LABORATORY SIMULATION OF A COMPLEX INSTITUTIONAL ENVIRONMENT IS JUSTIFIED LARGELY IN TERMS OF EXPLANATIONS OF HUMAN CONDUCT THAT ARE EFFECTIVELY PROVIDED. STUDIES OF THE EFFECT OF REAL INSTITUTIONS ARE SUSCEPTIBLE TO THE CHARGE THAT THE PEOPLE STUDIED ARE DIFFERENT FROM THE GENERAL POPULATION; THUS, THE IMPACT OF VARIABLES IS MINIMIZED BY SPECULATION ABOUT HOW DIFFERENT KINDS OF PEOPLE DISTRIBUTE THEMSELVES THROUGHOUT THE SYSTEM, CAUSING BEHAVIORAL PATHOLOGY TO BE EXPLAINED IN PERSONAL RATHER THAN OBJECTIVE TERMS. LABORATORY SIMULATION ALLOWS THE SCREENING OUT OF POSSIBLY PATHOLOGICAL PERSONS AND RANDOM ASSIGNMENT OF THE REMAINING INDIVIDUALS TO DIFFERENT TREATMENT CONDITIONS. THE INSTITUTIONAL NATURE OF THE LABORATORY SUPPORTS EXPERIMENTAL SIMULATION OF INSTITUTIONAL REALITY; EXPERIMENTAL CONTROL CAN BE VIEWED AS A SUBSET OF INSTITUTIONAL CONTROL. SOCIAL INSTITUTIONS THEMSELVES CAN BE VIEWED AS GRAND EXPERIMENTS, WHICH MAY NOT BE THE ONLY OR EVEN THE MOST DESIRABLE WAYS OF PARTITIONING SOCIETAL REALITY. METHODOLOGICAL PREFERENCES, CHOICE OF VARIABLES, AND LABELING CAN ALL INFLUENCE THE RESULTS OF SOCIAL SCIENCE EXPERIMENTATION, BUT THE DEGREE TO WHICH SOCIAL SCIENCE TECHNOLOGY IS OBJECTIVE OR VALUE FREE MAY BE LARGELY IRRELEVANT GIVEN THE STRONG VALUE COMMITMENTS OF THOSE IN A POSITION TO APPLY IT. IT MUST BE QUESTIONED ALSO WHETHER THE CONDUCT OF SOCIAL SCIENCE RESEARCH ITSELF DOES NOT CREATE A SOCIAL CONTEXT CAPABLE OF SHAPING AND CONTROLLING THE BEHAVIOR OF ALL ITS PARTICIPANTS, INCLUDING THE RESEARCHERS WHO THEMSELVES ARE SUBJECT TO GENERAL PRINCIPLES OF BEHAVIOR. ISSUES INVOLVING DISCOMFORT RESULTING FROM EXPERIMENTATION AND WHETHER INFORMED CONSENT IS A SUFFICIENT ETHICAL PRECAUTION IN RESEARCH ARE DISCUSSED. A SHORT EVALUATION OF THE STANFORD PRISON STUDY IS INCLUDED. FOR INFORMATION ON THIS STUDY, SEE NCJ-10301. (KJM)