Affective commitment is the core of organizational commitments (Mercurio, 2015; Noraazian and Khalip, 2016), it refers to employees’ (emotional) dependence on the organization, and the extent to which employees have the same values and objectives as the organization leads (Haider et al., 2019). Employees high in affective commitment believe in and accept the organization’s goals and values, intend to stay with the organization in a long-term, and commit to providing quality service on behalf of the organization (Shao et al., 2022).

Previous research reveals that servant leadership plays an important role in retaining employees’ affective commitment among various challenges (He et al., 2008; Mercurio, 2015; Goestjahjanti et al., 2022). More importantly, social exchange theory (SET) proposes that social exchange processes involve a series of interdependent and contingent interactions that evokes workplace obligation (Blau, 1974; Richard and Emerson, 1976), for which leadership plays an important role (Liden et al., 1997).

Servant leaders care for their followers’ personal needs and interests, and focus on the benefits of organizations and communities rather than their own interests (Eva et al., 2019). Compared to others (e.g., transformational leadership), this type of leadership is more likely to create social exchange relationships between leaders and their followers. According to SET, it is expected that when servant leaders provided employees with good working atmospheres, opportunities to improve professional skills, emotional support in need, in exchange, employees might develop a sense of identity and belonging to the organization (Cropanzano and Mitchell, 2005; Dahleez et al., 2021). Hence, it was hypothesized that servant leadership positively affected affective commitment (H1). In addition, this positive association could be mediated by (increasing) psychological safety and/or (decreasing) job burnout in workplace. Rationales were articulated below.

Psychological safety refers to the belief that workplace is safe for interpersonal risk-taking such that presenting of self-image, career, or status is free from the fear of adverse effects (Edmondson, 1999; Frazier et al., 2017). When employees have a high level of psychological safety, they feel free to speak their minds, actively exchange work-related knowledge and information, and are energized, creative, and effective in performance even in rapidly changing environments such as healthcare revolution (Edmondson and Lei, 2014).

Servant leadership has a positive effect on the attitudes and behaviors of employees, with leaders putting employees’ interests first, helping them develop and grow, and building good relationships with their followers. The recognition of being cared for, respected, and helped by their leaders might also grant the psychological safety to take challenges, seize opportunities, even taking risks to introduce different ideas; should there be any concern for adverse consequences related, employees are less fearful as they trust their servant leaders to “get their back” (Yan and Xiao, 2016). Therefore, it was presumed that servant leadership could increase employees’ psychological safety. According to SET, exchange relationships are characterized by the mutual caring of both parties’ interests (cf. Clark). With psychological safety granted, employees would be more willingly in taking responsibilities, stepping-out of comfort zones, and even bringing-about beneficial changes in the organization (Edmondson and Lei, 2014). In other words, employees who felt high in psychological safety in workplace would trust, identify, and dedicate more to the organization, resulting in both personal and organizational growth that could reinforce their affective bonding to the organization (Chandrasekaran and Mishra, 2012; Frazier et al., 2017; Li et al., 2022). Therefore, it was hypothesized that servant leadership could boost affective commitment by enhancing the experience of psychological safety (H2).

Job burnout is a progressive psychological response to chronic job-related stress and includes three components: emotional exhaustion, which refers to the excessive depletion of one’s emotional resources, resulting in feelings of emotional and physical burden and strain; depersonalization, which refers to a decreased ability to empathize with clients, and a cold, overly distant attitude; and a decreased sense of personal accomplishment, which refers to a loss of job satisfaction and competency (Lombardero-Posada et al., 2022; Yıldırım et al., 2023). Healthcare providers are exposed to many stressors including but not limited to challenges of clinical work, competing demands, conflicting roles, and relationships with leadership (Demin et al., 2017), which resulted in high prevalence of job burnout in this industry (Hall et al., 2016; Chirico et al., 2023). Especially with the shortage of human resources after the global Covid-19 pandemic, occupational hardship involving burnout, anxiety, and lack of career-related psychological security had become more and more prevalent(Chirico et al., 2022).

Studies have shown that servant leadership is negatively associated with job burnout, especially among medical staffs (Cai et al., 2013; Harju et al., 2018; Ma et al., 2021). Servant leadership is very likely to minimize the chances of having emotional exhaustion, depersonalization, or lack of personal accomplishment, because leaders of this type of care not only the performance but also the well-being of their followers’. More importantly, employees with less job burnout have more resources to reward the organization, such as passionate, devoted, and skilled workplace performance. In such healthy social exchange processes, chances are high that employees would like to serve the organization in long-term, as working here becomes rewarding and enjoyable (Blau, 1974; Richard and Emerson, 1976). In other words, employees with servant leaders are less likely to have their organizational commitment jeopardized by the violation of their identification to the organization (Spence Laschinger et al., 2009). Therefore, it was hypothesized that servant leadership could boost affective commitment by reducing the experience of job burnout (H3).

