Wednesday, October 08, 2014

Cognitive destratification?

Heartiste does a little shadowboxing in the cognitive stratification ring:
So there will be no genetic überwench class. This isn’t to say an evolved cognitive elite is impossible; rather, what appears to be happening is less IQ stratification than a perverse reiteration of the patented CH BOSSS (Boss-Secretary Sexual Strategy) sexual market mechanism to reduce wealth and class inequality. The high school grad secretary of yore has been replaced by the college grad secretary of today. And as long as she stays thin and pretty, she’ll catch the eye of that high status man, and GSS data will erroneously pick this up as mate sorting primarily based on college experience or IQ.
As he correctly asserts, it's likely not as dire a situation as many on the dissident right who haven't looked at the relevant data assume.

Rather than accentuating the putative dysgenic problem, as he insinuates the GSS does, though, survey data actually attenuate it. The GSS shows that among men there is simply no dysgenic trend to speak of. There is, however, such a trend among women. The simultaneous existence of the two diverging trends suggests that assortative mating assumptions are exaggerated (don't just take it from this amateur's reading of the numbers; Henry Harpending found the same). The silver tongue's descriptive phrase "boss-secretary sexual strategy" is an empirically valid one.

Heartiste's recognition of the relevance of credentialism is similarly perspicacious:
Conflating runaway credentialism with IQ misses the fact that today’s paper pushing woman with a communications degree was yesterday’s equally competent secretary with a high school degree, and perhaps even yesteryear’s farmhand mother with sharp instincts for survival.
Indeed, even many wise men fail to grasp this. Education is a far stronger predictor of female fertility than IQ is. Intelligence and educational attainment are correlated, of course, but once educational attainment is controlled for, there's very little difference in fertility across the intelligence spectrum. This isn't just the case at the national level; it's descriptive power is global in scope.

Another way to gauge whether or not cognitive stratification is occurring is to look at the standard deviation values in wordsum scores over time. A flattening out of the bell curve distribution--more low- and high-end scores and fewer middling scores--would suggest an increase in cognitive stratification over time while a spike in the distribution's hump--more middling scores, fewer extreme scores--would indicate a decrease in stratification.

The following graph shows standard deviation values in wordsum scores among native-born whites aged 30-50 by year of participation extending back to the earliest years of the GSS:

There is no evidence for cognitive stratification here. On the contrary, there appears to have been a gentle converging in intelligence the last several decades (at least as measured by vocabulary among whites, anyway). In 1978, 7.9% of respondents scored a 3 or lower (out of 10) and 20.0% scored a 9 or 10, with the remaining 72.1% falling in between. In 2012, the latest year for which data are available, 4.1% scored a 3 or lower and 12.2% a 9 or 10, with 83.7% falling in between. A substantially higher percentage of whites fall somewhere in the middle today than did a generation ago.

GSS variables used: YEAR, WORDSUM, RACE(1), BORN(1), AGE(30-50)


Anonymous said...

So bright people are being bred out of society?


Even easier for the eventual chinese takeover.

Anonymous said...

It might be that cramming more people into college causes them to pick up incidentals like vocabulary that they would not come across if they were working or home making. This is not a silver lining worth celebrating much tho.

JayMan said...

There appears to be a Flynn effect in the Wordsum scores. That, and the fairly low ceiling (since it only has 10 questions) might be artificially narrowing the variance.

Highly related: Henry Harpending just put of a draft paper that covers this very topic:

Draft of paper about Amish | West Hunter

Anonymous said...

Pretty much. What has changed is educational assortment, because Wordsum now has a greater correlation with a wide range of educational ranks. Wordsum assortment its hasn't changed (but then we can't test this because the GSS never included "spouse's wordsum", and couldn't).

Still, assortment for a personality that likes high formal education see,s more likely to me has changed. There are probably White families today, admittedly of few children, with an almost Chinese (beyond a merely German) enthusiasm and propensity for high education net of IQ that did not exist in the past.

Intelligence and educational attainment are correlated, of course, but once educational attainment is controlled for, there's very little difference in fertility across the intelligence spectrum.

There is some evidence that smart women had few children even before much enthusiasm for mass female education, in the form of Swedish records from the late 19th and early 20th century. I'd say probably educational attainment is what has low correlation with fertility after wordsum is accounted for.

anon said...

wonder what effect electronic media has on vocabulary. time was you didnt hear much fancy talk