In my Taki’s Magazine review of Carl Zimmer’s new book on heredity, She Has Her Mother’s Laugh, I trot out my dusty old casino analogy response to Richard Lewontin’s famous 85-15 argument about why race more or less doesn’t exist genetically:
Zimmer’s better argument against the genetic reality of race relies upon 46-year-old warmed-over Lewontinism. In 1972, Lewontin argued that 85 percent of total genetic diversity exists among individuals within racial groups, while only 15 percent tracks to geographic ancestry.
This might seem surprising at first, but stop and think about how different people are just within races. Consider the current NBA finals. The Cleveland Cavaliers’ top scorer LeBron James is massive, while the Golden State Warriors’ top scorer Kevin Durant is spindly.
Durant, who scored 43 in game 3, isn’t as spindly as he used to be. But still …
Therefore, ask Lewontin and Zimmer, what possible importance could there be in that measly 15 percent?
Zimmer quotes Lewontin on p. 209: “Since such racial classification is now seen to be of virtually no genetic or taxonomic significance either, no justification can be offered for its continuance.”
Just ignore it. “Where ignorance is bliss, ’tis folly to be wise.”
Of course, both basketball stars have significant sub-Saharan ancestry, as do most of their teammates. That 15 percent of genetic diversity that is not random apparently matters in helping cause the famously unequal racial makeup of the NBA.
A thought experiment: Imagine there is an Indian reservation casino where 85 percent of the spins of the roulette wheel are random, but the other 15 percent of the spins come up red if your croupier is an American Indian or black if your croupier is an African-American. Would you like to know that fact?
Yes, you would. A 15-percentage-point edge in making predictions is huge.
I first used this casino thought experiment way back in 2000. My question is: Did I get something really wrong with this? (I doubt that my arithmetic logic is all that accurate, but did I get it right within an order of magnitude?)
If not, why are intelligent people still promoting Lewontin’s 85-15 Ratio as the Be All and End All about race?
Or is just that there’s no market for a realistic perspective that the race glass is part full and part empty?
Or is it too easy to understand so it sounds lowbrow?
Update: Commenter res calculates that even sex accounts for less than half of the variation in height:
So if anyone is interested I went ahead and ran this analysis for about 8,000 adults in NHANES III:
NHANES is a scientific survey the feds conduct periodically for the benefit of the clothing industry to find out what sizes American consumers are
The result was a Cohen’s d of just over 1.75 for the two groups and the sex variable explained a little over 44% of the height variance.
Of course male or female is a very big deal in predicting height. But even the sex-height glass is less than half full.
Basically, making accurate predictions is hard. Therefore, every little edge helps.
So, the figure for height:sex (44%) is about 3 times as big as Lewontin’s figure for race overall (15%).
It’s important to note that while races can differ very sharply on some traits, there are often other traits where they don’t differ much at all: for example, in the US, whites and blacks tend to be quite similar in height.
Judging from the NHANES figures I looked at a decade ago for the height of blacks and whites, race only is minimally useful in predicting height: white males age 20-39 are something like 0.4 inches taller than blacks at the median, but blacks have a greater standard deviation. So there are a few more Kevin Durant sized blacks than whites on average.
But mostly knowing whether somebody is black or white doesn’t help you much in predicting his height. I’d guess that race only matters a few percentage points at most for predicting height among blacks and whites.
(Hispanic ethnicity is fairly important, though, for predicting height, although some of that is not genetic and can change over the generations.)
Also, keep in mind that just because whites and blacks in America are mostly about the same height doesn’t mean that they necessarily have the same gene variants for height. It could be that they tend to have different genes, but they just happen to come out about equal.
On the other hand, there are non-height related racial differences that also matter in the NBA. From my Taki’s Magazine review of Barack Obama’s favorite HBD book, David Epstein’s The Sports Gene:
BYU economist Joseph Price provided Epstein with some intriguing data on NBA players:
…the average white American NBA player was 6’7.5” with a wingspan of 6’10.” The average African-American NBA player was 6’5.5” with 6’11” wingspan; shorter but longer.
Epstein adds that the average African American in the NBA can jump 29.6” versus 27.3” for whites. Combined with the extra inch of reach, that helps explain the preponderance of blacks in a game where the single most important metric is how high in the air you can get your hand. One scientist told Epstein, “So maybe it’s not so much that white men can’t jump. White men just can’t reach high.”