The power few: Experimental criminology and the reduction of harm : t McCord Prize Lecture

The promise of experimental criminology is finding ways to reduce harm from crime and injustice. The problem of experimental criminology is that so few experiments produce evidence of big effects from the interventions they test. One solution to this problem may be concentrating scarce resources for experiments on the "power few:" the small percentage of places, victims, offenders, police officers or other units in any distribution of crime or injustice which produces the greatest amount of harm. By increasing the homogeneity and base rates of the samples enrolled in each experiment, the power few hypothesis predicts increased statistical power to detect program effects. With greater investment of resources, and possibly less variant responses to greater dosages of intervention-especially interventions of support, as distinct from punishment-we may also increase our chances of finding politically acceptable interventions that will work.