Abstract
This study aimed to explore predictors of Chinese university faculty’s occupational well-being in the reshaped work environment. To achieve this aim, the job demands-resources model was utilized to test the relationships of job demands (work–family conflict) and job resources (leader support) to occupational well-being among 375 university faculty (145 males and 230 females) at a comprehensive research university in China. We further intended to extend the theory by incorporating personal demands (the perfectionism personality) within the research model. Results indicated work–family conflict was indirectly related to job satisfaction via the mediator of emotional exhaustion. Leader support was indirectly related to job satisfaction via the mediators of emotional exhaustion and work engagement. The two perfectionism dimensions (concerns and strivings) functioned differently in the model. The concerns dimension positively predicted exhaustion but was nonsignificant for engagement. By contrast, the strivings dimension positively predicted engagement but was nonsignificant for exhaustion.
Keywords
During the past 2 decades, the Chinese higher education landscape has been reshaped due in part to large-scale massification, intensified internationalization, excessive workloads, and growing emphasis on teaching and research performance (Guo et al., 2019; Xu, 2019). Particularly, the Chinese government implemented in 2017 a higher education reform, called the “double first-class initiative,” with 42 elite universities selected into the “world-class university” initiative and another 95 universities selected into the “world-class discipline” initiative. One main feature of this national reform is the dynamic access mechanism based on the evaluation organized every 5 years (Peters & Besley, 2018). Based on the evaluation, all Chinese universities and their disciplines have chances of being selected into the initiative if they achieve high-level development and meet world-class standards. This reform can potentially promote the quality and international competitiveness of Chinese higher education. Meanwhile, it may impose extra work demands and pressures on faculty members because the universities already in the initiative and many others striving to enter the initiative have to compete with each other and thus require their faculty to be more productive and innovative in teaching and scientific research (Q. Liu et al., 2019).
The new work environment may cause some negative consequences for university faculty since high job demands can be associated with poor mental and physical health (Minnotte & Yucel, 2018) and reduced job satisfaction (Yeh, 2015). Hence, it can be vital to examine the relationships between factors of work environment and occupational well-being among Chinese university faculty. Situated in the job demands-resources model (JD-R; Bakker & Demerouti, 2007; Demerouti et al., 2001), our study has two main objectives. First, applying the JD-R, we examine roles of job demands (i.e., work–family conflict) and job resources (i.e., leader support) in predicting different aspects of occupational well-being: work-related emotional exhaustion (negative emotion), work engagement (positive emotion), and job satisfaction (cognitive evaluation; Han et al., 2020; Mudrak et al., 2018). Second, extending the JD-R, we intend to be the first to include one specific type of personal demand (the perfectionism personality) in the theory and examine its relationships with occupational well-being above work–family conflict and leader support.
Theoretical Framework and Empirical Evidence
JD-R and the intended extension by this study
The JD-R has acted as a highly influential theory for researching antecedents of employee well-being due to its flexibility and broad scope. For instance, some studies under its guidance contribute to our understanding of occupational well-being of university faculty by identifying teaching–research conflict, job insecurity, interpersonal conflicts, social support, and job autonomy as its predictors (Mudrak et al., 2018; Torp et al., 2018; Xu, 2019).
The JD-R distinguishes two broad categories of job characteristics (job demands and job resources) and proposes that they can activate two processes (the health impairment process and the motivational process; Bakker & Demerouti, 2007; Schaufeli & Taris, 2014). Job demands are defined as various aspects of a job requiring persistent effort and thus associated with physiological and psychological costs (Demerouti et al., 2001). Examples include job role conflict, job insecurity, and excessive workload. High job demands may instigate the health impairment process, causing health-related symptoms (i.e., emotional exhaustion, burnout, and strain). The underlying reason is that when dealing with challenging job demands, employees often make sustained efforts and experience high pressures that exhaust them physically or mentally over time, thus possibly causing health problems (Bakker & Demerouti, 2007). This argument has been empirically supported by both cross-sectional (Mudrak et al., 2018; Xu, 2019) and longitudinal studies (Prieto et al., 2008). Noteworthy is that the link between job demands and the motivation process (i.e., work engagement or motivation) is not hypothesized in the JD-R because this link may depend on the personal characteristics of employees and features of job demands (Schaufeli & Taris, 2014).
