Background: Despite known health benefits, the majority of children and adolescents do not meet recommended physical activity and sedentary behaviour guidelines. Amongst determinants of these behaviours, the influence of friends and peer groups remains largely understudied. In this thesis I investigated the influence of friends, peer groups and friendship network structure on physical activity and sedentary behaviours. Methods: Methods were drawn from a set of analytical tools known as 'Social Network Analysis', which use friendship nomination data, along with measures of physical activity and sedentary behaviour, to investigate the extent to which friendship network structure was related to the activity behaviours of individuals in the network. Results: There was strong and consistent evidence for similarities in physical activity and sedentary behaviour between friends and within friendship groups, both in children and adolescents. There was a positive correlation in physical activity behaviours between childhood friends, and the likelihood of friendship between two adolescents increased as their physical activity or sedentary behaviour became more similar. The extent of similarity in the behaviours between friends increased with age, and was strongest in late adolescence. Longitudinal analysis over a 7 year period indicated that friendship similarities in physical activity were partly explained by the influence of existing friends in changing an individual's behaviour, and partly by the selection of new friends with similar activity levels. In contrast friendship similarities in sedentary behaviour appear to be solely explained by selection of new similar friends, rather than the influence of existing friends. Conclusions: My findings indicate that friends and peer groups play an important role in shaping both physical activity and sedentary behaviour in children and adolescents, and that harnessing peer influence for behaviour change is a promising direction for future physical activity and sedentary behaviour interventions.