Professor Michael McDaniel of VCU recently had a paper published in Intelligence estimating average IQ by state based on NAEP testing results (easily accessible viewing of the estimates via Dienekes). You may recall that I attempted the same back in July. Well, we were after the same thing. Our results correlate almost perfectly at over .96. Doubtful a distinguished professor gets a confidence boost from the supportive results of some livingroom puke, but the puke sure does!
We did differ in some ways, however. McDaniel set the mean IQ at 100 with a standard deviation of 15 and averaged the mean results from NAEP reading and math scores by state for both the fourth and eighth grades. He also adjusted for the percentage of white children in each state attending non-public schools.
I took the regression equations produced by running the numbers in the data table put together by Richard Lynn in Race Differences in Intelligence where he correlates IQ scores with international math and science test scores (pp 173-175) and then adjusted the nominal test score values (by running an IQ of 98 through the regression equation produced by Lynn's numbers) on the international tests to the NAEP math and science tests in the US, applying an equal weight to eighth grade math and science NAEP scores.
I opted for science scores over reading scores for a few reasons: Lynn used math and science, scholastic science questions are more g-loaded than reading ones are (reading skills are more problematic at younger ages), and reading comprehension questions are more biased against newly-arrived immigrants than either science or math questions are. A minor drawback is that Kansas and Pennsylvania lack NAEP science results so I had to estimate using only the math results for these two states.
Whether math, science, or reading results are used is of mostly academic importance: math and science results correlate at .90, math and reading at .91, and (somewhat surprisingly) science and reading at .95 (all for eighth graders).
McDaniel probably improves on my estimates by taking non-public school attendance into account. He argues that private- and homeschooled children tend to be cognitively above average. Generally that makes sense, although about 7% of private schools in the US are devoted to special education, and with over three-fourths are religiously affiliated, questions of values and morality rather than just academic attributes have to be considered. The NCES estimates that privately-schooled eighth graders score the equivalent of 12.3 points better on NAEP math tests (not in terms of IQ--the max is 500). I'll adjust for the proportion of private school attendees accordingly in the near future to see if it might improve the estimates.
In his VDare column on McDaniel's work, Steve Sailer compares the professor's estimates with those of previous good-faith attempts at ascertaining average state IQs. I correlated mine with the same attempts. With McDaniel's in green and mine in blue, estimates correlate with the 1960 Project Talent at .63 and .60, a mid-eighties study of Vietnam Vets' IQ at .63 and .61, with a combined ACT/SAT estimate at .71 and .71, and with Tickle's averages at .53 and .52. Quite similar, although McDaniel's are a bit more vigorous. I suspect that is due to the non-public schooling adjustment.
McDaniel's work is long overdue, as it dispels the spurious estimates of state IQ that have bubbled up in the past. And his academic courage is admirable. For example, in the discussion following his results, he writes:
IQ at the individual level has strong correlates with race. There are large and intractable mean racial differences in IQ at the person level... Because racial composition of the state is a large magnitude correlate of state IQ, one cannot expect meaningful changes in estimated state IQ as long as the state racial composition is relatively stable.Plenty of sharp Americans, as well as myself, have long advocated the institution of a merit immigration system to allow the US to glean the global cream of the crop (to increase the national IQ and standard of living, shrink the wealth gap, etc) rather than absorbing millions of destitute third-world liabilities as our current immigration policy does. McDaniel logically takes this same argument to the state level:
States might structure incentives to encourage those with high IQs to remain in the state. Likewise, a state may encourage high IQ individuals to have children. Over time, these policies should raise the average IQ of state residents.Without apology, he suggests different eugenic techniques, entertaining the 2,500 year old Platonic idea of state-permitted birthing. No less unapologetically, he points out that states might consider becoming "Jewish-friendly" to pull in buckets of Ashkenazi.
Business schools harp relentlessly on the idea of human capital, and yet the full scope of what this means is so rarely bantered about. Instead, education and training proxy loosely for IQ, but at great deadweight loss (a concept B-schools are also familiar with). I wish I could take a few classes with McDaniel--undoubtedly the lectures and conversations would be more fruitful than the typical blather that ignores human biodiversity.