AbstractObjectiveTo assess the relationship between obesity and sedentary behaviours, such as watching television or using a computer (‘screen-time’), and describe how this relationship varies between population subgroups.DesignCross-sectional analysis of the relationship between obesity (BMI ≥ 30 kg/m²) and screen-time, adjusted for age, sex, income and education and compared according to a range of personal characteristics.SettingNew South Wales, Australia.SubjectsA total of 91 266 men and women aged 45 years and above from the general population of New South Wales in 2006–2007 and providing self-reported information on height and weight and other factors.ResultsObesity prevalence was 21·4 %. Compared to individuals with <2 h of daily screen-time, the adjusted relative risks (RR) of obesity were 1·35 (95 % CI 1·26, 1·44), 1·70 (95 % CI 1·59, 1·82), 1·94 (95 % CI 1·81, 2·08) and 1·92 (95 % CI 1·80, 2·06) for 2–3, 4–5, 6–7 and ≥8 h, respectively. The increase in obesity with increasing screen-time was similar within categories of overall physical activity, but was attenuated in those in full-time paid work, compared to non-workers (P for interaction < 0·0001). Among non-workers, the overall obesity RR per 2 h increase in daily screen-time was 1·23 (95 % CI 1·21, 1·25) and was significantly elevated in all groups examined, ranging from 1·16 to 1·31 according to sex, level of age, education, income, smoking and fruit consumption. The RR did not differ significantly according to overall physical activity, region of residence and alcohol and vegetable consumption, but was substantially lower in disabled v. not-disabled individuals (P for interaction < 0·0001).ConclusionsObesity increases with increasing screen-time, independent of purposeful physical activity. This was observed in all population groups examined, although it is attenuated in full-time workers and disabled individuals.