This moral imperative emerges from researchers' professional obligation to provide valid answers to questions about the effectiveness of treatments, practices, and programs in the field of criminal justice. The scientific foundation of this moral imperative is the statistical principle that makes randomized experiments the preferred method for ruling out alternative causes (causes unrelated to the experimental conditions) of the outcomes measured. The continued use of nonexperimental methods in situations where randomized experimental studies are appropriate constitutes a serious violation of professional norms for researchers. It is also a violation of the professional responsibilities of the criminal justice agencies that sponsor evaluations of programs. Critics of randomized experiments have argued that it is unethical to make random assignments of individuals to a control condition instead of a treatment condition, because it deprives them of the benefits of treatment. This argument is undermined by one of the fundamental principles of experimental research, i.e., that experiments should only be conducted when treatment effects are unknown or at least uncertain. Thus, those assigned to a control condition are not being deprived of a treatment with known benefits, because the very purpose of the experiment is to determine whether or not the treatment does, in fact, have any benefits. The moral imperative is to use the most effective research methodology available to determine what works and does not work. The challenge for the criminal justice enterprise is to develop an infrastructure for the integration of program design, implementation, and research evaluation that effectively complies with this moral imperative. 1 figure, 59 references