The relationship between trait hostility and sleep quality remains underexplored despite the empirical indications that individuals with increased hostility experience more stress, a factor known to degrade sleep quality [
Pronounced stress reactions are characteristic of individuals who score highly on measures of trait hostility [
Numerous studies [
Stress is perhaps, the most studied psychosocial precipitant of sleep disturbance. Current models describe stress as a process with four basic aspects: the stress stimuli, perceived stress, the stress response (physiological, affective, and cognitive), and feedback from the stress response [
Observational and experimental studies link stress exposure to sleep quality, and it is widely acknowledged that stress plays a causal role in initiating sleep disruption and the onset of disordered sleeping [
Morin et al. [
Individuals who score high on measures of trait hostility share a striking number of characteristics with poor sleepers, such as pronounced reactions to stress, increased negative affect, and ruminative tendencies which can prolong arousal following stress [
The psychophysiological reactivity model hypothesizes that hostile individuals experience anger more intensely and for longer periods than controls, causing more sustained and more intense activation of the sympathetic nervous system [
To date, only one study has simultaneously examined sleep quality, hostility, and stress. Brissette and Cohen [
Models of sleep disturbance emphasize the importance of individual differences in stress responding [
A number of methodological improvements in the current study will extend previous work in this area. Specifically, this study utilizes multiple subscales of the Cook-Medley Hostility Scale [
A convenience sample of 73 undergraduate psychology students (26 men, 47 women) aged from 17 to 25 years
This survey was created for the purposes of this investigation. It includes questions about gender, ethnic identification, living arrangements, a variety of health questions, and other health variables known to influence cardiovascular function and arousal.
The PSQI is one of the most widely used measures of subjective sleep quality. The PSQI is a retrospective, self-report inventory that asks participants to report on their subjectively experienced sleep quality and disturbances over the last month [
The PSS is a retrospective, self report measure of stress appraisal that asks about participants perceptions of stress over the previous month [
The CMHo Scale is a widely used 50 item self-report measure of trait hostility [
The SDC, a 21 item self-report log, was developed by Levin and Fireman [
Prospective daily sleep quality (DSQ) was measured by asking participants “What was the quality of your sleep last night?” Participants responded using a 9-point Likert scale. All daily ratings were averaged across the two-week reporting period for use in analyses.
The DSI is a prospective, daily measure of the individualized impact of relatively minor stress events [
To decrease the daily time commitment for participants, 19 items were eliminated that were likely to have a low occurrence among an undergraduate sample. The final measure included 39 events and two blank spaces for additional write-in events. The measure outcomes include the number of events that are endorsed as having occurred (stress events per day) and the sum total of the impact rating of these events (stress per day). These were used to calculate the average impact rating an individual endorses when experiencing a stress event (stress per event), our prospective estimate of perceived stress.
The current study’s design is correlation based and employed both cross-sectional and prospective measures [
Data analyses were completing using a popular statistical software package. Initial analyses focused on establishing the psychometric properties of the collected data set. Internal consistency for measures was calculated using Cronbach’s alpha. Variables were also examined to ensure they met assumptions of any statistical tests for which they were utilized. A series of Pearson’s correlations describe the relationship between retrospective and prospective measures of sleep and stress to establish convergent validity. Convergent validity was also examined by replicating previously found associations between relevant variables. Examples include a positive association between cognitive hostility and some stress variables and negative associations between stress variables and sleep quality. A series of Pearson’s correlations and regression analyses were used to test the studies main hypotheses. All mediation analyses were conducted using the process outlined in Baron and Kenny [
The relationship between trait hostility and sleep quality was initially investigated using a series of Pearson’s product-moment correlation coefficients. Cognitive hostility (as measured by the cynicism and hostile attribution subscales of the CMHo Scale) and behavioral hostility (as measured by the Aggressive Responding subscale) were correlated with the sleep quality component scores from the PSQI and the SDC.
Significant negative correlations were found between cognitive hostility and both measures of sleep quality such that increased hostility related to decreases in both retrospective sleep quality
The relationship between trait hostility and the three stress variables was investigated using Pearson’s correlations. The cognitive and behavioral subscales of the Cook-Medley Hostility Scale were compared with a measure of retrospective perceived stress, the PSS, as well as the primary measures from the DSI: stress impact per event and stress events per day. There was a positive correlation between trait cognitive hostility and prospectively measured perceived stress,
The sleep quality and stress variables were also compared using a series of Pearson’s correlations. There was a strong negative correlation between the PSQI and PSS
Prospective measures of sleep were also associated with levels of daily and retrospective perceived stress such that increased stress experience predicted poorer sleep. DSQ was strongly associated with the average severity ratings per stress event
To test our hypothesis that perceived stress mediates the relationship between trait hostility and self-reported sleep quality, a series of regression analyses were run using the procedure outlined by Baron and Kenny [
The first model examined whether the relationship between cognitive hostility and DSQ was mediated by the tendency to rate daily stress events as more severe. As can be seen in Table
The pearson correlations.
Variable | PSQI | DSQ | CogHo | BHo | PSS |
---|---|---|---|---|---|
PSQI | |||||
DSQ | .282* | ||||
CogHo | −.421** | −.260* | |||
BHo | −.050 | −.042 | .208** | ||
PSS | −.584*** | −.515*** | .352* | −.104 | |
SPE | −.432*** | −.517*** | .321* | −.025 | .613*** |
*
Cognitive hostility and daily SQ mediated by stress per event.
