Validity and reliability of activity measures in African-American Girls for GEMS

Purpose: To determine the reliability and validity of physical activity monitors and self-report instruments suitable for young African-American girls. Methods: A validation study was conducted by the Girls health Enrichment Multi-site Studies (GEMS) research team to compare an accelerometer with a pedometer and two self-report instruments for assessing physical activity in African-American girls, age 8-9 yr. Girls (N = 68) attended two clinic visits spaced 4 d apart. Each girl wore a MTI/CSA accelerometer (used as the criterion standard for validity) and a pedometer simultaneously for four consecutive days. Girls completed on two occasions a 24-h physical activity checklist of yesterday and usual activities, including sedentary activities (GEMS Activity Questionnaire, GAQ), and a 3-d computerized self-report instrument (Activitygram). Results: Girls were (mean ± SD) 9.0 ± 0.6 yr old and had a body mass index of 19.4 kg.m^-2. Reliability measured by intraclass correlations (ICC) and Pearson correlation coefficients (r) were calculated for the MTI/CSA (ICC = 0.37, P < 0.0001), pedometer (ICC = 0.08, P = 0.094), Activitygram (ICC = 0.24) (P = 0.005), and GAQ for physical (r = 0.80, P < 0.0001) and sedentary (r = 0.3-0.5, P < 0.005) activities. Significant Pearson correlations between the MTI/CSA and the other instruments, as a measure of validity, were observed for the 4-d average pedometer score (r = 0.47, P < 0.0001), 3-d average Activitygram score (r = 0.37, P = 0.002), and the average of the two yesterday and two usual GAQ activity scores for a subset of 18 physical activities questions (r = 0.27, P = 0.03; and r = 0.29, P = 0.02, respectively). The MTI/CSA was uncorrelated with single day scores from the three other instruments. Conclusion: The reliability of the instruments tested was acceptable, except the pedometer. Validity correlations were significant when more than one day was used. Self-report instruments need further development for improved reliability and validity.