Conservation of resources theory (Hobfoll, 1989) assumes that individuals acquire additional resources to meet challenging tasks when they perceive potential personal growth and resource gains. Resources aid the process of growth and gain, because initial resource promotes future gain, thus generating “gain spirals” (Chen et al., 2015 p. 97). Based on COR theory, we argue that job complexity is positively related to approach crafting. Job complexity facilitates thinking skills and triggers employees’ initiatives to acquire more knowledge, information, and support from colleagues (Sung et al., 2017). Thus, complex jobs will most likely help employees develop their skills and gain additional resources from their work environment. In this view, job complexity motivates and legitimates employees’ self-job redesign and self-management, such as job crafting (Hornung et al., 2010; Ohly and Fritz, 2010; Petrou et al., 2012). Complex jobs also increase task motivation and enable employees to exhibit skills and novel approaches to solve problems. Employees will be prompted to craft their job to fulfill the required tasks (Frese et al., 2007; Tims et al., 2013; Kuijpers et al., 2020).

Tims and Bakker (2010) suggested that complicated tasks will increase employees’ ability to identify alternative opportunities and strategies to obtain additional resources. Under challenging job conditions, employees focus on acquiring resources and on investing their resources to gain additional resources through job crafting (Nielsen, 2013; Harju et al., 2016). For example, employees need to process information and experience new problems, which often create constructive interaction with supervisors or other colleagues that provide help, support, and guidance. Thus, approach crafting aiming at acquiring resources may represent a favorable way to handle complex jobs.

However, some employees with complex jobs may employ an avoidance strategy. According to the resource loss process in COR theory, employees exposed to a stressful situation tend to have a more negative work attitude (Samma et al., 2020). Job complexity prompts employees to experience strain and use avoidance crafting to reduce their strains. Complex jobs create extensive responsibilities and strain employees’ time and efforts, depleting their energy and resources (Ito and Brotheridge, 2003). Employees may want to reduce demands to reduce resource loss. Therefore, employees with highly complex jobs may employ escape/avoidance strategies to ease their strains.

Building on COR theory, we posit that approach crafting represents a resource gain process as employees deal with challenging job demands that motivate them, whereas avoidance crafting constitutes a resource loss process when employees feel that such job demands are overly stressful. Employees are inclined to choose tasks at which they will perform well and that are not too difficult (Elliot and Thrash, 2001). Approach crafting seeks to obtain and retain the positive aspects of work and gives employees an opportunity to demonstrate their competence (Elliott and Dweck, 1988; Harju et al., 2021). However, avoidance crafting occurs when employees try to avoid potential failure when handling certain tasks which may have them appear incompetent (Harju et al., 2021). Thus, employees may avoid certain aspects of complex jobs while approaching others. We suggest the following hypothesis:

Unlike routinized jobs, complex jobs provide employees with more opportunities to explore, exercise control, and be responsible for outcomes (Pierce et al., 2009), generating positive states of vigor, dedication, and absorption (i.e., work engagement). Job complexity satisfies individuals’ desire to learn and achieve at work since complex jobs offer the kinds of opportunities and internal rewards they value. Specifically, complex jobs are more malleable and accessible in making changes and employees are encouraged to consider alternative solutions to handle them (Pierce et al., 2009). Thus, employees are more likely to invest more of themselves (e.g., their resources, time, and efforts) into the job (Brown et al., 2014). Moreover, complex jobs are stimulating and challenging, which may trigger employees to invest additional efforts to fulfill achievements and lead to the motivational process (Chung-Yan, 2010). Such resource gain process triggered by job complexity results in the positive effects that include personal growth and development (LePine et al., 2005). Consequently, those employees may experience a high level of job meaningfulness (Bunderson and Thompson, 2009), which in turn will be reflected in a heightened sense of dedication.

Further, job complexity can enhance employees’ absorption in their jobs because challenging jobs require a high level of information-processing capacity (Gardner and Cummings, 1988; Bledow et al., 2011). Complex jobs require individuals to become both self-absorbed and self-revealing to allocate work-related resources actively toward the tasks (Rich et al., 2010). In contrast, simple or routine jobs may cause boredom, distract employees from their tasks, and reduce their interest in tasks (Fisherl, 1993), resulting in a lower level of work engagement (Gorgievski and Hobfoll, 2008). As such, job complexity triggers resource gains and fosters work engagement.