Step | IV | DV |
|
SE |
|
|
Adj |
sr2 |
---|---|---|---|---|---|---|---|---|
1 | CogHo | DSQ | −.092 | .043 | −.290 | .068 | .053 | .053* |
2 | CogHo | SPE | .092 | .034 | .321 | .103 | .089 | .089* |
3 | CogHo | DSQ | −.037 | .040 | −.105 | .277 | .254 | .010 |
SPE | DSQ | −.596 | .140 | −.483 | .277 | .254 | .210* | |
Standardized indirect effect = −.0550, |
||||||||
95% CI = −.108; −.013 |
*
The second mediation analysis examined whether the relationship between cognitive hostility and retrospective sleep quality (PSQI) is mediated by retrospective perceived stress. As seen in Table
Cognitive hostility and the PSQI mediated by PSS.
Step | IV | DV |
|
SE |
|
|
Adj |
sr2 |
---|---|---|---|---|---|---|---|---|
1 | CogHo | PSQI | .388 | .105 | .421 | .177 | .164 | .177* |
2 | CogHo | PSS | .718 | .239 | .352 | .124 | .110 | .124* |
3 | CogHo | PSQI | .227 | .097 | .246 | .393 | .374 | .053* |
PSS | PSQI | .225 | .047 | .497 | .393 | .374 | .216* | |
Standardized indirect effect = .1613, |
||||||||
95% CI = .072; .279 |
PSS: Perceived Stress Scale; PSQI: Pittsburg Sleep Quality Index, CogHo: cognitive hostility.
*
The present study found that increased trait hostility is associated with decreased retrospective and prospectively measured sleep quality and that this relationship is significantly mediated by one’s response to stress. Significantly, only the cognitive component of hostility was associated with heightened stress and sleep quality. This is consistent with previous studies which found differential associations between the components of hostility and both stress and sleep. For instance, Wilkinson [
Individuals who scored highly on cognitive hostility reported more daily and retrospective perceived stress compared to participants who scored low on the subscale. Importantly, while hostility was unrelated to stress event frequency, high hostility subjects reported heightened reactivity to their stressors and rated their stress as more severe than low hostile participants. This is consistent with Williams et al. [
Our study replicated previous findings associating increased stress with poor sleep quality [
The subjective measures of sleep quality had a consistently strong, negative association with perceived stress. These data support earlier findings by Morin et al. [
The studies primary hypothesis was that increased reactivity to stress would account for a significant portion of the relationship between hostility and sleep quality. Consistent with the hypothesis, mediation analyses indicated that hostility is related to sleep primarily via increases in stress experience. This finding was true for both prospective and retrospectively measured sleep quality and perceived stress and indicates that trait hostility is a risk factor for stress-related sleep disruption.
This hypothesis was primarily based on two theories. The first theory, the psychophysiological reactivity model, asserts that individuals high in hostility are cognitively and somatically hyperreactive to stress [
One possible explanation for these findings is that individuals high in cognitive hostility attend to and ruminate more on their internal responses to stress. The cognitive model of insomnia prioritizes this type of repetitive, negatively toned cognitive activity as a major pathway through which sleep can be disrupted [
While the behavioral component of trait hostility was unrelated to sleep quality, cognitive hostility was associated with subjective sleep quality. Previous studies have found similar associations [
Overall, the findings support our hypothesis that heightened trait hostility acts as a risk factor for poor sleep. However, behavioral hostility was not associated with any measure of sleep quality. Previous studies exploring aggression and sleep have generally reported a negative association between these two variables [
Causality cannot be directly addressed in the present study. The question of whether increased hostility leads to poor sleep, is caused by poor sleep, or some combination of the two will have to be resolved using a design appropriate for establishing causality. For instance, future investigators may attempt to manipulate levels of hostility and measure any subsequent changes in sleep quality. Studies already exist in which interventions targeting sleep quality impact variables associated with the hostility construct, such as aggression [
An interesting possibility is that the relationship between hostility and sleep is reciprocal. Trait hostility may actively degrade sleep via increased arousal as we suspect, and poor sleep may exacerbate hostile responding and stress responses. Conversely, good sleep might serve to diminish hostility, even in individuals who are high in trait hostility. Good sleep may therefore minimize the likelihood of negative outcomes associated with high trait hostility, such as coronary heart disease. Additionally, if arousal proves to be an important mechanism through which cognitive hostility impacts sleep and degrades health, then there are multiple points at which that process might be disrupted through intervention. One could actively target cognitive hostility or perceived stress through counseling.
A number of methodological and design issues in the present study suggest caution in interpreting our findings. The current study utilized self-report measures of hostility, stress, and sleep, which raises concerns of biased responding and shared method variance. Objective measures of sleep, hostility, and stress would be valuable supplements to any self-report instruments utilized in future studies. Additionally, the current study had a relatively small sample size consisting primarily of young, Caucasian college students. Future studies in this area should utilize broader samples. Last, we did not directly measure presleep arousal or utilize a design that allowed us to establish directionality. Important extensions of any subsequent studies will be to explicitly test for presleep arousal levels, establish causality, and look for the presence of proposed causal mechanisms such as rumination and possible moderators such as coping style.
Our findings suggest a number of fruitful avenues for clinical intervention. Interventions targeting at reducing hostile cognitions and behaviors thought to exacerbate stress responding in this group, such as rumination, might diminish the impact of hostility on sleep quality and health. Conversely, behavioral sleep interventions are generally low risk and highly effective. If poor sleep quality does exacerbate hostility, improving sleep quality might help to minimize some of the negative social and health effects trait hostility has been linked to. In conclusion, the current study provides evidence that increased trait cognitive hostility is associated with poorer subjective sleep via increases in perceived stress. Additional work is required to address issues of causality and directionality, as there are possible implications for the treatment and health of both poor sleepers and individuals high in hostility.