Validity of Pedometers to Measure Step Counts During Dance

Background/Purpose: To establish validity evidence for measurement tools is an endless process because validity is context specific. Pedometers are considered a valid measurement tool of step counts during controlled and free-living situations, however it is of common belief that pedometers cannot measure step counts during dancing activity. With this in mind, the purpose of this study was to establish validity evidence for the measurement of step counts by pedometer during different styles of dance. Methods: A total of 30 adults of a college dance class participated in this study. The participants completed a trial of 3 or 5 min of different styles of dance which included shag, salsa, meringue, waltz, shuffle, mixer, and freestyle. While dancing, the participants wore a Walk4life W4L MVP pedometer on the right hip and had their number of steps counted by a research assistant. To determine the accuracy of pedometer determined steps, Pearson correlation, paired t-test, absolute percent error (APE), and bias were calculated between actual step counts and pedometer step counts for the 3 and 5-min dances. Analysis/Results: For the 3-min dances the actual step count was 223 ± 98 and the pedometer count was 211 ± 113. For the 5-min dances the actual step count was 367 ± 115 and the pedometer count was 359 ± 130. The correlations were .92 and .77 for the 3 and 5 min dances, respectively. T-test showed no significant differences for the 3 (t(29) = 1.46, p = .16) and 5-min dances (t(29) = 0.53, p = .60). However, APE was extremely high, 17.7 and 19.4% for the 3 and 5-min dances, respectively. Bias was -8.0% and -0.1%, respectively. Conclusions: The overall averages of actual step counts and pedometer determined step counts were not significantly different. However at the individual level there was a large error, indicating that the pedometers did not follow a patter of over or under estimating the step counts. This demonstrates that pedometers can be used while dancing to estimate the number of steps taken by a group, but not for individuals.