Are international students relatively resilient? Comparing international and domestic students’ levels of self-compassion, mental health and wellbeing

ABSTRACT Because international students encounter unique stressors in addition to those experienced by all higher education students they are often assumed to be at heightened risk for experiencing mental health difficulties. However, empirical studies have consistently found that international students report comparable or lower rates of depression and anxiety than their domestic counterparts. The present study investigates one possible explanation for these unexpected results: are international students relatively resilient? More particularly, we investigate self-compassion (SC) – an emotion-regulation strategy known to buffer the impact of stressors on mental health. While SC has been associated with positive psycho-social outcomes in the higher education context, existing studies have not disaggregated international and domestic students. The present study addresses that gap by comparing the levels and estimating the effects of SC among international and domestic students. Analyses found that international students on average reported significantly higher SC than their domestic counterparts. Without the protection afforded by those higher levels of SC, we estimated that international students would report higher psychological distress and lower psychological wellbeing than domestic students. Implications for fostering and supporting the mental health and wellbeing of all higher education students are discussed.


Introduction
University students are known to be, on average, a 'high risk' group for experiencing mental health difficulties as a result of the various developmental, social, financial and academic challenges they experience (Orygen, 2017).In that context, it is vital to know which students are at higher risk than others so that institutions can provide appropriate services and support, and direct their limited resources effectively.There is a widely shared view that international students are particularly vulnerable to experiencing poor mental health and wellbeing while studying outside their home countries (e.g., Orygen, 2017;Shadowen et al., 2019) given the unique psychosocial and academic challenges associated with adjusting to new academic practices and culturesoften studying in a non-native language, and generally experiencing some degree of culture shock as well as a loss of social and family support networks (Alharbi & Smith, 2018;Cemalcilar & Falbo, 2008;Shadowen et al., 2019).International students were also particularly affected by the many disruptions to higher education wrought by the Covid-19 pandemic, including sudden and changing travel restrictionsleaving them either locked out of or stranded in their host countriesas well as the introduction of online learning and the closure of campus-based services (see e.g., Chen et al., 2020;King et al., 2020).
Given these unique and additional challenges, one would expect international students to report higher levels of psychological distress and lower levels of psychological wellbeing than their domestic (host country) counterparts.Yet that is generally not the case.Prior to the Covid-19 pandemic, empirical studies in US, Canadian, New Zealand and Australian universities found that international students reported comparable or lower levels of depression and anxiety symptoms (e.g., King et al., 2021;Larcombe et al., 2016), psychological distress (Skromanis et al., 2018) and perceived stress (Gardner et al., 2014) when compared to domestic students.Even during Covid-19 lockdownswith on-campus learning suspended, travel restricted and social isolation requirements in forcetwo separate Australian studies found that levels of psychological wellbeing did not vary significantly between international and domestic students (Dodd et al., 2021;Liu et al., 2021).How can these unexpectedly 'positive' results for international students be explained? 1 Several explanations are possiblefor example, stressors experienced by international students may have been over-rated, or the measurement instruments used in empirical studies may under-estimate international students' true levels of psychological distress due to cultural differences in interpretation and expression.While we consider the latter possibility more likely than the former, the present study remains agnostic on both those possibilities while exploring a third option: namely, that international students on the whole are relatively resilient, meaning that they are 'coping' with the stressors they encounter.Within a 'stress and coping' framework (Lazarus & Folkman, 1984), psychological distress or strain is understood to arise when perceived stressors from life changes and other challenges overwhelm or exceed the individual's coping strategies and psychological resources (Biggs et al., 2017).If international students engage adaptive or positive coping strategies to manage the various stressors they encounter (see, e.g., Russell et al., 2010) that is, if they are relatively 'resilient'this would explain why their (true) levels of psychological strain and poor wellbeing are no higher than domestic students' levels.This study explores that possibility by investigating one such adaptive coping strategyself-compassion (SC).By exploring international and domestic students' levels of SC, and its effects on their mental health and wellbeing, this study affords new insight into the adaptive coping strategies that international students may develop or apply to manage the academic and psychosocial challenges of studying abroad.The implications for fostering and supporting the mental health and wellbeing of both international and domestic university students are discussed.

