Erve Chambers is Assistant Professor of Anthropology at the University of South Florida, Tampa. The article is based on work done with the U.S. Department of Housing and Urban Development's Experimental Housing Allowance Program, while employed as an on-site observer by Ab t Associates, Inc., Cambridge, Massachusetts. The writer is indebted t o Dr. Harry Wolcott and Chris Gaffney, University o f Oregon, and t o Dr. Ev Glatt, Program Development and Research, U.S. Department of Housing and Urban Development, Washington, D. C. D URING THE LAST few years the U.S. Department of Housing and Urban Development (HUD) has sponsored major research on a new housing allowance program. One aspect of the research was the experimental operation of eight administrative agencies, located in various communities across the country.' These were fully operative housing agencies, each funded to enroll and provide financial assistance to several hundred low and middle income families. The program offered a unique opportunity to follow a social innovation from its beginning. There have been few previous opportunities to design and initiate comparative research on the activities of public service agencies, their impact on the populations they serve, and their impact on the comnlunity at large. Most research on similar organizations has been formulated around a set of preexisting administrative conditions. In this case, service agencies were created around an experimental design. It was as close as one might expect to come to creating a laboratory in the field. One feature of the research design for HUD's experimental housing allowance program which should be of particular interest to anthropologists was the use of experienced fieldworkers in on-site observation. Each administrative agency was assigned a full-time observer who would live in the community, visit the agency daily, work with a limited number of program participants, and perform specified data collection and reporting duties. In six of the eight sites, the individuals contracted to do this work had extensive training in anthropology. This was a unique form of employment, not only for the anthropologists actually on site with the experiment, but for anthropology in general. The following report will provide a brief discussion of the work, for the most part derived from personal experience.* There also will be an attempt to place the work in the larger context of current anthropological research, particularly as regards areas of policy relevant and urban studies.