Steve Levitt (of Freakonomics fame) is at it again. Using a new data set, he and coauthor Roland Fryer have just written a paper, "The black-white test score gap through third grade". It will definitely add more fire to the debate surrounding the question of why test scores of African American students consistently underperform those of other groups.
Levitt and Fryer find that little differences in test scores between black and white children in the fall of their kindergarten years (blacks score less than 1/10 of 1 standard deviation below whites on math scores at kindergarten age and actually are slightly above whites in reading scores). However, with each year of schooling, black children fall further behind.
It's an interesting piece, and I predict it'll get cited in the popular press often in the next few months. If you want the paper (in PDF format) click here.