Job resources are defined as factors that function in facilitating achieving work goals and reducing physiological and psychological costs (Demerouti et al., 2001). Examples include social support, job autonomy, and performance feedback. According to the definition, job resources may instigate both the motivational process and the health impairment process (Schaufeli & Taris, 2014). The underlying reason is that when employees have adequate job resources, they tend to possess stronger motivation to set higher work goals and pursue better job performance. When employees lack necessary job resources, they may feel overburdened or become recovered more slowly from the energy depletion, which leads to stress, burnout, and health problems. In other words, job resources may positively predict work motivation or engagement and negatively predict exhaustion or burnout. In addition, the JD-R proposes that exhaustion (or burnout) stemming from high job demands and/or low job resources negatively predicts job satisfaction or performance due to the gradual draining of mental resources or withdrawal behaviors; however, work engagement (or motivation) stemming from high job resources positively predicts job satisfaction or performance due to the willingness to dedicate efforts and abilities to work tasks (Schaufeli & Taris, 2014). This proposition also offers mediation perspectives: Exhaustion (or burnout) and work engagement (or motivation) may mediate the relationships of job demands and resources to job satisfaction or performance (Bakker & Demerouti, 2007; Schaufeli & Taris, 2014).
Despite the high popularity of the JD-R, Bakker and Demerouti (2017) emphasize that there is still scope to expand the theory by outlining some new issues for future investigation. They recommend research to include personal demands as extra predictors of employee well-being within the JD-R. To date, such empirical evidence is scarcely available in the context of the JD-R. Personal demands are defined as “the requirements that individuals set for their own performance and behavior that force them to invest effort in their work and are therefore associated with physical and psychological costs” (Barbier et al., 2013, p. 751). Examples include workaholism, goal setting, and perfectionism (Bakker & Demerouti, 2017; Prieto et al., 2008). In our study, the personality of perfectionism is selected for two reasons. First, Ozbilir et al. (2015) and Stoeber and Damian (2016) recommend more research on the predictive roles of the perfectionism trait in the workplace. Furthermore, Rice et al. (2016) also note that the effects of perfectionism are not well understood in academic settings (e.g., students, administrators, and teachers/professors). Taken together, there is an obvious need of exploring the perfectionism trait among university faculty. Second, compared to other personal demands, perfectionism is a more complex factor including both adaptive (e.g., concerns over making mistakes and fear of negative evaluation) and maladaptive (e.g., striving for perfection and setting high standards) aspects (Stoeber & Damian, 2016). Thus, examining this factor in the JD-R may add more insightful knowledge to the occupational psychology literature. Theoretically inspired by these arguments, our study tests how this enduring personality trait can be related to the dual processes within the JD-R.
Work–family conflict and occupational well-being
Due to time and energy constraints, many working individuals face problems of balancing the dual roles in work and family domains. Stressors or negative moods at work may spill over into family domains, thus resulting in work–family conflict (Minnotte & Yucel, 2018), defined as “a form of inter-role conflict in which the role pressures from the work and family domains are mutually incompatible in some respect” (Greenhaus & Beutell, 1985, p. 77). Following the JD-R, work–family conflict is treated in our study as job demands that may be related to emotional exhaustion and job satisfaction. Senécal et al. (2001) found that work–family conflict increased emotional exhaustion among a sample of French Canadians. Utilizing conservation of resources theory, Ito and Brotheridge (2012) revealed that work–family conflict positively predicted emotional exhaustion and negatively predicted job satisfaction among Canadian employees. Positive linkages between work–family conflict and poor mental/physical health were also reported among Americans (Minnotte & Yucel, 2018). In European contexts, Italian workers’ work–family conflict was positively associated with emotional exhaustion (Vignoli et al., 2016). Utilizing the JD-R, Mudrak et al. (2018) found that work–family conflict was indirectly related to poor job satisfaction via job stress among Czech academics.
Leader support and occupational well-being
As a specific type of environmental support, leader support is conceptualized as the extent to which Chinese university faculty can receive work-related support from their supervisors. Selecting leader support as job resources can be particularly important in Chinese high power distance cultures characterized by high levels of societal acceptance of inequality in power and wealth distribution (Hofstede, 2001). In the workplace, employees from such societies highly value how they are treated by authority figures (e.g., receiving support from supervisors; Tyler et al., 2000). Since interpersonal relationships between supervisors and subordinates are culturally bound, the workplace in China can be an ideal context to demonstrate the impact of leader support on occupational well-being. Nonetheless, its impact on occupational well-being in Chinese sociocultural contexts is still underresearched, particularly within the JD-R. Li et al. (2020) can somewhat support our argument by targeting Chinese employees and finding that leader support can be a driving force for self-management motivation and behaviors. Charoensukmongkol et al. (2016) revealed that leader support positively predicted job satisfaction and negatively predicted all three dimensions of job burnout measured by the Maslach Burnout Inventory (MBI; Maslach et al., 1996) among American university faculty. According to Suan and Nasurdin (2016), employees with adequate leader support can have more work-related resources, motivating them to be more engaged, set higher work goals, and strive for better performance.
