There have been many pieces of research on psychological resilience. Most of them focus on the antecedents and consequences of psychological resilience or take psychological resilience as an intermediary or regulatory variable. The psychological resilience of this paper refers to the individual in the workplace situation. There is no uniform definition of psychological resilience. However, they all believe that the experience of adversity is the first defining element (24). This paper defines psychological resilience from the perspective of ability (25), explicitly referring to the ability of expatriates to deal with stressors and adjust themselves to normal status. Fisher, Ragsdale (26) point stressors at work may be short-term and sudden high-risk events (e.g., public safety events) or long-term continuous circumstances (e.g., work stress). For expatriates from international construction companies, they have to face the public workplace stressors mentioned above. At the same time, due to the high risk of the international construction industry and the working environment of uprooting, especially the impact of COVID-19, stressors include both from the pandemic (27), family (28), and workplace (29).

In workplace, stressors are regarded as an adversity, and individual's protective resources are first invoked in response to stress (30). Protective resources are directly related to psychological resilience (31). If there are too many stress events and the resources that individuals can use are insufficient, the psychological elasticity will be worse. Taking mental health as an outcome of psychological resilience, scholars found that employees' resilience is positively related to their mental health (32), and negatively affects the expression of burnout and emotional exhaustion (33–35). McLarnon and Rothstein (36) pointed out that half of the indicators in the Workplace Resilience Scale were negatively correlated with depression. Moreover, Ferris, Sinclair (37) emphasized that the lack of individual psychological resilience caused physical and psychological stress, such as low emotional, easy to fatigue, and poor attention. Previous studies have started investigating the individual resilience as a mediation factor. The challenge-hindrance stressors model posits that workplace stressors can be grouped into two categories. Hindrance stressors will interfere with performance or goals, while challenge stressors contribute to performance opportunities (38). Based on the model, Crane and Searle (39) found resilience played a full mediation effect between the negative relationship of challenge stressors and strain and the positive relationship of hindrance stressors and strain. Also, Kinman and Grant (32) found resilience played a full mediation effect between the negative relationship of emotional intelligence and mental distress.

Since the outbreak of COVID-19, some studies focus on the relationship between resilience and mental health. However, researchers have regarded resilience as one of the components of psychological capital, which has been found to be related with factors in workplace and employee's mental health (25). Lawal et al. (40) used the standard scale to test the mental health of ordinary people in COVID-19. Their research pointed out that the individuals' psychological distress, depression, and anxiety were significantly higher than those of normal ones due to the influence of COVID-19. In the early study, researchers paid more attention to doctors, nurses, and other people who had direct contact with COVID-19 (41). Rossi et al. (42) analyzed the mental health status of Italian residents during the closed period. They found that stressors in COVID-19 significantly impacted depression, anxiety, and stress perception. Barzilay, Moore (43) found that people with higher psychological resilience were not prone to depression and anxiety. As the pandemic continues, we believe this phenomenon may also be present among international construction expatriates. Therefore, the hypothesis is proposed.

Many studies have confirmed that psychological resilience among individuals differs, which depends on many factors. Personal resources are considered one of the most critical factors (44). Employees' professional knowledge about the work (45) or technology related to work (46) is positively related to psychological resilience. Although there is no direct research on the relationship between work experience and psychological resilience, like workability, work experience is also an essential resource in the individual workplace. Experienced workers know better how to deal with difficulties and perform better under stressors (47). The positive state and emotion could improve employees' psychological resilience, while the negative emotions had an opposite effect in case of an organizational crisis (48). Therefore, we believe that it is reasonable that expatriate experience, as a unique resource, is related to individual psychological resilience. Therefore, the hypothesis is proposed.

Resilience is a protective mechanism when people are facing of adversity which is usually associated with lower mental distress (49). However, no thorough research points out that work experience is directly related to individual mental health statuses such as depression and anxiety. However, Wiseman, Curtis (50) found that individuals with different life experiences have different manifestations of depression and anxiety. Stress levels were higher for the general public than those working directly (front line nurses) with COVID patients, possibly related to experience and confidence (51).

Moreover, Rossi et al. (12) and Nwachukwu et al. (52) found that in the elderly group, psychological resilience has a more significant impact on individual mental health states such as depression and anxiety. It is found that age is the regulatory factor between psychological resilience and mental health. Furthermore, generally, older people have more work experience. Therefore, the hypothesizes are proposed.

