In many states, standardized tests are used to hold schools accountable for student academic achievement. To motivate improvement in test scores, financial awards are given to teachers and administrators in schools that show the greatest gains. However, failure to adjust for initial conditions may put awards out of the reach of some schools and fail to produce the desired incentives. In this paper, we examine factors that influence gains in test scores using school-level data from California (1999-2003). We find evidence that validates the existing California award system--test scores improved the most for weak schools.