"The Negative Effects of Privilege on Educational Attainment: Gender, Race, Class, and the Bachelor's Degree"

William Mangino, Social Science Quarterly (http://onlinelibrary.wiley.com/doi/10.1111/ssqu.12003/abstract)

Basically, what it seems to boil down to is that:

1. Women born between 1974 and 1983 attend college and, among those who attend college, complete a 4-year degree, at higher rates than men in the same cohort. I think this is a well-known phenomenon, though I don't know if there are good explanations for it.
2. College attendance and completion is highly correlated with parental income, but at very high levels (as in, like $500,000+, though from the presentation of data it's not quite clear) the rate of completion (among individuals who began some kind of post-secondary education) falls from close to 100% to something like 4/5. The thing is, that there are only 72 individuals with parental income of 250k or above (with a mean parental income of 450k), so these effects could be being driven by a very small number of cases. (If we improbably assume 450k is the median as well as the mean, then the difference between 97% and 83% is only 5 trust fund babies deciding to write the Great American Novel instead of finishing college, or something.)
3. Blacks are less likely than whites to attend any post-secondary education, and less likely to complete it, though this latter effect (at least) is attenuated by controlling for parental income. If, in addition to parental income, you control for household composition variables (2 parents, # siblings, parental education levels), the effect switches direction (though isn't statistically significant). If you add controls for living in the NE or in an urban area and for "social capital," the effect gets a bit bigger and becomes significant at 10%. The effect becomes highly significant and substantial (larger than the coefficient for a dummy for Asian ethnicity) if you also add a control for "academic orientation."

So, it seems like (1) isn't news, (2) isn't terribly substantial, and (3) is ambiguous at best. I think it amounts to this: black kids will on average say they don't value academic achievement as much as white kids with otherwise similar demographic variables, but they will perform much better than one would expect based on the value they claim to put on academic achievement. I suppose one could spin that as weighing against the "burden of acting white" argument, but I don't know.