Abstract:
Educational tracking seeks to group students by unobserved ability using measures of observable acquired skills. In a model where individuals have differential skills prior to beginning formal education due to differences in early childhood development (e.g. linguistic, cultural, or nutritional disadvantages), we show that color-blind tracking systematically underplaces minorities. As a result, minorities have, in expectation, higher abilities than non-minorities assigned to the same track regardless of track. A counterintuitive empirical implication of the model is that, conditional on tracking score and track, minorities will outperform non-minorities in subsequent testing following tracking. Affirmative action policies seeking to equalize posttracking outcomes share similar flaws to color-blind standards in that the average ability of minorities assigned to the upper track remains higher than for non-minorities.