Table 5. Significant indirect associations estimated from autoregressive cross-lagged path models.

Indirect path	b (95% CI)	P-value
Screen time → Total sleep timea
Screen time(day 1) → Screen time(day 2) → Total sleep time (day 2)	-2.727 (-4.981, -0.473)	.018
Screen time(day 1) → Screen time(day 2) → Total sleep time (day 2) → Total sleep time(day 3)	-0.919 (-1.764, -0.074)	.033
Screen time(day 2) → Total sleep time(day 2) → Total sleep time(day 3)	-1.868 (-3.540, -0.196)	.028
Activity counts → Total sleep timea
Activity counts(day 1) → Activity counts(day 2) → Activity counts(day 3) → Total sleep time(day 3)	-0.077 (-0.142, -0.012)	.018
Activity counts(day 2) → Activity counts(day 3) → Total sleep time(day 3)	-0.153 (-0.276, -0.030)	.015
Total sleep time → Activity countsa
Total sleep time(day 1) → Activity counts(day 2) → Activity counts(day 3)	-0.057 (-0.090, -0.024)	.001
Screen time → Sleep efficiencyb
Screen time(day 1) → Screen time(day 2) → Sleep efficiency(day 3)	-0.318 (-0.561, -0.075)	.010
Screen time(day 1) → Screen time(day 2) → Sleep efficiency(day 2) → Sleep efficiency(day 3)	-0.139 (-0.253, -0.025)	.017
Screen time(day 2) → Sleep efficiency(day 2) → Sleep efficiency(day 3)	-0.285 (-0.510, -0.060)	.013
Screen time → Sleep fragmentc
Screen time(day 1) → Screen time(day 2) → Sleep fragment(day 2)	0.350 (0.019, 0.681)	.039
Screen time → Sleep qualityd
Screen time(day 1) → Screen time(day 2) → Sleep quality(day 2)	-0.007 (-0.013, -0.001)	.025
Screen time(day 1) → Screen time(day 2) → Sleep quality(day 2) → Sleep quality(day 3)	-0.004 (-0.008, -0.0001)	.028
Screen time(day 2) → Sleep quality(day 2) → Sleep quality(day 3)	-0.008 (-0.016, -0.0002)	.024

b = unstandardized path coefficient; CI = confidence interval.

^a estimated from the model (a) presented in the Fig 1.

^b estimated from the model (b) presented in the Fig 1.

^c estimated from the model (c) presented in the Fig 1.

^d estimated from the model (d) presented in the Fig 1.