LAUSR

ABSTRACT Free-living comparability of wGT3X-BT and GT9X ActiGraphs has not been verified. Participants (n = 34, 35% females, 7–17 y) wore hip-worn monitors for one week. Vector magnitude (VM; 15-s epoch), Mean Amplitude Deviation (MAD) and Euclidean Norm Minus One (ENMO) (5-s epoch), and percent of time spent in each activity intensity (using cut-points) were calculated. At the epoch-level, correlations were strong (r = 0.822–0.963), mean absolute percent difference (MAPD) was 41.5 ± 62.2% (VM), 40.8 ± 52.7% (MAD), and 110.9 ± 74.0%, with moderate (ENMO κ = 0.664–0.789) to strong (VM κ = 0.915) agreement for activity intensity. When collapsed to the individual level, MAPD was 4.2 ± 3.4% (VM), 6.8 ± 3.7% (MAD), and 35.9 ± 26.1% (ENMO), correlations were moderate (ENMO r = 0.618) to high (MAD and VM r = 0.996), but ENMO, MAD, and time spent in activity intensities using ENMO cut-points were not statistically equivalent (using 2 one-sided tests) between monitors. While time spent in activity intensities was often comparable between models, caution is warranted for epoch-level comparisons.