Finally, job burnout could be a function of age or gender given employees at different life stages or in different family roles could face various life work conflicts, psychological safety could be a function of profession or career stage due to different organizational roles (Artz et al., 2022). Possibilities that the proposed mediation effects could be conditioned by demographic and/or professional characteristics could not be rule out in advance. Therefore, this study intended to explore whether the proposed model differed at demographic and/or professional features.

The objectives of the current study were to (a) examine the proposed parallel mediation effects of psychological safety and job burnout on the relationship from servant leadership to employee’s affective commitment; and (b) to explore the potential differences of the above associations among different demographic and professional characteristics. In short, this study served as a possible extension of the social exchange theory, such that it examined whether servant leadership could evoke beneficial exchange processes via increasing psychological safety (the willingness to exchange) and decreasing job burnout (the resources to exchange).

A convenience sampling was conducted at a local comprehensive, second-grade hospital in Shanghai, China November 2021. This hospital consisted of various clinical departments¹ it was typically representative of the general hospitals of its level in China. All staffs (n = 1925) were invited to participate, and a total of 931 valid responses (M_age = 35.58 years, SD = 9.26; 80.42% were female) were collected. All participants provided informed consent.

This Research was approved by the Research Ethics Committee of Shanghai Jiao Tong University School of Medicine (reference ID: SJTUPN-202202). Hospital department heads were contacted in advance to introduce the purpose and procedure of this study. They were also instructed to encourage employees to participate in this anonymous, online survey distributed via a mature Chinese online survey platform (Wenjuanxing, Changsha, China). Multiple procedures were adopted to prevent common method bias frequently observed in self-report surveys, including, (a) the randomization of question orders, and (b) the insertion of reverse coded questions. In addition, the response window was specified as 3 days after the release of the survey link, all respondents were encouraged to response in a one-time and in privacy. Hospital staffs were further required to not to discuss the study or their own responses with each other until the study ended.

Well established psychological scales were adopted to measure servant leadership, psychological safety, job burnout, and affective commitment. The original English versions were translated into Chinese following standard procedures of translation and back-translation. Wording was modified to fit the Chinese healthcare environment, when necessary. All translated scales were tested for content and construct validity, and three items were removed for low factor loadings and cross-loadings on multiple dimensions (see Supplementary Table 1 for details). For each construct, a composite score was created by averaging all items such that higher scores indicate higher levels of that construct.

All analyses were conducted in SPSS 26.0, Jamovi 2.3.21, and SPSS AMOS 26.0. Psychometric properties of the translated scales were examined before hypothesis testing. Reliability was examined via internal consistency, and a McDonald’s ω > 0.700 was considered acceptable (Trizano-Hermosilla and Alvarado, 2016). Validity was examined via confirmatory factor analysis (CFA). Scales were individually analyzed according to theoretical structures, and model fits and factor loadings were evaluated in conjunction to assess structural validities against pre-determined criteria (namely, goodness-of-fit index (GFI), comparative fit index (CFI), and Tucker-Lewis index (TLI) > 0.900; χ²/df < 3, and root mean square error of approximation (RMSEA) < 0.080) (Lt and Bentler, 1996; Schermelleh-Engel et al., 2003). In addition, scales were modeled together upon their own factor models to examine the convergent and discriminate validities via average variance extracted (Fornell and Larcker, 1981) [AVE; which was expected to be larger than the correlation coefficients among the corresponding factors and composite reliability (CR > 0.700) (Peterson and Kim, 2013)]. Harman’s single-factor test was used to probe common method bias (Podsakoff et al., 2003; Tehseen et al., 2017).

For hypothesis testing, the proposed parallel mediation effects were examined using SPSS PROCESS macros 22.0 (Model 4; bootstrap samples = 5,000) (Hayes, 2017). Variance inflation factor (VIF) was adopted to evaluate multicollinearity, and given the cross-sectional design, a relatively strict cut-off value was determined prior to data analysis such that a VIF smaller than 2.5 was considered acceptable.

