In examining the production of crime and justice field experiments during the 1990s, this study focused on the characteristics of criminological experiments conducted during the period from 1991 through 2000 under funding from the National Institute of Justice (NIJ), the principal research agency of the U.S. Department of Justice.
During this period, NIJ supported 990 distinct research projects under grants of just over $225 million. Of these projects, only 21 used experimental methodologies, specifically, the random assignment of cases to alternative treatments. The award for the average experiment was approximately $100,000, which was 43-percent higher than other types of research projects. The issues studied by these experiments were concentrated in the areas of treatments/sanctions for illegal drug use (eight studies) and treatments/sanctions for domestic violence (seven). There were also two studies of policing practices, two on pretrial-release decisionmaking, and two on parole outcomes. The 21 experiments were conducted in 20 different agencies. Although never a substantial proportion of the NIJ research budget, NIJ financial support for randomized experimentation declined from 1991 to 2000. Neither the availability of funds, the nature of policy issues, implementation difficulties, average costs, nor the availability of interested and willing organizations or researchers apparently accounted for NIJ's decreasing funding and promotion of randomized experiments. Apparently, the limited use of rigorous randomized research by NIJ was due primarily to a lack of widespread energetic support for this type of research methodology among personnel involved in policy decisions bearing upon NIJ grant awards. This may be changing, as evidenced by language included in two recent NIJ research solicitations that explicitly encourage the use of experimental design, i.e., a multiple site evaluation of drug courts and a national evaluation of federally supported prisoner reentry programs. 2 tables, 7 notes, and 53 references