Tutek Joshua, Molzof Hylton E, Lichstein Kenneth L
a The University of Alabama , Department of Psychology, Tuscaloosa, United States of America.
Chronobiol Int. 2017;34(10):1401-1412. doi: 10.1080/07420528.2017.1373118. Epub 2017 Oct 24.
Nonrestorative sleep, a form of subjective sleep disturbance that has been largely neglected in the literature, is newly accessible to researchers via the validated restorative sleep questionnaire (RSQ). The daily version of the RSQ allows for analysis of within-subjects variation in restorative sleep across repeated samplings, and such day-to-day regularity in sleep variables has been highlighted as an important new direction for research. The present study used a sophisticated statistical approach, multilevel modeling, to examine the contributions of circadian chronotype, calendar day of questionnaire completion (weekends versus weekdays), and their interaction in explaining both interindividual and intraindividual variance in restorative sleep. Analyses were conducted using an archival dataset of college undergraduates who continuously completed daily RSQs over a 14-day sampling period. In the final multilevel model, possessing an evening type predicted lower restorative sleep between subjects, while sampling on weekdays predicted lower restorative sleep within subjects. Furthermore, a cross-level interaction was observed, such that the difference in restorative sleep on weekends versus weekdays was more pronounced among those with greater evening circadian preference. All of the effects were maintained after accounting for the significant influence of gender (women had less restorative sleep than men). These results are theoretically consistent with findings that evening types display stronger disparities in sleep schedules across free and workdays (i.e., social jet lag), and attest to the usefulness of multilevel models for statistically investigating how stable traits interact with factors that vary day to day (e.g., work or school schedules) in influencing sleep outcomes.
非恢复性睡眠是一种在文献中很大程度上被忽视的主观睡眠障碍形式，研究人员通过经过验证的恢复性睡眠问卷（RSQ）新近能够对其进行研究。RSQ的每日版本允许分析重复采样中恢复性睡眠的个体内变化，而睡眠变量的这种日常规律性已被视为一个重要的新研究方向。本研究采用了一种复杂的统计方法——多层建模，来检验昼夜节律类型、完成问卷的日历日（周末与工作日）及其相互作用对解释恢复性睡眠的个体间和个体内差异的贡献。分析使用了一个档案数据集，该数据集来自在14天采样期内连续完成每日RSQ的大学生。在最终的多层模型中，属于晚睡型的人在个体间预测恢复性睡眠较低，而在工作日采样在个体内预测恢复性睡眠较低。此外，还观察到一个跨层次的相互作用，即周末与工作日恢复性睡眠的差异在昼夜节律偏好更强的人中更为明显。在考虑了性别的显著影响（女性的恢复性睡眠比男性少）后，所有效应均保持不变。这些结果在理论上与以下发现一致，即晚睡型在自由日和工作日的睡眠时间表上表现出更强的差异（即社会时差），并证明了多层模型在统计研究稳定特质与每日变化的因素（如工作或学校时间表）如何相互作用以影响睡眠结果方面的有用性。
