Sunday, January 07, 2007

Short guys: thank environmental heterogeneity   posted by agnostic @ 1/07/2007 05:49:00 PM

At Steve Sailer's blog, there's an interesting discussion on height in music stars, with some evidence that country singers are taller on average than rock musicians. The mean for rock stars is ~5'10", although personal observations in the comments suggest exaggerated heights among the shorter rock stars like Mick Jagger. I looked through the CelebHeights website for height data on "pretty boys" (e.g., Johnny Depp), and they too appear to be at or below the male population average of 5'10" (see list at the end of this post). If the apparent trend is real, it would be a nice illustration of why all males aren't the same ideal height: there may be some ideal height, but those falling on either side of ideal will still have some niche to fill. Thus, selection is likely more of the balancing type -- it keeps heights within a tolerable range, in this case normally distributed.

Just because you're within 1 SD of the mean on the below-average side doesn't mean you're doomed -- you can excel in areas where being short is more advantageous, such as bouncing around on stage as a musician, generating pretty boy / heartthrob appeal, or honing your skills as a dancer (which becomes more difficult as your center-of-gravity increases and your limbs become longer). This assumes the below-average individual possesses independent other appealing traits, which is why the height distribution isn't uniform -- guys who are 5'7 and otherwise unappealing will be selected against, so 5'7 males will be less frequent than 5'10 males, but enough of the former can rely on other qualities to find mates that their frequency won't be close to zero.

Now, it may sound strange to lump being a rockstar and being a good dancer into the same category, but that just proves the point: just one generation ago during the disco era, dancing skills were highly valued in males. Now, not really. Stochastic environments tend to result in a more diverse range of phenotypes -- unlike, say, the constant environment of oxygen in the air, which will weed out human lungs designed to process anything other than oxygen. So, it seems that predicting fitness based on a guy's height is chancy enough that non-ideal phenotypes aren't mercilessly purged from the genepool.

This scenario predicts that human height will show moderate-high heritability, since directional selection isn't at work and so doesn't exhaust genetic variance in the trait, and since height is not so fitness-neutral that there is nearly unconstrained variance among individuals. Sure enough, h^2 = 0.65 [see Note1]. All this said, I'm no expert on life history theory, so there are surely some subtleties that I'm missing.

As a final thought, complex modern societies open up many more niches to be exploited by yesteryear's outcasts (asocial introverts for one), and the accelerated pace at which aspects of social life change in such societies -- especially due to technological changes -- introduces greater environmental stochasticity. Aside from the disco example, consider the present-day greater fitness of lesser-IQ individuals compared to higher-individuals: in the mid-19th C., who could've rationally predicted that the less intelligent would turn the Darwinian tables on the more brainy? If we ended welfare state policies that support large families among lower-IQ individuals, the trend could snap back to the way it was in 1850. This greater unpredictability of life compared to that of hunter-gatherer societies tells us that phenotypic variance should have exploded not long after the transition to agriculture roughly 10,000 years ago. Concluding where we began with height, Greg Cochran coined a nice mneumonic for remembering the take-home lesson here: "The bow begat the Bushmen."

Pretty Boys

Gael Garcia Bernal 5'6.5
Tom Cruise 5'7
Wilmer Valderrama 5'7
Scott Wolf 5'7
Johnny Depp 5'8
Ryan Phillippe 5'8
Jon Bon Jovi 5'9
Jared Leto 5'9
Dave Navarro 5'9
Scott Baio 5'10
Orlando Bloom 5'10
Leonardo DiCaprio 5'10
Matt Damon 5'10
Colin Farrell 5'10
Jude Law 5'11
Brad Pitt 5'11
Mark McGrath 5'11
Jake Gyllenhaal 6'
Freddie Prinze Jr. 6'1
Gavin Rossdale 6'1
Josh Hartnett 6'3

Median = 5'10. Sex appeal doesn't tail off as height decreases in this sample: look who's 5'9 or shorter. I'm sure there are other data points, but I'm only going to tolerate looking up so much data on pretty boys in the interests of science. I randomly thought of as many as I could, then Googled websites showcasing pretty boys, so if anyone is going to add more data, try to make it random rather than only looking for confirming or disconfirming data points. I think I covered the real heartthrobs, though, which are the most important data. I invite those more inclined to study this -- such as our legions of teenage girl readers -- to pick up where I'm leaving off.

Note1: Falconer & Mackay give two references for heritability of human height, one of which is here, and the other of which is:

Huntley (1966). Heritability of intelligence. pp. 201-18 in Meade & Parkes (eds.), Genetic and Environmental Factors in Human Ability. Oliver and Boyd: Edinburgh.