I combine sociological and economic research to test a new theoretical model of the causes and consequences of teacher responses to students' track location. I examine the impact of teacher reward structures on educational inequality by analyzing how grading practices affect students' effort and achievement across tracks. Differences in grading practices determine incentive structures for student behavior and educational investments and thus may be an important mechanism in explaining track effects on academic achievement. I apply student fixed effects models across tracks to the NELS:88 and find that, first, track placement affects achievement, second, although grading practices affect achievement, they only explain a minor part of the track effect, and, third, teacher expectations and perceived class ability level explain the positive track effect for high-track students. These findings suggest that high-track students have higher achievement because their teachers perceive them as better students.