Perceived Stress, but Not Rumination, Mediates the Relationship Between Trait Mindfulness and Sleep Quality in Young Adults

Introduction

Poor sleep among the general population is prevalent and rising.^1,2 Notably, up to 50% of young people rate their sleep quality as poor, and a majority are chronically sleep deprived.³ The negative mental health outcomes of poor sleep have been consistently documented, including links to anxiety, mood disorders, substance use, impulsivity, and aggression,^3–6 alongside links to physical health risks such as lower immunity, cancer, cardiovascular disease, and type 2 diabetes.^7,8 Current National Institute for Health and Care Excellence (NICE) guidelines recommend face-to-face Cognitive Behavioral Therapy for Insomnia (CBT-I) as first-line treatment for insomnia.⁹ Yet, this is currently not readily available in the UK, and there is no direct evidence comparing alternative online self-help programs eg Sleepio with face-to-face CBT-I.⁹ Additionally, CBT-I has a non-response rate of up to 40% and a high remission rate.¹⁰ Alternatively, short-term pharmacological treatment for insomnia is advised;¹¹ short-term effects are comparable to CBT-I, and combined, the two treatments have increased efficacy.¹² However, pharmacological treatment comes with disadvantages including a variety of side effects and risk of dependency.^9,12 Furthermore, insomnia typically returns following withdrawal of pharmacological treatment and a meta-analysis found that the effects of long-term pharmacological treatment may decrease over time.¹² Hence, there is growing interest in alternative therapies for reducing sleep problems, such as mindfulness-based interventions (MBIs), which have shown efficacy comparable to CBT-I with high adherence, and no adverse side effects.^13,14

Higher trait mindfulness – an individual’s innate tendency to purposefully maintain awareness of the present moment in an open-minded and non-evaluative manner¹⁵ – has been linked to a range of positive physical and mental health outcomes including better sleep in a recent meta-analysis.¹⁶ There is also evidence that sleep quality serves as a mediator in the relationship between trait mindfulness and better wellbeing, and other research has identified sleep quality as a mediator in the relationship between trait mindfulness and physical health.^17,18 These findings highlight the important role of sleep quality in explaining the influence of trait mindfulness on both mental and physical health; however, the relationship between sleep quality and trait mindfulness is not well understood. Cross-sectional^19–21 and experimental²² studies investigating the relationship between trait mindfulness and sleep quality have identified psychological distress, eg perceived stress, as a mediating variable. However, this has only been investigated using subjective sleep measures, and not in young adults specifically. There is evidence indicating that subjective and objective sleep measures do not always correlate, perhaps due to underreported sleep duration and overestimated sleep latency.^23,24 Given the prevalence and implications of poor sleep among young adults in particular, the relationship between trait mindfulness and sleep quality, and the potential mediating variables, requires further attention.

Factor analyses have led to the conceptualization of trait mindfulness as comprising five facets; mindfulness is the act of observing or attending to internal and external experiences and describing these with awareness of the present moment. This is done with non-judgement of thoughts and feelings and non-reactivity; allowing these to come and go.²⁵ The Five Facets of Mindfulness Questionnaire (FFMQ) was designed to measure mindfulness at the facet level and is widely used.^25,26 In adults with insomnia, higher levels of awareness have been linked to less severe symptoms of insomnia, less sleep disturbance, and fewer dysfunctional beliefs about sleep; non-judging was also associated with fewer dysfunctional beliefs about sleep, and both non-judging and non-reactivity were linked to less sleep difficulty.²⁷ Likewise, Xie et al found that awareness, non-judging, and non-reactivity were associated with better sleep health.²⁸ These findings suggest that awareness, non-judging, and non-reactivity play key roles in influencing sleep quality. It is proposed that awareness is necessary to allow modification of thoughts.²⁹ Ong et al’s Meta-cognitive Model of Insomnia posits that primary arousal, involving negative beliefs and attitudes about sleep, directly inhibits sleep.³⁰ The theory describes an iterative process whereby secondary arousal, including awareness of negative beliefs and attitudes, and attachment to them, further precipitates negative cognition, ultimately resulting in further sleep difficulties. Acceptance (non-judging and non-reactivity) deactivates hyperarousal, shifting meta-cognition to a more objective and adaptive state, thus reducing sleep difficulties.

