Abstract
The present study examined the configurations, or profiles, taken by distinct global and specific facets of job engagement and burnout (by relying on a bifactor operationalization of these constructs) among a nationally representative sample of Canadian Defence employees (
Keywords
According to Kahn (1990), job engagement occurs when employees’ personal resources are actively channeled toward the realization of their work. High levels of job engagement facilitate the accomplishment of organizationally valued behaviors, support behaviors that are more focused and vigilant, and help workers meet the emotional demands of their roles (Kahn, 1990). Job engagement is a precursor of desirable outcomes for the organization (e.g., lower turnover intentions, better performance; Rich, LePine, & Crawford, 2010) and the employee (e.g., higher job satisfaction; Haynie, Mossholder, & Harris, 2016). Conversely, burnout is characterized by high levels of emotional exhaustion and negative attitudes toward work (Maslach, Schaufeli, & Leiter, 2001). Employees suffering from burnout feel disillusioned, helpless, irritated, and worn out. They have lost connection with their work and distance themselves mentally and emotionally from their work activities (Leiter & Maslach, 2016). Burnout is associated with high levels of turnover intentions (Cheng, Bartram, Karimi, & Leggat, 2016) and depressive symptoms (Hatch, Potter, Martus, Rose, & Freude, 2019).
Despite abundant research (Bakker & Demerouti, 2017; Laughman et al., 2016) supporting the benefits associated with job engagement components (physical, cognitive, and emotional; Rich et al., 2010) and the undesirable outcomes of burnout components (emotional exhaustion and disengagement; Demerouti, Mostert, & Bakker, 2010), little is known about their combined impact (Mäkikangas, Hyvönen, & Feldt, 2017). Two very distinct, yet complementary, types of analyses can help to better comprehend this combined impact. On the one hand, variable-centered analyses can help to understand the nature of associations between a subset of variables and other subsets of variables. Unfortunately, these analyses are unable to reveal a clear picture of the combined effect of more than two or three variables. On the other hand, person-centered analyses consider the configurations taken by a set of variables among discrete subpopulations, or profiles, of employees.
In the present study, we rely on person-centered analyses (i.e., latent profile analyses—LPA) to identify the configurations of burnout and job engagement among different profiles of employees in a way that would be impossible to achieve using variable-centered analyses. More precisely, the resulting profiles would represent discrete subpopulations of employees characterized by qualitatively distinct configurations of burnout and job engagement (e.g., such as a profile of employees experiencing high levels of burnout coupled with low levels of job engagement, a profile characterized by high levels of engagement coupled with high levels of emotional exhaustion, or even a profile dominated by specific dimensions of burnout and/or job engagement). This approach helps to achieve a more integrative, or holistic, understanding of the reality of job engagement and burnout profiles as it is truly experienced among employees, and in a way that shares clear connections with our (i.e., researchers, managers, and practitioners) tendency to think of employees as members of discrete categories (Zyphur, 2009). More precisely, whereas a nomothetic variable-centered approach considers job engagement and burnout components as separate interrelated constructs, a more idiographic person-centered approach rather focuses on how all of these components are experienced together by different types of employees (Marsh, Lüdtke, Trautwein, & Morin, 2009; Meyer & Morin, 2016).
Emerging person-centered research has started to look at how job engagement and burnout components combine within specific individuals (Abós, Sevil-Serrano, Haerens, Aelterman, & García-González, 2019; Mäkikangas, Feldt, Kinnunen, & Tolvanen, 2012, 2014, 2017; Moeller, Ivcevic, White, Menges, & Brackett, 2018; Salmela-Aro, Hietajärvi, & Lonka, 2019; see Table S1 in the online supplements for an overview of the profiles identified in these studies, as well as their associations with a variety of predictors and outcomes). However, no research has done so while considering the specificity of all theoretical facets of job engagement, or by simultaneously considering employees’ global and specific levels of job engagement (physical, cognitive, and emotional; e.g., Gillet, Morin, Jeoffrion, & Fouquereau, 2020c) and burnout (emotional exhaustion and disengagement; Isoard-Gautheur et al., 2018). Information is thus lacking regarding the nature of employees’ job engagement and burnout configurations based on a complete theoretical coverage of the inherent multidimensionality of both constructs.
