Introduction

The mental health of university students has become a worldwide concern [1]. It is estimated that around 35% of university students meet the diagnostic criteria for a mental disorder [2]. In fact, their risk for developing a mental disorder has significantly increased up to around 50% as a result of the COVID-19 pandemic [3,4]. Overall, university students are particularly vulnerable to stress, which in this population is strongly related to academic underperformance, i.e., failure to fulfil academic obligations, and problematic health behaviors, such as substance use [5,6]. In addition, higher perceived stress levels in university students are associated with poorer quality of life, well-being, and sleep quality [7].

Given the high prevalence of stress among university students, there has been a burgeoning interest in detecting potential protective skills that could be enhanced to improve their mental health, such as mindfulness, self-compassion, and psychological flexibility [8,9]. Many studies have consistently indicated the negative association of mindfulness, compassion, and psychological flexibility with perceived stress in general populations [10–12]. Below is a brief conceptualization of them and a summary of the research studying its association with perceived stress among university students.

Mindfulness can be defined as a two-component mood that involves the self-regulation of attention, i.e., paying attention to the immediate experience (including internal experiences such as thoughts), and a curious, open, and accepting orientation towards it [13]. The cultivation of mindfulness skills leads to improvements in attentional control, acceptance of one’s experience, and non-reactivity to acute stressors [14,15]. Thus, high-level mindfulness individuals frequently respond to acute stressors with non-reactive acceptance instead of showing a pattern of excessive and blunted responses that could lead to a host of poor health outcomes [16]. This is reflected in previous studies involving university students. For instance, Burger et al. [17] investigated the associations between psychological distress and mindfulness in a sample of 174 university students and found that higher levels of mindfulness (especially the facet of acting with awareness) significantly predicted lower levels of negative affect, fatigue, nervousness, and agitation. Similarly, Muro et al. [18] found that mindfulness was associated with higher levels of life satisfaction in a sample of Spanish university students, a result which is consistent with the multicentre cross-sectional study of Salvarani et al. [19], who found that undergraduate nursing students (n = 622) with higher dispositional mindfulness scores had lower levels of stress symptoms.

Compassion is a cognitive, affective, and behavioural process that involves recognising suffering, common humanity, empathy, tolerance for uncomfortable feelings, and motivation to act or acting to alleviate the suffering of others and/or oneself [20]. Being self-compassionate during stressful life circumstances enables a pool of emotion-regulation strategies (i.e., healthy reappraisals, emotional acceptance, and self-soothing) that protects against distress [21]. In this line, Rahmandani et al. [22] found a significant negative association between self-compassion and several dimensions of distress (e.g., loss of confidence, anxiety), suggesting that high self-compassion might explain a lower level of stress in university students. In addition, Chan et al. [23] have found in a sample of 536 university students that compassion for others and self-compassion were linked to lower levels of perceived stress.

Finally, psychological flexibility is defined as the ability to consciously make contact with the present moment and behave in a way that serves valued goals [24]. Previous research about this construct have focused on experiential avoidance instead, which is a process aimed at trying to facilitate psychological flexibility. Experiential avoidance is defined as a “phenomenon that occurs when a person is unwilling to remain in contact with particular private experiences (e.g., bodily sensations, emotions, thoughts, memories, behavioural predispositions) and takes steps to alter the form or frequency of these events and the contexts that occasion them” ([24], p. 1154). Individuals who tend to rigidly avoid uncomfortable or unwanted internal experiences are more vulnerable to stress because this avoidant response pattern inhibits effective behaviors to cope with stress and might even increase the frequency of these unwanted internal experiences [25]. For instance, Chou et al. [26] found in a sample of 500 university students that experiential avoidance was a significant predictor of Internet addiction, significant depression, and suicidality. Moreover, Farr et al. [27] found that experiential avoidance was a significant predictor of depression, anxiety, and stress, this positive association being moderated by self-compassion. Specifically, these authors showed that individuals high in self-compassion reported lower levels of depressive symptoms across low to high experiential avoidance levels than individuals with low self-compassion levels [27].

Mindfulness, self-compassion, and experiential avoidance are three interrelated processes. Both compassion for others and self-compassion have consistently been found to be positively related to trait mindfulness (e.g. [28]). In addition, experiential avoidance and self-compassion are strongly and negatively associated [29]. Similarly, mindfulness has been found to negatively correlate with experiential avoidance [30]. According to Neff [31], there is a reciprocal association between mindfulness and self-compassion, in which each one is necessary for facilitating and enhancing the other. For example, a self-compassionate response requires a mindful approach to painful thoughts and feelings. However, an important distinction between them is that mindfulness is applied to pleasant, neutral, or unpleasant experiences, whereas self-compassion is focused on suffering. Moreover, self-compassion is focused on the global self, meanwhile mindfulness can be applied to the self as well as to thoughts, feelings, and sensations [32]. Furthermore, being unjudgmental and acceptance (central facets of mindfulness) can be considered the beneficial counterparts of experiential avoidance [30]. Likewise, mindfulness and experiential avoidance are drawn from very different theoretical contexts, i.e., Eastern philosophy and functional contextualism, respectively.