On the other hand, it is well established that high levels of perceived stress negatively impact sleep quality.^31–33 Perceived stress refers to the extent to which an individual finds their life unpredictable, uncontrollable, and overloaded.³⁴ Extensive cross-sectional^35,36 and longitudinal³⁷ evidence demonstrates a relationship between perceived stress and poor sleep quality, using a variety of self-report and objective measures.^38–41 Perceived stress has also been associated with lower levels of trait mindfulness.⁴² It is thought that mindfulness mitigates stress appraisals and stress reactivity responses, in turn enhancing sleep quality, and other health outcomes.^43–45 This may explain why perceived stress appears to mediate the relationship between trait mindfulness and sleep quality,^19–21 and provides a rationale for exploring this in more detail, and with inclusion of objective sleep measures.

There is also considerable evidence that rumination, defined by Trapnell & Campbell as “a self-attentiveness motivated by perceived threats, losses or injustices to the self”⁴⁶ negatively impacts sleep quality due to heightened cognitive and physiological activity prior to sleep onset.^47–49 This has been attributed to the negatively toned cognitive arousal associated with rumination, eg repeatedly contemplating the causes and outcomes of a stressful event after it has occurred, with negative self-focus.^50,51 Like perceived stress, previous studies have shown that rumination impacts sleep quality based on both subjective,⁵² and objective measures,^53,54 affecting sleep latency,⁵⁵ disturbance,⁵⁶ duration,⁵⁷ and efficiency,⁵⁸ consequently increasing daytime dysfunction.⁵⁹ The opposing roles of mindfulness and rumination are not surprising since these are theoretically opposing constructs, which are negatively correlated.⁶⁰ The Monitor and Acceptance Theory suggests that mindfulness may counteract rumination: attention monitoring (observing), alongside acceptance (non-judging and non-reactivity) can inhibit the negative emotional and cognitive mechanisms linked to rumination.⁶¹ Given the interrelations between trait mindfulness, perceived stress, and rumination, this provides a rationale for exploring rumination as another possible mediator in the relationship between trait mindfulness and sleep quality. Rumination has been found to mediate the relationship between mindfulness and poor sleep quality in a previous study.⁶² The present study is the first to consider both rumination and perceived stress as potential mediators in the relationship between trait mindfulness and sleep quality.

Present Study

Within the literature, there is compelling evidence suggesting that mindfulness, perceived stress, and rumination each play a role in sleep quality; however, there are limited studies looking into the interaction between these variables and the impact on sleep quality. To date, no study has investigated the mediating effects of both perceived stress and rumination on the relationship between trait mindfulness and sleep quality in young adults, nor has this been investigated with the use of objective sleep measures.

Thus, we sought to add mechanistic detail to the literature. Data were collected from a sample of UK-based healthy young adults. Both subjective (the Pittsburgh Sleep Quality Index [PSQI]⁶³ and the Consensus Sleep Diary)⁶⁴ and objective (wrist-worn actigraphy) measures of sleep quality were employed.

It is hypothesized that i) higher self-reported mindfulness will be associated with lower self-reported perceived stress, lower self-reported rumination, and better subjective sleep quality (measured by the PSQI); ii) higher self-reported rumination will be associated with higher self-reported perceived stress, and iii) both self-reported perceived stress and rumination will be associated with poorer subjective sleep quality (measured by the PSQI). We further hypothesized that iv) perceived stress and rumination will mediate the relationship between trait mindfulness and subjective sleep quality (measured by the PSQI). In a secondary sample, we explored whether these hypotheses held when sleep quality was assessed using sleep diary- and actigraphy-based methods.

