Yes and yes. We live in a world where people disregard qualifiers, so if you say “on tests of mathematical ability, men have higher variance in test scores, so the most talented mathematicians are disproportionately men” people will hear “men are better at math” and assume that average men are better than average women at math (this might also be true, but is not what you said). Basically, some people don’t distinguish between “most a are b” and “most b are a”, so you end up with people drawing conclusions that hurt other people with no real benefit. So as a general rule, we pretend that there are no between-group differences because if we don’t, people have a tendency to focus exclusively on between group differences and ignore within-group differences, which is worse.
We live in a world where people disregard qualifiers, so if you say “on tests of mathematical ability, men have higher variance in test scores, so the most talented mathematicians are disproportionately men” people will hear “men are better at math” and assume that average men are better than average women at math (this might also be true, but is not what you said).
I could make similar argument about a lot of things we do here, e.g., people hear “consequentialism” and think “the ends justify the means”, that doesn’t stop LW from promoting consequentialism.
So as a general rule, we pretend that there are no between-group differences because if we don’t, people have a tendency to focus exclusively on between group differences and ignore within-group differences, which is worse.
For example, suppose I want to hire the best mathematicians for a project, they’ll likely be disproportionately White or Asian men. Someone who followed your advise looking at the mathematicians I hire would conclude that I was racist and sexist in my hiring and we live in a society where the courts might very well back them. Thus the only way for me to avoid being considered a racist and sexist is to intentionally fudge the numbers based on race and sex, which itself requires that I know the truth about racial and gender differences so I know which way to fudge.
I could make similar argument about a lot of things we do here, e.g., people hear “consequentialism” and think “the ends justify the means”, that doesn’t stop LW from promoting consequentialism.
Nope, and some people will express disapproval of LWers who promote consequentialism. Being right doesn’t make you immune to social stigma.
Intentionally believing false things always carries a cost.
Yes, it does. So does unintentionally believing false things. This is definitely not a one-sided issue, as much as people like to pretend that is it. Anti-discrimination policies reduce one cost at the expense of raising another.
For example, suppose I want to hire the best mathematicians for a project, they’ll likely be disproportionately White or Asian men.
In the case that you both want to hire and are able to hire exceptional mathematicians, anti-discrimination policies are likely to hurt both parties involved. (In theory, laws regarding disparate impact wouldn’t actually affect you if you were hiring based on demonstrable mathematical prowess, but in practice business necessity would be hard to prove). The mathematicians are actually likely to be hurt considerably more, because without anti-discrimination policies, they would probably be in higher demand and thus able to ask for much higher pay.
The real problem comes in when employers decide that they need exceptional people but can’t actually identify these exceptional people. If filtering based on race was allowed, employers would use that (the best mathematicians are disproportionately white and asian, therefore if I hire a white or asian I’ll get an above-average mathematician).
Basically, you’re right except for the problem where humans mix up p(a|b) and p(b|a), which causes people to do stupid things (most of the people who win the lottery buy lots of tickets, so if I buy lots of tickets I’m likely to win the lottery). If you actually know what you’re hiring based on, anti-discrimination policies will prevent you from having 100% of your workforce be the very best, but even if only whites and asians had the required skills, you’re still looking at 77% of the population in the US, so it falls in the category of “annoyance” not “business killer”. In terms of fudging, you can detect statistically significant deviations just as well as someone looking at your hiring data. You don’t need to know beforehand.
Of course, if these things weren’t the case you’d still face social stigma for saying anything that sounds vaguely racist. Because while these two societal tendencies have strong effects in opposite directions, they’re not there by virtue of reasoned argument, and so removing one but not the other is likely to cause more harm than good (probably, I have no idea how one would go about removing either societal tendency to test that hypothesis). If both tendencies could be eliminated, that would be best, and here you probably can talk about it without much social stigma, but if you ask those questions in everyday life, you will be labeled as a racist.
Intentionally believing false things always carries a cost.
Yes, it does. So does unintentionally believing false things.
The difference is that if you unintentionally believe something false, you can update when you find new evidence; whereas once you start intentionally believing false things, you’ve declare all truth your enemy.
