Background Epidemiologic research suggests that physical activity is associated with decreased prevalence of depression. However, the relationship between physical activity accumulated in various domains and depression remains unclear. Further, previous population-based studies have predominantly utilized self-reported measures of physical activity and depression symptom subscales. Associations between physical activity in various domains (leisure, work, active commuting, yard/household) and depression were examined using both subjective and objective measures of physical activity and a diagnostic measure of depression. Methods Analyses (conducted in 2007) included data from 1995 young adults participating in a national study (2004–2006). Physical activity was measured by self-report (International Physical Activity Questionnaire) and objectively as pedometer steps/day. Depression (DSM-IV 12-month diagnosis of major depression or dysthymic disorder) was assessed using the Composite International Diagnostic Interview. Results For women, moderate levels of ambulatory activity (≥7500 steps/day) were associated with ∼50% lower prevalence of depression compared with being sedentary (<5000 steps/day) ( p trend=0.005). Relatively low durations of leisure physical activity (≥1.25 hours/week) were associated with ∼45% lower prevalence compared with the sedentary group (0 hours/week) ( p trend=0.003). In contrast, high durations of work physical activity (≥10 hours/week) were associated with an approximate twofold higher prevalence of depression compared with being sedentary (0 hours/week) ( p trend=0.005). No significant associations were observed for steps/day in men or for other types of self-reported activity including total physical activity in both men and women. Conclusions These findings indicate that the context in which physical activity is assessed and the measurement methods utilized are important considerations when investigating associations between physical activity and depression.