Perfectionism and occupational well-being
As noted earlier, we intended to include perfectionism within the JD-R as personal demand and investigate its relationship with occupational well-being. Perfectionism is a personality trait describing individuals who pursue flawlessness and perfection, set high-performance standards, and possess tendencies for overly critical evaluation (Stoeber & Damian, 2016). Researchers mainly differentiate its two distinctive dimensions: perfectionistic concerns and perfectionistic strivings (Flaxman et al., 2018; Ozbilir et al., 2015; Stoeber & Damian, 2016). Perfectionistic concerns are regarded as maladaptive due to its features of self-criticism and excessive worries about potential mistakes and failures (Frost et al., 1990). Individuals high in this dimension tend to suffer from some health-related problems (e.g., obsessive-compulsive disorder, emotional conflict, and depression; Frost et al., 1990) and display some affective and behavioral features (e.g., undue worries about mistakes and failures, fear of negative evaluations by others, and negative reactions to imperfection; Stoeber & Damian, 2016). Perfectionistic strivings are relatively adaptive for being featured by strivings for typically self-imposed and high personal standards, thus possibly energizing and activating individuals’ self-monitoring motivation to meet the standards (Damian et al., 2017; Flaxman et al., 2018).
The studies examining the roles of perfectionism in the Chinese higher education environment were mostly conducted among Chinese university students. For instance, Y. Zhang et al. (2007) found that perfectionistic concerns were positively associated with academic burnout, and perfectionistic strivings were positively associated with academic engagement. In another instance, Smith et al. (2015) differentiated functional roles of both perfectionistic concerns and strivings in predicting negative emotionality (i.e., depression, anxiety, and stress) between Canadian and Chinese university students. Using a multigroup analysis, the authors revealed that perfectionistic strivings were negatively related to negative emotionality for the Canadian group, but this relationship was nonsignificant for the Chinese group. Perfectionistic concerns were positively related to negative emotionality for both groups. In contrast, an extensive review of the literature indicated that there seemed very few studies examining Chinese university faculty’s perfectionism and its impact.
Therefore, we have to rely on other samples to support our hypothesized relationships of perfectionism with work engagement, emotional exhaustion, and job satisfaction. Although work engagement has attracted much scholarly attention in the past decade, very little research examines the relationship between perfectionism and work engagement. Stoeber and Damian (2016) summarized that there may be only four articles that had investigated this relationship by then. All four studies (Childs & Stoeber, 2010; Ozbilir et al., 2015; Tziner & Tanami, 2013; Wojdylo et al., 2013) found perfectionistic strivings to be positively associated with work engagement among employees from various work sectors. By contrast, only Ozbilir et al. (2015) established the negative relationship between perfectionistic concerns and work engagement, while the other three found no relationship between them. In more recent years, Flaxman et al. (2018) offered extra support for Ozbilir et al.’s (2015) finding by revealing that perfectionistic concerns can be negatively and indirectly related to work engagement via the mediator of work-related worry. In our work, it is assumed that perfectionistic concerns negatively predict engagement because this dimension entails avoidance-oriented motivation (Flaxman et al., 2018). Additionally, both concerns and strivings dimensions were found to be positively associated with emotional exhaustion (Lee & Anderman, 2020; Mitchelson & Burns, 1998). As for the relationship between perfectionism and satisfaction, Lee and Anderman (2020) found that perfectionistic concerns negatively predicted school satisfaction among American undergraduates. Similarly, Mitchelson and Burns (1998) revealed the negative relationship between perfectionistic concerns and job satisfaction, and Hochwarter and Byrne (2010) revealed the positive relationship between perfectionistic strivings and job satisfaction.
Emotional exhaustion and work engagement as mediators
Sticking to the JD-R, our study examines whether work–family conflict (job demands) can be negatively related to job satisfaction via the mediator of emotional exhaustion and leader support (job resources) can be positively related to job satisfaction via the dual mediators of emotional exhaustion and work engagement. Work-related emotional exhaustion refers to a feeling of chronic fatigue and strain stemming from overtaxing work (Skaalvik & Skaalvik, 2011). Work engagement measures “a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption” (Schaufeli et al., 2002, p. 74). The two constructs have been evidenced as important antecedents of work attitudes and performance. For instance, highly exhausted teachers may shape negative attitudes toward the job, thus reducing job satisfaction (Skaalvik & Skaalvik, 2009). Yan et al. (2019) found that work engagement positively predicted job satisfaction because highly engaged employees tended to utilize various job resources and achieve work goals.