If you actually know what you’re hiring based on, anti-discrimination policies will prevent you from having 100% of your workforce be the very best, but even if only whites and asians had the required skills, you’re still looking at 77% of the population in the US, so it falls in the category of “annoyance” not “business killer”.
Depends on the size of the business and your margin. Most small businesses can’t afford to have 23% of there employees be dead weight, especially if they have to pay them the same as the others to avoid looking like they have racist pay policies.
Most small businesses can’t afford to have 23% of there employees be dead weight, especially if they have to pay them the same as the others to avoid looking like they have racist pay policies.
Most small businesses don’t need to hire the top 0.01% in any given skillset. The small businesses that do need to hire that exclusively and the small businesses that are strapped for cash are generally two distinct sets. In any case, without those policies, the top 0.01% could demand more money, and so the business wouldn’t be in much better of a position. It’s really the top 0.01% of workers who bear the majority of the cost of anti-discrimination policies, because they could negotiate better pay if the policies weren’t in place.
It is a tradeoff. Empirically, societies that oppose discrimination tend to do better (though there are obvious confounds and this doesn’t necessarily mean that the anti-discrimination policies improve outcomes—it may just mean that richer people prefer egalitarian policies more). In American culture, at least, you will generally be labeled as a racist if you imply that there might be between-group differences, whether or not you can back that up with good arguments.
The difference is that if you unintentionally believe something false, you can update when you find new evidence; whereas once you start intentionally believing false things, you’ve declare all truth your enemy.
By all means, keep in mind that the social fiction of perfect equality in ability across groups is unlikely to be true. But also keep in mind that it’s a polite fiction and you will be stigmatized if you point out that it’s unlikely to be true. The term “racist” usually refers to someone who doesn’t respect that social convention, and both of the statements you were questioning go against that social norm. “Racist” doesn’t mean “factually incorrect”, it means “low status and icky”.
Most small businesses don’t need to hire the top 0.01% in any given skillset.
The same logic applies if you want to hire people in the top 10%. Yes, there may very well be enough blacks in the 10% that if you had first choice among them you could hire enough to comply with disparate impact. However, in reality you’re competing for the few blacks in the top 10% with all the other businesses who also need to hire the top 10% and there aren’t enough to go around.
By all means, keep in mind that the social fiction of perfect equality in ability across groups is unlikely to be true. But also keep in mind that it’s a polite fiction and you will be stigmatized if you point out that it’s unlikely to be true. The term “racist” usually refers to someone who doesn’t respect that social convention, and both of the statements you were questioning go against that social norm. “Racist” doesn’t mean “factually incorrect”, it means “low status and icky”.
Yes and at LW our goal is to raise the sanity waterline.
How about also considering the costs, benefits, and comparative advantages when dealing with various topics? One does not get extra points for doing things the hard way. Instead of dealing with some topics directly, it would be better to discuss more meta, e.g. to teach people about the necessity of doing experiments and evaluating data statistically. This will prepare the way for people who will later try to deal with the problem more directly.
Now it may seem that when I see people doing a mistake, and I don’t immediately jump there and correct them, it is as if I lied by omission. But there are thousands of mistakes humans make, any my resources are limited, so I will end ignoring some mistakes either way.
Make sure you pick your battles because you believe you can win them and the gains will be worth it. Instead of picking the most difficult battle there is, simply because choosing the most difficult battle feels high-status… until you lose it.
Not all of them, it’s just that there aren’t enough non-dead weight non-white non-asians to go around for all the businesses who need competent employees while complying with disparate impact.
Not all of them, it’s just that there aren’t enough non-dead weight non-white non-asians to go around for all the businesses who need competent employees
How do you know? Not every business is a silicon valley start up that needs to be staffed almost entirely super smart people. The typical company is much more pyramidal. A lot of employers want a lot of employees who will happily work for the minimum wage.
while complying with disparate impact.
