Recently, OERI’s director, Grover Whitehurst (2002) and other educational researchers (see, for example, Mosteller &Boruch, 2002) have been arguing that nothing less than randomized experiments will do for evaluations of educational interventions and policies. The strong emphasis on randomized experiments is welcome, but ironic. After many years of relative policy indifference to experiments of any kind, OERI is leaping over the rigorously matched experiment to demand randomized experiments.