Objectives. To examine: 1) the relationship between residents' responses toward caring for underserved children and families during residency and their perceptions of their continuity clinic experience; and 2) which characteristics are related to continuing to care for the underserved after completion of training. Design. Cross-sectional descriptive study. Methods. A 49-item questionnaire was mailed to 210 third-year pediatric residents at 12 urban training programs in the Northeast in May 1995. Information was collected about residents' emotional responses toward caring for underserved families, their assessments of clinic operations, their sense of effectiveness in caring for underserved patients in continuity clinic, preresidency experiences with the underserved, and their intent to care for the underserved after training. Results. Of 210 surveys mailed, 71% were returned. Thirty-six percent of residents planned to pursue a career in primary care, 53% did not, and 11% did not answer or were planning a year as chief resident. Fifty-seven percent of all residents planned to devote a portion of their practice toward caring for the underserved after training. Residents whose emotional responses toward caring for the underserved included: 1) not worrying that they had become numb to children's psychosocial difficulties, 2) not feeling angry with how families cared for their children, and 3) feeling more empathy with the underserved had significantly higher mean scores on both their assessment of clinic operations and their sense of effectiveness. The only demographic characteristic associated with a greater sense of effectiveness was being black. To better characterize which residents planned to care for the underserved after training, we examined a subsample of 46 residents who recalled an interest in caring for the underserved during residency training and who were pursuing a career in primary care. Residents that did not recall an interest in caring for the underserved at the onset of residency training were unlikely to have plans to care for the underserved after the completion of training. Within this group residents who planned to care for the underserved after training differed significantly from residents who did not plan to continue this work by feeling a greater sense of effectiveness in clinic, feeling less worried about becoming numb, and having greater empathy for underserved families. Conclusions. There are a number of identifiable emotional responses residents develop toward caring for the underserved that relate to their perceptions of continuity clinic and whether a resident chooses to continue to care for the underserved after training.