Agreement Between Physical Activity Measurement Devices: Steps/Minute and Activity Time

Purpose: The aims of this study were to compare the relative and absolute agreement between the Walk4Life Duo and Yamax SW651 pedometers on the measure of steps/min, and the Walk4Life Duo and observed PA time (min) in high school physical education. Agreement between the Walk4Life pedometer and criterion measures was hypothesized to not demonstrate clinical acceptance (absolute value percent error [AVPE]≤10%), with steps/min significantly underestimated and PA time significantly overestimated.

Methods: High school students (9th-12th grades, N=195) from three Midwestern U.S. schools were participants. Relative agreement was analyzed using correlation, and difference tests were used to examine absolute agreement. Absolute agreement was examined overall; across PA, BMI (kg/m²), and stature (cm) quintiles; and across schedule types (Block vs. Traditional) and lesson content themes (invasion, rock wall/ropes, dance/gymnastics, treadmill/elliptical, and cross-training activities). Alpha was set at .05.

Analysis/Results: Relative agreement for both steps/min and PA time was significant and strong (r>.95). Overall mean difference scores suggest an underestimation of steps/min (3.59±5.75, p<.05) and an overestimation of PA time (-10.52±6.48, p<.05) by the Walk4Life pedometer. The overall AVPE scores fell outside the limits of clinical acceptance for steps/min (11.50±11.20%) and PA time (74.11±37.91%) measure comparisons. Steps/min AVPE scores across PA, BMI and stature quintiles suggest that: (a) the upper PA level quintile reached clinical acceptance (7.59±5.92%) and was significantly different than the lower quintile (17.08±14.04%), (b) the lower three BMI quintile level AVPE scores were clinically acceptable and were significantly different than the upper quintile level, and (c) stature AVPE scores across quintile levels were not significantly different; however, the upper quintile level was clinically acceptable. PA time AVPE scores were not significantly different between PA, BMI, and stature quintile levels, and were not clinically acceptable (62.08±23.35% to 85.77±63.42%). Schedule type (Block vs. Traditional) did not significantly influence AVPE scores for steps/min (10.84±10.62% vs 12.07±11.71%) and PA time (72.43±19.42% vs. 75.21±46.345) comparisons. Steps/min AVPE scores were clinically acceptable for treadmill/elliptical and invasion activities (1.25±.83% and 9.67±10.30%), and PA time AVPE scores were clinically unacceptable for all lesson content themes (59.57±17.34% to 86.65±69.23%), except treadmill/elliptical (8.23±3.97%).

Conclusions: The hypotheses of this study were supported, in that the Walk4Life steps/min and PA time AVPE scores were clinically unacceptable, steps/min was underestimated, and PA time was overestimated. Furthermore, the output functions of steps/min and PA time of the Walk4Life pedometer were influenced by individual and contextual variables.