Related Experiment Video
Updated: Mar 24, 2026
Physical Activity Measurement in Children Accepting Table Tennis Training
Published on: July 27, 2022
Effects of Varying Epoch Lengths, Wear Time Algorithms, and Activity Cut-Points on Estimates of Child Sedentary
Jorge A Banda1,2, K Farish Haydel1,3, Tania Davila1,3
1Stanford Solutions Science Lab, Department of Pediatrics and Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, California, United States of America.
Objective:
To examine the effects of accelerometer epoch lengths, wear time (WT) algorithms, and activity cut-points on estimates of WT, sedentary behavior (SB), and physical activity (PA).
Methods:
268 7-11 year-olds with BMI ≥ 85th percentile for age and sex wore accelerometers on their right hips for 4-7 days. Data were processed and analyzed at epoch lengths of 1-, 5-, 10-, 15-, 30-, and 60-seconds. For each epoch length, WT minutes/day was determined using three common WT algorithms, and minutes/day and percent time spent in SB, light (LPA), moderate (MPA), and vigorous (VPA) PA were determined using five common activity cut-points. ANOVA tested differences in WT, SB, LPA, MPA, VPA, and MVPA when using the different epoch lengths, WT algorithms, and activity cut-points.
Results:
WT minutes/day varied significantly by epoch length when using the NHANES WT algorithm (p < .0001), but did not vary significantly by epoch length when using the ≥ 20 minute consecutive zero or Choi WT algorithms. Minutes/day and percent time spent in SB, LPA, MPA, VPA, and MVPA varied significantly by epoch length for all sets of activity cut-points tested with all three WT algorithms (all p < .0001). Across all epoch lengths, minutes/day and percent time spent in SB, LPA, MPA, VPA, and MVPA also varied significantly across all sets of activity cut-points with all three WT algorithms (all p < .0001).
Conclusions:
The common practice of converting WT algorithms and activity cut-point definitions to match different epoch lengths may introduce significant errors. Estimates of SB and PA from studies that process and analyze data using different epoch lengths, WT algorithms, and/or activity cut-points are not comparable, potentially leading to very different results, interpretations, and conclusions, misleading research and public policy.
