Background: Current interventions to enhance patient self-efficacy, a key mediator of health behavior, have limited primary care application. Objective: To explore the effectiveness of an office-based intervention for training resident physicians to use self-efficacy-enhancing interviewing techniques (SEE IT). Design: Randomized controlled trial. Participants: Family medicine and internal medicine resident physicians (N∈=∈64) at an academic medical center. Measurements: Resident use of SEE IT (a count of ten possible behaviors) was coded from audio recordings of the physician-patient portion of two standardized patient (SP) instructor training visits and two unannounced post-training SP visits, all involving common physical and mental health conditions and behavior change issues. One post-training SP visit involved health conditions similar to those experienced in training, while the other involved new conditions. Results: Experimental group residents demonstrated significantly greater use of SEE IT than controls, starting after the first training visit and sustained through the final post-training visit. The mean effect of the intervention was significant [adjusted incidence rate ratio for increased use of SEE IT∈=∈1.94 (95% confidence interval = 1.34, 2.79; p∈<∈0.001)]. There were no significant effects of resident gender, race/ethnicity, specialty, training level, or SP health conditions. Conclusions: SP instructors can teach resident physicians to apply SEE IT during SP office visits, and the effects extend to health conditions beyond those used for training. Future studies should explore the effects of the intervention on practicing physicians, physician use of SEE IT during actual patient visits, and its influence on patient health behaviors and outcomes.