Edit: A comment by Steve led me to take another look at the data, this time using sample weights, which correct for missing data. Based on the weighted data, the hybrid performance is consistent with a genetic hypothesis but the adoption data is much less so. The departure from prediction is mainly due to the low performance of Whites adopted by Whites; Blacks adopted Whites due little better than Blacks overall. The patterns of differences indicate that the math assessment used is significantly influenced by home environmental factors, which calls into question the use of it as an index of IQ. I included tables with the weighted values below. The Ns are the weight values, not sample size.
Below is my preliminary analysis of the High school Longitudinal study of 2009, which is a nationally representative study of American high school students. All students were give the uniquely designed HSLS math test, which I used as my index of cognitive ability. Both students and parents reported their race, and I used these reports co-jointly as an index of race. Hispanics were excluded from all analyses.
Using the mother/father, biological/adopted variables, I identified 99 traditionally adopted Whites, 8 transracially adopted Whites, 13 traditionally Blacks, and 18 transracially adopted Blacks. As can be seen in the first table, the transracially adopted Whites performed as the traditionally adopted Whites and mutatis mutandis for Blacks. I also identified 5412 Whites with two biological White parents, 474 Blacks with two biological Black parents, and 62 individuals of mixed biological parentage. As can be seen in the first table, the biracial students performed intermediate to the mono-racial ones. The above analysis was restrictive, as individuals were only included if both mother and father data was available and if the parental racial identification matched student racial identification (e.g., both biological parents identify as White, student identified as White). I also looked at the performance of self-identifying Biracials. As can be seen, they perform intermediate to self-identifying monoracial. In all analyses I calculated MIQs using the full sample standard deviation of 10.08.
The second table gives a three way comparison between adopted Black, White, and Asian students. Included also are students with transracial guardians or stepparents.
The following tables include sample weights (Ns are the sample weights not the sample size):