Every correlation here is with the questionnaire total, on the same 4,810 teenagers. r² is the share of the differences between them that the measure accounts for.

measure	r	r²
screen_time_index	+0.1344	1.8%
sleep_quality_index	+0.3806	14.5%
avg_sleep_hours	−0.2603

Sleep quality accounts for 8× the variance screen time does.

how much of the screen effect travels with sleep

held constant	screen × it	partial r	absorbed
avg_sleep_hours	−0.2563	+0.0725	46%
midsleep_weekend_hours	+0.2865	+0.0968	28%
sleep_quality_index	+0.0923	+0.1078	20%
social_jetlag_hours	+0.1820	+0.1237	8%

Starting from +0.1344. Holding hours slept constant absorbs 46% of it. This is a decomposition, not a causal claim — see the limits note on the piece.

the association is the same in both sexes

group	n	pearson	95% CI
Boy	2,446	+0.1376	+0.0985 – +0.1763
Girl	2,364	+0.1409	+0.1012 – +0.1802