Methods
Study design and subjects
The present study was a cross-sectional design and a cluster sampling was adopted. Data were collected from the international students of China Medical University in November 2020. The inclusion criteria of the potential participants were (1) able to get access to internet, (2) a current student of the University, (3) able to read, fully understand and answer the survey questions. One thousand and fifteen students who met the inclusion criteria were initially contacted via electronic email. Then, in the online survey, there was a brief explanation about the study, and the participants were asked to complete an informed consent agreement, in which they were made aware that participation was completely voluntary. Research Ethics Committee of China Medical University approved our study (2020–25), and the study was performed in accordance with the Declaration of Helsinki. Finally, a total of 543 international students participated, and 512 of them were able to complete the questionnaires. The overall response rate was 50.44%.
Measurements
Measurement of anxiety symptoms
The anxiety symptoms were measured with the Generalized Anxiety Disorder 7-item (GAD-7) questionnaire. The questionnaire consists of seven items for observing the frequency of anxiety symptoms with a four-point Likert scale from 0 “not at all” to 3 “almost every day” [
28]. The anxiety level is reflected by the total score, where higher scores indicate more symptoms of anxiety. Scores of 5, 10 and 15 represent the cutoffs for mild, moderate and severe anxiety symptoms, respectively [
29]. Previous studies have demonstrated the GAD-7 has high reliability as well as good criterion and construct validity [
30‐
33]. The Cronbach’s alpha for this sample was 0.92.
Measurement of stressors in university life
Stressors in university life of international students were measured by 7 questions regarding (1) health problems, (2) financial pressure, (3) academic difficulties, (4) interpersonal relation difficulties, (5) daily life difficulties, (6) adverse life events and (7) language barrier [
11,
12]. Participants answered 1 (not at all) to 4 (very serious) to the questions. The total score represents the severity of the stressors perceived by the participant. A Cronbach’s alpha of 0.80 was found for the scale in this study.
Measurement of self-efficacy
Self-efficacy was assessed with the General Self-efficacy Scale (GSES), which is a 10-item measure of an individual’s confidence in his or her ability to deal with stressful situations [
34]. Items are scored on a four-point Likert-type scale ranges from 1 (not at all true) to 4 (exactly true), and responses are calculated to yield a total score of all item scores where higher scores indicate higher levels of self-efficacy. The GSES has good psychometric properties, and many studies have confirmed its internal consistency reliability, convergent and discriminant validity [
35,
36], demonstrating that GSES is a reliable and valid measurement. The Cronbach’s alpha for GSES in the present study was 0.95.
Measurement of perceived stress
Perceived stress was evaluated by the 10-item version of Perceived Stress Scale (PSS-10), which is a self-report measure designed to assess the extent to which participants appraise their lives to be stressful [
37]. Each item is rated on a 0 (never) to 4 (very often) Likert scale by the respondent to indicate how often the participant experienced specific feelings or thoughts. The total scores of the measure are obtained by adding the score of each item (4 items are reverse-scored) to provide a continuous measure of perceived stress, and higher scores indicate greater perceived stress. PSS-10 has demonstrated strong psychometrics. Its coefficient alpha reliability ranged between 0.84 and 0.91 in previous studies [
6,
38], and in this study it was 0.87.
Demographic characteristics
Age, gender, current place of residence (Asia/Africa/North America/Europe/Oceania) and educational background were investigated for demographic characteristics.
Statistical analysis
SPSS 17.0 was used for the statistical analysis. Descriptive statistics including frequency distributions for the nominally scaled demographic variables provided a profile of the sample. We found the scores of GAD-7 were not normal distribution after testing the normality for continuous variables. Therefore, Mann–whitney U test was conducted to determine if the groups were statistically equivalent on anxiety symptoms. Spearman’s rank correlation coefficients were used to examine relationships between continuous variables.
The sequential mediation was tested using PROCESS macro program for SPSS [
39], which facilitated path analysis-based mediation analyses. We verified the hypothesis model by the bias-corrected percentile bootstrap method, with 5000 resampled samples. 95% confidence intervals for the mediation effects were estimated and the results were considered significant when the 95% confidence interval did not include zero. We generated direct effect of stressors in university life on anxiety symptoms and indirect effects of stressors in university life on anxiety symptoms through the mediators (self-efficacy and perceived stress) in the mediation using the model 6 of PROCESS. There were three routes of indirect effects in the sequential mediation model. When the direct effect became non-significant but the indirect effect was significant, full mediation was established. Partial mediation was confirmed if both effects are significant [
40]. Continuous variables were all centralized before the model was validated to avoid multicollinearity. Two-tailed alpha 0.05 was used for significance testing purposes.
Discussion
Previous studies mainly focused on the cross-cultural adaptation of international students, but less on the stressors and related stress responses. In our study sample, the mean scores of PSS-10 was 16.53, which is lower than the local students in Turkey (18.03) [
41], Saudi Arabia (20.10) [
42] and China (21.13) [
43]. In addition, 28.71% of the international students had anxiety symptoms in the present study, which is also lower than the domestic Chinese students (46.85%) [
43] and Libyan students (64.50%) [
44]. A possible reason may be that the participants in our study all come from a medical university, and they may already have certain amount of knowledge on mental health. It is also possible that the measures taken by their university to manage the stress have been effective.
