Reliability and Validity of Pedometers in a Free-living Environment

• Main Author: Brown, Ernest Leroy
• Format: Others
• Published: PDXScholar 1995
• Subjects: Walking > Measurement; Health and Physical Education
• Online Access: https://pdxscholar.library.pdx.edu/open_access_etds/4870; https://pdxscholar.library.pdx.edu/cgi/viewcontent.cgi?article=5942&context=open_access_etds

In the field of exercise science there exists no single best method, or tool, for the measurement of physical activity, in particular, activity in everyday free-living conditions. The pedometer, a tool for recording the number of steps taken by an individual, could potentially measure this important component of free-living physical activity. To establish the reliability and validity of the pedometer, 40 subjects wore two pedometers (same brand) in two consecutive I 0-minute trials during normal daily activity. Both trials were videotaped. Each videotape segment was replayed, the number of steps were counted and this count served as the criterion measure of steps. In order to evaluate the reliability of the criterion measure the researcher recounted ten of the forty trials a second time and performed an intraclass reliability estimate and follow-up ANOVA comparing the two separate counts. This yielded an intra-observer reliability estimate of R=0.99 (F=l .36, p=.27). Data analyses included trial-to-trial comparisons of pedometer recordings, left-toright comparisons of pedometer recordings, and comparisons of pedometer recordings to the established criterion scores. Results of trial-to-trial comparisons yielded intraclass reliability estimates of R=0.87 (F=l .51, p=.23) for the left side pedometer and R=0.90 (F=.97, p=.33) for the right side pedometer; no significant differences were found. Estimates of pedometer consistency (left versus right pedometer) yielded a correlation ofR=0.96, with follow-up ANOVA (F=6.46 and p=.02) indicating significant differences between left and right side pedometers. Comparisons of pedometers to the established criterion scores (validity) yielded correlations ofR=0.84 (F=l .85, p=.18) for the left pedometer and R=O. 79 (F=S. 71, p=.02) for the right pedometer. Follow-up ANOVA indicated a significant difference between pedometer and criterion scores for the right pedometer but not the left. Under the conditions of this study, the pedometer worn at the waist level directly above the left leg provided reliable and valid measures of walking steps taken during typical everyday activities. The pedometer worn on the right side of the body underestimated the number of steps taken. Further research on the influence of leg dominance, surface, shoe type, pedometer brand, and gait is needed.