Cognitive arousal mediates the relationship between perceived ostracism and sleep quality but it is not moderated by experiential avoidance

Previous research suggests that longer-term perceived ostracism is related to poor sleep quality. In this study, we investigated the mediating effect of cognitive arousal on the perceived ostracism – sleep quality relationship. We also investigated whether experiential avoidance was a moderator of the cognitive arousal – sleep quality relationship. Participants ( N = 251) were recruited through online research portals to take part in an online survey. A path analysis was used to test a moderated mediation effect between variables. It was found that cognitive arousal mediated the perceived ostracism – sleep quality relationship; however, experiential avoidance was not a significant moderator. These findings suggest that further research needs to be conducted to elucidate the mechanism of experiential avoidance to account for when it may impact sleep quality. Moreover, treatment interventions targeted at reducing cognitive arousal (e.g., cognitive behavioural therapy) prior to sleep are likely to bear some fruit.

Temporal Need Threat Model (TNTM), humans evolved a reflexive ostracism-detection mechanism to immediately react to perceived threats of rejection. Indeed, if one were excluded from the group as a hunter-gatherer in the past, this would risk death from predation or starvation. Therefore, according to the TNTM, when ostracized, people will rapidly become distressed as their fundamental psychological needs (i.e., belonging, control, self-esteem, meaningful existence) are depleted (i.e., the reflexive stage). Several experimental studies have replicated the consistent finding that ostracism can increase psychological distress as a function of depleting fundamental needs (e.g., Wesselmann, Wirth, Mroczek, & Williams, 2012;Williams, Cheung, & Choi, 2000;Zadro, Boland, & Richardson, 2006;Zadro, Williams, & Richardson, 2004).
However, after a period of reflection (e.g., up to 45 min in experimental studies), individuals typically recover quickly from such threats following appraisal (i.e., the reflective stage). Williams (2009) purports that individual differences occur within the reflective stage which can influence how quickly one copes with their ostracism. Several moderators of the short-term effects of ostracism have been reported in the literature including, but not limited to: social anxiety (Zadro et al., 2006), attachment style (Hermann, Skulborstad, & Wirth, 2014;Yaakobi & Williams, 2016), temporal perspective (Garczynski & Brown, 2014) and self-construal .
However, of relevance to the present study, these moderators have not been investigated in the context of impact of ostracism on sleep.
According to Williams (2009), when individuals fail to cope with their ostracism, or their ostracism persists (i.e., the longer-term effects), they then enter the chronic (resignation) stage whereby they become resigned to their ostracism and experienced prolonged psychological distress (e.g., depression). Research has shown that longerterm ostracism is associated with high levels of psychological distress (e.g., Tyndall, Waldeck, Riva, Wesselmann, & Pancani, 2018;Wu, Yim, Kwan, & Zhang, 2012). There is, however, little empirical research into the longer-term effects of ostracism, and importantly, the mediators of such effects.
Although the association between ostracism and general psychological distress has been well supported in the literature, there is limited research on how ostracism is related to sleeping quality, which is the focus of the present paper. As stress can predict sleeping problems (e.g., Linton, 2004), and being ostracized is a powerful and ubiquitous social stressor (Riva, Wesselmann, Wirth, Carter-Sowell, & Williams, 2014;Williams, 2009;Williamson, Thomas, Eisenberger, & Stanton, 2018), the experiences of such events may also impact upon sleep. According to the TNTM (Williams, 2009), individuals often ruminate and worry on their ostracism (i.e., they are cognitively aroused) which may delay sleep onset and impair sleep quality. Indeed, such perceived events can lead to temporary impairment in executive functioning (e.g., working memory, decision making) as people dwell on their ostracism (Buelow, Okdie, Brunell, & Trost, 2015). Moreover, when people ruminate on their ostracism, this has been found to lead to more frequent reports of psychological distress (Wesselmann, Ren, Swim, & Williams, 2013). Such increases in cognitive arousal (e.g., presleep arousal) are known contributors to poor sleeping quality (Clancy, Prestwich, Caperon, Tsipa, & O'Connor, 2020;Guastella & Moulds, 2007;Smith, Perlis, Smith, Giles, & Carmody, 2000;Tousigant, Taylor, Suvak, & Fireman, 2019).
To the best of our knowledge, there are only two empirical studies that examined the predictive relationship between ostracism and sleeping quality (Chen & Li, 2019;Pereira et al., 2013). Pereira et al. (2013) recruited 99 employees (from various industries) to participate in 2-week data collection of daily self-reported levels of perceived ostracism and a variety of sleep measures. It was found that daily workplace ostracism predicted sleep fragmentation in the following night. However, it was also shown that sleep-onset latency, sleep efficiency and self-reported sleep quality were unrelated to daily workplace ostracism (Pereira et al., 2013). Therefore, in the shortterm, ostracism appears to have mixed effects on sleeping behaviour.
Despite not being central to their hypotheses, Pereira et al. also tested the effects of chronic ostracism (i.e., longer-term self-reports) and the impact on sleeping quality. It was found that in contrast to the shortterm effects, longer-term ostracism was significantly associated with poorer sleeping quality. It must be noted, however, that Pereira