Whatever that means.. If you think US affiirmative action, or something, is the issue, then it cancels within the US. If you think it makes the US less competitive than polities that don’t have AA, then that’s only part of a bigger problem, because, given your assumptions, the US would be at a severe disadvantage compared to any given Asian
nation anyway. But it doesn’t appear to , so maybe factors other than DNA are important.. Who knows? We can only try to deduce what you might be saying from your hints and allegations.
The thing is society doesn’t “carry on as though people are equal”. Society, at least the more functional parts of society, treat things like affirmative action and disparate impact, as things to be routed around as much as possible because that’s necessary to get things done efficiently.
It would have been helpful to answer the question as stated. Not all societies have affirmative action and my polity doesn’t. Depending on ones background assumptions, affirmative action could be seen as restoring equality, or creating inequality. You seem to have assumed a take on that without arguing it. It would have been helpful to argue it, and not to treat “society” as synonymous with “US society”.
The real problem comes in when employers decide that they need exceptional people but can’t actually identify these exceptional people. If filtering based on race was allowed, employers would use that (the best mathematicians are disproportionately white and asian, therefore if I hire a white or asian I’ll get an above-average mathematician).
Basically, you’re right except for the problem where humans mix up p(a|b) and p(b|a),
“The best mathematicians are disproportionately white and asian, therefore if I hire a white or asian I’ll get an above-average mathematician” is Bayesianly correct if the race is the only thing you know about the candidates; but it isn’t (a randomly-chosen white or Asian person is very unlikely to be a decent mathematician), and the other information you have about the candidates most likely mostly screens off the information that race gives you about maths skills.
Hmm, so E(the Math SAT score that X deserves|the Math SAT score that X got is 800, and X is male) is just 4 points more than E(the Math SAT score that X deserves|the Math SAT score that X got is 800, and X is female). That doesn’t sound like terribly much to me, and I’d guess there are plenty of people who, due to corrupted mindware and stuff, would treat a male who got 800 better than a female who got 800 by a much greater extent than justified by that 4-point difference in the Bayesian posterior expected values. (Cf the person who told whowhowho that Obama must be dumber than Bush—surely we know much more about them than their races?)
I’m not sure if this is correct, but I sometimes wonder given how they’re surrounded by spin-doctors and other image manipulators how much we really know about prominent politicians, especially when the politician in question is new so you can’t look at his record.
My conclusion still holds if you simply need mathematicians in the top 10%, for example, only the analysis is slightly more complicated.
More than 1% of the world is racist.
Also, taboo “racist” unless you agree with faul_sname’s definition, in which case whether being a “racist” is a bad thing is precisely the question under discussion.
My conclusion still holds if you simply need mathematicians in the top 10%, for example, only the analysis is slightly more complicated.
So you agree that, in the original example, you’re more likely than not just being a racist? Because you certainly seem to be moving the goal post over to “top 10%” …
faul_sname’s definition
That link does not appear to point to a definition.
Yes and yes. We live in a world where people disregard qualifiers, so if you say “on tests of mathematical ability, men have higher variance in test scores, so the most talented mathematicians are disproportionately men” people will hear “men are better at math” and assume that average men are better than average women at math (this might also be true, but is not what you said). Basically, some people don’t distinguish between “most a are b” and “most b are a”, so you end up with people drawing conclusions that hurt other people with no real benefit. So as a general rule, we pretend that there are no between-group differences because if we don’t, people have a tendency to focus exclusively on between group differences and ignore within-group differences, which is worse.
I could make similar argument about a lot of things we do here, e.g., people hear “consequentialism” and think “the ends justify the means”, that doesn’t stop LW from promoting consequentialism.
Intentionally believing false things always carries a cost.
For example, suppose I want to hire the best mathematicians for a project, they’ll likely be disproportionately White or Asian men. Someone who followed your advise looking at the mathematicians I hire would conclude that I was racist and sexist in my hiring and we live in a society where the courts might very well back them. Thus the only way for me to avoid being considered a racist and sexist is to intentionally fudge the numbers based on race and sex, which itself requires that I know the truth about racial and gender differences so I know which way to fudge.
Nope, and some people will express disapproval of LWers who promote consequentialism. Being right doesn’t make you immune to social stigma.
