4 ms·

Well I'm talking totally theoretically, just to show that it's possible to discriminate in order to achieve a better end result for the business. But theoretic

by edanm 9y ago

Well I'm talking totally theoretically, just to show that it's possible to discriminate in order to achieve a better end result for the business. But theoretically, if a large employer were to collect the statistics to calculate the correlation between GPA and job performance, and then were to discover that an equivalent GPA between, say, men and women lead to higher job performance in women, then faced with two candidates with the same GPA, you'd prefer the woman. This is not vague and has nothing to do with the underlying mechanics - you as the employer theoretically don't care, all you care is that same GPA woman leads to a better job performance as same GPA man. An employer like Google could discover things like this, if they exist, and might want to, for purely business reasons, do this. (No idea on the legality of any of this, seems to me like it's illegal but I have no idea).

manigandham 9y ago
That doesn't make sense. Reviewing by motivation + results already delivers you the best people, which is the best for the business. Why use some random group dimension? It is absolutely vague because that is only relevant when talking about averages of the whole group, but the point of fairness and equal opportunity is to treat people like individuals and judge them accordingly. Instead of using some strange GPA proxy for job performance across a giant group, just use the actual job performance (history) of the individual and see if they're good for your business. You cannot use generic group attributes to judge a single person. It sounds "good" but never works out and is the cause of the most of the mess in the first place.
- edanm 9y ago
  Look, I don't think we really disagree on anything fundamental here. Of course you should judge people as individuals. Of course looking at past performance is probably better, and fairer, than anything else. I'm completely with you on that. The only thing I'm saying is, there are logically plausible situations, in which you might want to use generic group attributes. Especially in cases where you have no prior knowledge - e.g. everything you say is great, but how do I look at past history when judging college grads with no work experience? I mean, let's take an extreme case and see if we can at least agree on that. One of the things that some economists often claim about capitalism is that it helps with diversity, since if for example everyone is refusing to hire black people because of racial discrimination (think 1930s), then an employer will come along and offer to hire black people for cheaper, and will therefore outcompete other companies, who will have to also hire black people at slightly higher wages, etc. If we were in such a (1930s-like) world where every employer refuses to hire black people, and we were deciding who to hire, and our normal practice would be only to hire people with 2 years' worth of job experience. We would never hire a black person, because none of them would fit our requirements. In such a world, wouldn't it make sense, from a purely self-interested economic perspective, to change the requirements for black people to compensate for the screwed-up biased world we find ourselves in? If not, then is the classic economic argument wrong?
  - manigandham 9y ago
    > there are logically plausible situations, in which you might want to use generic group attributes. I do not see what the logical situation is, because it is overruled completely by the individual's data which you already have right in front of you. It just sounds like applying biases by guessing with group membership instead of looking at the person. Anytime the group membership overrules the individuals' attributes, it's going the wrong way. Also while I understand the context, some things just don't matter because life is fundamentally unfair and full of chaos. Yes everyone works differently at the same task even with similar results. Maybe they're smarter, or maybe they had a bad year, or got sick, or something else... how can you possibly account for any or all of that in a reasonable sense? > but how do I look at past history when judging college grads with no work experience? Judge their school experience. Give them responsibility commensurate with their skills and measure continuously to establish the history, then promote or fire. > One of the things that some economists often claim about capitalism is that it helps with diversity Yes, it does. However capitalism isn't perfect and is still up to legal, political, and generally irrational whims of the people. If you're talking about 1930's then there were actual laws segregating people. If you move forward to WWII, then it was the political climate where it was accepted to not work with others regardless of the money. If anything, I think a common love for profit actually does a lot to remove these barriers once people realize they're all better off generating and spending wealth instead of hating each other over looks and other useless attributes.

by edanm 9y ago

manigandham 9y ago

That doesn't make sense. Reviewing by motivation + results already delivers you the best people, which is the best for the business. Why use some random group dimension? It is absolutely vague because that is only relevant when talking about averages of the whole group, but the point of fairness and equal opportunity is to treat people like individuals and judge them accordingly. Instead of using some strange GPA proxy for job performance across a giant group, just use the actual job performance (history) of the individual and see if they're good for your business. You cannot use generic group attributes to judge a single person. It sounds "good" but never works out and is the cause of the most of the mess in the first place.

edanm 9y ago
Look, I don't think we really disagree on anything fundamental here. Of course you should judge people as individuals. Of course looking at past performance is probably better, and fairer, than anything else. I'm completely with you on that. The only thing I'm saying is, there are logically plausible situations, in which you might want to use generic group attributes. Especially in cases where you have no prior knowledge - e.g. everything you say is great, but how do I look at past history when judging college grads with no work experience? I mean, let's take an extreme case and see if we can at least agree on that. One of the things that some economists often claim about capitalism is that it helps with diversity, since if for example everyone is refusing to hire black people because of racial discrimination (think 1930s), then an employer will come along and offer to hire black people for cheaper, and will therefore outcompete other companies, who will have to also hire black people at slightly higher wages, etc. If we were in such a (1930s-like) world where every employer refuses to hire black people, and we were deciding who to hire, and our normal practice would be only to hire people with 2 years' worth of job experience. We would never hire a black person, because none of them would fit our requirements. In such a world, wouldn't it make sense, from a purely self-interested economic perspective, to change the requirements for black people to compensate for the screwed-up biased world we find ourselves in? If not, then is the classic economic argument wrong?
- manigandham 9y ago
  > there are logically plausible situations, in which you might want to use generic group attributes. I do not see what the logical situation is, because it is overruled completely by the individual's data which you already have right in front of you. It just sounds like applying biases by guessing with group membership instead of looking at the person. Anytime the group membership overrules the individuals' attributes, it's going the wrong way. Also while I understand the context, some things just don't matter because life is fundamentally unfair and full of chaos. Yes everyone works differently at the same task even with similar results. Maybe they're smarter, or maybe they had a bad year, or got sick, or something else... how can you possibly account for any or all of that in a reasonable sense? > but how do I look at past history when judging college grads with no work experience? Judge their school experience. Give them responsibility commensurate with their skills and measure continuously to establish the history, then promote or fire. > One of the things that some economists often claim about capitalism is that it helps with diversity Yes, it does. However capitalism isn't perfect and is still up to legal, political, and generally irrational whims of the people. If you're talking about 1930's then there were actual laws segregating people. If you move forward to WWII, then it was the political climate where it was accepted to not work with others regardless of the money. If anything, I think a common love for profit actually does a lot to remove these barriers once people realize they're all better off generating and spending wealth instead of hating each other over looks and other useless attributes.