Linear regression analyses predicting body mass index (BMI) from daily TV viewing time.

a. Only TV viewing time entered the model as independent variable. b. Model was adjusted for gender and age. c. Model was adjusted for gender, age, employment status, marital status, education level. d. Model was adjusted for gender, age, employment status, marital status, education level and vigorous physical activity. e. Models were adjusted for age, employment status, marital status, education level and vigorous physical activity. The interaction terms were constructed for each moderator. f. Models were adjusted for age, employment status, marital status, education level and vigorous physical activity. g. Models were adjusted for gender, employment status, marital status, education level and vigorous physical activity.