Examining the Reciprocal Association Between Worry and Sleep: Disaggregating Between- and Within-Person Effects

Studies investigating the reciprocal association between worry and sleep have yielded mixed findings and suffered from methodological limitations. The purpose of this study was to assess the reciprocal association between worry and various indices of sleep (i.e., sleep onset latency [SOL], total sleep duration [TSD], sleep quality [SQ], and morning lingering [ML]). In a longitudinal daily assessment study, participants (N = 182) completed a sleep diary and a measure of worrisome thinking three times per day for 7 days. We used linear mixed effects models to examine worry predicting sleep, and Tobit regression models to examine sleep predicting worry, and controlled for previous day report of the outcome variable and person- and day-level effects. Results indicate that greater person-level worry, but not day-level worry, was significantly associated with longer SOL, worse SQ, and longer ML. Furthermore, greater person-level SOL, TSD, and ML was associated with greater worry, and greater person-level SQ was associated with less worry. Finally, we found that shorter day-level TSD and lower day-level SQ was associated with greater next-day worry. These findings support a reciprocal person-level association between worry and sleep (except for TSD), and a unidirectional day-level association between TSD and SQ as predictors of worry.