# Meter = Yard

First, you keep saying that “there are no human races. If you compare genetic variation within each of the so-called races with other populations, there is more vriation within group than between groups.”

I love this point. Such wisdom. But when I tried it out today on a fellow worker, I got confused. He happened to mention that men are on average a few inches taller than the average woman. What a know-nothing, I thought, and then I used your argument: “If you look at variation in height among men, you find some men who are 4 feet tall and others who are over 7 feet tall. Thus, there is more variation among men than there is between men and women.” Pretty good point, no? But my co-worker didn’t seem as shock-and-awed as I expected. He just said, “Sure, there’s lots of variation among men, but there’s still an overall average difference between male and female height, even if it’s comparatively small.”

I was stumped, and I have to ask for your assistance here. What do I say when some hack tries to suggest that there can be differences between groups even if each individual group has internal variation? I mean, clearly it’s a fact of anthropology that [variation within a group] = [no possible differences between that group and another group]. Just no way it can be. But I can’t seem to get that point across very effectively.

Maybe an analogy would help. Such as, “There’s more variation between inch 1 and inch 36 of a yardstick than there is between a yardstick and a meter. Therefore 1 yard = 1 meter.” Would a simple example like that help? Or maybe you can think of others.

— Commenter John Brown

You’ll have to read the post and the thread to get the full context, but its pretty funny. Also, look at the comment directly below the one linked to.

## 4 thoughts on “Meter = Yard”

1. swong says:

Looking at similar debates happening on other sites, I’ve noticed a pattern:

Statement: Group A’s average on is slightly higher than group B’s average on .

Dissenter: Ha! Here’s a example in Group B whose score is higher than average! Your claim is faulty!

Example: Men tend to be taller than women. However, I know a tall woman. Therefore men aren’t taller than women. No one on either side of the debate seems to notice the fallacy in this reasoning.

I wonder how the researchers classified subjects by race. Genetic analysis? Self-identification by test subjects? Skin color chart?

2. Well, that’s just it. With enough data you can find statistically significant differences between any two *arbitrarily* chosen groups.

The degree of the difference is what’s important. That black people get certain diseases highly disproportionately suggests “race” is an important category in that narrow context. But “race” is used as a category in circumstances where its just not important.

I’m not sure IQ differences by race is interesting, for example. Its not surprising to me at all that IQ has a genetic component. Given genes aren’t randomly distributed through populations (i.e. that you have one gene makes it more likely you’ll have another gene correlated with it) because groups of people evolved semi-separately, it also doesn’t surprise me that genes determining skin color might correlated with genes helping to determine IQ. Different groups of people had to contend with different environments through evolutionary time.

But so what. A couple IQ points is peanuts for those using these results for statistical discrimination… Will you not hire a black person because his group has 4 IQ points less on average than the other candidates. Even if in the unlikely case that 4-points translates to significant differences in worker quality, won’t you have more information available to assess the individual’s intelligence (e.g. letters of recommendation, GPA, verbal skills demonstrated in the interview, etc) to make conditional means based on race moot? And does the hiring firm have the conditional mean for social classes? for gender? for feet sizes? for whatever else?

But would the black candidate even get a job interview? There’s some recent studies that show black sounding names on resumes are discriminated against. Perhaps this reflects the differences in average IQ’s translated by hiring firms to potential worker quality (i.e. lower group IQ means that group might, on average, be worse employees). If this is the case, wouldn’t the long-term equilibrium be for there to be no “black sounding” names? If you care about maximizing your children’s economic outcomes, you wouldn’t burden them with a name that makes it harder for them to get jobs. If this isn’t the equilibrium and “black sounding” names persist, then maybe its because black parents have preferences different from maximizing their children’s economic well-being. In which case, I’m not sure if there’s any efficiency improving policy and I’m back to thinking race doesn’t matter.

The next move is to get all paternalistic and say “well, black parents should care about their children’s economic well-being”. To which, I respond, “mind your own business and just because you have preferences such that money outcomes are all-important that doesn’t mean the rest of us should have those preferences.”

The empirical studies I like best are the ones that discover new important categories; they find “economic significance” where people hadn’t looked before. I hate the studies that use “statistical significance” to perpetuate tired old categories. And race is one of the tiredest and oldest categories.

3. swong says:

“Its not surprising to me at all that IQ has a genetic component.”

I agree. I believe we side with a tiny minority in this belief. The concept that intelligence can be derived from anything other than free personal choice seems deeply offensive across the political spectrum. I’ve always felt that getting angry at a person for being stupid makes as much sense as getting angry at a person for being short.

I wouldn’t worry about the effects of this study on hiring practices. Anyone willing to ignore race in hiring will continue to ignore race in hiring, as they’ve already found that it really doesn’t matter (unless they’re famous geneticists or something). Anyone who makes an issue out of race will continue to make an issue out of race.

Besides… Jews are a tiny tiny tiny minority in the world. You can hire one average Jew, with his 112 IQ, or you can hire 30 Blacks for the same price, who would add up to like a 3000 IQ. I dunno about you, but I’d sure like to hire an African village to generate MD5 collisions in real time.