OBJECTIVE To examine the effectiveness of a formal tracking system for residents' clinical experiences. METHODS We examined three shifts, selected at random, for each resident (without residents' knowledge) during emergency rotations. Information from patient charts was compared with residents' computerized records for rotation (location and preceptor) and patient (age, sex, diagnosis, and procedure) information. METHODS The Northeastern Ontario Family Medicine Program, a program designed to provide remote, rural, and northern resident experience. METHODS First-, second-, and third-year residents on emergency rotations in the academic years 1992 to 1994. METHODS Compliance, reliability, and validity of tracking records. RESULTS Residents recorded patient encounters 88% of the time. Compliance with rotation information was high (100% rotation, 94% preceptor). Agreement on patient age and sex was high. Procedure compliance was somewhat lower (83%). Intrarater reliability (91%) and inter-rater reliability (78%) are acceptably high, as is validity when compared with a gold standard entry (88%). CONCLUSIONS Regular entry of reliable and valid data is facilitated by the computerized resident-patient encounter tracking program. This computer tool should prove useful for multilevel program evaluation in the future.