This paper examines the feasibility of distributing Elementary and Secondary Education Act Title I aid among school districts according to the educational achievement of pupils in each district. The "predicted achievement method" would permit the distribution of aid on the basis of population characteristics in a way that, on average, directs more funds to communities with more low achievement pupils. The evidence for Massachusetts indicates that the multiple regression predicted achievement method would provide a significantly better targeting of funds to low achievement districts than the current method of distributing funds according to the number of children in poverty. The first section of the paper discusses the logic of the predicted achievement method and the way that it would be implemented in practice. Section 2 describes the Massachusetts data used in the analysis. The third and fourth sections present results for mean achievement scores and low achievement scores. (Author/IRT)