Background
Self-compassion (SC) is an adaptive coping strategy known to moderate or buffer the impacts of stressors on psychological health (Chang et al., 2017;Zessin et al., 2015).As conceptualised by Neff (2003aNeff ( , 2003b)), SC is characterised by three inter-related practices: treating oneself with the kindness one would extend to a friend in similar circumstances, rather than forming negative self-judgements (Self-Kindness); understanding that difficulties and failings are part of the human condition and so connect us to others, rather than feeling isolated by negative feelings when experiencing set-backs (Common Humanity); and holding the present suffering in 'balanced perspective' rather than allowing oneself to be carried away by (or over-identifying with) painful experiences and emotions (Mindfulness) (Neff, 2003a(Neff, , 2003b)).
Studies among university students have shown that SC predicts higher psychological wellbeing (Saricaoglu & Arslan, 2013), lower depression and anxiety (Chang et al., 2017;Kroshus et al., 2021), and less homesickness and depression (Terry et al., 2013), as well as greater vitality (Gunnell et al., 2017) and life satisfaction (Hope et al., 2014).More particularly, SC has been shown to moderate or buffer the negative effects of academic stress on mental health (Lee & Lee, 2020;Poots & Cassidy, 2020), meaning that among students who report similarly high levels of academic stress or burnout, those with higher SC report significantly fewer symptoms of psychological distress than students with lower SC.Not only is SC protective; it is also malleable.Understood as an adaptive response to hardship, rather than a dispositional trait (such as optimism), the ways of thinking that characterise SC have the potential to be taught or fostered (Kroshus et al., 2021;Neff, 2015).Even short interventions designed to develop students' SC have been shown to improve emotional self-regulation and reduce depression and anxiety in higher education settings (e.g., Dundas et al., 2017).
Despite this evidence that SC plays an important role in university students' psychosocial adjustment and mental wellbeing, none of these studies has disaggregated international and domestic students.As a result, it is not known whether levels of SC are similar among international and domestic students; nor is it known whether SC has the same influence on rates of psychological distress and wellbeing in the two groups.The present study addresses those gaps and aims to advance understanding of international students' mental health and wellbeing by comparing the levels and estimating the effects of SC among international and domestic students.The specific research questions answered by this article are: RQ1: What levels of SC are reported by international and domestic students?RQ2: To what extent is SC protective of international and domestic students' mental health and wellbeing?RQ3: Do international students' rates of poor MHW change if SC is controlled for?

Method
Participants and procedure Participants were Bachelor, Master or research higher degree (RHD) students from a large, metropolitan Australian university.Australia is the third largest host country for international students globally (https://uis.unesco.org/en/uis-student-flow). At the study university, in the years of data collection (2017)(2018), international students made up approximately 35% of the total student population and were predominantly from China, Indonesia, Malaysia and India.
A total of 7518 students (6159 coursework and 1359 RHD) participated in a large, survey-based research project investigating students' wellbeing and course experiences, for which approval had been obtained from the university's Human Research Ethics Committee (HREC #1239160.2;#1239160.3).Study participants were recruited using email invitations, advertisements on university websites and a snowball technique promoted by relevant student associations.Informed consent was given by proceeding with the anonymous survey.Coursework students were surveyed in second semester, 2017, while RHD students were surveyed in second semester, 2018.Gift voucher prize draws (150 x $150 for coursework students and 50 x $100 for RHD) were offered as incentives to participate.Response rates ranged from 15% to 78% across the participating coursework programs; the response rate for RHD students was 27%.In general, the respondent samples reflected well the populations in the participating programs at the time of data collection, although females were slightly overrepresented in the respondent sample (for further detail, see Larcombe et al., 2022;Larcombe et al., 2021).Importantly, international students were well represented, comprising 27.4% of the coursework respondents and 38.5% of the RHD participants.The present study analyses responses from 5403 students (29.8% international) who answered all relevant survey questions.The socio-demographic and enrolment characteristics of international and domestic students in this sample are detailed in Table 1.

Measures
The larger project collected information about university students' wellbeing, course experiences, socio-demographic characteristics, and enrolment details.Only scales relevant to the present analyses are described here.