| Experiential avoidance
A promising potential moderator of sleep quality is that of experiential avoidance (Hayes, Strosahl, & Wilson, 1999). Experiential avoidance (a component of the acceptance and commitment therapy [ACT] model of psychological flexibility/inflexibility) is defined as behaviour that attempts to 'alter the frequency or form of unwanted private events, including thoughts, memories and bodily sensations, even when doing so causes personal harm' (Hayes, Pistorello, & Levin, 2012, p. 981). Hayes et al. (2012) posit that experiential avoidance is implicated in the onset and perpetuation of psychopathology as habitually avoiding uncomfortable private events can prevent people from working towards their living in accordance with their core values. For example, if a son refused to speak to their mother after they perceived that their mother was ignoring them, but valued a close relationship with her, then this may lead to temporary relief (e.g., distraction away from feelings of anger or sadness). Indeed, experiential avoidance may be adaptive in the short-term in coping with stressful events (Hayes et al., 2012). However, if the son valued a close relationship with their mother, and continued to not live in accordance with their values (e.g., avoiding going to family events despite wanting close connection with his family) this may cause prolonged psychological distress. Attempts to suppress one's unwanted thoughts (e.g., attempting to alter the frequency or form of one's thoughts associated with ostracism) has been demonstrated to paradoxically increase the frequency and intensity of those thoughts (Wenzlaff & Wegner, 2000). Moreover, researchers have consistently shown that experiential avoidance can be a moderator or mediator of psychological distress (e.g., Bardeen, 2015;Bardeen, Fergus, & Orcutt, 2013;Kashdan & Kane, 2011;Pickett, Bardeen, & Orcutt, 2011;Roush, Brown, Mitchell, & Cukrowicz, 2019;Spinhoven, Drost, de Rooij, van Hemert, & Pennix, 2014). Tyndall et al. (2018) found that experiential avoidance appeared to moderate the relationship between perceived ostracism and psychological distress. Specifically, at high levels of experiential avoidance, the relationship between ostracism and distress was significant; however, at low levels, the relationship was no longer significant.
Thus, at low levels of experiential avoidance, people appeared to cope with or recover quicker from events such as ostracism. Such findings support the ACT model of behaviour change (Hayes et al., 2012). ACT is purported to help promote psychological flexibility (A-Tjak et al., 2015;Hacker, Stone, & MacBeth, 2016;Lee, An, Levin, & Twohig, 2015;Levin et al., 2014) and reduce reliance on experiential avoidance (i.e., increasing emotional self-regulation), to facilitate clients to work towards living in accordance with their values (e.g., the son contacting his mother in spite of feeling angry and uncomfortable in order to attempt reconciliation).

