Several studies have shown socioeconomic differences in leisure-time physical activity. One explanation may be socioeconomic differences in relevant psychosocial conditions. The Malm? Diet and Cancer Study is a prospective cohort study including inhabitants in Malm?, Sweden. The baseline questionnaire used in this cross-sectional study was completed by the 11,837 participants born 1926-1945 in 1992-1994. Leisure-time physical activity was measured by an item presenting a variety of activities. These activities were aggregated into a summary measure of leisure-time physical activity that takes both the intensity and duration of each specific activity into consideration. The effects of the psychosocial variables on the socioeconomic differences in leisure-time physical activity were calculated in a multivariate logistic regression analysis. The quartile with the lowest degree of leisure-time physical activity was not evenly distributed between the socioeconomic groups. Socioeconomic differences were seen as odds ratios 1.5 for skilled and 1.5 for unskilled male manual workers, compared to the high level non-manual employees. An OR 1.6 was observed for female unskilled manual workers. Self-employed men and female pensioners also had a significantly increased risk of low leisure-time physical activity. Adjustment for age, country of origin and previous/current diseases had no effect on these SES differences. Finally, adjusting for social participation almost completely erased the SES differences. Among the psychosocial variables, social participation was the strongest predictor of low physical activity, and a strong predictor for socioeconomic differences in low leisure-time physical activity. Social participation measures the individual's social activities in, for example political parties and organisations. It therefore seems possible that some of the socioeconomic differences in leisure-time physical activity are due to differing social capital between socioeconomic groups.