ABSTRACT
Social experiments randomly assign people (or sometimes sets of people, i.e., neighborhoods or communities) either to a group that is subject to one or more policy “treatments” or to one that continues to be subject to the prevailing policy norm (“controls”). For example, a social experiment might test the effi cacy of a welfare-to-work program by randomly assigning welfare applicants to the new program (perhaps an intensive, coached job search combined with the provision of services like transportation assistance and subsidized child care) and to the old standard, which leaves the initiative to fi nd a job nearly completely up to the welfare recipient.1
