Evidence continues to accumulate showing that the average IQ of Ivy League students is 122.  In a previous post, I cited a study showing that a sample of Harvard students tested in 2003 averaged Wechsler IQs of 122 (U.S. white norms); 124 (U.S. norms), and now I’ve analyzed data from another sample of Ivy League students circa 1971.5 who also averaged Wechsler IQs of 122.

In this study, which a commenter informed me of, six samples of  seniors from  the extremely prestigious Dartmouth (the 12th most selective university in America) averaged 1357 on the pre-1974 SAT, and a full-scale IQ of 127.88 on the WAIS (U.S. white norms); 128.95 (U.S. norms).

However because the WAIS was normed in 1953.5, and the full-scale Flynn effect on the WAIS from 1953.5 to 1978 was 0.306 IQ points per year (U.S. norms)(see page 240 of Are We Getting Smarter?), then WAIS IQs measured in 1971.5 would be inflated by 5.51 points.

This reduces the their IQs from 128.95 to 123.44 (U.S. norms); 122 (U.S. white norms).

How does this compare to their SAT scores?  If all American late teens (not just the college bound elite) took the SAT in the 1970s, the average combined score would have been about 770 (see The Bell Curve, pg 422), suggesting 770 equated to IQ 100 (U.S. norms).  Meanwhile, Mensa requires SAT scores obtained before 1974 to be at least 1300, suggesting 1300 = IQ 130 (U.S. norms).  From these two data points, we might guess that 1357 before 1974 equaled IQ 133 (U.S. norms); 132 (U.S. white norms).

So Dartmouth students, largely selected based on SAT scores, averaged 33 IQ points above the U.S. mean on the SAT, but regressed to 23.44 points above the U.S. mean on the Wechsler.  This would seem to imply a 0.7 correlation (23.44/33) between the Wechsler and the pre-1974 SAT.

This is much higher than the 0.53 correlation between the Wechsler and the post-1995 SAT I estimated based on the regression of Harvard students.  More research is needed to determine whether this is just chance fluctuation from one study to another, or whether it reflects an actual reduction in the correlation in recent decades.  Another possibility is that because the regression slope was estimated from a much higher point in Harvard students, it may simply reflect a reduction in the correlation at higher levels, either because Spearman’s Law of Diminishing Returns strongly exists, or because of ceiling bumping, which should be more acute on the new SAT.