We tested the hypothesis model with a sample of expatriates in international construction enterprises. In the study, expatriates were defined as “Citizens of the home country or third country whom international construction companies appoint to work in the host country, among which the citizens of the home country who work in the host country are mainly.” Therefore, we cooperated with SINOPEC Engineering (Group) Co. Ltd. (SEG), which was listed in the Top 250 contractors of ENR in 2020. We used the online questionnaire platform of Wenjuanxing to collect data. The HR department sent the questionnaire to all the expatriates by the inner system of SEG from May 5 to May 25 in 2020. In the survey, 3,091 valid questionnaires were received, including managers, workers, subcontractors, and others. The characteristics, including gender, age, position, and international assignments experience, were collected in the demographic information (Table 1). The expatriates of enterprises participating in the research are informed in advance that the results are only used for academic research and participate voluntarily.

Stressors (SE) during COVID-19 were assessed using a checklist of stress events developed by this research. The list of stress events was obtained through a literature review and employee interviews. Therefore, the checklist explores thirteen different stressors in Table 2. In the questionnaire, each item has a yes/no response as a binary variable. 0 = “feel no stress due to this during COVID-19” and 1 = “feel stress due to this during COVID-19.” The Cronbach's alpha was 0.888 in this research.

Psychological resilience (PR) was measured by 10-item Connor–Davidson Resilience Scale (CD-RISC-10). The CD-RISC is used widely to assess resilience (55). The original version of the CD-RISC-10 was created by Campbell-Sills and Stein (56) includes ten items. The response scale has a 5-point range: 1 (not true at all), 2 (rarely true), 3 (sometimes true), 4 (often true), and 5 (true nearly all of the time). In the present study, Cronbach's alpha for the overall CD-RISC-10 was 0.937.

Depression was measured by the 9-item Patient Health Questionnaire (PHQ-9) in a Chinese version (57). PHQ-9 contains nine items measured by a 4-point Likert scale. The total score of PHQ-9 was analyzed as a continuous variable. PHQ-9 is used as a tool for screening depression in many countries around the world. The Cronbach's alpha was 0.938 in this research.

Anxiety was measured by the 7-item Generalized Anxiety Disorder questionnaire (GAD-7) in a Chinese version (58). GAD-7 contains seven items measured by a 4-point Likert scale. The total score of GAD-7 was analyzed as a continuous variable. Many countries use GAD-7 for anxiety screening. In this research, Cronbach's alpha was 0.936.

Perceived stress was assessed by the Chinese version of the 10-item Perceived Stress Scale (PSS-10) (59). PSS includes ten items rated on a 5-point Likert scale. The Cronbach's alpha was 0.922 in this research.

International assignments experience was divided into five groups, and respondents could choose the answer according to their situation. This variable was analyzed as a ordinal variable with five scores (1 = 0–1 years, 2 = 1–3 years, 3 = 3–5 years, 4 = 5–10 years, 5 = 10 years and over.) The scales and question items are listed in Appendix A.

After standardized the data of PHQ and GAD scales into 5-point, the analysis process followed the steps below. Firstly, we used Harman's single-factor test to check the common method biases of all of the items in the four scales. Moreover, confirmatory factor analyses (CFA) were conducted using AMOS 26.0. Secondly, we used the SPSS 26.0 to conduct descriptive statistics and correlation analysis. Thirdly, verify the mediating effect between stressors and mental distress with psychological resilience as the mediating variable. The significance of mediation effect was judged by checking whether the confidence intervals of 95% bootstrap repeated 5,000 times included zero. Finally, conditional indirect effects of COVID-19-related stressors on mental distress, which was mediated by psychological resilience and moderated by international assignment experience was tested. This moderated mediation model is based on Hayes's Model 58. Moreover, we conducted the simple slope test to determine how the international assignment experience moderates the relationship between COVID-19 related stressors and psychological resilience.

We randomly divided the sample into two parts. Half of the samples were used for exploratory factor analysis (EFA) without rotation. The results of Harman's single-factor test extracted ten factors with eigenvalues above one. Before the rotation the variance explained by the leading common factor was 37.34%, which was less than the 40% required by the critical criteria (60). Therefore, this study does not consider the influence of common method bias.