Overall, these are three psychological constructs of “third wave” cognitive behavioral therapies, which focus on the relationship between the individual and his/her thoughts and emotions regardless of the content [12]. As mentioned, university students show increased risk for distress and focusing on specifically how they respond to the internal experiences (thoughts, emotions, urges, sensations, memories, etc.) associated with their stressful circumstances might provide further insights about the relevance of applying third-wave approaches (e.g., mindfulness-, compassion-based interventions or acceptance and commitment therapy). As a matter of fact, it seems that increasing mindfulness facets, compassion, or psychological flexibility (by reducing experiential avoidance) may help to improve psychological distress in university students (e.g. [17,19,27]). However, very few studies have simultaneously studied the ability of trait mindfulness, compassion, and experiential avoidance, along with sociodemographic and academic variables, to explain levels of psychological distress in university students. Including all those variables in the same model would add specific information about which of these psychological constructs should be more emphasized in mental health interventions for university students.

In this sense, Martínez-Rubio et al. [33] showed that experiential avoidance, self-compassion and the mindfulness facets of observing and acting with awareness were significantly related to lower burnout symptom (‘overload’, ‘lack of development’, and ‘neglect’) levels in a sample of 644 undergraduate nursing and psychology students, but the potential influence of sociodemographic and academic factors in that association was not controlled. Results showed that experiential avoidance was the strongest predictor of the burnout dimensions of ‘overload’ and ‘lack of development’, meanwhile the dimension of ‘neglect’ was better explained by acting with awareness [33].

Therefore, the aim of the current study was to explore the relative contributions of three processes that are therapeutically relevant to perceived stress (i.e., trait mindfulness, self-compassion, and experiential avoidance), a critical construct among university students who are at risk of burnout, depression, and anxiety. According to the literature, we expected that mindfulness, self-compassion, and experiential avoidance would explain a significant proportion of the variance of perceived stress after controlling for relevant sociodemographic and academic variables such as age, gender, or having a scholarship. The direction of the relationship between mindfulness/self-compassion and perceived stress was expected to be negative (inverse). Alternatively, a positive relationship was expected between experiential avoidance and stress. No prior hypotheses were established on the strength of the associations due to the scarce number of previous studies.

Materials and methods

Participants

Table 1 displays all the sociodemographic, academic, and psychological characteristics for the total sample. The study sample was composed of 589 university teaching students from the San Vicente Mártir Catholic University of Valencia (Spain), of whom 81.2% were female, and the mean age was 22.13 years old (SD = 3.90; range: 18–48). All participants were White. The inclusion criteria were: (1) being ≥18 years old, (2) being an undergraduate student in the field of education, and (3) being able to understand written Spanish.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sociodemographic, academic, and psychological characteristics of the sample (n = 589).

https://doi.org/10.1371/journal.pone.0280791.t001

Results

Table 2 shows the summary of the results from the hierarchical multiple regression analysis. The complete output is shown in S2 Table. According to the sample size (n = 589) and number of predictors (17 variables), the regression would be sensitive to effects of f² = 0.03 (small) with 80% power (alpha = .05, two-tailed). Sociodemographic and academic variables were entered at Step 1, explaining 12% of the variance in PSQ-SF scores (f² = 0.14). After entering the FFMQ-SF, SCS-SF, and AAQ-II scores at Step 2 the total variance explained by the model was 50%, F(17, 542) = 32.29, p < .001, f² = 1. In other words, the psychological measures (FFMQ-SF, SCS-SF, and AAQ-II) explained an additional 38% in perceived stress levels (PSQ) after controlling for sociodemographic and academic variables, R squared change = .381, F change (3, 542) = 139.41, p < .001.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Hierarchical multiple regression analysis summary predicting total perceived stress scores with sociodemographic, academic, and psychological variables.

https://doi.org/10.1371/journal.pone.0280791.t002

Being a woman (sr = 0.16, p < .001), perceived family support (sr = -0.21, p < .001), left home (sr = 0.08, p = .043), and study hours per week (sr = 0.12, p = .004) were significant predictors at Step 1. In the final model, among the psychological measurements, only SCS-SF and AAQ-II significantly explained the PSQ-SF scores, with the AAQ-II recording a higher semi-partial correlation value (sr = 0.31, p < .001) than the SCS (sr = -0.17, p < .001). That is, the AAQ-II and SCS-SF scores uniquely explained around 10% and 3%, respectively, of the variance in PSQ-SF scores. In addition, perceived family support (sr = -0.16, p < .001), study hours per week (sr = 0.12, p < .001), having a partner (sr = 0.09, p = .004), being female (sr = 0.08, p = .010), and age (sr = 0.07, p = .015) made a unique statistically significant contribution to explaining PSQ-SF scores. When the Benjamini–Hochberg correction was applied to correct for multiple comparisons, all significant effects remained significant.