Mental health and wellbeing
Well-established and widely used instruments that have been validated with cross-cultural samples were used to measure respondents' mental health and wellbeing.Specifically, symptoms of depression and anxietyindicative of psychological distresswere measured using the relevant sub-scales from the short-form of the Depression, Anxiety and Stress Scales (DASS-21; Lovibond & Lovibond, 1995).Respondents were asked to rate the extent to which items applied to them over the past two weeks on a 4-point Likert-type scale ranging from 0 (did not apply to me at all) to 3 (applied to me most of the time).The DASS-21 has been shown to have good reliability as well as construct and discriminant validity across different racial groups (e.g., Bibi et al., 2020).The 7-item Depression (α = .91)and Anxiety (α = .84)sub-scales had good internal consistency in the current sample.Respondents' scores on the DASS-21 Depression and Anxiety sub-scales were categorised using cut-offs prescribed by the scale developers (Lovibond & Lovibond, 1995).Scores in the 'Severe' or 'Extremely severe' range for DASS-21 Depression or Anxiety are described here as High.
Psychological wellbeing was assessed using the 18-item Ryff's scales of Psychological Well-Being (PWB; Ryff & Keyes, 1995), which measures six dimensions of positive psychological functioning: Autonomy; Environmental mastery; Personal growth; Positive relations with others; Purpose in life; and Self-acceptance.Respondents were asked to rate their level of agreement with items on a 7-point Likert-type scale ranging from 'Strongly agree' to 'Strongly disagree'.The factor structure of Ryff's PWB scale, with six correlated first-order factors and a single second-order factorgeneral psychological wellbeinghas been shown to be consistent in diverse socio-cultural and linguistic contexts (Sirigatti et al., 2012).Cross-cultural testing of the 18-item PWB scale also provides support for the construct validity of the scale with different national groups (Sirigatti et al., 2012).The PWB (total score) had good internal consistency in the current sample (α = 0.82).As the scale developers have not prescribed cut-offs for discrete levels of psychological wellbeing, PWB scores more than one SD below Mean (i.e., scores < 4.3) are described here as Low.
High Depression, High Anxiety and Low PWB are described as indicators of poor mental health and wellbeing (MHW).

Self-Compassion
Self-compassion was measured using the 12-item short-form version of Neff's Self-Compassion Scale (SCS-SF; Raes et al., 2011).Respondents were asked to rate items on a 5- point frequency response scale from 1 (almost never) to 5 (almost always) and an average was calculated.Investigations of the psychometric properties of the SCS across diverse populations found invariance across linguistic groups, indicating that people with different linguistic backgrounds conceptualise self-compassion similarly (Tóth-Király & Neff, 2020).The scale had good internal consistency in the current sample (α = .85).
In line with the scale developer's guidelines (https://self-compassion.org/selfcompassion-test/), SC scores <2.5 are described here as Low, scores of 2.5-3.5 are described as Moderate, and scores >3.5 are described as High.

Results
Preliminary findings on international and domestic students' MHW Table 2 reports international and domestic students' mean scores for the three MHW measures.A series of t-tests showed that international students, on average, reported significantly lower Depression than domestic students, t(3079.25)= 4.781, p < .001;however, the effect size (d = 0.14) was very small.There was no statistically significant difference between the mean scores of international and domestic students for Anxiety, t(3182.05)= -.839,p = .41.On average, international students reported significantly lower PWB scores than domestic students, t(5186) = 4.487, p < .001;however, again, the effect size (d = 0.14) was very small.
Table 2 also reports correlations between study variables for both international and domestic student cohorts.All correlations were statistically significant.The effect size is large for the correlations between Depression and SC and between PWB and SC, and medium-to-large for the correlations between Anxiety and SC.
When MHW scores were categorised (see Methods section for cut-offs), similar proportions of international and domestic students met the criteria for High Depression, High Anxiety and Low PWB (Table 3), with international (cf domestic) students having a slightly reduced risk of High Depression (Risk ratio 0.82, Table 3) and a slightly increased risk of Low PWB (Risk ratio 1.14, Table 3).
RQ1: what levels of SC are reported by international and domestic students?
Preliminary investigations of SC scores by level of education found that, in both the international and domestic cohorts, any differences were explained by age.As a result, students were not disaggregated by level of education in subsequent investigations of SC.The mean SC score for the (combined) international student cohort was 3.08 (SD 0.66) while domestic students' mean SC score was 2.89 (SD 0.75).A t-test revealed that this difference was significant, with a small-to-medium effect size, t(3443.32)= −9.363,p < .001,d = −0.26.Further analyses (chi-square tests of association) showed that the international students' higher mean score was mainly the result of fewer international students reporting Low SC (Table 4)only 16.3% of the international students reported Low SC compared to 29.6% of the domestic students.Calculated as a risk ratio (Table 4), this means that international students are approximately half (0.55 times) as likely as domestic students to report Low SC.