Method

Participants

Data were obtained from 170 participants, aged 18–37 (M = 20.8, SD = 2.9); 145 self-identified as females, 23 self-identified as males, and 2 self-identified as other (see Table 1 for demographics and Supplementary Table 1 for further details). The dataset comprised 2 subsamples: the primary sample (N = 140) who completed a subjective sleep questionnaire only, and the secondary sample (N = 30) who completed sleep diaries and actigraphy measures in addition to the measures completed by the primary sample. An a priori G*Power analysis with a significance criterion of α = 0.05 and power = 0.80 indicated an ideal sample of 192 participants for a medium effect size (f² = 0.0638) in a mediation analysis with 4 predictors. Participants were recruited through volunteer sampling via email and flyers distributed across the University of Surrey campus. Inclusion criteria were age range (18–40 years old), and the secondary sample excluded those who had practiced mindfulness regularly before.

Table 1 Demographics of All Participants (N = 170)

Measures

Subjective sleep quality was assessed through the PSQI⁶³ in both samples, and in the secondary sample through the Consensus Sleep Diary;⁶⁴ objective sleep quality was measured in the secondary sample through wrist-worn actigraphy.

The PSQI is a 19-item self-report questionnaire that retrospectively assesses seven components of subjective sleep (perceived quality, latency, duration, efficiency, disturbance, medication, and daytime dysfunction) over the past month. A global sleep quality score is calculated which ranges from 0 to 21; scores above 5 indicate poor sleep quality.⁶³ High internal validity for this measure has previously been reported using Cronbach’s alpha (α = 0.83).⁶³

The Consensus Sleep Diary captures subjective sleep quality through bedtime, time tried to go to sleep, sleep latency (time taken to fall asleep, minutes), number and length of awakenings (minutes), wake time, rise time and self-rated sleep quality.⁶⁴ Sleep quality is rated on a 5-point Likert scale, 0 (very good), 1 (good), 2 (fair), 3 (poor), and 4 (very poor). The Consensus Sleep Diary was completed directly after awakening for seven days.

Participants in the secondary sample wore a tri-axial GENEActiv actigraphy device (Activinsights Ltd., Kimbolton, England) for seven days. The devices were placed on the non-dominant wrist. The data were summarized into 30-second epochs and analyzed using the GENEActiv Software based on gross motor movement, body temperature, and light exposure, to estimate total elapsed sleep time (time between falling asleep and waking, minutes), sleep duration (actual time spent asleep, minutes), number of activity periods after sleep onset, median activity period (minutes), and sleep efficiency (%).

Mindfulness was assessed through the 24-item short form of the Five Facets of Mindfulness Questionnaire (FFMQ-SF) which measures five facets of mindfulness (observing, describing, acting with awareness, non-judging of inner experience, and non-reactivity to inner experience) on a 5-point Likert scale, 1 (never or very rarely true), 2 (rarely true), 3 (sometimes true), 4 (often true), and 5 (very often or always true).⁶⁵ All five facets have good internal validity with Cronbach’s alpha coefficients ranging from 0.73 to 0.91.⁶⁵ Respondents are scored on each facet as well as a total mindfulness score that ranges from 24 to 120; high scores indicate high levels of mindfulness.

Perceived stress was measured through the Perceived Stress Scale (PSS), which is a widely used 10-item measure of stress perception with high internal validity (α = 0.78).³⁴ The measure uses a 5-point Likert scale, 0 (never), 1 (almost never), 2 (sometimes), 3 (fairly often), and 4 (very often), to assess perceived stress levels over the last month. This is based on a total score ranging from 0 to 40 with high scores indicating high levels of perceived stress.

The Rumination-Reflection Questionnaire (RRQ) is a 21-item measure of rumination and reflection. Only the rumination sub-scale was utilized in this study, which has exceptionally high internal validity with a Cronbach’s alpha coefficient of 0.90.⁴⁶ Items are scored on a 5-point Likert scale, 1 (strongly disagree), 2 (disagree), 3 (neutral), 4 (agree), and 5 (strongly agree), and averaged; high scores indicate high levels of rumination.

