Coming together

As I mentioned in my last post, Charles Murray holds that social class differences in intelligence are increasing.  The General Social Survey has measures of vocabulary knowledge, which is a surprisingly good measure of IQ, and occupation.  The GSS computes "occupational prestige" scores from a person's occupation.  I computed a measure of the gap between classes (the coefficient in a regression of prestige on vocabulary score.  The results:

 That looks like a clear decline, not an increase.  In fairness, I should point out that they changed the prestige scores (indicated by the different colors) and based on the two years for which they computed both, the relationship is weaker with the second form.  If you adjust for that, then the evidence for a decline becomes weaker.  But no matter how you look at it, the relationship isn't getting stronger.   While this is just a quick analysis based on one short test, Herrnstein and Murray's case was just a plausible story illustrated with some anecdotes.  I haven't read Murray's new book yet, but unless he's found some better evidence, it seems like intelligence is one way in which America is not coming apart.