This excluded 1.3–2.2% of cases from each continuous variable. Owing to a substantial selleck proportion of cases with at least one missing value in at least one covariable or exposure variable (22–28% depending on the exposure variable) we performed multiple imputation. IBM SPSS V.20 was used to conduct the multiple imputation, missing values were imputed

for all covariables and exposures, with observed maximum and minimum values used as constraints. Outcome variables did not have missing values imputed, but were included in the imputation models to predict missing values in other variables. Linear regression was used as the type of imputation, and five cycles of imputation were conducted resulting in five imputed data sets. Results from these five

data sets were combined using the multiple imputation module in SPSS to provide pooled results. The imputed sample size is limited to the number of valid observations for each outcome variable (2289 for accelerometry-measured ST, 2279 for TV time, 2253 for non-TV sitting time and 1170 for occupational sitting time). Non-imputed results are presented in the appendix. Statistical analysis Analyses were weighted for non-response to give a sample that was representative of adults living in England. The associations between each of the socioeconomic indicators (household income, social class, education, SEP score and area deprivation,) and each individual ST indicator (TV time, non-TV sitting tine, occupational sitting/standing and accelerometry-measured ST) were examined using generalised linear models (GLM), and by multiple linear regression to determine linear trend p values. Results are presented for the whole week, the weekday/weekend day-specific results can be found in the online

appendix. We also repeated the SEP score analyses stratified by economic activity (employed/self-employed vs non-economically active). SPSS V.21 was used for all analyses. For all multivariate analyses Dacomitinib we used the complex samples GLM procedure to take into account the complex survey design. Different models were adjusted for: (1) age and sex; (2) additionally for BMI, limiting long standing illness, difficulty with usual activities, car ownership, drinking frequency, smoking status, and other socioeconomic indicators (household income, social class, area deprivation); (3) additionally for time spent in self-reported MVPA or accelerometry-measured MVPA as appropriate, and average accelerometer wear time on valid days. Models 2 and 3 with accelerometry-measured ST as the outcome were also adjusted for average accelerometer wear time on each valid day. This work conforms with the STROBE statement for observational studies.