Household income was not associated with non-TV sitting time. Figure 3 Multivariate-adjusted average daily sedentary time by household income quartile. Model 1: adjusted for age and sex; model 2: further adjustments for area deprivation, body mass index (BMI), limiting long

standing illness, difficulty with usual activities, currently … Educational attainment and ST Figure 4 presents the associations between the highest educational qualification and each measure of ST. Educational attainment was positively associated with accelerometry-measured ST and inversely associated with TV time in all models. Occupational sitting/standing time was inversely associated with education but the association did not appear to be linear (it was evident across the lowest three educational levels only) and was attenuated to the null following adjustments for potential confounders. There was a weak positive association between education and non-TV sitting time, following adjustments for potential confounders in models 2 and 3. Figure 4 Multivariate-adjusted difference average daily sedentary time by highest qualification. Model 1: adjusted for age and sex; model 2: further adjustments for area deprivation,

body mass index (BMI), limiting long standing illness, difficulty with usual … Occupational social class and ST As shown in figure 5, occupational social class was positively associated with accelerometry-measured ST and occupational sitting/standing. The initial inverse association with TV time (model 1) was attenuated to the null following adjustments for potential confounders. Similarly to SEP score

and income, social class was not associated with non-TV sitting time. Figure 5 Multivariate-adjusted average daily sedentary time by occupational social class. Model 1: adjusted for age and sex; model 2: further adjustments for area deprivation, body mass index (BMI), limiting long standing illness, difficulty with usual activities, … Area deprivation and ST Area-level deprivation was positively associated with TV time (the lower the deprivation the lower the TV time) but these associations did not persist in the adjusted models (figure Entinostat 6). Area deprivation was not associated with any other measures of ST (figure 6). Figure 6 Multivariate-adjusted sedentary time by area deprivation quintile. Model 1: adjusted for age and sex; model 2: further adjustments for area deprivation, body mass index (BMI), limiting long standing illness, difficulty with usual activities, car ownership, … Differential associations between imputed and non-imputed data There were no differences between the imputed and non-imputed models describing the associations between SEP score and ST indicators, although the 95% CIs were slightly broader in the unimputed models due to the lower sample size (see online supplementary figure S2).