Time with friends and physical activity as mechanisms linking obesity and television viewing among youth

BACKGROUND: Though bivariate relationships between childhood obesity, physical activity, friendships and television viewing are well documented, empirical assessment of the extent to which links between obesity and television may be mediated by these factors is scarce. This study examines the possibility that time with friends and physical activity are potential mechanisms linking overweight/obesity to television viewing in youth. METHODS: Data were drawn from children ages 10-18 years old (M = 13.81, SD = 2.55) participating in the 2002 wave of Child Development Supplement (CDS) to the Panel Study of Income Dynamics (PSID) (n = 1,545). Data were collected both directly and via self-report from children and their parents. Path analysis was employed to examine a model whereby the relationships between youth overweight/obesity and television viewing were mediated by time spent with friends and moderate-to-vigorous physical activity (MVPA). RESULTS: Overweight/obesity was directly related to less time spent with friends, but not to MVPA. Time spent with friends was directly and positively related to MVPA, and directly and negatively related to time spent watching television without friends. In turn, MVPA was directly and negatively related to watching television without friends. There were significant indirect effects of both overweight/obesity and time with friends on television viewing through MVPA, and of overweight/obesity on MVPA through time with friends. Net of any indirect effects, the direct effect of overweight/obesity on television viewing remained. The final model fit the data extremely well (χ(2) = 5.77, df = 5, p<0.0001, RMSEA = 0.01, CFI = 0.99, TLI =0.99). CONCLUSIONS: We found good evidence that the positive relationships between time with friends and physical activity are important mediators of links between overweight/obesity and television viewing in youth. These findings highlight the importance of moving from examinations of bivariate relationships between weight status and television viewing to more nuanced explanatory models which attempt to identify and unpack the possible mechanisms linking them.