I'm thinking about those tables in No Excuses again. The ones comparing blacks to whites to asians? I already noted one reason why these simple comparisons reflected ignorance on the part of researchers--no controls for other factors. Not region, not education level of parents, not amount of school spending (which ITSELF must be measured carefully, as old schools spend a LOT less on students than newer schools do because old schools have infrastructural problems), etc. This would be akin to measuring voting patterns of blacks and whites in the fifties and coming up with the conclusion that black folk don't give a damn about politics because they didn't vote at the same rates.
There is another significant problem in making blanket comparisons between Asian Americans and African Americans. The Asian American population is a conglomerate of several different ethnicities. Thai, Laotians, Vietnamese, Japanese, Chinese, Korean, Indian, plus several others...all fall in the same pot. These groups have very different trajectories, and by lumping them together we are ignoring significant differences. I am fairly sure that scholars actually CAN disaggregate the "Asian" population they choose not to. Perhaps this is because only by aggregating them are the populations SOMEWHAT comparable in size (black population is around 40 million, the asian population is around 12 million). I also think though that aggregating them (particularly without controls) presents a statistical picture that is in line with the "model minority" myth. If for example, every Asian ethnicity other than the Japanese were shown to have lower incomes, and educational levels than whites, what would that do for the idea that Asians have this culture that causes them to succeed?