Yes, it does. So does unintentionally believing false things. This is definitely not a one-sided issue, as much as people like to pretend that is it. Anti-discrimination policies reduce one cost at the expense of raising another.
In the case that you both want to hire and are able to hire exceptional mathematicians, anti-discrimination policies are likely to hurt both parties involved. (In theory, laws regarding disparate impact wouldn’t actually affect you if you were hiring based on demonstrable mathematical prowess, but in practice business necessity would be hard to prove). The mathematicians are actually likely to be hurt considerably more, because without anti-discrimination policies, they would probably be in higher demand and thus able to ask for much higher pay.
The real problem comes in when employers decide that they need exceptional people but can’t actually identify these exceptional people. If filtering based on race was allowed, employers would use that (the best mathematicians are disproportionately white and asian, therefore if I hire a white or asian I’ll get an above-average mathematician).
Basically, you’re right except for the problem where humans mix up p(a|b) and p(b|a), which causes people to do stupid things (most of the people who win the lottery buy lots of tickets, so if I buy lots of tickets I’m likely to win the lottery). If you actually know what you’re hiring based on, anti-discrimination policies will prevent you from having 100% of your workforce be the very best, but even if only whites and asians had the required skills, you’re still looking at 77% of the population in the US, so it falls in the category of “annoyance” not “business killer”. In terms of fudging, you can detect statistically significant deviations just as well as someone looking at your hiring data. You don’t need to know beforehand.
Of course, if these things weren’t the case you’d still face social stigma for saying anything that sounds vaguely racist. Because while these two societal tendencies have strong effects in opposite directions, they’re not there by virtue of reasoned argument, and so removing one but not the other is likely to cause more harm than good (probably, I have no idea how one would go about removing either societal tendency to test that hypothesis). If both tendencies could be eliminated, that would be best, and here you probably can talk about it without much social stigma, but if you ask those questions in everyday life, you will be labeled as a racist.
When does that occur? What happened to resume″s, qualifications and tests?
The difference is that if you unintentionally believe something false, you can update when you find new evidence; whereas once you start intentionally believing false things, you’ve declare all truth your enemy.
Depends on the size of the business and your margin. Most small businesses can’t afford to have 23% of there employees be dead weight, especially if they have to pay them the same as the others to avoid looking like they have racist pay policies.
Most small businesses don’t need to hire the top 0.01% in any given skillset. The small businesses that do need to hire that exclusively and the small businesses that are strapped for cash are generally two distinct sets. In any case, without those policies, the top 0.01% could demand more money, and so the business wouldn’t be in much better of a position. It’s really the top 0.01% of workers who bear the majority of the cost of anti-discrimination policies, because they could negotiate better pay if the policies weren’t in place.
It is a tradeoff. Empirically, societies that oppose discrimination tend to do better (though there are obvious confounds and this doesn’t necessarily mean that the anti-discrimination policies improve outcomes—it may just mean that richer people prefer egalitarian policies more). In American culture, at least, you will generally be labeled as a racist if you imply that there might be between-group differences, whether or not you can back that up with good arguments.
By all means, keep in mind that the social fiction of perfect equality in ability across groups is unlikely to be true. But also keep in mind that it’s a polite fiction and you will be stigmatized if you point out that it’s unlikely to be true. The term “racist” usually refers to someone who doesn’t respect that social convention, and both of the statements you were questioning go against that social norm. “Racist” doesn’t mean “factually incorrect”, it means “low status and icky”.
The same logic applies if you want to hire people in the top 10%. Yes, there may very well be enough blacks in the 10% that if you had first choice among them you could hire enough to comply with disparate impact. However, in reality you’re competing for the few blacks in the top 10% with all the other businesses who also need to hire the top 10% and there aren’t enough to go around.
Yes and at LW our goal is to raise the sanity waterline.
Yes, it is.
How about also considering the costs, benefits, and comparative advantages when dealing with various topics? One does not get extra points for doing things the hard way. Instead of dealing with some topics directly, it would be better to discuss more meta, e.g. to teach people about the necessity of doing experiments and evaluating data statistically. This will prepare the way for people who will later try to deal with the problem more directly.
