Using a factorial design, four aspects of an educational program for 160 hypertensive patients were manipulated: number of meetings, patient responsibility and participation, directiveness of the intervention, and emphasis on negative consequences of uncontrolled hypertension. Validity checks on the manipulations included content analysis of the nurse-patient interaction and interview-based measures of the patient's responsibility, participation, and awareness of dangers. Outcome variables included repeated measures of patient knowledge, assessment by the nurse of patient attainment of identified goals, and reduction of the patient's mean arterial blood pressure. High indirect interventions tended to lead to higher goal attainment, particularly in the psychosocial area. Emphasis on negative consequences tended to promote learning for patients with long standing diagnoses, but to retard learning for recently diagnosed patients. Additional meetings and emphasis on patient responsibility were not helpful alone, but in combination they tended to lead to greater learning. Although as a whole, patients in the program tended to reduce their blood pressures, there were no statistically significant main effects or interaction effects of the educational approach variables on blood pressure reduction.