[I have a newer post on this. See: Blacks, IQ, and color in the NLSY97. The authors of this paper found a smaller effect than I did because they a) excluded mixed race Blacks (i.e. people that have one Black and one White parent), b)
lumped the upper end of the continuum together (1-3), and c) used a smaller sample size then I did. In the larger sample, the difference between darker and lighter “Blacks” is larger and is more linear.)
I just came across a prepublished paper on colorism. The authors found only a weak association between skin-color and AFQT scores in the National Longitudinal Study of Youth 1997 cohort.
The 0.25 SD or so difference between the upper and lower ends of the African American color continuum is a far cry from the 0.5 SD difference that Lynn (2002) found in the 1982 GSS sample.
As I noted over at Inductivist, as indexed by skin color, the genetic difference in white admixture between the most and least caucasian Blacks is maybe .6 (restriction of range) x .45. (reliability of skin color as an index of African Ancestry). If the difference in genotypic IQ between Whites and Blacks, who on average are 25% caucasian, is 1 SD, the difference between pure African Americans — which is not to say Africans because the former did not exactly hail from a representative sample of the latter — and Whites would be 1.3 SD. The difference between the most and least admixed blacks as indexed by color, then, would then be .4SD. Correcting for the unreliability of both measures of IQ and skin color, the expected difference might be about .33.
Kreisman & Range. (Unpublished). “Effect of Skin Color __ on Wages and __ Employment for __ Black Males”r