| Experiential avoidance and sleep quality
There is little direct empirical evidence that has examined experiential avoidance and sleeping quality. However, some researchers have theorized the role of experiential avoidance (within the broader framework of the ACT model of psychological flexibility) and its impact on sleep quality within the context of those who suffer with chronic pain (e.g., Borthelius, Jernelov, Fredrikson, McCracken, & Kaldo, 2015;McCracken, Williams, & Tang, 2011), and college students with depression symptoms (Peltz, Rogge, Bodenlos, Kingery, & Pigeon, 2020). It seems that there is promise for applying ACT-based therapeutic interventions to increase psychological flexibility and reduce experiential avoidance in those who present with sleep difficulties (Daly-Eichenhardt, Scott, Howard-Jones, In a somewhat related vein, researchers have examined the effects of thought suppression (i.e., a common strategy when using experiential avoidance to avoid unwanted painful thoughts) on sleeping behaviour. For example, Harvey (2003) showed that people who were asked to suppress a self-selected (unwanted) thought before sleep, reported significantly poorer sleep quality and longer sleep-onset latency than those who were given nonsuppression instructions (i.e., asked to allow painful thoughts to come and go). This effect was observed for both people diagnosed with insomnia and those with no sleeping disorder. Similarly, people with insomnia or comorbid insomnia (e.g., also diagnosed with an anxiety disorder) tend to report that such thought control strategies are not effective in helping to increase sleeping quality in comparison to 'good sleepers' (Bélanger, Morin, Gendron, & Blais, 2005). Therefore, the use of experiential avoidance (e.g., thought suppression) may impair sleep quality (Daly-Eichenhardt et al., 2016) as attempts to avoid unwanted psychological events (e.g., thoughts, feelings, sensations relating to one's ostracism) could perpetuate one's distress. This seems particularly pertinent given the cardinal role that sleep plays in processing our emotions (Goldstein & Walker, 2014;Kahn, Sheppes, & Sadeh, 2013;Tempesta, Socci, De Gennaro, & Ferrara, 2018;Walker & van der Helm, 2009), and subsequent capacity to regulate our own emotions (Gujan et al., 2011;Mauss et al., 2013).

| A moderated mediation model
There is some evidence that experiential avoidance as a construct considerably overlaps with factors that influence cognitive arousal such as rumination (e.g., Schut & Boelen, 2017;Williams, 2008) and neuroticism (Rochefort, Baldwin, & Chmielewski, 2018). Therefore, to examine whether experiential avoidance would potentially moderate the relationship between cognitive arousal and quality of sleep we developed a path model utilizing a structural equation modelling procedure ( Figure 1). We proposed that the positive relationship between perceived ostracism and sleep quality would be mediated by cognitive arousal (Hypothesis 1).
Moreover, the cognitive arousal-sleep quality relationship would be moderated by experiential avoidance such that the relationship would be weaker for individuals with lower levels of experiential avoidance than for those with higher experiential avoidance (Hypothesis 2). We used a modified version of the 10-item WOS (Ferris et al., 2008; α = .0.94) on a scale from 1 (never) to 7 (always). Participants responded to statements about everyday ostracism episodes that occurred within the last 6 months without being restricted to a specific context (e.g., 'others ignored you' as opposed to 'others ignored you at work'). The modified WOS has been utilised as a measure of general perceived ostracism within the social exclusion literature (Waldeck, Tyndall, Riva, & Chmiel, 2017). An exploratory factor analysis revealed a unidimensional factor structure with the range of factor loadings .70-.86) similar in magnitude to the factor loadings reported by Ferris et al. (2008).
Sample items include, 'The key to a good life is never feeling any pain' and 'I would give up a lot not to feel bad'.

