The purpose of this paper is to use the New Jersey Experiment (NJE) as a case study context in which to discuss the special problems social policy experimentation poses for the survey research community. Section I presents a brief overview of the NJE's purposes, design, and operations. Section II discusses problems and issues in measurement strategy and instrument design. Section III discusses issues in field operations, and Section IV discusses issues in response error, including validity, reliability, experimental effects and other methodological problems. Section V presents a summary and conclusion.