BACKGROUND: In Pakistan, there is a lack of information about the work environment of residency programs. This lack is a major impediment in their improvement. One of the approaches for improvement in these programs can be directed through the residents' own perception of their working conditions. Therefore, we collected data which would reflect working conditions of residents. AIM: To assess the perceived status of "work environment" in different specialities. MATERIALS AND METHODS: A cross-sectional survey was conducted in four teaching hospitals of Karachi from July 1999 to January 2000. Residents from selected programs were grouped into four broad groups: specialist, medical, surgical and multidisciplinary. Responses of residents were obtained on a Likert scale of 0 to 4. Indices were formed for two components of work environment: academic and mistreatment. STATISTICAL ANALYSIS: Differences between residents' groups were assessed through analysis of variance (ANOVA) at 5% significance. RESULTS: A total of 341 registered residents responded (response rate: 75%). Surgical residents were working more than 80 h/week and this was more than the other three groups. Medical residents were spending the highest actual time on research and teaching activities (10% and 14%, respectively). Academic index score was highest for surgical group (15.81, SD = 4.69) and lowest for multidisciplinary group (11.82, SD = 4.80). Medical group had the highest perceived mistreatment index score (5.56, SD = 4.57). CONCLUSIONS: In a study of work environment of residency programs, differential impact was found for the four groups on work environment perceptions. Most of the residents recognized undergraduate teaching, grand rounds, patient rounds and seminars or workshops as contributing to their academic learning. Reporting of sexual harassment was low, indicating either underreporting or cultural dynamics of our setting.