want a study for?

The claim that race-blind procedures generate identical incentives for identical black/asian students? That's not a study, it's a 2 line proof that could go on any econ 101 midterm.

If outcome(x.copy(race=asian))=outcome(x.copy(race=black)) for all x, then U(outcome(x.copy(race=asian)))=U(outcome(x.copy(race=black))) for any utility function U, thus marginal utilities (i.e. incentives) must also be identical.

Similarly, I'd be unable to come up with a study proving that d/dx x^2 = 2x. I guess you'd disagree with that claim too?

First, drop the bad attitude. I already know calculus, so obviously I won't disagree with what I already know.

But this:

> If the message sent is "we let anyone qualified in without even looking at who you are", that provides incentives for anyone to acquire education.

and this:

> Given that Asians can't get accepted because of their race, yet keep trying

I don't know to be true. So yeah, you'd need to prove them to me before I agree to them.

I gave you the proof of the former in the message you just replied to. It's not a citation, it's merely two lines of econ 101 math.

I'm not going to bother googling for the latter, since I'm sure that will be ignored as well.

Your equation doesn't actually capture human behavior, you know.

We aren't talking behavior, we are talking incentive. If you want to argue that different races behave differently in response to the same incentives, be my guest.

Prove that there is in fact a 'bias' and not simply a reflection of the number of blacks and latinos who have the aptitude to become a doctor.

There is nothing about being a doctor that requires one to be a particular race or gender.

Reply on news.ycombinator.com