A version of the research act, termed naturalistic behaviorism, is presented and compared to existing formulations of the method. A sensitizing framework for organizing naturalistic studies is presented. Special attention is given the problems of sampling, measurement, and causal analysis. Examples from an ongoing study of early childhood socialization are presented as tentative solutions to the sampling and analysis issues. Central weight is given to introspective-investigator accounts of social process.