ABSTRACT
Randomized experiments can be typecast as a means of program evaluation, feeding the deceptive perception that they are more useful for testing specific policy options than they are for building criminological theory. The history of science, however, is marked by “crucial experiments.” These key turning points have occurred when theoretical principles were either confirmed or disproven, and our knowledge of the world suddenly sprinted forward. There is no reason why this pattern of advancement should be any less applicable to criminology than it is to other scientific disciplines. What remains unclear are the features of an experimental research design that can make it most useful for playing this leading role.