Additionally, prior studies have revealed mediating roles of the two constructs. Supportive evidence can be offered by Skaalvik and Skaalvik (2011) who found that emotional exhaustion mediated the relationships of time pressure and discipline problems (job demands) to job satisfaction among teacher samples. More relevant to our research objectives is the study by Karatepe (2013) in which exhaustion mediated the work–family conflict–job performance relationship. Researchers document that individuals with adequate job resources (e.g., social support) may have more psychological capital and positive work experiences, thus potentially reducing job stress and stimulating motivation/engagement (Schaufeli & Taris, 2014), which subsequently facilitates promoting satisfaction and performance (Karatepe, 2013; Skaalvik & Skaalvik, 2009). Supporting the argument, Mudrak et al. (2018) revealed job stress and teacher engagement as dual mediators linking leader support and job satisfaction.
Finally, we assume that the relationships between perfectionism and job satisfaction may be mediated by exhaustion and engagement. Based on the earlier discussion, there might be two distinct pathways. First, there might be positive links from the two perfectionism dimensions to exhaustion which might in turn be negatively linked to job satisfaction. Second, we expect a negative link from concerns to work engagement and a positive link from strivings to work engagement, and engagement might in turn be positively linked to job satisfaction. Although very few scholars explicitly examine the two pathways, we can rely on prior studies that focus on relevant personality traits or pathways to support our mediation assumptions. For instance, work engagement was found to mediate the relationship between proactive personality (i.e., actively manipulating work environment) and job satisfaction (Yan et al., 2019). Targeting university undergraduates, Miquelon et al. (2005) found that academic motivation mediated the relationship between perfectionism and psychological adjustment. Similarly, Gaudreau and Antl (2008) identified goal-achieving motivation and coping styles as chain mediators in the relationship between perfectionism and life satisfaction among French-Canadian athletes. In addition, emotional exhaustion was revealed as a mediator in the relationship between workaholism and psychological diseases measured by Goldberg’s (1992) General Health Questionnaire among Italian university faculty (Converso et al., 2019). Given a significant lack of explicit empirical evidence, we will offer research questions concerning mediating roles of exhaustion and engagement in the perfectionism–job satisfaction relationships rather than offer specific hypothesis.
The current study
Anchored in the JD-R, a conceptual model is outlined to present our research aims (see Figure. 1). Accordingly, the following hypotheses and research question centering on the mediated relationships of job demands, job resources, and personal demands to job satisfaction are formulated:
The conceptual research model. Note. The covariance lines among the exogenous variables are not drawn for model parsimony; “+” hypothesizes a direct and positive relationship; “-” hypothesizes a direct and negative relationship.
Method
Participants and Procedure
Participants of this study were faculty at a prominent and comprehensive research university in China, one of the forty-two in the “First-Class University” Initiative. It has typically highly demanding and insecure work contexts (e.g., research productivity and short-term contracts). Thus, this university is suitable for our study examining predictors of occupational well-being.
Data collection was assisted by several administrators at the university. They helped the authors distribute the online survey, confidentiality, voluntary participation, and selection criterion to potential participants through friend groups of a popular social network site which consist of almost all members at their respective faculties. The selection criterion mandated that participants must be teaching staff working full-time at the university. This procedure brought a final analysis sample of 375 participants who completed the online survey as well as the written informed consent. The participants’ average age was 39 years (SD = 7.2), 230 (61.3%) were female, and 145 (38.7%) were male. Of the total, 185 (49.3%) faculty reported to conduct research in social sciences and humanities and 190 (50.7%) in natural sciences and engineering. Finally, 84 (22.4%) faculty identified themselves as professors, 122 (32.5%) as associate professors, 148 (39.5%) as lecturers, and 21 (5.6%) as assistant professors.
Measures
Using the back-translation method, two bilingual experts in the occupational psychology research translated all measures from English to Chinese to avoid linguistic bias.
Work–family conflict
The Work–Family Conflict Scale (Netemeyer et al., 1996) was used to assess to what extent faculty perceived that fulfilling work roles interfered with family roles. Netemeyer et al. (1996) focused on three different samples and demonstrated the reliability (Cronbach’s αs were .88, .89, and .88, respectively, for the three samples) and the validity of the measure by conducting confirmative factor analysis. Che et al. (2017) applied this measure among Chinese full-time nurses and confirmed its good reliability and validity. It included five items (e.g., Things I want to do at home do not get done because of the demands my job puts on me). Scale responses ranged from 1 = strongly disagree to 5 = strongly agree. Cronbach’s α for the current study was 0.85, indicating good reliability (Koo & Li, 2016).
Leader support
J. Zhang et al. (2019) developed a five-item Leader Support Scale based on Jain and Nair’s (2017) Work Support Scale, by replacing “my co-workers” with “my supervisors” in the questionnaire. Further, J. Zhang et al. (2019) sampled Chinese university faculty and demonstrated its reliability (Cronbach’s α was 0.95) as well as its validity by conducting confirmative factor analysis. The five items were used in our study (e.g., My supervisors give me aid in making work-related decisions). Scale responses ranged from 1 = strongly disagree to 5 = strongly agree. Cronbach’s α for the current study was .92, indicating excellent reliability.