Procedure

Prior to data collection, favorable ethical opinion was obtained from the London-Surrey NHS Research Ethics Committee and the University of Surrey Ethics Committee. All participants were provided with information sheets outlining the aims of the study, and informed consent was acquired prior to data collection. All participants completed the PSQI, FFMQ-SF, PSS, and RRQ via a Qualtrics online survey. Participants in the primary sample were able to access the survey through a link in the recruitment email, while those in the secondary sample visited the Research Centre to complete the online survey and collect a GENEActiv watch and a copy of the Consensus Sleep Diary, along with instructions. All participants were debriefed upon completion of the study.

Data Analysis Plan

Data were analyzed using IBM SPSS version 27; a p value of <0.05 was considered statistically significant. There were no outliers within the dataset, with all z-scores falling within the cut-off points of ±3.29.⁶⁶ Questionnaire measures were completed in full by all participants and for analyses of questionnaire measures the two samples were combined (N = 170). Pearson’s correlations were used to assess relationships between the questionnaire measures as all z-scores for skewness and kurtosis were within the cut-off points of ±3.29.⁶⁷ From the secondary sample (N = 30), only 23 participants provided seven valid days of actigraphy data, and only 21 participants provided complete sleep diaries; incomplete responses were excluded from analyses. To assess relationships between diary- and actigraphy-based sleep quality and the other measures, both Spearman’s and Pearson’s correlations were used, as diary-based sleep latency, efficiency and self-rated quality, and actigraphy-based efficiency were non-normally distributed (see Supplementary Table 2). Bonferroni correction was used to control for family-wise error in multiple correlations. Mediations were assessed using model four in PROCESS version 3.5 for SPSS. The mediation model is depicted in Figure 1. Assumptions of accuracy, normality, linearity, multicollinearity, homogeneity, and homoscedasticity were met. The indirect effects and 95% confidence intervals were calculated based on 5000 bootstrap samples; effects were considered significant if confidence intervals did not overlap with zero. Mediations based on diary- and actigraphy-based sleep quality could not be assessed due to a lack of significant correlations with trait mindfulness and perceived stress.

Figure 1 Hypothesized mediation between trait mindfulness, sleep quality, perceived stress, and rumination.

Results

Data Cleaning

From the sleep diaries, sleep duration and efficiency were calculated (see Reed & Sacco for equation);⁶⁸ seven-day averages were created for latency, duration, number of awakenings, efficiency, and self-rated quality. Bedtime and wake time reported in the sleep diaries were used to correct the estimated times from the actigraphy devices when the disparity was more than 30 minutes, subsequently adjusting actigraphy-based elapsed sleep time, duration, and efficiency. Actigraphy data for duration, number of activity periods and efficiency were averaged across the seven nights.

Descriptive Statistics

For participant characteristics, see Table 1 (further details in Supplementary Table 1).

Overall, only 64.71% of participants scored <5 out of 21 (indicating good sleep quality) on the PSQI, see Table 2. In the secondary sample, who completed sleep diaries, 70.8% of participants self-rated their sleep quality as fair or poor (averaged across the 7 days), despite the average reported sleep duration being nearly 7.5 hours and sleep efficiency averaging >85%. However, the average objective sleep duration and efficiency from actigraphy was lower, indicating that participants overestimated their actual sleep time and efficiency. See Supplementary Tables 3–5 for minimums, maximums, means, standard deviations, medians, interquartile ranges, and gender differences for all measures.

Table 2 Minimums, Maximum, Means, and Standard Deviations (SD) for All Self-Report Measures

Main Analysis

Pearson’s correlation analyses showed that trait mindfulness was negatively correlated with perceived stress, r(168) = −0.63, p < 0.001, rumination, r(168) = −0.55, p < 0.001, and global PSQI scores, r(168) = −0.50, p < 0.001. There was a moderate positive association between perceived stress and rumination, r(168) = 0.55, p < 0.001. Both perceived stress, r(168) = 0.62, p < 0.001, and rumination, r(168) = 0.41, p < 0.001, were positively associated with global PSQI scores. All six comparisons remained significant at a Bonferroni’s adjusted alpha level of 0.008.

