Race, genes, and sports.

Science, technology, and life.
Dec. 5 2008 7:30 AM

White Men Can't Jump?

Race, genes, and sports.

A few days ago, I wrote about a test, now being marketed in the United States, that predicts whether your toddler has more potential as a power athlete or as an endurance athlete. The test examines ACTN3, a gene that affects fast generation of muscular force. Fray poster Andrea Freiboden isn't impressed. "What a lot of crap. Just look at the race of the athlete," she writes:

Generally, people of West African origin have more fast twitch muscles which allow intense bursts of power. This is why running backs, defensive linemen, and receivers are almost all black. We don't need any expensive test. All you have to do is look at the physique. Blacks in basketball are lean and musularly [sic] hard. Whites have softer muscles, which is why white basketball players have to rely more on skill than blacks who have the advantage of skill + great speed/strength.

William Saletan William Saletan

Will Saletan writes about politics, science, technology, and other stuff for Slate. He’s the author of Bearing Right.

Oy. I've been through this wringer before. It's true that some racial averages differ in part for biological reasons. It's also true that that this is one of them. But Freiboden is exactly wrong. Race is a less, not more, reliable gauge of physical characteristics than genes are. In fact, that's one of the chief consolations of nontherapeutic genetic testing: No matter how inaccurate genes are as a predictor of this or that ability, they're more accurate than predictions based on race. And the sooner we get past judging by race, the better.

ACTN3 has two key variants: R and X. To recap:

Roughly speaking, the more copies of the R variant you have, as opposed to the X variant, the more likely you are to excel at sports requiring power or speed. (You can be RR, RX, or XX.) The testing company, Atlas Sports Genetics, cites studies that support this pattern. A 2003 analysis of hundreds of athletes who had represented Australia at international meets found that 53 percent of the male competitors in sprinting or power events were RR—nearly twice the prevalence of this genotype in a less-athletic population sample. None of the 35 female sprinters were XX. Nor were any of the 25 male Olympic sprinters. Subsequent studies show the same basic pattern in Finland, Greece, and Russia.

Few genes are known to be decisive in determining life outcomes. Nutrition, training, and other genes matter. But the evidence that this gene significantly influences athletic ability is strong.

Now look at the frequency of the R and X variants in different populations. According to data published seven years ago in Human Molecular Genetics, the relative frequency of the X allele is 0.52 in Asians, 0.42 in whites, 0.27 in African-Americans, and 0.16 in Africans. If you break out the data further, the frequency of the XX genotype is 0.25 in Asians, 0.20 in European whites, 0.13 in African-Americans, and 0.01 in African Bantu. Conversely, the frequency of RR (the genotype for speed and power) is 0.25 in Asians, 0.36 in European whites, 0.60 in African-Americans, and 0.81 in African Bantu. Among Asians, you can expect to find one RR for every XX. Among whites, you can expect nearly two RRs for every XX. Among African-Americans, you can expect more than four RRs for every XX.

So, yes, all other things being equal, you can expect this gene to cause Africans and African-Americans to be disproportionately represented at the highest levels of speed and power sports. And you can expect the opposite for Asians. But contrary to Freiboden's claim, you can't expect what we actually find in, say, basketball. Five years ago, 77 percent of NBA players were black; only 21 percent were white. According to a study reported last year, black players earned 83 percent of the league's court time. Now contrast that with genetic and population data. Compared with whites, black Americans aren't even twice as likely to be RR, and they're more than half as likely to be XX. Furthermore, among American men aged 20 to 35, there are about five times as many whites as blacks. In sum, blacks are about twice as dominant in the NBA as ACTN3 alone would predict. Something else must be going on: culture, resources, differential treatment, other genes … you name it.

Basketball has lots of confounding factors. It favors height, stamina, and court vision in addition to speed. So, let's look at a sport where explosive force alone is decisive: sprinting. Several years ago, Jon Entine, the author of Taboo, summarized the data:

There are no sprinters of note from Asia, even with more than 50 percent of the world's population, a Confucian and Tao tradition of discipline, and an authoritarian sports system in place in the most populous country, China. No white sprinter can be found on the list of 100-meter sprinters; the best time by a white, 10 seconds, ranks more than 200th on the all-time list. … All of the 32 finalists in the last four Olympic men's 100-meter races are of West African descent. The likelihood of that happening based on population numbers alone—blacks from that region, now living around the globe, represent approximately 8 percent of the world's population—is 0.0000000000000000000000000000000001 percent.

Note the distinction: West Africans dominate sprinting. East Africans do better at distance running. So already, the evidence points beyond race toward a more precise category: population. And with ACTN3, we're beginning to advance from population comparisons toward the salient level of analysis: genes. Remember, as we noted about ACTN3 in general, differences at the elite level probably exceed differences among the rest of us. But they're still real.

I've had my share of arguments with people who deny that race is biologically meaningful. Many of them are dedicated to the proposition that all humans are created equal, not just in the sense of moral worth or treating each person on his merits, but literally, in the sense that no genetically based difference can be admitted in average ability between populations. That kind of egalitarian literalism—I call it liberal creationism—becomes harder and harder to sustain in the face of evidence such as the data on ACTN3.

On the other hand, those of us who accept such differences must understand them accurately and describe them responsibly. As Fray poster Njuzu puts it, "Race is a very inexact and unreliable proxy for genetics." Race is not a causal unit. There's no such thing as having fast-twitch muscle fiber because you're black. The causal unit is a gene, or a network of genes, or a network of genes and environmental factors. Being black only makes you more likely to have a genetic variant that makes you more likely to have extra fast-twitch fiber. That's a lot of "likelies," not certainties. And you can eliminate part of the uncertainty by testing for ACTN3, which takes you past the crudeness of race to the relative precision of genetics. I'm not exactly thrilled about a world in which kids' futures are projected from being RR or XX. But it's a hell of a lot better than a world in which they're projected from being black or white.

(Now playing at the Human Nature blog: 1. The myth of morning-after pills. 2. In praise of lethal rationing. 3. Armed robotry.)