Perfectionism personality
We used the Multidimensional Perfectionism Scale–Brief validated by Cox et al. (2002) to measure perfectionism. In Cox et al.’s (2002) study, this brief scale was assessed among both university students and adult outpatients and revealed to have good reliability (Cronbach’s α was .85 for university students and .83 for adult outpatients) and validity. Smith et al. (2015) employed this scale to differentiate effects of perfectionism on negative emotionality between Canadian and Chinese students and showed that this scale had good psychometric properties for both groups. We assessed the two dimensions: perfectionistic concerns (six items; e.g., If I fail partly, it is as bad as being a complete failure) and perfectionistic strivings (five items; e.g., I set higher goals than most people). Scale responses ranged from 1 = strongly disagree to 5 = strongly agree. Cronbach’s αs for the current study were .91 for perfectionistic concerns indicating excellent reliability and .85 for perfectionistic strivings indicating good reliability.
Work-related emotional exhaustion
We used the six-item Emotional Exhaustion Scale from the MBI (Maslach et al., 1996). This scale has been widely used in the occupational psychology literature for its good psychometric properties in various sociocultural contexts. For example, Wang et al. (2021) used it in Chinese contexts and demonstrated its reliability (Cronbach’s α was 0.78) and validity with both exploratory and confirmative factor analyses. An example was “I feel used up at the end of the workday,” and scale responses ranged from 1 = strongly disagree to 5 = strongly agree. Cronbach’s α for the current study was .92, indicating excellent reliability.
Work engagement
The Utrecht Work Engagement Scale was developed by Schaufeli et al. (2002). Its short form (nine items; Seppälä et al., 2009) with good psychometric properties was used to assess work engagement. It included three subscales: Vigor (three items; e.g., At my job, I feel strong and vigorous), Dedication (three items; e.g., My job inspires me), and Absorption (three items; e.g., I feel happy when I am working intensely). This construct has been assessed among Chinese employees and shown to have good reliability (Cronbach’s αs ranged from .76 to .79 for the three subscales) and validity (Slemp et al., 2021). Scale responses ranged from 1 = never to 5 = always. Reliability was good for the current study, with Cronbach’s αs being .90 for Vigor, .86 for Dedication, and .81 for Absorption.
Job satisfaction
Six items were used to assess faculty’s satisfaction with job (Schriesheim & Tsui, 1980). These items respectively evaluated satisfaction with nature of work, supervision, relationships with colleagues, salaries, promotion, and job in general (e.g., To what extent am I satisfied with promotion opportunities). Meng (2020) employed this scale among Chinese university faculty, demonstrating its good reliability (Cronbach’s α was 0.87) and validity by confirmative factor analysis. Scale responses ranged from 1 = strongly dissatisfied to 5 = strongly satisfied. Cronbach’s α for the current study was .90, indicating good reliability.
Statistical Analyses
Analysis of Moment Structures (AMOS) Version 22.0 was used to test the research model by means of maximum likelihood method in structural equation modeling (SEM). Before testing the model, we constructed three-item parcels respectively for the six variables of work–family conflict, leader support, perfectionistic concerns and strivings, emotional exhaustion, and job satisfaction. The parceling strategy is increasingly used for its psychometric advantages for SEM models including many multi-item variables (Little et al., 2002). According to Little et al. (2002), the precondition for item parceling is that scales have to be unidimensional. Therefore, we conducted exploratory factor analyses separately for the six variables. The results showed that the items of these respective variables all loaded on a single factor (χ2 ranged from 772.79 to 1,631.51; df ranged from 10 to 15; all ps < .001). These results allowed us to perform the item parceling. Regarding work engagement, we used its three subscales as the indicators. The following model–data fit indices were used for model evaluation: χ2/df ratio (<3), comparative fit index (CFI > .95), Tucker–Lewis index (TLI > .90), root mean square error of approximation (RMSEA < .08), and Standardized Root Mean Square Residual (SRMR <.08) (Hu & Bentler, 1999).
Results
Preliminary Analysis
Table 1 presents descriptive statistics and correlations among the focal variables and demographics (i.e., gender, age, research field, and professional titles). The table showed positive or negative correlations among the focal variables as well as a negative correlation between the demographic variable of professional titles and perfectionistic concerns. We conducted a one-way ANOVA which indicated a significant difference in the concerns dimension across groups of professional titles (F = 7.262, p < .001). Specifically, assistant professors (M = 3.12 and SD = 1.40) seemed to have higher levels of perfectionistic concerns than lecturers (M = 2.30 and SD = 0.91), associate professors (M = 2.19 and SD = 0.86), and professors (M = 2.11 and SD = 0.85). Since none of the demographic variables were correlated with the mediator or outcome variables, they will not be included in the research model for further analyses.