Our study showed that financial pressure and language barrier were the most serious stressors in university life among international students, which were different from the findings demonstrating that academic difficulties were the primary sources of stress in university students. As pointed out by Grable and Joo, the students who face financial crisis tend to be more likely to drop out of the university or achieve lower grades than others [
45], which may cause serious stress to the students. The importance of financial pressure to international students in our study was in line with the previous studies on international students which indicated that the financial pressure was a particular concern and at higher risk for problem of mental health [
46,
47]. Language insufficiency has also been found to be a critical stressor that international students encounter in other studies, because language proficiency was essential in international students’ sociocultural adjustment [
48,
49]. In this situation, the students may face concomitant problems such as lack of confidence and low self-efficacy, again causing higher level of stress. This finding consisted with the results in previous studies which proved language deficit was a significant source of stress among international students [
50,
51]. Furthermore, in our study, stressors in university life were found positively associated with anxiety symptoms of international students (
β = 0.23,
t = 5.83,
p < 0.01), which supported H1 and was consistent with other studies [
52]. Since the students are exposed to various stressors in university life to different extent and it may not be possible to remove the stressors from their roots, understanding the internal mechanism becomes very important in order to reduce the adverse effect of stressors in university life and maintain the mental health of students.
In the present study, stressors in university life were negatively associated with self-efficacy (
β = -0.24,
t = -5.51,
p < 0.01) and positively associated with perceived stress (
β = 0.38,
t = 9.68,
p < 0.01), which supported H2 and H5. Self-efficacy was negatively associated with perceived stress (
β = -0.24,
t = -6.13,
p < 0.01), and perceived stress was positively associated with anxiety symptoms (
β = 0.51,
t = 12.85,
p < 0.01), which supported H8 and H6. Unexpectedly, sequential mediation model testing didn’t show a direct effect of self-efficacy on anxiety symptoms nor an indirect effect of self-efficacy in the association between stressors in university life and anxiety symptoms. Therefore, H3 and H4 were not supported, which indicated that the association of self-efficacy with anxiety symptoms was not direct, similar with the findings from a study on medical college students in Philippines [
53]. Instead, self-efficacy played a sequential mediating role with perceived stress in the association between stressors in university life and anxiety symptoms, which supported H9 and indicated self-efficacy’s direct relationship with perceived stress rather than anxiety symptoms. Although the sequential mediation effect accounted only for 12.5% of the total mediation effect, it still implied that the impact of self-efficacy on anxiety symptoms was generated through perceived stress. This result supported the transactional model of stress. It also indicated that self-efficacy was an effective protective factor against stress. Individuals who have lower levers of self-efficacy do not have enough confidence and the ability to cope with the external and internal environment. They will perceive more severe stressors and stress, and are more prone to show anxiety symptoms. Self-efficacy improvement interventions in previous researches have shown that the methods were effective in empowering participants to cope with stress [
22,
24,
54‐
56].
Another finding of our study was the partial mediation effect of perceived stress in the association of stressors in university life and anxiety symptoms among international students, which supported H7. Perceived stress alone accounted for 87.50% of the total mediation effect and 43.75% of the total effect. Its strong effect indicated its important role in facilitating translation of stressors in university life into anxiety symptoms, and this is in line with other studies that assumed appraisals were important determinants of adjustment to stressful encounters [
57]. Previous empirical researches have shown similar findings of the mediation effect of perceived stress [
25‐
27]. Combined with our findings of self-efficacy as a protective factor against stress, interventions can be considered using self-efficacy training to alleviate perceived stress and promote the positive appraisal on stressors in university life to reduce anxiety symptoms. There already have been some researches of stress management among university students using cognitive behavioral therapy which have achieved a significant reduction in perceived stress and anxiety symptoms after the intervention, with the enhancement of self-efficacy as well [
17].
Limitations
Our study has several limitations. Given the cross-sectional design, it’s unable to make any assertions regarding causation. A further experimental design of study in the future should be employed to determine causal relationships. Another limitation is that there may have been response biases in the self-report of the individuals completing the measures. Finally, as most of the participants were from Asia, the results in this study may not apply equally well to the students in other part of the world. Future research could expand the diversity of the university types to better capture the students from other part of the world.
Despite of the limitations, this study has discovered the sequential mediating role of self-efficacy and perceived stress in the association between stressors in university life and anxiety symptoms, and provided a new perspective on how to maintain mental health for international students. The sequential mediators provides a deeper insight into the underlying mechanism of stressors in university life towards anxiety symptoms among international students. At the same time, this study has broadened the application scope of self-efficacy in the field of stress research, and is also an empirical contribution to the theory of transactional model of stress using in the population of international students. In addition, our study shows that identification and evaluation of stressors in university life are important, and financial pressure and language barrier should be given more attention for international students. This would be valuable in ensuring implementation of stress reduction programs to effectively support students. Given that few studies on university life stressors exist in the literature of international students, our study is very important because it has filled in the gap. As for practical implications, our study findings may apply to all the international students who have poor self-efficacy or perceive higher levels of stress or are struggling with anxiety. Therefore, counselling focusing on financial pressure and language barrier, as well as introduction of specific interventions into university campus for international students should be encouraged, and the university educators should utilize self-efficacy improvement and stress reduction measures in the training programs to support students.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.