When the hierarchical multiple regression was re-tested with the SCS-SF total score calculated without adding the mindfulness and overidentification items scores (S3 Table), FFMQ-SF did significantly explain PSQ-SF scores (sr = -0.07; p = .015), along with the aforementioned variables. When the Benjamini–Hochberg correction was applied to correct for multiple comparisons, all significant effects remained significant except for age (adjusted p = .056). In particular, AAQ-II, SCS-SF, and FFMQ-SF scores uniquely explained around 11%, 2%, and 0.55%, respectively, of the variance in PSQ-SF scores.

Discussion

This study examined the explanatory power of mindfulness, self-compassion, and experiential avoidance with regards to perceived stress in a sample of university students while controlling for the influence of relevant sociodemographic and academic variables. When all of them were taken into consideration as potential explanatory variables of perceived stress in the regression model, mindfulness, self-compassion, and experiential avoidance explained a greater variance of perceived stress than all sociodemographic and academic variables together.

In particular, experiential avoidance was by far the most explanatory variable of levels of perceived stress (direct association). Self-compassion was also a significant explanatory variable that was inversely associated with perceived stress, but to a lower degree. Finally, trait mindfulness showed a non-significant additive explanatory power in the regression, but when self-compassion scores did not account for mindfulness and overidentification items, mindfulness was shown as a significant explanatory variable inversely associated with perceived stress. Similarly, experiential avoidance significantly explained the levels ‘overload’ and ‘lack of development’ burnout dimensions to a greater degree than self-compassion and mindfulness in our previous study [33]. Thus, both results may be further evidence of a stronger connection between psychopathology and experiential avoidance than between psychopathology and other third-waves constructs (e.g., mindfulness or compassion) in university students.

In our sample, perceived stress was positively associated with experiential avoidance, which could be regarded as a psychological variable with potential risk. In contrast, perceived stress was negatively correlated with self-compassion and mindfulness, which, in turn, could be seen as protective factors. The same scenario was shown regarding the association between experiential avoidance, mindfulness, and self-compassion with burnout dimensions in our previous study [33]. These results can be explained taking the emotion dysregulation model of psychological distress into account [45], which posits that the ways individuals experience and respond to emotions can lead or not to psychological distress. According to this model, motion dysregulation (or misregulation) can occur when the selected emotion regulation strategies do not match the situation or keep the person in an endless struggle to free himself of the unwanted emotions, being experiential avoidance the flagship example among maladaptive strategies. In fact, it is considered a strong transdiagnostic predictor of psychological distress [12]. In contrast, higher mindfulness and self-compassion levels (as a state or trait) are typically related to less psychological distress. For instance, it has been found that low-level mindfulness individuals frequently respond to acute stressors showing a pattern of excessive and blunted responses that could lead to a host of poor health outcomes [16]. In addition, the negative components of self-compassion showed greater associations with psychological distress than the positive counterparts in a recent meta-analytic study [46].

Our findings provide further evidence on the association between experiential avoidance and psychological disturbances [47] and would indicate that stressed students present a tendency to escape or avoid private events (e.g., emotions or memories), regardless of their sociodemographic and academic profile. In this line, previous studies have reported significant associations between experiential avoidance and depression or suicidality in Taiwanese university students [26], as well as between experiential avoidance and burnout syndrome in Spanish undergraduate nursing students [33]. Experiential avoidance would also seem to moderate the positive relationship between anxiety sensitivity and perceived stress in a large community sample of university students from the United States [48]. It is also an important construct in anxiety disorders [49]. In fact, there is a certain amount of anxiety research that demonstrates the mental health costs associated with inflexible development of an avoidance response style, such as the permanence of the experience of bodily arousal in people with panic disorder, being concerned about openly exposing and communicating intense emotional experiences to other people, and the fear of strong emotional impulses in anxiety disorders. The abovementioned mental health costs associated with inflexible psychological features perpetuate anxiety disorders and are related to an experiential avoidance style [49].

Self-compassion was the second-best explanatory variable in explaining perceived stress in our sample. Its significant association with perceived stress would indicate that stressed students tend to be self-critical rather than self-kind against failure and adversity, regardless of their sociodemographic and academic profile. These findings are also consistent with previous studies on university students [50,51], in which self-compassion mediated the negative relationship between perceived stress and anxiety and depression symptoms in a large sample of German university students. Furthermore, it was suggested that self-compassion could reduce anxiety and depression levels [50,51], and may increase students’ capacity for managing the emotional demands of their studies [52].

