The number of randomized experiments on crime and justice with a minimum of 100 participants more than doubled between 1957 and 1981, when there were 37, and between 1982 and 2004, when there were 85. There was an increase in very large, multisite replication experiments and in experiments in which the unit of randomization was the area. Results in the second period were generally more encouraging, showing that some interventions reduced offending. The main advantage of a randomized experiment is its high internal validity. Randomization ensures that the average person or place in one condition is equivalent on all measured and unmeasured variables to the average in another condition, within the limits of statistical fluctuation. Observed differences in outcomes can be attributed to differences in interventions rather than to pre-existing differences. Randomized experiments have formidable practical and ethical problems but are often feasible and should be used to test causal hypotheses and evaluate well-defined technologies wherever possible. (Published Abstract)