RQ2: to what extent is SC protective of international and domestic students' mental health and wellbeing?
To quantify the protective influence of SC among the international and domestic cohorts, we first developed hierarchical multivariate regression models for DASS-Depression, DASS-Anxiety and PWB, adding the demographic and enrolment variables (reported in Table 1) in the first step and SC in the second step (see Supplementary material).This means that the models tell us how much of the variance in the MHW outcomes is predicted by SC after all the demographic and enrolment variables have been taken into account.As Table 5 shows, the combined demographic variables predicted less than 6% of the variance in MHW outcomes while SC predicted between 14 and 35% of variance even when added at a subsequent step.This confirms that SC is strongly protective of MHW for both international and domestic cohorts.For each MHW outcome, the ΔR 2 for step 2 is slightly higher for the domestic (cf international) cohort, indicating that SC explains slightly more variance in MHW levels in that group.
We then investigated the effects on MHW outcomes of Low SC levels specifically.The relations between Low SC and poor MHW indicators were analysed by calculating risk ratios (RRs; Andrade, 2015).All RRs (Table 6) are higher than two confirming that Low SC is a strong risk factor for poor MHW outcomes in both international and domestic cohorts.Among international students, those with Low SC are 4.14 times as likely as those with Moderate or High (Mod-High) SC to report High Depression; 3.49 times as likely to report Low PWB; and 2.07 times as likely to report High Anxiety.The RRs for domestic students are even higher, indicating that Mod-High levels of SC afford slightly stronger protection against poor MHW for domestic students.Supporting this, the rates of poor MHW among the Mod-High SC international students (Table 6) are slightly higher than among Mod-High SC domestic students, indicating that factors unrelated to SC make a greater contribution to international (cf domestic) students' poor MHW outcomes.
RQ3: do international students' rates of poor MHW change if SC is controlled for?
Finally, to test whether SC might account for lower than anticipated rates of poor MHW among international students, we investigated how their rates of High Depression, High  Anxiety and Low PWB change if SC is controlled for.More particularly, we estimated what international students' rates of poor MHW would be if their rates of Low SC matched those of domestic studentsthat is, if 29.6% of international students reported Low SC, rather than the observed rate of 16.3%.This was done by calculating the potential impact fraction (PIF)an epidemiological measure of the proportional change in average incidence of a health outcome after a change in the level of exposure to a risk factor (Barendregt & Veerman, 2010) in this case, Low SC.As shown in Table 7, if international students reported Low SC at the same rate as domestic students (29.6%), we could expect to see, for every 100 international students, an additional 4.2 cases of High depression, 2.8 cases of High Anxiety and 4.3 cases of Low PWB, all attributable to the 13.3% increase in the incidence of Low SC.These additional cases would result in higher proportions of international students reporting poor MHW compared to domestic students: 19.3% of international students would then report High Depression, compared to 18.5% of domestic students; 26.3% of international students would report High Anxiety, compared to 22.8% of domestic students; and 22.6% of the international students would report Low PWB, compared to 16.0% of domestic students.

