Validity and Reliability of Activity Measures in African-American Girls for GEMS
- Added on December 7, 2010
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 x m. 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, = 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.
Link to Abstract: http://www.ncbi.nlm.nih.gov/pubmed/12618587