http://www.umich.edu/~rdytolrn/pathwaysconference/presentati...
In other words, you want to choose a definition of SES that predicts how the parents interact with the children before age five, as it's not (mostly) income or race that creates the achievement gap. E.g. just separate the parents into welfare, working class, or professional, like Hart & Risley do in Meaningful Differences. I think in order to understand how SES effects the achievement gap, you need to choose a definition of SES that is broad and qualitative rather than quantitative and limited.
And race does correlate with peer SES, as whites (and asians, and most hispanics, and well-off blacks) flee schools once there are too many low-SES blacks.
The question is - how big a problem is "white flight" / segregation. If it is seen as a huge cause of inequality, then there are ways to reduce it.