FACTOR STRUCTURE AND CORRELATES OF SEDENTARY BEHAVIOR OF CHILDREN IN GRADES 6-8

779 Research on sedentary behavior in youth has focused on the relation between television watching and body composition. Few studies have examined other sedentary behaviors that may contribute to inactivity. We examined the factor structure of 7 sedentary behaviors in 1626 11-13 yr. youth from 24 middle schools using a 7-day recall checklist (response rate = 72%). Test-retest reliability estimates (ICC) for the sedentary behaviors ranged from R = 0.3 to 0.96. Additional correlates of sedentary behavior were also examined. The sample was comprised of 58% girls, with 53% white, 16% Latino, 12% multi-ethnic, 11% Asian/Pacific Islander, and 7% African-American. Principal components analysis with varimax rotation yielded a 3 component solution that explained 57% of the variance in youth sedentary behavior. Sedentary Entertainment (watching TV/video, playing video games) explained the largest variance (21%), followed by Socializing (18%; talking on phone, talking with friends), and Educational Activities (18%; homework, computer/internet use, reading). Girls engaged in more Socializing than boys (1.8 vs. 1.1 hrs/day, respectively; p<.001), less Entertainment (1.6 vs. 2.3 hrs/day, p<.001), and similar amounts of Educational activity (1.6 vs. 1.5 hrs/day). Adjusting for household education, total physical activity (PA) correlated positively with all 3 components of sedentary behavior in boys, but only Educational activities and Socializing in girls. For girls, liking of school physical education was negatively correlated with sedentary Entertainment (r = −.12, p<.001). Sedentary Socializing was also associated with selected unhealthful dietary habits of boys and girls. Middle school students spend a large amount of time in sedentary behaviors (mean = 4.73 hrs/day), and 3 factors were identified. Surprisingly high correlations with PA were found, suggesting either that there is time for both sedentary behavior and PA, or there is a response bias. Supported by NIH HL54564