More specifically, we hypothesize that work engagement plays a mediating role between job complexity and approach crafting behavior. According to COR theory, employees are motivated to accumulate and obtain valuable resources to deal with challenging job demands (i.e., resource gain process of COR theory). Complex jobs allow employees to personalize their tasks because confronting tough tasks enables them to seek new solutions to complete their jobs (Brown et al., 2014). Therefore, individuals experiencing complex jobs might craft their jobs more proactively because they can acquire additional resources to deal with challenging job demands through approach crafting. As dedicated employees are more likely to exert additional efforts to seek resources and information, employees with positive psychological functioning (i.e., high work engagement) are able to enhance their job crafting (Bakker and Demerouti, 2017). Approach crafting enables employees to accumulate additional tangible and intangible resources to reduce uncertainty and gain a strong social support network (Robinson and Griffiths, 2005); this enhances their productivity and enables them to cope with demanding and complex jobs.

Further, employees fully immersed in their work can use their personal resources more efficiently and tend to be more focused on their work (Chang et al., 2013). Employees immersed in their jobs are more likely to seek out new perspectives, challenges, and solutions. Besides, their immersion motivates employees to focus on work activities and be more persistent to achieve their goals. As a result, individuals feel motivated to invest themselves in their work and acquire the resources needed to overcome potential challenges (Aubé et al., 2009); all these endeavors facilitate approach crafting at work. Moreover, when individuals with challenging jobs are vigorous, they are more likely to consider problem-solving choices and actions to overcome their challenges (Barsade, 2002). This augmented positive cognitive state is an essential motivation for approach crafting because it allows individuals to build a new association between the job and the required resources.

Employees tend to acquire and invest resources to maximize their psychological energy resources (Halbesleben et al., 2014). Those with larger psychological resources can achieve their goals and thrive through the resource gain process. Employees with a high level of engagement will be motivated to gain additional resources through approach crafting actions, such as incorporating challenging tasks or seeking other social resources (Tims et al., 2012; Niessen et al., 2016). In line with the above discussion, we hypothesize that there exists a relationship between job complexity and approach crafting through work engagement.

Complex jobs do not always lead to positive psychological functioning but can lead to a negative psychological state, such as energy depletion. Employees with complex jobs may attempt to conserve energy in order to recover resource losses states (Samma et al., 2020). When employees fail to achieve a challenging goal, they will seek to conserve resources by reducing efforts for tasks that consume their resources (Hobfoll, 1989) and protect themselves from strains by reducing resource losses when employees feel hard to meet the requirements of challenging jobs and gain resources (LePine et al., 2005). Complex jobs may impose psychological and cognitive pressure on employees (Sung et al., 2017), leading to resource loss and energy depletion which most likely occurs when individuals possess inadequate resources or insufficient resource gain to meet work demands. Thus, job complexity potentially drains valuable resources, which leads them to experience physical and psychological exhaustion (Chung-Yan, 2010). Based on this reasoning, we propose that stressful situations caused by complex jobs are likely to wear out an employee’s psychological energy resources.

Our study also proposes that job complexity results in a resources loss process, leading to energy depletion. According to COR theory’s resource loss process, individuals with complex jobs may craft their jobs less proactively to conserve resources. As workload, accumulation can cause employees feel overwhelmed and unable to deal with their tasks. As a result, they will feel frustrated and exhausted (Samma et al., 2020). By reducing their workload or withdrawing from their work role, they may then try to reduce strain and protect their valuable resources (Tims et al., 2012). Exhausted employees tend to display avoidance crafting because they cannot regulate their energy successfully (Demerouti et al., 2005) or adapt to their depleting resources (Wright and Hobfoll, 2004). Individuals who are in the condition of energy depletion might be more likely to withdraw from their job role to protect existing personal resources.

