The study found that screening for domestic violence did not occur regularly in the sample of patient encounters, despite the initial enthusiasm for the project shown by students who volunteered for the study. Although 11 percent of the patient encounters included a discussion of domestic violence, only one student conducted a substantial amount of screening, which accounted for 82 percent of all cases in which domestic violence was discussed. The most common reason given for not asking patients about domestic violence was the perception that it seemed unrelated to the patient's chief complaint or reason for the visit. Domestic violence was never discussed with male patients, despite students having been previously informed of evidence that violence in intimate relationship is often mutual. The modeling of routine screening practices for domestic violence in teaching situations may be an effective means of helping students to learn appropriate ways of introducing this sensitive and typically secret subject in screening for a patient's mental and physical health. Out of 112 students eligible to participate in the study, 26 volunteered to participate, and 20 subsequently dropped out for various reasons, leaving 6 students involved for the duration of the project. The study was conducted for 5 weeks (June 1-July 5, 1999). Students were instructed to act in their clinical practice as they had been instructed in the domestic-violence education they had received at medical school, their own comfort levels, and the guidance of their supervising physician. Following each patient encounter, students completed a patient encounter form regarding any discussion of domestic violence that had occurred. 3 tables and 16 references