Facets of Mindfulness and Perceived Stress

Non-judging, r(168) = −0.45, p < 0.001, non-reactivity, r(168) = −0.48, p < 0.001, awareness, r(168) = −0.46, p < 0.001, and describing, r(168) = −0.47, p < 0.001, were significantly correlated with perceived stress in the expected negative direction, at a Bonferroni’s adjusted alpha level of 0.008. However, observing was not significantly correlated with perceived stress.

Facets of Mindfulness and Rumination

Non-judging, r(168) = −0.48, p < 0.001, non-reactivity, r(168) = −0.54, p < 0.001, awareness, r(168) = −0.44, p < 0.001, and describing, r(168) = −0.37, p < 0.001, were negatively correlated with rumination. However, observing was positively correlated with rumination, r(168) = 0.22, p = 0.005. All correlations were significant with a Bonferroni’s adjusted alpha level of 0.008.

Facets of Mindfulness and Sleep Quality

Total PSQI scores were negatively correlated with non-judging, r(168) = −0.37, p < 0.001, non-reactivity, r(168) = −0.36, p < 0.001, awareness, r(168) = −0.39, p < 0.001, and describing, r(168) = −0.39, p < 0.001. All relationships remained significant at a Bonferroni’s adjusted alpha level of 0.008.

Mediation Analysis

A mediation analyzing the effect of trait mindfulness on sleep quality, measured by the PSQI, with perceived stress and rumination as mediators and age and gender as covariates was conducted (Figure 1). High levels of mindfulness were negatively associated with perceived stress, b = −0.37, t(166) = −10.50, p < 0.001, and high perceived stress was positively related to poor subjective sleep quality, b = 0.24, t(164) = 5.98, p < 0.001. High levels of mindfulness were also negatively associated with rumination, b = −0.03, t(166) = −8.83, p < 0.001; however, rumination was not significantly associated with subjective sleep quality, b = 0.19, t(164) = 0.46, p = 0.64. The indirect effect through perceived stress was significant, effect size = 0.08, 95% CI [−0.12, −0.06], standardized coefficient β = −0.30, 95% CI [−0.41, −0.21]. However, the indirect effect through rumination was not, effect size = −0.01, 95% CI [−0.03, 0.02], standardized coefficient β = −0.02, 95% CI [−0.11, 0.07]. This indicates that perceived stress fully mediated the relationship between trait mindfulness and subjective sleep quality. Gender was a significant covariate in the relationship between mindfulness and rumination, b = 0.35, t(166) = 3.03, p = 0.0028. Thus, results confirmed the hypothesis that trait mindfulness is indirectly associated with subjective sleep quality through perceived stress but did not indicate any role of rumination in the relationship between trait mindfulness and subjective sleep quality (see Figure 2).

Figure 2 The mediating effect of perceived stress in the relationship between trait mindfulness and subjective sleep quality. The negative association between trait mindfulness and sleep quality was fully mediated by perceived stress. Note: *p < 0.001.

Secondary Sample Analysis

These results should be regarded as exploratory given the small sample that provided these data.

Sleep diaries: Analyses of the secondary sample revealed an uncorrected positive correlation between diary-based sleep ratings and trait mindfulness, however, none of the other diary-based measures were correlated with trait mindfulness, perceived stress, or rumination (see Table 3).

Actigraphy: There were no significant correlations between perceived stress, rumination, or trait mindfulness, and any of the actigraphy-based measures (see Table 4).