Discussion
It is vital to better understand the MHW of international students in higher education so that institutions can provide appropriate services and targeted interventions.Given the unique and additional stressors that international students are known to face (Alharbi & Smith, 2018;Chen et al., 2020), it is puzzling that empirical evidence comparing international and domestic students' levels of MHW does not confirm researchers' expectations that international students are at heightened risk for experiencing mental health difficulties (e.g., King et al., 2021).While this may be a result of measurement error, given the consistency of findings in studies using different research instruments, the present study explored an alternative explanation: are international students distinctly resilient?More particularly, we explored SC and whether this adaptive coping strategy might account for international students' lower than expected rates of poor MHW.To our knowledge, this is the first study to compare international and domestic students' SC levels and to estimate the influence of SC on observed rates of depression, anxiety and psychological wellbeing in the two cohorts.
As a preliminary step, we compared MHW outcomes for international and domestic students.Our data showed that international students reported lower average Depression scores than domestic students and similar average Anxiety scores.This was consistent with previous Australian and international studies (e.g., King et al., 2021;Skromanis et al., 2018) finding that international students are not at higher risk for experiencing psychological distress.However, international students did report lower average levels of PWB than domestic students in the present studypossibly because Ryff's 18-item PWB scale is more sensitive than the 5-item World Health Organisation Wellbeing scale used in recent Australian research (Dodd et al., 2021;Liu et al., 2021) finding no differences in the wellbeing levels of international and domestic students.Further research on positive dimensions of wellbeing among international students is needed to confirm the present result and balance the focus on psychological distress (rather than wellbeing) in research to date.
There are three key findings from the present study which have important implications for understanding and supporting the MHW of both international and domestic students.First, answering RQ1, we found that international students' average SC scores were significantly higher than domestic students' scores.This difference was largely because international students were approximately half (0.55 times) as likely as domestic students to report Low SC.This means that more international (cf domestic) students are responding to the psychosocial difficulties, disappointments and set-backs they experience at university with the combination of self-kindness ('It's okay not to be perfect'), common humanity ('It's not just me'), and mindful awareness ('This is a set-back, not the end of the world') that distinguishes a protective, self-compassionate mindset (Neff, 2003a(Neff, , 2015)).Why might this be the case?Our data cannot answer this question.It is possible that international students' higher rates of SC reflect cultural differences in coping responses (see, e.g., Tóth-Király & Neff, 2020), or the attributes of students who choose to study abroad (a form of selection bias in the population).It is also possible that the framing of the 'international student experience' (Russell et al., 2010;Shadowen et al., 2019) activates and fosters SC.
International students are generally advised (pre-departure or during orientation) that studying abroad will not be easy: they should expect to experience some negative emotionsincluding homesickness and lonelinessas well as academic set-backs as they transition into and adjust to conditions in their host country.Normalising such difficulties, and advising that they will ease as familiarity with the host culture increases, may assist international students to suspend negative self-judgements when they encounter personal limitations and academic difficulties.Moreover, advice that all international students experience such difficulties to some degree may encourage individuals to recognise the shared features of their experience, rather than personalising negative feelings and self-isolating when difficulties are encountered.The time-limited nature of studying abroad (Alharbi & Smith, 2018) may also assist international students to maintain a balanced perspective as they can view their negative emotions and hardships as temporary, and consequent on the settings and circumstances in the host country, rather than personalising and over-identifying with them.
Further research is needed to investigate whether it is this framing of the psycho-social challenges inherent in international education that assists to foster a protective, self-compassionate mindset among international students.If it does, this suggests that re-framing the domestic 'student experience' in terms that identify more strongly the difficulties and challenges inherent in higher education ('this is hard'), as well as their universality ('all students experience academic disappointments and set-backs at some point') and transitory nature ('it will get easier with practice and persistence') may also help to foster higher levels of SC among domestic students.
Second, answering RQ2, our study found that SC is strongly protective of MHW outcomes in both international and domestic cohorts.While our results indicate that SC affords even stronger protection to domestic (cf international) students, SC predicted substantial variance in the MHW outcomes of both international and domestic students.Moreover, even moderate levels of SC were found to be protective against poor MHW in both cohorts: it was Low SC that significantly increased the risks of High depression, High anxiety, and Low PWB (all risk ratios higher than two for both cohorts).These results support research by Tóth-Király and Neff (2020) indicating that people with different linguistic backgrounds conceptualise SC similarly and can derive similar benefits from a self-compassionate mindset.They also indicate that institutions could improve student MHW by embedding SC 'training' (e.g., Dundas et al., 2017) moreover, our results indicate that such interventions only need to reduce rates of Low SC to be effective, they do not need to foster High SC.
We must add a caveat here, however.Our data indicate that SC interventions would make more difference to the MHW of domestic (cf international) students for two reasons.First, almost twice as many domestic (cf international) students reported Low SC: the fact that 83.7% of international students already report Mod-High SC means there is less scope to improve MHW in this cohort through interventions designed to reduce Low SC.Second, we found that Mod-High levels of SC were even more protective for domestic (cf international) students.This is likely because SC is useful for managing temporary, external stressors and difficult emotions; research suggests it is not able to buffer chronic or persistent stressors such as financial strain, discrimination and harassment (Kroshus et al., 2021).Thus, while our results show that SC strongly influences MHW, given that most international students in the present study were already relatively self-compassionate, institutional interventions to support and foster MHW in this cohort may have wider reach if they look beyond SC and target the stressors that international students typically experience, including discrimination, loneliness, language difficulties and academic adjustment issues (see Alharbi & Smith, 2018).
Third, answering RQ3, our study found that, if SC is controlled forspecifically, if international students' rates of Low SC are matched to those of domestic studentswe would see the higher levels of poor MHW among international (cf domestic) students that researchers expect.More particularly, with matching levels of Low SC, international students' rates of High Depression would be at least comparable to those of domestic students (rather than being lower) and international students' rates of High Anxiety as well as Low PWB would be substantially higher than those of domestic students.This finding indicates that SC is at least a partial explanation for the unexpected comparability of international and domestic students' levels of MHW in empirical studies.This is important not only because it solves a research 'puzzle'.The alternative explanation of 'relative resilience' that we have identified may allay speculation that previous research has over-stated the stressors affecting international students, or that the instruments used to measure MHW underestimate international students' levels of distress.
Our SC findings also afford new insight into international students' experiences of studying abroad.The fact that most are employing a self-compassionate mindset to adaptively manage the various stressors they encounter is testament to their resilience.However, 'coping' with stressors is taxing and ultimately unsustainable without positive psycho-social experiencessuch as feeling connected with others, a sense of personal growth and purpose, self-acceptance and autonomy (Ryff & Keyes, 1995) to replenish reserves of PWB.Moreover, 'coping' can only ever mitigate poor MHW; it cannot foster good MHW.As a result, for both international and domestic cohorts, SC interventions should be only one component of a multi-faceted 'whole-of-institution' strategy (Orygen, 2017) that seeks to support student MHW not only by fostering individuals' resilience but also by enhancing the environmental psycho-social resources that are vital to student wellbeingsuch as peer social connections, a sense of institutional belonging, and autonomous motivation for learning (Larcombe et al., 2022).