Means, Standard Deviations, and Correlations.
*p < .05. **p < .01.
Assessment of the Measurement Model and Common Method Variance
A measurement model with seven latent variables was tested by confirmatory factor analysis. The results showed that the model fitted the data well: χ2 (168, N = 375) = 328.048, p < .001, χ2/df = 1.953, SRMR = .047, RMSEA = .050, confidence intervals (CI) = [.042, .59], CFI = .972, and TLI = .966. In addition, factor loadings of the observed indicators on their respective latent variables were significant at p < .001.
Studies using self-report measures need to check common method bias. As suggested by Podsakoff et al. (2003), we performed Harman’s single-factor test by forcing all 42 items to load on a single unrotated factor. This single factor extracted only 26% of the total variance, far below the warning criterion of 50%. Testing a one-factor model revealed very poor model indices: χ2 (189, N = 375) = 4,492.943, p < .001, χ2/df = 23.772, SRMR = .256, RMSEA = .247, CI = [.241, .253], CFI = .259, and TLI = .177. These results indicated that common method bias was not a problem for our study.
Assessment of the Structural Model
After the measurement model was confirmed to be acceptable, we assessed the structural model that received a good model–data fit: χ2 (169, N = 375) = 334.517, p < .001, χ2/df = 1.979, SRMR = .050, RMSEA = .051, CI [.043, .059], CFI = .972, and TLI = .965. The standardized coefficient paths are presented in Figure. 2. The figure showed that work–family conflict (β = .21, p < .001) and perfectionistic concerns (β = .43, p < .001) positively predicted emotional exhaustion, while leader support (β = −.13, p = .004) negatively predicted emotional exhaustion. In contrast, perfectionistic strivings (β = −.02, p = .683) were nonsignificant for emotional exhaustion. Leader support (β = .28, p < .001) and perfectionistic strivings (β = .18, p = .011) positively predicted work engagement, while perfectionistic concerns (β = −.05, p = .509) were nonsignificant for work engagement. Finally, work engagement (β = .30, p < .001) and leader support (β = .53, p < .001) directly and positively predicted job satisfaction, while emotional exhaustion (β = −.12, p = .020) directly and negatively predicted job satisfaction.
Results of the structural model. Note. The covariance lines among the exogenous variables are not drawn for model parsimony; the solid lines indicate significant coefficient paths; the dotted lines indicate nonsignificant coefficient paths. *p < .05. **p < .01. ***p < .001.
Assessment of the Mediating Roles
Bootstrapping method in SEM was employed to evaluate the mediating effects. Specifically, 95% CI without zero can inform that mediation is statistically significant, whereas 95% CI including zero can inform that mediation is not statistically significant (Cheung & Lau, 2008).
Based on 5,000 bootstrap data samples, the results indicated that the indirect effect of work–family conflict (95% CI [−.064, −.004]) on job satisfaction via emotional exhaustion was significant. Thus, Hypothesis 1 was supported by the result. Similarly, the indirect effect of leader support (95% CI [.047, .172]) on job satisfaction via emotional exhaustion and work engagement was also significant. Thus, Hypothesis 2 was supported by the result. Regarding the research question, we found that indirect effects of both perfectionistic concerns (95% CI [−.143, .003]) and strivings (95% CI [−0.003, 0.146]) on job satisfaction were not significant. Thus, exhaustion and engagement did not mediate the relationship between perfectionism and job satisfaction.
Discussion and Recommendations for Future Research
Guided by the JD-R, the present study achieved the general objective of examining predictors of occupational well-being among Chinese university faculty. It was revealed that job demands and resources had distinctive relationships with the dual processes. More importantly, our study expanded the occupational psychology research for being among the first to incorporate personal demands within the JD-R.
The Roles of Work–Family Conflict and Leader Support
In our work, specific types of job demands (work–family conflict) and job resources (leader support) were selected, and the results exactly conformed to the key predictions of the JD-R. Specifically, work–family conflict positively predicted the health impairment process, while leader support negatively predicted the health impairment process and positively predicted the motivational process.