| RESULTS
All questions were subject to forced-choice responding. Therefore, there was no missing data within the dataset. Means, SD and Spearman's rho bivariate correlations are among the key variables are presented in Table 1. As indicated, perceived ostracism (rs = .39, p < .001), and experiential avoidance (rs = .34, p < .001) had a significant moderate positive correlation with sleeping quality.
A Harmon-one factor test was conducted to examine if common method variance was a concern for this study. It was found that a one factor model accounted for 24.8% of the variance which is lower than the 25% recommended by Williams, Cote, and Buckley (1989). Therefore, common method bias is unlikely to a pervasive problem for this study.   (Hooper, Coughlan, & Mullen, 2008;Kline, 2005).

| Path analysis findings
Ostracism yielded a statistically significant path to sleep quality (explaining 15% of the variance). The path coefficient from ostracism to cognitive arousal was statistically significant (explaining 39% of variance). Cognitive arousal yielded a statistically significant path to sleep quality (explaining 57% variance). Additional sensitivity analyses were conducted by exploring the effects of the model with known covariates. Age, gender and ethnicity were added to the model as covariates and did not reveal a significant association with the variables.

| DISCUSSION
The goal of the present study was to examine a proposed moderated mediation model whereby cognitive arousal may mediate the relationship between perceived ostracism and sleep quality, and experiential avoidance is a moderator of the effects of arousal. Consistent with prior literature (e.g., Chen & Li, 2019;Pereira et al., 2013), ostracism was found to be related to poor sleep quality. Moreover, we found that cognitive arousal appeared to mediate the perceived ostracismsleep quality relationship, which supports prior theory (TNTM; Williams, 2009) and literature to suggest that arousal (e.g., worry, rumination) can maintain the negative effects of ostracism (e.g., Wesselmann et al., 2013). Therefore, one tentative argument could be that some ostracized individuals have poor sleep quality as their needs are still thwarted, and as such, significant resources are devoted to self-regulate themselves (e.g., dwelling on the reasons behind one's ostracism). Future researchers may consider exploring the experience of threatened psychological needs further by using qualitative methods (e.g., interpretative phenomenological analysis [Smith & Shinebourne, 2012]) with ostracized individuals who suffer sleep disorders.
Contrary to our expectations, however, experiential avoidance was not a significant moderator of sleep quality. This was somewhat unexpected given the negative impact of engaging in experiential avoidance-like techniques (e.g., thought suppression; thought control; Bélanger et al., 2005;Harvey, 2003;Harvey, Tang, & Browning, 2005) on sleep. Similarly, despite not being central to our hypotheses, we also found that experiential avoidance was not a significant moderator of the perceived ostracism-sleep quality relationship. This was somewhat surprising considering Tyndall et al. (2018) reported that experiential avoidance moderated effects of perceived ostracism on psychological distress over time. Given that no specific measure of psychological distress (e.g., depression, anxiety, or stress) was administered in the present study we cannot, of course, conclude that experiential avoidance does not moderate psychological distress from perceived ostracism. Rather, within the context of sleep quality, level of self-reported experiential avoidance does not appear to significantly influence quality of sleep.
Experiential avoidance and rumination have been postulated to 'load on the same latent factor of perceived inability to control negative thinking' (Schut & Boelen, 2017, p. 11; see also Spinhoven, Drost, de Rooij, van Hemert, & Pennix, 2016). Similarly, Williams (2008) viewed rumination and experiential avoidance as similar modes of mind, or mode of mental processing, and thus it is a noteworthy finding that experiential avoidance did not predict or moderate difficulty Note: Root mean square error of approximation (RMSEA), goodness of fit index (GFI), adjusted GFI (AGFI) and comparative fit index (CFI) > 0.9 indicate adequate fit.
in falling asleep in the present study. It should be acknowledged that as the majority of research that found a link between experiential avoidance and psychological distress used the Acceptance and Action Questionnaire-II (AAQ-II; Bond et al., 2011), whereas we employed the BEAQ (Gámez et al., 2014) in the present study, it could be argued that it is not a faithful test of an association between experiential avoidance and quality of sleep. However, there is emerging consensus that the AAQ-II lacks discriminant validity as a measure of experiential avoidance (e.g., Rochefort et al., 2018;Tyndall et al., 2019;Wolgast, 2014) as distinct from psychological distress or negative emotionality, and thus the BEAQ served as a more focused measure of the construct.
Our study has some strengths to report. One strength is that this study, to the best of our knowledge, is the first to examine the role of experiential avoidance as a moderator of sleeping behaviour, as well as accounting for the predictive role of perceived ostracism. Second, we used validated measures of perceived ostracism and sleeping quality as recommended by Pereira et al. (2013). However, it could be noted that as we have modified the WOS (Ferris et al., 2008), this potentially could be considered to be a new measure, and thus unvalidated. The modified WOS was found to be reliable and held to a single factor structure within the context of this study. Furthermore, the modified WOS (to measure general perceived ostracism) has been utilised in previous studies within the ostracism literature (e.g., Tyndall et al., 2018;Waldeck et al., 2017).
Third, we obtained an adequate sample size for the purpose of this exploratory study as it was higher than the acceptable sample to parameter ratio (10:1) in a path-analysis (Jackson, 2003).
However, our study has some limitations that should be noted.
First, as all data came from the same source at one timepoint, this increases the risk of common method bias. However, we do not believe this to be a pervasive problem as the Harman one-factor test yielded a relatively low percentage of variance explained for a common method factor. Second, there were potential confounding and/or moderating variables which were not accounted for in this study (e.g., participants with mental health diagnoses, sleep disorders). Indeed, some individuals (e.g., those diagnosed with social anxiety disorder; Zadro et al., 2006) are more likely to be more sensitive to and catastrophise their ostracism than others. We therefore propose that future researchers control for