Employees with challenging jobs may experience a depletion of resources and will attempt to make their situations less overwhelming in the dysfunctional state of energy exhaustion (Hobfoll, 1989). Employees in the energetically depleted state may have difficulties recognizing opportunities and challenges but may also avoid challenges (Swider and Zimmerman, 2010). Avoidance crafting represents any efforts by employees to evade challenging job demands. It constitutes a resource loss process, where employees seek to mitigate the straining effects of work to protect their resources and improve their wellbeing (Harju et al., 2021). Since approach crafting requires additional efforts (Zhang and Parker, 2019), employees may adopt the avoidance strategy to relieve stress. The resource loss process of COR theory suggests that individuals with demanding jobs might be less likely to engage in behavior that consumes their resources (Hobfoll, 1989). We suggest the existence of an indirect relationship between job complexity and avoidance crafting through energy depletion. Figure 1 summarizes the proposed research model of this study.

We tested our hypotheses with a sample of employees from six companies covering a variety of occupations and industries in Macau. Participants were recruited by contacting the human resource managers of the participating companies, requesting their assistance in our study. Using a questionnaire-based survey (e.g., Rasool et al., 2020), we collected data at three time points to reduce common method bias (Podsakoff et al., 2003). The questionnaire included the following parts: (i) a mention of the study purpose and privacy statements, (ii) the possible answers on a 5-point Likert scale, and (iii) the respondents’ demographic information. Respondents were notified that participation in the survey was voluntary, confidential, and anonymous, as we did not use clear names but identification codes known only to the authors.

At Time 1 (T1), we distributed the paper-and-pencil survey to 400 employees and asked them to rate their job complexity level, receiving the completed survey from 367 participants (91.8% response rate). Approximately 2 weeks later after the participants had completed the Time 1 survey (at Time 2), they were asked to rate their work engagement and energy depletion. We chose to separate both questionnaires by a 2-weeks interval because our model deals with a psychological process and its effects on employee behavioral outcomes (for a similar approach see Deng et al., 2018). A total of 323 employees (88.0% response rate) returned the T2 survey. After an additional 2 weeks (at time T3), we sent out the last questionnaires to obtain information regarding the levels of the participants’ approach and avoidance crafting. From the 292 returned questionnaires (90.4% response rate), 41 had to be discarded due to missing or randomly filled data (25) and unmatched responses (16), resulting in an overall response rate of 62.8% (i.e., 251 completed data sets from 400 questionnaires that had initially been sent out.

The 251 included participants had an average age of 42.12 years (SD=12.37), an average organizational tenure of 6.95 years (SD=6.15), and 51% of them were females. Their highest education levels were high school diploma (17%), a 2-year college or undergraduate degree (65%), or graduate degrees (18%). They were employed in a variety of occupations, including hotel frontline employees (35%), marketing and sales (24%), education (14%), finance (8%), and others (19%). Detailed descriptions of the samples are presented in Table 1.

The survey measurement was translated into Chinese from the original construct in English. Back-translation procedures (Brislin, 1980) were applied to improve measurement reliability and validity in a different language.

To test this dual-mediation model, we used structural equation modeling (SEM) to investigate the hypothesized relationships using the Mplus software (v.8.3). Indirect effects were tested using the bootstrapping method. Following Edwards and Lambert (2007), we used moderated path analysis and bootstrapping to test the moderated mediation hypotheses. Given the relatively small sample size compared to the number of items, we modeled the variables as latent variables with parcels (Marsh et al., 1998). For example, for the latent construct approach crafting, five parcels were created for each approach crafting dimension. Parceling can make the measurement models more reliable (Little et al., 2002). Also, the reliability of the parcels was checked, and all reliability estimates were above 0.84.

Table 2 reports means, standard deviations, correlations, and their Cronbach’s alpha values of all study variables. The scales used were highly reliable as the value of Cronbach’s alpha was above the standard.70.

Given that data were collected through a single source, we conducted Harman’s single-factor analysis to test whether common method bias can be a potential concern in our study (see Podsakoff and Organ, 1986). The first factor in the unrotated structure had an eigenvalue of 10.90 and only accounted for 25.96 per cent of the variance. This suggests that common method bias is less likely to influence the results of this study.

