Shared sleep fragmentation in undergraduate roommates: dyadic actigraphy in a dense urban campus setting

Introduction Dyadic sleep research has focused on romantic partners, leaving open whether nonromantic people who share a sleep space show related sleep. We examined sleep continuity and roommate concordance among undergraduate roommate dyads on a dense urban campus bordered by an interstate highway, freight-rail corridors, and an overnight air-cargo hub. The setting is described as context; no environmental exposure was measured. Methods Baseline data were analyzed from 29 roommate dyads (58 participants aged 18–19 years) enrolling in a mindfulness program at the University of Louisville in 2017. Participants were instructed to wear wrist actigraphs for four nights and completed one-week versions of the Pittsburgh Sleep Quality Index (PSQI) and Perceived Stress Scale. The planned analysis estimated roommate concordance with intraclass correlations (ICC); analyses of same-night covariation, bedroom sharing, and perceived stress were exploratory. To separate roommate similarity from shared calendar timing, roommates were compared with same-wave non-roommates (pseudo-dyads), and night-level analyses were adjusted for day of week and week of semester. Results Participants averaged 6.98 h of sleep, 66 min of wake after sleep onset (WASO), and 18.6 actigraphy-detected awakenings per night. Roommate concordance was clearest for awakening frequency (ICC = 0.38, 95% CI [0.04, 0.68]), which exceeded pseudo-dyad concordance (mean ICC = 0.09); after centering on data-collection wave it was lower and imprecise (ICC = 0.25 [−0.11, 0.59]). Six other actigraphy outcomes had smaller estimates, with intervals including zero. Roommates’ night-to-night deviations in WASO were correlated ( r = 0.29 [0.07, 0.44]); this association persisted after day-of-week and week-of-semester adjustment and exceeded 95% of pseudo-dyad re-pairings, whereas same-night covariation in awakenings did not. Bedroom-sharing comparisons (18 vs. 10 dyads) were too imprecise to be informative. Perceived stress was associated with participants’ own PSQI scores ( β = 0.55 [0.39, 0.71]) and, more weakly, their own WASO ( β = 0.23 [0.04, 0.43]). Conclusion Undergraduate roommates showed modest similarity in sleep fragmentation that was not fully explained by shared calendar timing, extending dyadic sleep research beyond romantic partners. Because environmental exposures were not measured, the findings cannot be attributed to the urban setting; studies pairing dyadic sleep measurement with synchronized environmental monitoring are needed.