Table 3 Summary of Pearson’s and Spearman’s Correlations Between Sleep Diary and Questionnaire Measures

Table 4 Summary of Pearson’s and Spearman’s Correlations Between Actigraphy and Questionnaire Measures

Discussion

This study sought to investigate the interacting effects of trait mindfulness, perceived stress, and rumination in relation to subjective and objective sleep quality. Better sleep quality has been consistently linked to trait mindfulness, however, knowledge around the mediating variables, particularly in young adults, is limited. As expected, high levels of trait mindfulness were associated with low levels of perceived stress, rumination, and better subjective sleep quality (measured by the PSQI). Both rumination and perceived stress were associated with poorer subjective sleep quality on the PSQI. Importantly, perceived stress was found to fully mediate the relationship between trait mindfulness and subjective sleep quality; however, no mediating role was found for rumination. Other than the positive (uncorrected) association between trait mindfulness and diary-based sleep ratings, none of the sleep diary- or actigraphy-based measures were related to any of the other variables, but these results should be regarded as preliminary given the small secondary sample who provided these data. We also explored the effects of specific mindfulness facets.

The negative association between trait mindfulness and rumination is consistent with previous studies and the fact that they are theoretically opposed constructs.⁶⁰ The Monitor and Acceptance Theory suggests that mindful individuals are more likely to disengage from ruminative thoughts, resulting first from observing, followed by non-judging and non-reactivity, thus reducing stress and stress-related outcomes.^19,60,61 While heightened observing of thoughts is a characteristic of high ruminators (a positive association with rumination was found here), the other facets (including non-judging and non-reactivity) were negatively associated with rumination, highlighting the importance of these other factors for counteracting ruminative tendencies. In turn, rumination was negatively related to subjective sleep quality, as is well demonstrated in the literature.⁴⁷ The heightened cognitive and physiological arousal caused by rumination is thought to underlie this.

Trait mindfulness, and the facets of non-judging, non-reactivity, awareness, and describing, were negatively correlated with perceived stress in the present study. Higher trait mindfulness is consistently associated with lower perceived stress,^42,69,70 although fewer studies have explored the effects of individual facets. Here, observing was not associated with perceived stress. This finding is of value as it provides support for previous studies that found likewise. Both Ballantyne et al and McBride et al found that while the other facets were linked to reduced perceived stress, observing was associated with higher stress and physical symptoms of stress.^71,72 Again, in relation to Lindsay & Creswell’s Monitor and Acceptance Theory, it seems that attention monitoring alone is insufficient to reduce stress and might in fact be detrimental.⁶¹ Without the ability to respond to present moment experience in a non-judgmental/non-reactive way, higher levels of observing may lead to distress and maladaptive coping.⁷³ Other negative effects of observing have been noted, eg the meta-analysis by Sala et al found that higher observing was associated with higher substance use.¹⁶

As expected, perceived stress and trait mindfulness were inversely related to subjective sleep quality. Sala et al found that acting with awareness and describing were the most important facets regarding sleep quality while observing was not related, possibly due to higher engagement in health risk behaviors, eg substance use.¹⁶ Here, non-judging, non-reactivity, awareness, and describing were seen to be of benefit, while observing had no relationship with sleep quality. In other studies, in undergraduate samples, acting with awareness and non-judgement were associated with higher sleep quality on the PSQI, while observing had negative effects.^18,74 Recent work in an undergraduate student sample found that the associations between observing and sleep health were moderated by non-reactivity: observing was associated with worse sleep at lower levels of non-reactivity, but better sleep health at higher levels of non-reactivity.²⁸

Trait mindfulness is known to be a buffer against psychological distress and perceived stress, which in turn impacts health behaviors such as sleep quality.^19–21 Simione et al found that, in a general adult sample, mindfulness was negatively related to stress and this effect fully mediated the positive relationship between mindfulness and sleep quality.²¹ Lau et al and Roberts & Danoff-Burg found a partial mediation effect in undergraduate samples.^19,20 The present study sought to clarify this relationship, focusing on young adults, and by including rumination as an additional mediator. This is of value given the well-characterized relationships between mindfulness and rumination, and between rumination and sleep. Thus, reduced rumination is a potential mechanism through which mindfulness could be linked to better sleep. This is the first study to test such a model. Results showed that the relationship between trait mindfulness and subjective sleep quality is fully mediated by perceived stress. However, no mediating effect of rumination was observed.