Antecedents and/or consequents of work–family conflict have been adequately researched in the Western work environment (e.g., Dettmers, 2017; Minnotte & Yucel, 2018; Torp et al., 2018; Vignoli et al., 2016). Beyond those studies, our study examined this construct in Chinese sociocultural contexts and confirmed that Chinese faculty suffering from higher levels of this conflict tended to feel more exhausted emotionally. Reasonably, alienation from family roles stemming from high job demands appeared to have eroded faculty’s emotional balance and resulted in exhaustion. We can further interpret this relationship from cultural perspectives. Two main Chinese cultural features are collectivism and Confucianism. The former suggests that Chinese people prefer viewing themselves as in-group members of social networks and highly value harmonious interpersonal relationships within the networks (Hofstede, 2001). The latter nurtures a climate for behavioral features of “saving face” and avoiding shame (Hofstede, 2001) and coping by “forbearance” (C. Liu et al., 2008). The emphasis on harmony and forbearance may make Chinese faculty feel reluctant to initiate direct conflicts with supervisors for reducing excessive workloads. As a result, more negative emotions and fewer family responsibilities stemming from the workloads may cause higher levels of emotional exhaustion. Hence, our work can not only advance the research on work–family conflict by examining its relationship with exhaustion among Chinese academics but also provide recommendations for researchers to conduct cross-cultural studies to gain a more informed understanding of this relationship in collectivist and individualist cultures.
Leader support was found to be an important factor for faculty’s overall occupational well-being due to its negative relationship with exhaustion and positive relationships with work engagement and job satisfaction. These findings were consistent with prior research (Charoensukmongkol et al., 2016; Suan & Nasurdin, 2016). Early research usually emphasized social support in general, while recent research focused more on social support from specific resources (e.g., support from family, friends, colleagues, and leaders). Our study selected leader support due to Chinese high power distance cultures where working individuals highly value how they are treated by leaders or authority figures (Tyler et al., 2000). Indeed, our results indicated that leader support can be functional for faculty to stay more engaged in work, reduce psychological costs, and promote job satisfaction. Leader support can be particularly important for the sharply increasing number of Chinese academics to compete for the limited resources (e.g., funding projects, training programs, and promotion chances). As such, perceptions of this work-related support can be highly valued and eagerly pursued by faculty as a way leading to career development.
Our findings also confirmed the mediation perspectives proposed in the JD-R (Bakker & Demerouti, 2007; Schaufeli & Taris, 2014), revealing complexities of the relationships of work–family conflict and leader support to occupational well-being. Specifically, work–family conflict was indirectly related to job satisfaction via the mediator of emotional exhaustion. Since the direct path from the conflict to satisfaction was not significant (see Figure. 2), we can conclude that this was a full mediation. It suggested that this relationship can be described exclusively in terms of indirect effects via emotional exhaustion. Leader support was indirectly related to job satisfaction via the dual mediators of exhaustion and engagement. The mediation suggested the instrumental values of leader support, functioning as a means to enhancing job satisfaction through stimulating engagement and reducing exhaustion. Further, since the direct path from leader support to job satisfaction was still significant (see Figure. 2), we can conclude that this was a partial mediation. It implied that there might be other variables that can also play important mediating roles. Given that university faculty mostly work as both teachers and researchers, we recommend future research to examine teaching engagement and research engagement as potential mediators to offer more insightful knowledge.
The Roles of the Perfectionism Trait
Our study differentiated predictive roles of two perfectionism dimensions in occupational well-being among university faculty, a sample receiving very little attention in perfectionism research. The concerns dimension was found to positively predict emotional exhaustion, consistent with Garratt-Reed et al. (2018) and Lee and Anderman (2020). This finding further confirmed the well-documented maladaptive nature of perfectionistic concerns (e.g., Flaxman et al., 2018). As discussed previously, excessive worries of mistakes and fear of negative evaluations exhibited by individuals high in perfectionistic concerns can potentially make them feel emotionally “drained” (Stoeber & Damian, 2016). The strivings dimension was found to positively predict work engagement, which confirmed the adaptive nature of perfectionistic strivings. The finding echoed the argument that setting high standards, though they may be exceedingly high, has the potentials to energize and regulate individuals’ motivation for performance and engagement in goal-achieving processes (Damian et al., 2017).
Of interest, we did not find the relationship between perfectionistic strivings and exhaustion, contradicting prior research that consistently revealed their positive relationship (Lee & Anderman, 2020; Mitchelson & Burns, 1998). We may interpret this inconsistency by taking sample features into consideration. One feature of university faculty is possessing high levels of academic degrees (mostly doctoral degrees). When they were students, many of them tended to set high academic standards for themselves such as the challenging and arduous goal of pursuing doctoral degrees. When working as teachers, many of them still have to struggle to meet high self-imposed expectations (e.g., innovative research and publications in competitive journals). Presumably, setting high standards may have become an essential part of their academic and work domains such that they may be accustomed to the standards and would not feel emotionally exhausted by the standards. An alternative explanation can be offered by Bakker and Demerouti (2017). They argue that employees in socially respected occupations (e.g., professors and top managers) often face many challenging responsibilities, but they tend to have many job resources at their disposal to fulfill the responsibilities. This argument may hold true for university academics with various support resources and personal skills to facilitate meeting the high standards and thus would not feel much emotional exhaustion. Therefore, it may be an inspiring research idea to conduct comparative studies on the relationship between the two perfectionism dimensions and emotional exhaustion among employees in occupations with different levels of social status or prestige.
