BACKGROUND: Few studies have examined the causal relationship between transport mode and body mass index (BMI). METHODS: We examined between-person differences and within-person changes in BMI by transport mode over four time points between 2007 and 2013. Data were from the How Areas in Brisbane Influence HealTh and AcTivity project, a population-representative study of persons aged 40-65 in 2007 (baseline) residing in 200 neighbourhoods in Brisbane, Australia. The analytic sample comprised 9931 respondents who reported on their main transport for all travel purposes (work-related and non-work-related). Transport mode was measured as private motor vehicle (PMV), public transport, walking and cycling. Self-reported height and weight were used to derive BMI. Sex-specific analyses were conducted using multilevel hybrid regression before and after adjustment for time-varying and time-invariant confounders. RESULTS: Independent of transport mode and after adjustment for confounders, average BMI increased significantly and linearly across the four time points for both men and women. Men and women who walked or cycled had a significantly lower BMI than their counterparts who used a PMV. BMI was nearly always lower during the time men and women walked or cycled than when they used a PMV; however, few statistically significant differences were observed. For women, BMI was significantly higher during the time they used public transport than when using a PMV. CONCLUSION: The findings suggest a causal association between transport mode and BMI and support calls from health authorities to promote walking and cycling for transport as a way of incorporating physical activity into everyday life to reduce the risk of chronic disease.