In addition to more familiar research issues, intervention research projects in naturalistic settings present the investigator with a number of practical challenges including gaining access to potentially resistant populations, maximizing participation rates in the face of weak incentives for cooperation, getting valid answers to sensitive questions, and meeting ethical obligations when health or legal problems are discovered in the course of study. Generalizable approaches to these challenges are addressed in the context of a retrospective evaluation of the implementation of OSHA's 1984 ethylene oxide standard in Massachusetts hospitals. In the evaluation study, enthusiastic cooperation was secured, a 96% participation rate was realized, sensitive questions were posed successfully, and worker health risks discovered in the course of study received attention without having to wait for the write-up of the study results. Key elements in the study population's receptivity appear to have been (I) the investigator's familiarity with the hazard, its setting, respondent concerns and needs, and (2) reciprocity in the form of providing follow-up consulting services as an integral part of the research process, delivered to each hospital at the conclusion of data collection. Specific techniques used in the study are presented in the hopes of aiding other investigators facing similar practical challenges in occupational health and safety intervention research.