| CONCLUSION
The present study supported the findings in the literature that perceived ostracism is a predictor of poor sleep quality (Chen & Li, 2019;Pereira et al., 2013). Moreover, we found that cognitive arousal is a mediator of this relationship. However, experiential avoidance was not a moderator of sleep quality despite a large body of literature highlighting key moderating and mediating effects of experiential avoidance in self-reported distress. Further research needs to be conducted to explore the mechanism of experiential avoidance in the context of sleep, given a focus in some studies on increasing psychological flexibility and acceptance in chronic pain patients concerning their sleep difficulties (Daly-Eichenhardt et al., 2016;McCracken et al., 2011), and college students with depressive symptoms (Peltz et al., 2020), as the present data suggests that interventions designed to reduce cognitive arousal prior to sleep might be most promising.
Indeed, it is perhaps unsurprising that cognitive behavioural therapy (e.g., CBT-i for insomnia) is consistently regarded in the literature as the gold-standard nonpharmocological treatment for sleep difficulties (Koffel, Koffel, & Gehrman, 2015;Murtagh & Greenwood, 1995;Van Straten et al., 2018), with particular focus on intervention techniques such as cognitive restructuring designed to target processes such as presleep cognitive arousal (Bélanger et al., 2005;Lancee et al., 2019).
Through a specific focus on process-based therapy (Hofmann & Hayes, 2019), future researchers can more systematically identify the key mechanisms that mediate sleep difficulties to serve as targets for therapeutic intervention in either CBT-or ACT-based treatment packages.

CONFLICT OF INTEREST
The authors declare no potential conflict of interest.

DATA AVAILABILITY STATEMENT
The dataset for this paper is not part of a larger dataset. The dataset for this paper is archived on the Open Science Framework Repository via the following link: https://osf.io/tvxsy/?view_only=ae4713acbbf14d2 6a75ef6148ec30a3e.

Ian Tyndall
https://orcid.org/0000-0003-2108-9203 ENDNOTE 1 We ran an additional analysis to examine whether experiential avoidance was a moderator of the relationship between perceived ostracism and sleeping quality. Similarly, the results showed a non-significant path for experiential avoidance as a moderator (β = 0.04, p = .46).