Abstract: | Gender has been neglected in models of the social determinants of health. We use walking as a case study to demonstrate how gender might be incorporated into multilevel social determinants of health frameworks to investigate health behaviours. We found that while men and women had some similar individual (e.g. confidence in doing regular physical activity) and environmental (e.g. presence of destinations) predictors of walking there were also gender differences in the associations found at both of these levels. For example, low levels of education were only associated with men's walking time while having people in the household who made walking easy or hard was only associated with women's walking time. Likewise, having a variety of places to walk to was important for women's walking but not men's. These results indicate that both universal and gender‐specific approaches to health education, health promotion and planning might be needed to improve walking levels. |