Table 8.

Calibration and sharpness on the test set.

Source	$ECE(\widehat{Q})$	$IQR(\widehat{Q})$
Apple Watch (N=726)	0.037930	3.904
Garmin (N=53)	0.037934	3.994

Mean expected calibration error and interquartile range computed separately for test samples collected using Apple Watch devices and Garmin devices.