Limitations, Theoretical, and Practical Implications
Some research limitations of this study need to be acknowledged. First, our cross-sectional design requires caution to refer to the studied relationships as causal relationships. Second, the sampling was conducted in a single Chinese university though it has typically demanding work environment. Future research can collect data from samples that are more representative of the whole population. Third, the demographic variables were associated with neither the mediators nor the outcome variable, which may have restricted our understanding of university faculty’s occupational well-being. Future research may need to collect more relevant demographics (e.g., years of work, teaching hours per week, and research publications).
Despite these limitations, our findings can have several theoretical implications. First, we tested the influential JD-R among an underresearched sample (i.e., university faculty). Further, the rather few exceptions only considered a part of the JD-R, either examining antecedents and/or consequences of job demands (Torp et al., 2018; Xu, 2019) or neglecting organizational outcomes (Han et al., 2020). Moving beyond prior studies, our study more comprehensively emphasized how job demands and resources were related to university faculty’s occupational well-being. Second, we expanded the JD-R by empirically supporting Bakker and Demerouti’s (2017) recent recommendation that personal demands should be researched within the theory. More importantly, our findings yielded some distinctive functional roles of personal demands from those of job demands and resources. According to the key predictions of the JD-R (Bakker & Demerouti, 2007; Schaufeli & Taris, 2014), job demands positively predict the health impairment process, while job resources negatively predict the health impairment process and positively predict the motivational process. In contrast, our findings evidenced that personal demands (e.g., the two perfectionism dimensions) may positively predict both of the two processes. In other words, personal demands may be a double-edged sword because they may involve both self-imposed burdens associated with physiopsychological costs and self-determining motivation associated with work efforts and engagement. As such, our study offered new insights into the dual processes of the JD-R through which the two perfectionism dimensions seemed to operate. Third, these recommendations may be important based on our findings: (a) examining other types of personal demands (e.g., workaholism, goal setting, and performance expectations) within the JD-R to generalize common and distinctive features and (b) extending this research line to different sociocultural contexts and work sectors to gain more nuanced understanding.
From the applied perspective, our findings suggest that university management needs to consider balancing faculty’s work–family interface. It can be of great importance that faculty are relieved from administrative work duties (e.g., writing administrative reports and preparing evaluation materials) as much as possible and devote their primary attention to teacher and researcher roles. Beyond that, prior studies provided ways of creating family-supportive work environment without reducing work demands such as flexible work arrangements, onsite child care services, and providing boundary-spanning resources (Haar et al., 2014; Karatepe, 2013). Given the enormous benefits of leader support, the management should create a climate where seeking work-related support is encouraged, whether it be emotional or instrumental support. Emotionally, positive feedback and encouragement are highly appreciated especially when faculty achieve successes or meet hardships. Such emotional support is not only motivational for employees to stay engaged but also helpful to alleviate work stress or recover from the stress (Charoensukmongkol et al., 2016). Instrumentally, training programs and work-related advice can be effective to promote skills variety and self-efficacy beliefs that are important sources of work motivation, engagement, and performance. Finally, work-related personality assessment (e.g., perfectionism, positive affect, or core self-evaluation) can be implemented regularly through observations, feedbacks, or surveys. The strategy, however, has been largely ignored in Chinese educational contexts although it provides useful information for designing interventions (Cao & Meng, 2020). In perfectionism assessment, it is important to differentiate the concerns and strivings dimensions because they functioned differently in the dual processes. Based on the assessment results, cognitive behavioral therapy interventions can be offered to address some problematic aspects of the personality (e.g., perfectionistic concerns), thus enhancing the person–job fit (Flaxman et al., 2018). Such interventions can be particularly useful for faculty at their early career stages (e.g., assistant professors) because our one-way ANOVA revealed that they scored higher in perfectionistic concerns than lecturers, associate professors, and professors. It makes sense that young faculty tend to hold a disadvantageous position within the university hierarchy, which may exacerbate their worries about mistakes and fear of negative evaluations from supervisors. This argument can be supported by Mudrak et al.’s (2018) study which found that university faculty with low professional titles possessed less influence over work and suffered from more job insecurity. To handle this issue, university faculty high in the concerns dimension may be trained to refrain from “punishing” for making trivial mistakes. To conclude, our findings based on applying and extending the JD-R can expand the occupational psychology literature and help design and upgrade service programs to promote university faculty’s occupational well-being.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study is funded by the Social Sciences Funding Project of Jilin Province, China (Grant Number: 2019B168 and 2018B126), and Fellowship of China Postdoctoral Science Foundation (Grant Number: 2020T130089).