The other half of samples was conducted a confirmatory factor analysis in Amos 26.0. The expatriates' international assignments experience and stressors questionnaire answers objective facts and does not test an implicit variable. Therefore, the scale's content validity and discriminant validity are not tested. As shown in Table 3, all factor loadings were above 0.6 and significant, indicating that the measured item validity was acceptable. The composite reliability (CR) for each construct's was >0.7, indicating that CR was acceptable. Moreover, each construct's average variance extracted (AVE) is more significant than 0.5, indicating that convergence validity was acceptable. In addition, the square root of AVE of each construct is larger than the correlation coefficient between any two constructs. This shows that the discrimination of each structure is significant. Therefore, the validity of each measure is acceptable. And the RMSEA value of the model is 0.062 (< 0.01), CFI value is 0.937 (>0.09), and NFI value is 0.933 (>0.09). Thus, the model fit is acceptable.

Means, standard deviations, and correlations of the main study variables were presented in Table 4. Correlation analyses showed that expatriates' psychological resilience is significantly associated with stressors (r = −0.374, p < 0.01), PHQ (r = −0.665, p < 0.01), GAD (r = −0.619, p < 0.01), and PSS (r = −0.510, p < 0.01). This satisfied the prerequisites of mediation analysis.

For the control variables, age correlated significantly with IAE (r = 0.423, p < 0.01), PR (r = 0.120, p < 0.01), PHQ (r = −0.155, p < 0.01), GAD (r = −0.118, p < 0.01), and PSS (r = −0.055, p < 0.01); gender correlated significantly with IAE (r = −0.086, p < 0.01) and PR (r = 0.044, p < 0.05). In the tests of relevant hypotheses, age and gender were controlled (61).

According to the correlation analysis results, they meet the conditions for establishing an intermediary relationship between the three factors. The mediating model (Hypothesis 1) was tested by Model 4 of PROCESS. The results in Figure 2 show that stressors significantly negatively predicted psychological resilience (t = −1.601, p < 0.001). When stressors and psychological resilience were considered in the regression equation, psychological resilience was significantly negative with PHQ (t = −0.569, p < 0.001). The indirect effect was tested by 5,000 resampling bootstraps. Results in Table 5 showed the mediation effect of psychological resilience between COVID-19 stressors and PHQ is significant (Effect = 0.911, SE = 0.056, 95% boot CI = [0.803, 1.022]). The mediation effect's 95% bootstrap confidence interval (CI) did not contain zero, and the indirect effect accounted for 63.26% of the total effect. Therefore, psychological resilience mediated the relationship between stressors and PHQ. Thus, Hypothesis 1a was supported.

Also, psychological resilience significantly negatively predicted GAD (t = −0.633, p < 0.001). Results in Table 5 showed the mediation effect of psychological resilience between COVID-19 stressors and GAD is significant (Effect = 1.013, SE = 0.059, 95% boot CI = [0.898, 1.128]). The mediation effect's 95% bootstrap confidence interval (CI) did not contain zero, and the indirect effect accounted for 57.66% of the total effect. Therefore, psychological resilience mediated the relationship between stressors and GAD. Thus, Hypothesis 1b was supported.

Finally, psychological resilience significantly negatively predicted PSS (t = −0.473, p < 0.001). Results in Table 5 showed the mediation effect of psychological resilience between COVID-19 stressors and PSS is significant (Effect = 0.757, SE = 0.052, 95% boot CI = [0.660, 0.864]). The mediation effect's 95% bootstrap confidence interval (CI) did not contain zero, and the indirect effect accounted for 53.05% of the total effect. Therefore, psychological resilience mediated the relationship between stressors and PSS. Thus, Hypothesis 1c was supported. The effect value of each path is shown in Figure 2.

Moderating effects are hypothesized at two stages of the mediation relationship. The first stage is the effect of stressors on psychological resilience. International assignment experience moderates the relationship between stressors during COVID-19 and psychological resilience. When expatriates' international assignment experience is more prosperous, the negative effect of stressors during COVID-19 on expatriates' psychological resilience is weaker. The second stage is the effect of psychological resilience on mental health outcomes. International assignment experience moderates the indirect effect of stressors during COVID-19 on expatriates' mental health via psychological resilience. When expatriates' international assignment experience is less prosperous, the negative effect of expatriates' psychological resilience on mental health is more robust. PROCESS model 58 was conducted to test Hypotheses 2, 3a, 3b, and 3c, respectively.

