Table 2

Binomial logistic regressions (adjusting only for age and sex)

Low: <1 hourHigh: 3 to <5 hoursVery high: 5+ hours
OR (95% CI)P valueOR (95% CI)P valueOR (95% CI)P value
Late sleep onset
(school days)
0.63 (0.53 to 0.75)<0.0011.38 (1.16 to 1.65)<0.0012.75 (2.38 to 3.18)<0.001
Late sleep onset
(free days)
0.6 (0.51 to 0.7)<0.0011.44 (1.19 to 1.74)<0.0013.05 (2.63 to 3.54)<0.001
Late wake time
(school days)
1.12 (0.77 to 1.62)ns1.63 (1.05 to 2.53)<0.052.49 (1.62 to 3.83)<0.001
Late wake time
(free days)
0.81 (0.69 to 0.96)<0.051.16 (0.97 to 1.39) ns 1.82 (1.54 to 2.16)<0.001
Sleep onset latency >30 min0.92 (0.8 to 1.05)ns1.24 (1.04 to 1.49)<0.051.48 (1.27 to 1.71)<0.001
Frequent nighttime awakenings1.06 (0.9 to 1.26)ns1.31 (1.07 to 1.61)<0.052.11 (1.75 to 2.55)<0.001
  • See table 1 for criteria and prevalence of each sleep outcome. ORs measure how much higher or lower the odds of a given sleep outcome are for each category of social media user (low: <1 hour; high: 3 to <5 hours; very high: 5+ hours) compared with the reference category (average users: 1 to <3 hours). ORs and 95% CIs greater than 1 indicate higher odds; those below 1 indicate lower odds. ORs are adjusted to control for exact age and sex.

  • ns, not significant.