Even though the inverse association between dispositional mindfulness and psychopathological symptoms has been consistently reported [53], our results showed a non-significant relationship between mindfulness as a trait and perceived stress. One possible explanation for that unexpected result was the partial overlap between mindfulness (FFMQ) and self-compassion (SCS) measurements [50]. That is, Neff’s model [31] includes mindfulness (vs. overidentification) as one of the three positive components of self-compassion, which is mirrored in the SCS. In fact, when the overlap between these measurements was solved, mindfulness and perceived stress association were found to be statistically significant. Therefore, the non-significant association that was initially found was mostly related to psychometric issues that future research should address using other compassion measurements such as the Sussex-Oxford Compassion Scales [28]. Thus, our results would indicate that stressed students tend to be unaware of the present moment experience and to respond reactively to it. In this line, previous studies have shown the inverse association between mindfulness and perceived stress levels [17,19] as well as the positive association between mindfulness and satisfaction with life [18].

Regarding sociodemographic and academic variables, high total study hours per week, having a partner (vs. being single), being female (vs. being male), and being older were significantly associated with high levels of perceived stress. In addition, perceived family support was found to be a protective variable, a good/very good perception of family support being associated with lower levels of perceived stress. Finally, having children, being in employment, having left the family home, having a scholarship, the academic year, and the number of failed subjects were not significant contributors to higher levels of perceived stress.

As mentioned, higher perceived family support significantly explained lower perceived stress levels in our sample of university students. These findings may provide a basis for explaining how family support may be a relevant buffering variable in the perception of the stress that occurs among undergraduate students [54,55]. For example, in a sample of Chinese university students, family cohesion explained social adjustment by increasing a sense of security [55], altogether psychosocial factors that are known to buffer the harmful effects of stressors on well-being. Conversely, having a partner was positively associated with perceived stress. It is possible that the family is more supportive of difficulties during the university experience than a partner, who may be more of an obstacle or a source of tension during this period.

Regarding the significant positive association found between perceived stress and study hours per week, the perception of overload due to the amount of study hours has previously been associated with stress and also burnout [56,57]. Furthermore, the load of study hours per week constitutes not only an associated perceived stress variable, but also an explanatory factor that could impair academic performance among university students [58]. Finally, as in other studies, being female and of an older age also explained increased perceived stress levels [35,59,60].

Up to now, mindfulness-based interventions for university students have shown promising effectiveness [8,61], though our results suggest that comprehensive programmes more oriented at fostering other crucial process variables for stress reduction such as experiential avoidance and self-compassion are also needed in students showing greater levels of perceived stress. Therefore, these variables should be considered as targets in programs aimed at fostering well-being and mental health in university students. Programs aimed at increasing certain psychological skills such as Acceptance and Commitment Therapy [62,63] or the Mindful self-compassion program [64] should be considered when choosing interventions for this population. Indeed, it should be noted the importance of the behavioral components of these interventions as an addition to mindfulness. Given that self-compassion and avoidance (or acceptance) are behavioral in nature, this may be a critical addition to mindfulness-based interventions.

Finally, taking into account that our sample was composed of university teaching students it should be noted that psychological flexibility and self-compassion are important aspects of social and emotional competences in teaching professionals’ performance. In turn, the social and emotional competence of teachers is a key point to promote healthy student-teacher relationships, effective classroom management, and effective social and emotional learning implementation that contributes to a healthy classroom atmosphere. This healthy classroom atmosphere benefits students’ social, emotional and academic learning process which, in turn, has an indirect effect on teachers’ social/emotional well-being [65].

Conclusions

In conclusion, mindfulness, self-compassion, and experiential avoidance explained perceived stress levels in university students studying education after controlling for relevant sociodemographic and academic variables. The association found between these variables and perceived stress underlies the need to include them as potential intervention targets, especially in the current post-pandemic times. The higher explanatory power of experiential avoidance in our analysis demonstrates that although mindfulness-based programmes have been reported as key interventions to promote psychological well-being, other interventions focused on psychological flexibility should also be taken into consideration. Thus, our findings underpin the need for formative interventions that would aim firstly to reduce experiential avoidance levels, secondly to promote self-compassion, and thirdly to use mindfulness training as a psychotherapeutic aid. Adequate efforts should be made by higher education institutions to recognise and address university students’ mental health challenges and towards facilitating their mental health recovery, by reducing the long-term consequences of the pandemic on mental health. Thus, these three variables that can be trained and modified could be included as key factors in specific psychological interventions targeted at university students, as previous studies already suggested. Nevertheless, the causal/temporal relationships between experiential avoidance, self-compassion, and mindfulness skills and perceived stress should be explored in future research using more refined study designs.

Supporting information