Before hypotheses testing, we conducted a series of confirmatory procedures to examine the discriminant validity of the study’s key measures: job complexity, work engagement, approach, and avoidance crafting (see Table 3). We excluded energy deletion into our measurement model because the model failed to convergence (only one item scale to measure energy depletion). The results showed that proposed four-factor measurement model resulted in a good fit with the data (χ²=278.82, df=224, p<0.01; CFI=0.98; RMSEA=0.03). In addition, the standardized factor loadings ranged from 0.67 to 0.85, and all of them were significant at a level of p<0.001, demonstrating relatively good convergent validity. Table 3 shows that various alternative measurement models displayed a significantly worse fit than the hypothesized four-factor model (all Δχ² tests, p<0.01). The results provided support for the distinctiveness of the key measures in this study. Thus, we proceeded to hypothesis testing using the five study variables.

We tested our dual-research model in Mplus Version 8.3 (Muthén and Muthen, 2017). In Hypothesis 1, we expect that job complexity predicts approach crafting. The results showed that complex job was positively related to approach crafting (b=0.27, p<0.001; see Table 4), supporting Hypothesis 1. Hypothesis 2 suggests that job complexity is positively related to avoidance crafting. The results revealed that job complexity was positively related to avoidance crafting (b=0.22, p<0.001; see Table 4). Thus, Hypothesis 2 was supported.

Job complexity was significantly related to work engagement (b=0.18, p<0.05; see Figure 1), supporting Hypothesis 3. In Hypothesis 4, we expect that work engagement mediates the relationship between job complexity and approach crafting. We bootstrapped 10,000 samples and used the bootstrap estimates to construct bias-corrected confidence intervals (CI) for all significance tests reported in this study (Mooney et al., 1993; Shrout and Bolger, 2002). The bootstrap estimate (0.03) for the indirect effect fell within the 95% bias-corrected confidence interval 95% CI [0.01, 0.07], supporting the significance of work engagement as a mediator in the relationship between job complexity and approach crafting (see Table 4). As to Hypothesis 5, job complexity was significantly related to energy depletion (b=0.61, p<0.001; see Figure 1). In Hypothesis 6, the bootstrap estimate (0.22) for the indirect effect fell within the 95% bias-corrected confidence interval 95% CI [0.15, 0.32], supporting energy depletion fully mediating the relationship between job complexity and avoidance crafting (see Table 4). Thus, Hypothesis 6 was supported.

Our study makes several theoretical contributions to the field. Firstly, this study examines the underlying psychological mechanisms of job complexity and approach/avoidance crafting and contributes valuable knowledge to elucidate the psychological mechanisms linking the various antecedents to approach and avoidance crafting, a field of research that is only poorly explored (Ghitulescu, 2007). Individuals can either try to handle complex job demands by optimizing their resources investment through approach crafting, or they may try to cope with excessive demands by engaging in avoidance crafting (Zhang and Parker, 2019; Costantini et al., 2021). We suggest that the current resource state of an employee determines his/her adoption of approach versus avoidance crafting. Specifically, our study indicated that work engagement (i.e., resource gain process) and energy depletion (i.e., resource loss process) trigger approach and avoidance job crafting, respectively. This finding indicates that the emergence of different job crafting types from the same task design can instigate different resource states among employees.

Secondly, our results emphasize the importance of motivational and strain aspects. Complex jobs oriented on development stimulate higher levels of work motivation by creating critical psychological states in employees (i.e., work engagement) that enable them to engage in approach crafting. Job complexity leads to energy depletion which can lead to avoidance crafting. We therefore investigated how job complexity can trigger different employee workplace attitudes which can in turn prompt different forms of job crafting. Our results provide input for future research on the motivational and strain consequences of job complexity on employee functioning.

Thirdly, using COR theory as theoretical lens, our study offers empirical evidence that supports the resource gain and loss processes postulated in COR theory (Halbesleben et al., 2014). COR theory suggests that these competing processes help explain how individuals manage their resources in order to cope with demands (Dawson et al., 2016). However, prior studies mainly concentrated on the resource loss process, while neglecting resource gain (Hobfoll et al., 2018). The present study considers both processes in parallel, thereby facilitating a deeper understanding of both processes. Moreover, our results reveal that job complexity may display a “too-much-of-a-good-thing” effect, resulting in strain outcomes when taken too far (see Pierce and Aguinis, 2013, p. 315).

