To test the relationship between stress and sleep and the mod- erating role of coping style, we used a repeated-measure analysis of variance based on the general linear model (GLM) analysis (SAS, 1990), with gender and age as control variables and EFC, PFC, and DE as three linear independent measures.
All of the predictors and control variables were simultaneously entered into the model. The sleep measures on each testing period (low and high stress) were used as the dependent measures, with period as the independent repeating measure (with two levels for low- and high-stress measurement periods).
The test of our hypotheses regarding the moderating role of coping lies in the interactions between the period and the coping scale scores. As shown in Table 1, significant Period � EFC interaction effects were found for sleep period and true sleep time. These interactions indicate that high EFC scores were associated with reduction in sleep period in true sleep time (high-stress period in comparison with low-stress period). The Period � DE interac- tion was marginal and failed significance test with Bonferroni correction.
Significant main effects were found only for PFC for the sleep measures. High PFC scores were associated with longer sleep period and true sleep time (on both study periods). None of the main period effects were significant.
We retested the results of this model using a model of planned comparisons for our specific hypothesis regarding EFC and DE and obtained the same results. For an additional illustration of our findings, we divided the participants into two groups based on their EFC scores (high EFC for those above the median score and low EFC for those below the median score). Participants with low EFC score increased their true sleep time by 33 min (from the low- to the high-stress period), whereas true sleep time of participants with high EFC scores decreased by 23 min, respectively.
The subjective reports also yielded a significant Period � EFC interaction effect for the perceived sleep quality ratings of the
participants, F(1, 32) � 9.95, p � .005. Individuals with high EFC reported decreased perceived sleep quality during the high-stress period in comparison with the low-stress period, whereas the opposite was true for those with low EFC. Furthermore, a signif- icant correlation was found between the change in perceived sleep quality (from the low-stress to the high-stress period) and the change in level of stress between these periods, r � .45, p � .01. Increase in stress level was associated with significant decrease in perceived sleep quality.
Discussion
This study is, to the best of our knowledge, the first to assess, in a prospective manner, how individual coping styles moderate the response of the sleep–wake system to significant stress. The nat- uralistic design of the study enabled documenting sleep using objective measures in the regular sleep environment and subjective stress response using daily subjective ratings. However, the limi- tations of our study should also be emphasized. Our study relied on measuring stress and sleep during one low-stress period (regular school week) and one high-stress period (screening week), and some order effects cannot be ruled out. Although attempts were made to contact the participants to ensure completion of the daily reports, the compliance with the protocol of the daily reports could not be determined. In addition, the use of multivariate analysis and a number of different measures with a relatively small sample is also a limitation that should be noted, as it might have compro- mised the reliability of the findings. The significant increase in the stress level documented by the subjective daily stress ratings indicated that the period of evaluation for the clinical psychology graduate program was indeed a stressful period in comparison with the low-stress week that preceded it by 2–3 months.
In line with our hypotheses, changes in sleep were significantly moderated by the individual’s coping style. Individuals with high EFC shortened their sleep, whereas those with low EFC extended their sleep during the high-stress period. Furthermore, individuals with low EFC improved their perceived sleep quality during the high-stress period, whereas the opposite was true for those with high EFC. These results support earlier findings linking shortened sleep and insomnia with high EFC (Hicks et al., 1991; Morin et al., 2003). Thus, shortened sleep may reflect ineffective coping. This result is consistent with other findings suggesting that EFC is
Table 1 Multivariate Analysis Based on GLM Analysis: F Values and Significance of Effects (With Bonferroni Correction for Multiple Compariso
PLACE THIS ORDER OR A SIMILAR ORDER WITH NURSING TERM PAPERS TODAY AND GET AN AMAZING DISCOUNT
