Many studies show positive bidirectional associations between physical activity (PA) and sleep at the between-person level. There is an increased interest in investigating these associations at the within-person level. Few studies examined the effects of time-varying moderators on the within-person bidirectional associations between PA and sleep. This study aimed to examine the bidirectional within-person day-level associations between activity levels and self-reported sleep duration and explore the moderating effects of perceived stress on these day-level associations.
Data from 158 women that included 7-day free-living monitoring over 4 measurement periods was analyzed using multilevel modeling to explore the moderating effects of daily stress on the bidirectional, within-person associations between activity levels and self-reported sleep duration. Moderate-to-vigorous PA (MVPA) and sedentary behavior (SB) were estimated from a waist-worn accelerometer. Self-reported sleep duration and perceived stress were collected by ecological momentary assessment.
No significant within-person associations between MVPA minutes and self-reported sleep duration were found in either direction. However, engaging in more MVPA than one’s usual level was associated with longer sleep later that night when perceived stress was higher than usual (p = .04). Bidirectional negative within-person association between SB minutes and self-reported sleep duration was found (ps < .01). The negative association between SB and sleep duration later that night was stronger when perceived stress was lower than usual (p = .01).
Daily stress played an important role in the day-to-day associations between activity levels and sleep.