Now it may seem that when I see people doing a mistake, and I don’t immediately jump there and correct them, it is as if I lied by omission. But there are thousands of mistakes humans make, any my resources are limited, so I will end ignoring some mistakes either way.
Make sure you pick your battles because you believe you can win them and the gains will be worth it. Instead of picking the most difficult battle there is, simply because choosing the most difficult battle feels high-status… until you lose it.
It’s worth noting that many people also ignore the smallness of effects. It probably doesn’t end up mattering much, not worth arguing...
Exqueeze me, but since when did “not white or asian” equate to “dead weight”?
Not all of them, it’s just that there aren’t enough non-dead weight non-white non-asians to go around for all the businesses who need competent employees while complying with disparate impact.
So much for “23%”.
How do you know? Not every business is a silicon valley start up that needs to be staffed almost entirely super smart people. The typical company is much more pyramidal. A lot of employers want a lot of employees who will happily work for the minimum wage.
Whatever that means.. If you think US affiirmative action, or something, is the issue, then it cancels within the US. If you think it makes the US less competitive than polities that don’t have AA, then that’s only part of a bigger problem, because, given your assumptions, the US would be at a severe disadvantage compared to any given Asian nation anyway. But it doesn’t appear to , so maybe factors other than DNA are important.. Who knows? We can only try to deduce what you might be saying from your hints and allegations.
Why, is business an entirely zero-sum game within the US?
Some indeces of business performance are, such as relative rank.
Is the False Thing “people are equal” or “it is best for society to carry on as though people are equal”.?
The thing is society doesn’t “carry on as though people are equal”. Society, at least the more functional parts of society, treat things like affirmative action and disparate impact, as things to be routed around as much as possible because that’s necessary to get things done efficiently.
It would have been helpful to answer the question as stated. Not all societies have affirmative action and my polity doesn’t. Depending on ones background assumptions, affirmative action could be seen as restoring equality, or creating inequality. You seem to have assumed a take on that without arguing it. It would have been helpful to argue it, and not to treat “society” as synonymous with “US society”.
Ironically this is a case where p(a|b) is in fact a good proxy for p(b|a) and and the kind of filtering you’re objecting to is in fact the correct thing to do from a Bayesian perspective.
See also: Offended by conditional probability
“The best mathematicians are disproportionately white and asian, therefore if I hire a white or asian I’ll get an above-average mathematician” is Bayesianly correct if the race is the only thing you know about the candidates; but it isn’t (a randomly-chosen white or Asian person is very unlikely to be a decent mathematician), and the other information you have about the candidates most likely mostly screens off the information that race gives you about maths skills.
Read the comment I linked to and possibly subsequent discussion if you’re interested in these things.
Hmm, so E(the Math SAT score that X deserves|the Math SAT score that X got is 800, and X is male) is just 4 points more than E(the Math SAT score that X deserves|the Math SAT score that X got is 800, and X is female). That doesn’t sound like terribly much to me, and I’d guess there are plenty of people who, due to corrupted mindware and stuff, would treat a male who got 800 better than a female who got 800 by a much greater extent than justified by that 4-point difference in the Bayesian posterior expected values. (Cf the person who told whowhowho that Obama must be dumber than Bush—surely we know much more about them than their races?)
I’m not sure if this is correct, but I sometimes wonder given how they’re surrounded by spin-doctors and other image manipulators how much we really know about prominent politicians, especially when the politician in question is new so you can’t look at his record.
99% of projects do not need the top 1%. More than 1% of the world is racist.
Why should I believe you actually need the top 1%, when the statistics say there’s a greater than 50% chance that you’re actually just racist?
My conclusion still holds if you simply need mathematicians in the top 10%, for example, only the analysis is slightly more complicated.
Also, taboo “racist” unless you agree with faul_sname’s definition, in which case whether being a “racist” is a bad thing is precisely the question under discussion.
So you agree that, in the original example, you’re more likely than not just being a racist? Because you certainly seem to be moving the goal post over to “top 10%” …
That link does not appear to point to a definition.
You still haven’t defined what you mean by “racist”.