October 26, 2006
Hot, smart Brown girls gaining ground
Now that Project Runway is over, a new season of Top Chef has started. I don't like it as much as Project Runway since the chefs are typically under much tighter time constraints and thus don't have sufficient time to fully elaborate a creative product. You might think that another downside to Top Chef would be the lack of hot models (like Marilinda from PR) -- but you would be wrong. The host of the show is the first international supermodel to hail from India: Padma Lakshmi. Unlike other models, though, she graduated from college and has written several cookbooks -- all right, so not exactly proving a new theorem in math, but you can tell from her biography and the way she behaves on the show that she's got an IQ of at least 115 or so (+1 SD at least).
To revisit a couple biological and genetic themes, in the comments to my previous post on cross-assortative mating for good looks and intelligence, Jason refers me to a study on fruitflies which shows that sometimes traits which boost the fitness of males would incur fitness costs in a female, and vice versa. Therefore, cross-assortative mating for the most desirable male and female traits might be a complete wash for the children's fitness, or worse. However, human beings are not that sexually dimorphic, probably because paternal investment is high in our species, which requires females to choose good fathers and not simply the largest, sexiest males . Male fruitflies, by contrast, do a mating dance, leave their stuff, and then they're done. The same reasoning explains why male deer have such clunky antlers while does are not so burdened, as well as why peahens are drab compared to peacocks.
For IQ, there is no consistent difference in the means between males and females, though males show greater variance. There are non-trivial differences in the means for personality traits (as I reviewed here), though only two of the Big Five show noticeable differences. Agreeableness shows a 0.57 SD difference in means, such that males are less cooperative and more antagonistic than females; while females score 0.55 SD higher on Neuroticism. As for good looks, Kanazawa recently examined the attractiveness ratings of men vs women using a representative sample and found a difference between male and female means of 0.2 on a 5-point scale (3.6 female vs 3.4 male). He doesn't report the SD for this measure, but let's say it's 1, on the idea that the trait is normally distributed with a mean of 3 and that the 5-point scale is only sensitive to cover +/- 2 SD. (In fact, it looks like there is Lake Woebegone stuff going on, in that there are far many more attractive than unattractive people, but assume this for simplicity.) Then the male-female difference in means would be 0.2 SD, favoring females. So, for most things people care about when choosing a mate -- intelligence, attractiveness, and personality traits -- there is either no dimorphism or real but slight dimorphism, well under 1 SD . This dovetails with the high paternal investment of humans -- because males and females face pretty much the same problems in surviving and reproducing, we don't expect too many traits to be advantageous to one sex but not the other, or if so, the difference should not be extremely pronounced .
Therefore, human beings need not worry much about passing one male-typical or female-typical traits that might harm offspring of the opposite sex. While high IQ may not make females more desirable as mates, it certainly doesn't make them less desirable -- all complaints I've ever heard or read about high-IQ or high-status females had to do with their personalities, not the fact that they had larger vocabularies or were better at Tetris than the complaining male. And good looks certainly don't make males less desirable -- in fact, it will ensure that they are not only desirable but begin having sex earlier, not to mention the huge boost in confidence they'll enjoy as a result of being given the up-and-down look by girls throughout adolescence. (See here for a Google image search of "hot bollywood actor" and tell me if you think these guys are at a disadvantage in the dating & mating game.) The differences in Agreeableness, however, are more significant, so you may want to pick a partner to balance you out, as disagreeable females will be perceived as bossy and bitchy (though it would help males assert their authority). I don't see how high Neuroticism would be perceived as undesirable -- this doesn't refer to Woody Allenesque neuroticism, but rather emotional volatility or being on-edge. Since there is no assortative mating for personality traits, apparently being more emotionally labile doesn't harm one's dating prospects.
Which brings us back to the hot, smart Brown girls. As a result of the caste system, which has existed for at least 1000 years (and dates back perhaps to 800 BC, although the Wikipedia entry is not totally clear on this), there is extensive assortative mating for intelligence and status. One crucial effect of assortative mating is that it increases the narrow-sense heritability, making it regression toward the mean less severe. Recall that narrow-sense heritability measures the amount of variance in phenotype that is accounted for by additive genetic factors -- think of additive genetic factors as those that contribute "blindly" or "unconditionally" to the phenotype. For example, if there are 100 loci involved in IQ, and each locus can contribute either 0 or 1 points to the phenotype, additivity means that a "1 allele" at a certain locus will contribute its 1 point no matter what the value of the other allele at that locus is (i.e., dominance plays no role), and also regardless of what the values are at other loci (i.e., epistasis plays no role). If additive factors play a larger role, then less of the phenotype is accounted for by conditional and chancy factors, which may or may not occur. That makes it a safer bet that a child born to IQ 130 parents will have an IQ close to 130. The same is true for assortative mating for good looks -- it will make it a safer bet that the good-looking parents will have a child close to the parents' level. Thus, not only do upper caste South Asians have a high level of attractiveness and intelligence, which is reason enough to choose them as partners, but it's more likely that your children will remain close to the parental value, in comparison to mating with a smart, good-looking person of European or African descent, for example. All this goes to show that one must take care in comparing humans to species that exhibit low paternal investment, whose members have causes for concern that we should not fret over in our own lives.
 It's true, though, that human groups vary in the degree of paternal investment.
 Of course, even small differences between means, or greater variance assuming equal means, will have marked effects in the tails of the distribution (i.e., more males than females with IQ 160, more female than male bombshells).
 Note that this is the opposite of what is acceptable thinking in academia -- there, sex differences, though taboo, are more easily discussed and taken for granted, while discussion of ethnic or racial differences are dynamite. In reality, the most glaring inter-group differences are between populations adapted to different environments, since the males and females of a given environment face roughly the same pressures in human beings. That's why it's not uncommon to compare racial groups for IQ and find a 1 SD difference (or more) between means, while there is no such male-female difference anywhere.