According to Table 6, it was revealed that effect of stress events during COVID-19 on psychological resilience was significant (Effect = −1.587, SE = 0.071, 95% bootstrap CI = [−1.726, −1.448]), and more importantly, the effect was significantly moderated by expatriates' international assignments experience (Effect = 0.213, SE = 0.057, 95% bootstrap CI = [0.101, 0.324]). For clarity, we plotted stressors during COVID-19 on psychological resilience (Figure 3), separately for low and high stress (Mean-SD and Mean + SD, respectively). Simple slope tests indicate that for expatriates who have high international assignments experience, psychological resilience level is significantly descended as the increase of stressors during COVID-19 (Effect = −1.323, SE = 0.100, p < 0.001); for expatriates have low international assignments experience, psychological resilience level is also significantly descended as the increase of stressors during COVID-19 (Effect = −1.851, SE = 0.100, p < 0.001). Moderated by experience, the slope of the effect of stressors on psychological resilience changed from −1.851 to −1.323, reduced by about 0.5. It can be revealed that experience is a buffering factor when dealing with stressors during COVID-19, the psychological resilience of expatriates decreases slowly compared with those who have less experience (Figure 3). Thus, H2 was supported.

As shown in Table 7, the effect of psychological resilience on PHQ was significant when international assignments experience was high (Effect = −0.517, SE = 0.019, 95% boot CI = [−0.555, −0.479]) and low (Effect = −0.599, SE = 0.016, 95% boot CI = [−0.631, −0.567]). And, the of psychological resilience on GAD was significant when international assignments experience was high (Effect = −0.583, SE = 0.025, 95% boot CI = [−0.631, −0.535]) and low (Effect = −0.659, SE = 0.021, 95% boot CI = [−0.700, −0.618]). Also, the effect of psychological resilience on PSS was significant when international assignments experience was high (Effect = −0.403, SE = 0.025, 95% boot CI = [−0.452, −0.354]) and low (Effect = −0.527, SE = 0.021, 95% boot CI = [−0.569, −0.486]). Thus, it can be concluded from the simple slopes that expatriates' international assignments experience attenuated the effect of psychological resilience on mental health (Figures 4–6). Moderated by experience, the slope of the effect of psychological resilience on mental health outcomes reduced by about 0.1. This indicated that experience can be regarded as a buffering factor for individual psychological resilience of expatriates to regulate stress. Thus, H3a, H3b, and H3c were supported. While, the results indicate the buffering effect of experience differs in the two stages. It is more significant between stressors and psychological resilience.

Consistent with our expectation (Hypothesis 1), stressors during COVID-19 positively predicted PHQ, GAD, and PSS scores. Psychological resilience was a mediation factor in this relation, extending previous theory and empirical research. Previous researches showed that the COVID-19 pandemic had significantly increased the depression, anxiety, and stress perceived level (12, 42, 62). Our research findings also confirmed the positive association between stressors and mental distress. Moreover, other than directly affecting depression, anxiety, and stress perceived level, stressors during COVID-19 indirectly affect these three variables. This mediation model suggests a possible reason why the more stressors, the more prone to depression, anxiety, and stress may be beyond their resilience to deal with adversity.

Specifically, it advances our understanding of psychological resilience by applying the negative outcome of COVID-19-related stressors into depression, anxiety, and perceived stress. Although the antecedents and consequences of psychological resilience have been verified with various studies (45, 48, 63), there is no relevant evidence in the expatriates of multinational companies. Consistent with the results of Rossi's et al. (12) research, psychological resilience plays a mediation role between stressors during COVID-19 and mental health. After expatriates encounter COVID-19 related stressors, individual protective factors will play a role in helping expatriates recover from adversity. However, when the resilience is poor, the psychological state of expatriates will change and threaten their mental health. The mental health problems of expatriates in multinational enterprises are more prominent than their domestic employees (64), especially in the face of the outbreak and persistence of COVID-19. In our survey, 70% of the respondents were delayed in their return.

In addition, the two stages of the mediation process will be discussed separately. Consistent with previous reports, we find that COVID-19-related stressors decrease psychological resilience scores (65). According to the standard definition of psychological resilience, individuals experiencing stressors can inspire one's ability to cope with adversities (66). According to resource conservation theory (67, 68), psychological resilience is a positive conservation resource for individuals (69). The more the external stressors, the more resources the individual needs to recover from this state with a longer recovery process. The ability to recover is also reduced (68). Besides, psychological resilience scores are correlated with depression, anxiety, and stress perceived level. It indicates that higher psychological resilience can buffer the effect of COVID-19 related stressors on such depression, anxiety, and stress perceived (70). These researches are consistent with our results that people with a high level of psychological resilience have a more vital ability to recover from adversity. The probability of psychological problems will decrease, and their mental state will be healthier.

