BackgroundLittle is known about what happens to active commuting as children get older, and less is known about influences on changes in this behavior. This study examined predictors of increases in children's and adolescents' active commuting (walking or cycling) to/from school over a 2-year period.MethodsParticipants were initially recruited and assessed in 2001. Follow-up data were collected in 2004 and 2006 and analyzed in 2008. Participants were 121 children (aged 9.1±0.34 years in 2004) and 188 adolescents (aged 14.5±0.65 years in 2004) from Melbourne, Australia. Parents and adolescents reported their perceptions of individual-level factors and of the neighborhood social and physical environment. Weekly active commuting (walking or cycling) to/from school, ranging from 0 to 10 trips/week was also proxy- or self-reported at the initial measurement and again 2 years later. Logistic regression analyses examined predictors of increases in active commuting over time.ResultsChildren whose parents knew many people in their neighborhood were more likely to increase their active commuting (OR=2.6; CI=1.2, 5.9; p=0.02) compared with other children. Adolescents whose parents perceived there to be insufficient traffic lights and pedestrian crossings in their neighborhood were less likely to increase their active commuting over 2 years (OR=0.4; CI=0.2, 0.8; p=0.01), whereas adolescents of parents who were satisfied with the number of pedestrian crossings were more likely to increase their active commuting (OR=2.4; CI=1.1, 5.4; p=0.03) compared with other adolescents.ConclusionsSocial factors and physical environmental characteristics were the most important predictors of active commuting in children and adolescents, respectively.