The present findings also hold practical implications for managers. Jobs are becoming increasingly complex as they often involve more flexible and fluid employment arrangements (Okhuysen et al., 2013). Managers should be aware that the constantly changing demands of today’s complex jobs cannot wait for top-down job design solutions (Grant and Parker, 2009; Demerouti, 2014). Rather, managers should recognize the importance of bottom-up work design approaches (i.e., job crafting) and initiate interventions to channel employee job crafting efforts to a desired direction. Our findings suggest that inducing work engagement among employees (e.g., vigor, dedication, and absorption) can effectively promote approach crafting. In contrast, employees’ energy depletion may lead to avoidance crafting. Thus, managers should carefully increase additional resources, such as job autonomy, to allow employees to master their complex tasks. In the context of a stressful and demanding work, employees craft their jobs to acquire and protect resources. Thus, organizations and managers should provide employees with opportunities to craft their jobs to achieve a better balance between job demands and resources, which will lead to positive outcomes (Kristof-Brown et al., 2005).

Furthermore, instead of focusing on the beneficial and motivational effects of job complexity, organizations also need to adopt a more balanced view that also considers the potential strain effects of challenging jobs. This means that managers should monitor employees and ensure that task demand is reasonable, and control the flow of demands to ensure sufficient activation and prevent the depletion of the employees’ physical and emotional resources. As such, both top-down job design and bottom-up job crafting have significant implications for individual and organizational performance (Demerouti et al., 2019; Lichtenthaler and Fischbach, 2019).

The present study had several limitations that should be addressed in future research. Firstly, while we used self-reporting to assess the variables, which may lead to common method bias (Podsakoff et al., 2003), we adopt a potential procedural remedy in the form of three surveys conducted over a period of 4 weeks (Podsakoff et al., 2003). Our statistical analyses suggest that common method bias may not be a serious problem. Additionally, self-reported assessments of job crafting are appropriate because job crafting is self-initiated behavior. Thus, employees are in a better position than their supervisors and peers to rate their job crafting activities. Nevertheless, future research on job crafting may benefit from observing “actual job crafting,” e.g., measuring the amount of time spent on each task.

Secondly, although the dual pathway of the job crafting model is promising, other moderators, such as different personalities, should be explored to better understand when and how complex jobs influence job crafting. Future research could also examine the moderating effects of leader behavior on the job complexity-job crafting relationship (Thun and Bakker, 2018). Further, the potential moderating role of personal and job resources and their effects on job strain and work engagement should be explored in more detail (Bakker and Demerouti, 2017). Specifically, based on Job Demands and Resources theory, job resources (such as skill variety) and personal resources (such as self-efficacy) have a particular effect on motivation when job demands are high. Thus, employees with high job and personal resources may be more motivated to adopt job crafting.

Thirdly, our study does not consider the social, political, and relational elements of job characteristics which may influence how employees interpret their job tasks (Grant and Parker, 2009). Thus, future studies should examine whether other job-design characteristics are able to predict approach and avoidance crafting. Moreover, we did not consider the relationship between hindrance job demands (e.g., role conflict or role ambiguity) and job crafting. Therefore, future studies should examine a broader range of job demands to identify those that cause individuals to reduce and increase their job crafting efforts.

Fourthly, in an effort to avoid similarities and possible collinearities with work engagement, this study did not use the four items scale (Ryan and Frederick, 1997) to measure employees’ energy depletion. Instead, we adopted the item scale developed by Bakker and Oerlemans’s (2019) to assess employees’ positive energy available for purposive action. As this might cause validity issues, future research should consider other instruments to measure energy depletion, such as the 25-item state resource depletion scale by Ciarocco et al. (2007). Lastly, the extent to which our results can be generalized is unclear. Previous studies shown that low-skilled workers might implement job crafting differently from highly skilled professionals (Nielsen and Abildgaard, 2012). Future studies should verify our findings by using a more diverse group of individuals. Given that the average age of our sample was about 42 years, future research should examine and test our model in younger employees.