Mediation analysis was performed to examine whether staff members’ psychological safety and job burnout (in parallel) mediated the effect of employees’ perceived servant leadership on their affective commitment, with shift type and income level as control variables (Table 3). VIFs were below 2.00 for all independent variables, suggesting no substantial multicollinearity. Servant leadership was positively associated with affective commitment while psychological safety and job burnout simultaneously and partially mediated the above association. Specifically, after accounting for the indirect effects via psychological safety and job burnout, the direct effect remained significant. Together, this mediation effect accounted for 61.055% of the total effect from servant leadership to affective commitment. A further contrast of the two indirect effects suggested that the mediation effect via job burnout was stronger than the effect via psychological safety (Figure 2).

Multi-group analyses were conducted to examine whether the above mediation model held equivalently for (a) male versus female participants, (b) permanent staff versus temporary staff participants.³ Critical ratios for parameter differences were adopted for path comparisons. Results revealed that the negative effect of servant leadership on job burnout, the negative effect of job burnout on affective commitment, and the positive effect of psychological safety on affective commitment were greater in terms of the magnitudes for temporary staffs than for permanent staffs (Figure 3). However, the two mediation effects were not significantly different between the two groups (Bs < −0.002, p > 0.071). And accordingly, the total indirect effect was not significantly different (B = −0.162, p = 0.132, 95% CI = [−0.385, 0.047]) between the two groups, either. No gender difference was found across all pairs of paths (see Supplementary Table 2 for details).

Previous research on servant leadership mainly focused on the antecedences and consequences, little research attention was directed to the mechanisms why and how serving helps leading. As researchers pointed out, this field would benefit from in-depth examination of the mechanism(s) associated with servant leadership (Eva et al., 2019). In responding to which, this study examined and revealed that serving others’ needs brought benefits via not only organizational psychological climate but also followers’ affective well-being. Additionally, this study contributed to existing literature by providing non-western data for the examination of servant leadership.

A distinguishable feature of this study was that it examined behavioral tendencies (servant leadership and affective commitment) instead of concrete behaviors in social exchange processes (Cook et al., 2013). From leaders’ end, a servant orientation provides an umbrella test of resources such as love (i.e., an expression of affectionate regard, warmth, or comfort), status (i.e., an evaluative judgment that conveys prestige, regard, or esteem), and information (i.e., advice, opinions, instruction, or enlightenment) among the six types of critical resources proposed by the resource theory of social exchange (Tolin et al., 2003). From subordinates’ end, affective commitment is one of the socioemotional outcomes that address’s one’s social and esteem needs and conveys a sense of being valued and/or treated with dignity (Shore et al., 2001; Cook et al., 2013). While concrete behaviors (e.g., short-term monetary reward or cooperation) might be more convenient to measure, behavioral tendencies granted more generalizability in research findings. More importantly, the theoretical model tested here shed more light on those long-term, socioemotional outcomes rather than short-term, tangible ones, which were more important for the organizational development (Cook et al., 2013). Results added new empirical evidence of social exchange theory in leadership research. Using psychological safety and job burnout as proxies, results also highlighted that servant leadership contributed to beneficial social exchange processes via the securing of both willingness and resources from employee’s perspective.

Based on the results of this study, some practical implications are suggested. First, psychological safety and burnout significantly mediated the relationship between servant leadership and affective commitment, suggesting that managers need to appreciate how their servant leadership behaviors affect employees’ affective commitment. It is essential to clearly understand that servant leadership is an important part of employees’ increased affective commitment. The more followers perceive servant behavior, the more likely they are to commit to staying with the organization. In addition, this study shows that influencing affective commitment through burnout is stronger than psychological safety, suggesting that managers should prioritize reducing employee burnout. Finally, the mediation model holds across groups, suggesting that employees of different genders and employment forms can increase their affective commitment through servant leadership in the hospital context.

Several limitations need consideration. First, the cross-sectional design limited the examination of the causal relationships among the investigated variables, longitudinal studies are recommended to further the understanding of the mechanisms revealed here. Second, this study focused on the effects of servant leadership behaviors on individual-level mediators, while team-level mediators might play important roles as well, e.g., collaborative team atmosphere inspired by servant leadership’s other-interests orientation. Future studies might consider integrating mechanisms at both individual- and team- levels. Thirdly, authors encouraged some replication of the current findings in hospitals with different specialties and scales, as well as in other serving organizations to examine the generalizability of the theoretical model. Fourthly, though psychological safety and job burnout could proximate well the willingness and resources to exchange, authors recommended future research to adopt other variables (e.g., citizenship behaviors) for conceptual replication. Finally, this study focused on the perceived servant leadership, future studies could further examine actual levels (e.g., leaders’ self-report) to provide a comprehensive picture as well.