Limitations and further research
A strength of the present study was that it drew on a large data set including responses from international and domestic students enrolled at all levels of tertiary education (undergraduate, postgraduate coursework and doctoral research).However, further research is needed to test the replicability of results at other universities.Moreover, the data reported here were collected prior to the Covid-19 pandemic; further research is needed to establish whether international and domestic students' levels of psychological distress and wellbeing have returned to pre-pandemic levels now that travel restrictions have been lifted and campuses re-opened.The results reported here provide a baseline for such comparative studies.
The cross-sectional design of the current study also has limitations.For example, longitudinal research would be particularly useful to establish whether international students have higher SC prior to departurethat is, whether there is a 'selection bias' in the uptake of international educationor whether their SC levels increase as they respond adaptively to challenges and set-backs in the host country.Students' SC levels have been found to change across the first year of university (Gunnell et al., 2017); more experimental research is needed to understand the factors that influence SC levels in both international and domestic students.

Conclusion
Identifying the coping strategies that international students might draw on to manage the academic and psychosocial challenges they face studying abroad is essential to building a

Table 1 .
Select demographic and enrolment characteristics of international and domestic students in the sample for analysis.

Table 2 .
Means and standard deviations of measures, and correlations between scale scores for international and domestic students.

Table 3 .
Proportions of international and domestic students reporting poor mental health and wellbeing (MHW), with risk ratios and confidence intervals.

Table 4 .
Proportions of international and domestic students reporting Low, Moderate and High Self-Compassion scores, with the risk ratio for Low Self-Compassion.

Table 6 .
Proportions and risk ratios (RR) for poor MHW indicators for international and domestic students with Low or Moderate-High Self-Compassion.

Table 7 .
Potential impact factors (PIF) and estimated incidence of High Depression, High Anxiety and Low PWB among international students, consequent on an increase in the rate of Low Self-Compassion (SC).