Compared with Rossi et al.'s (12) research, in the expatriates' group, the path influence coefficient of stressors during COVID-19 on individual psychological resilience is significantly higher than that of the general group. The path coefficient of COVID-19-related stressors on depression, anxiety, and stress perception is also significantly higher than ordinary people.

The moderating effect on the two stages in our mediation model is discussed separately.

As predicted in our hypothesis 2, expatriates' international assignments experience moderates the association between stressors during COVID-19 and psychological resilience. Specifically, the negative predicting role of stressors during COVID-19 on psychological resilience is significant slowdown among those with richer international assignments experience; in contrast, for those with less international assignments experience, the relation between stressors during COVID-19 and psychological resilience is stronger. This is consistent with the point of view of resource conservation theory (67). Expatriate experience is an essential resource for individual international construction expatriates. When employees encounter stressors, the more experienced one can fully mobilize their resources (71). It is easier for them to find ways and attitudes to deal with the incident from their experience, thereby improving their ability to return to normal. Nevertheless, international construction is a highly uncertain environment (72); especially during the outbreak and duration of the COVID-19 pandemic, expatriates' feelings about home are amplified (73). This is also related to the feelings toward home in Chinese traditional culture (74). However, the continuation of the epidemic has made the road home extremely difficult. At this time, international assignments experience can fully adjust the ability of expatriates to regulate stressors and quickly recover to a well-adjusted state. This suggests that experience is one of the buffering factors for individual psychological resilience of expatriates to regulate stress.

As predicted in our hypothesis 3, expatriates' international assignments experience moderates the second mediation path from psychological resilience to PHQ, GAD, and PSS. The moderating role of expatriates' international assignments experience can be explained by the combined action of expatriates' international assignments experience and psychological resilience on PHQ, GAD, and PSS (60). Rossi et al. (12) find that age plays a moderating role between resilience, depression, and anxiety. Generally speaking, older people have more work experience. Expatriates with high psychological resilience are less likely to have depression and other emotions (75) because they have a more vital ability to adjust themselves to external events. This phenomenon is evident among experienced expatriates. These people can find similar experiences from their existing experiences and draw inferences from one instance (76). They know better how to improve their ability to adapt to crises to avoid unhealthy psychological states such as depression and anxiety. At the same time, the existing experience allows them to know the possible results after the occurrence of stressors to reduce their fear of the unknown future, which is very important for emotional stability and mental health.

When comparing the moderating effect in the two stages, it is more significant between stressors and psychological resilience. This result indicate that psychological resilience can be affect significantly by individual resources. Expatriates can take more measures to deal with adversities and adapt to health status. However, once the level of psychological resilience decreases and causes mental health problems, the moderating effect of experience is less significant.

Our research has practical value for international construction enterprises and managers of international construction projects, including three aspects. First, managers accurately understand the psychological health of expatriates under the normalization of COVID-19 and formulate intervention and help plans in time. Secondly, managers are more transparent that priority should be given to those with rich expatriate experience when selecting expatriates. In emergencies, experienced people can do an excellent job in self-psychological adjustment. Finally, managers should pay more attention to the psychological resilience of expatriates, which negatively affects the psychologically unhealthy states of expatriates, such as depression, anxiety, and stress. The individual's psychological resilience can be improved. Managers can prioritize those with high psychological resilience to work abroad and formulate improvement plans to help expatriates.

There are several limitations to this study. First, due to the limitations of cross-sectional data, the results of this study have limitations. The follow-up research can increase the data collection of profile and psychological experiment. Second, this study makes a subjective evaluation of self-perception. Although this method is widely used in a large number of studies, there may be a deviation between participants' self-perception and the actual state. In the future, physical measurement tools can be used to assist in the evaluation of psychological and emotional states. Finally, all the participants are Chinese. The cultural background of different countries may lead to differences in their psychological resilience. At the same time, the social environment of the host country may also have effect on physical and mental health of expatriates. Therefore, our findings may not be applicable to expatriates from other cultural backgrounds. In future research, we will cooperate with more international contractors to collect data to verify the research results and take specific measures to improve the psychological resilience of expatriates.