Reliability and Validity of Questions on the Youth Media Campaign Longitudinal Survey

Abstract

The study evaluated the reliability and validity of the physical activity questions in the Youth Media Campaign Longitudinal Survey (YMCLS), a nationally representative survey of 9- to 13-yr-old youth. The participants were 192 youth aged 9-13 yr (93 males and 99 females) in grades 4-8 from eight schools in a large, ethnically diverse school district. Participants completed two YMCLS phone interviews, which were administered 1 wk apart by trained interviewers. Reliability was examined by comparing data from two administrations of the survey. Validity was examined by comparing YMCLS measures from the second administration with temporally matched measures from an accelerometer and a detailed activity log. Reliability coefficients were high for estimates of organized activity (intraclass correlation coefficient (ICC)=0.78) and moderate for estimates of free-time activity (ICC=0.60) and total weekly activity (ICC=0.60). Estimates of total weekly activity sessions were significantly correlated with the accelerometer (r=0.24) and the activity log (r=0.46). Estimates of activity time and activity sessions on the previous day were also significantly correlated with the accelerometer (r=0.53 and 0.37, respectively) and the activity log (r=0.37 and 0.47, respectively). Correlations between the YMCLS and the activity log were higher for organized activity (r=0.72) than for free-time activity (r=0.46). Reliability and validity coefficients were similar for boys and girls, but older youth (11-13 yr) had higher coefficients than younger students (9-10 yr). The YMCLS has acceptable reliability and validity for estimating habitual physical activity in youth aged 9-13 yr.