This confirms that perceived stress plays a key role in the relationship. Thus, mindfulness appears to diminish the negative effects of stress, particularly in high-stress populations such as students, and when there is a direct impact of stress on the health outcome, eg sleep.⁴⁴ Trait mindfulness and mindfulness-based training have been shown to influence affect regulation, and mindfulness-based training has been shown to alter stress reactivity; this is thought to impact health outcomes by reducing physiological stress responses.^44,75,76 These findings provide valuable insight into the mechanisms linking trait mindfulness to sleep quality and are of high relevance to young adults, particularly students who often report poor sleep and high levels of stress;⁷⁷ stress appears to be the key mechanism and mediating factor (rather than rumination) in the relationship. This finding informs future research and intervention designs: it suggests that interventions should emphasize stress reduction within MBIs, for improving sleep quality among young adults. This is crucial given the high prevalence of sleep issues reported among young adult populations (as seen here and in previous research) and the important consequences of this for young adults’ daytime functioning, as well as longer-term physical and mental health outcomes.

A small secondary sample contributed objective sleep measures, allowing us to conduct exploratory analyses on these. This is of value given that the overwhelming majority of previous work has used subjective sleep measures only. Comparative studies in the sleep literature indicate that participants are often inaccurate when estimating their sleep duration and number of nighttime awakenings,¹³ with low correlations between self-reported sleep and actigraphy data.⁷⁸ Here, the objective actigraphy-based measures were not associated with perceived stress, rumination, or trait mindfulness. This accords with a recent meta-analysis by Clancy et al,⁷⁹ showing that links between rumination and sleep quality are much stronger for subjective, compared to objective measures of sleep quality. High rumination might create a cognitive bias for perceiving and reporting poor sleep quality, as theorized by Harvey et al.⁸⁰ Similar perception biases could serve to partially explain the relationships between stress, mindfulness, and subjective sleep quality as well. The lack of relationships with the objective sleep measures employed here would support this, but it is important to acknowledge that the non-significant findings may be due to the secondary sample being small and underpowered, potentially resulting in a Type II error. Thus, replication of the present study is required. A recent study in undergraduates derived a composite “sleep health” measure that combined actigraphy data with subjective sleep ratings; this measure correlated with higher mindfulness scores in awareness, non-reactivity, and non-judgment,²⁸ it would be of interest to explore whether this was also the case if only the objective ratings were included.

A weakness of the current study is that we relied on subjective measures of the traits under investigation, making the study susceptible to various biases including social desirability bias, demand characteristics, and order effects. The focus on trait-like characteristics is also a limitation. Pre-sleep rumination has the strongest negative impact on sleep;⁵² hence, future research should aim to employ measures of state rumination specifically assessing pre-sleep rumination, eg via experience sampling, to confirm the current findings.

Additionally, the study sample was not randomly selected and was recruited from one university, potentially skewing the findings, and limiting their generalizability. The sample was also predominantly female, and the small number of males prevented us from studying potential gender-specific effects. Although a female-dominated sample is not uncommon in the literature, future work should aim to explore the effects in a larger sample that has a more even gender balance.

Finally, a longitudinal design would have allowed inferences regarding causality. In the mediation analysis, trait mindfulness was input as the independent variable, and sleep quality as the dependent variable: this was based on evidence showing reductions in rumination and perceived stress, and improvements in sleep quality, following MBIs.^81–84 Still, due to the cross-sectional nature, the findings from this study are strictly correlational and the direction of causality cannot be inferred.

Conclusion

Overall, findings from this study indicate that perceived stress, rather than rumination, is a key mediating variable linking trait mindfulness to better subjective sleep quality in young adults. This adds important mechanistic detail to the literature, and the current findings have practical importance for addressing poor sleep quality, which is highly prevalent among young adults. Results provide a strong rationale for considering MBIs as an alternative to pharmacological treatment, as they have high efficacy and adherence, and no adverse side effects.¹³ Further, the findings also inform the design of MBIs, indicating that they should focus on reducing perceived stress, to optimize therapeutic outcomes.