To inform this issue, this study documents the job engagement and burnout configurations that best characterize members of a large-scale representative sample of Canadian Defence personnel, while accounting for the multidimensionality of job engagement and burnout. In doing so, we emphasize the importance of adopting a finer-grained representation of job engagement and burnout by simultaneously considering their global and specific components. This more refined perspective helps us to better understand the unique, and complementary, role played by each specific facet of burnout and job engagement beyond the role played by employees’ global levels of burnout and job engagement. This approach should help us to uncover whether profiles defined by more, or less, balanced configurations of burnout and job engagement across components may carry greater, or lower, risks for exposed employees. For instance, despite the generally assumed positive effects of job engagement, is it possible for some highly engaged employees to also experience higher than expected levels of emotional exhaustion? Alternatively, is job engagement enough to limit employees’ risk of turnover intentions, or would its benefits be maximized only for employees presenting some specific combinations of burnout and job engagement components?
This study also documents the construct validity (Muthén, 2003) of these profiles by considering their associations with job demands and resources (work overload, role ambiguity, interpersonal justice, transformational leadership, organizational support, and psychological empowerment), and turnover intentions. From a practical perspective, this study should help provide improved guidance for managers seeking to nurture, preserve, and improve psychological health among different types of employees (Zyphur, 2009). For instance, by documenting the implications of these profiles for employees’ turnover intentions, this study should help organizations determine which profiles should be prioritized from an intervention perspective. Likewise, understanding how different types of job demands and resources contribute to the emergence of different profiles should also provide guidance in relation to the identification of actionable levers of intervention. Importantly, a unique contribution of this study would be to examine these associations from a multilevel perspective. This multilevel perspective will allow us to verify whether, and how, the effects of employees’ perceptions of job demands and resources on their likelihood of profile membership would differ across the individual versus group levels of analyses, leading to a clearer understanding of the role played by employees’ shared perceptions of the job demands and resources present in their work unit, once properly disaggregated from their individual perceptions of these same job demands and resources.
A Multidimensional Representation of Burnout and Job Engagement
It is recognized that an assessment of job engagement should tap into the physical, cognitive, and emotional facets of this construct (Rich et al., 2010), just like an assessment of burnout should at least tap into its emotional exhaustion and disengagement components (Demerouti, Bakker, Vardakou, & Kantas, 2003). It has also been proposed that employees might experience job engagement and burnout in a more holistic manner as a function of two overarching dimensions (Alfes, Shantz, Truss, & Soane, 2013; Cheng et al., 2016). This global approach is supported by high correlations among ratings of physical, cognitive, and emotional job engagement (Shuck & Reio, 2014), and among ratings of emotional exhaustion and disengagement (Demerouti et al., 2010). Research has also shown that a higher-order representation of job engagement relates more strongly to antecedents and outcomes than its first-order components (Shuck, Nimon, & Zigarmi, 2017). However, research has also revealed well-differentiated associations between distinct components of job engagement and burnout, and a variety of outcomes, thus supporting the existence of conceptually distinct components of job engagement (Shuck et al., 2017) and burnout (Collie, Granziera, & Martin, 2018).
These options are not mutually exclusive as burnout might also exist as a global entity (i.e., burnout) reflecting commonalities among ratings of emotional exhaustion and disengagement, themselves including specificity unexplained by this global burnout entity (Barcza-Renner, Eklund, Morin, & Habeeb, 2016; Isoard-Gautheur et al., 2018). Likewise, job engagement can occur both as a global construct anchored in the variance shared across its dimensions (emotional, physical, and cognitive), themselves retaining some specificity (Gillet et al., 2020). Higher-order results support the idea that both constructs can be represented as global entities (Rich et al., 2010; Sinval, Queirós, Pasian, & Marôco, 2019). However, a remaining question is whether enough specificity exists in the physical, cognitive, and emotional components once global job engagement is considered, and in the emotional exhaustion and disengagement components once global burnout is considered (Gillet et al., 2020c; Sinval et al., 2019).
Although hierarchical models have often been used to address this question (e.g., Rich et al., 2010; Sinval et al., 2019), these models involve a stringent proportionality constraint in defining how the items relate to the higher-order factor and to the specific part of the first-order factors that is not explained by the higher-order factor (i.e., its disturbance; e.g., Chen et al., 2006). Indeed, in hierarchical models, items define first-order factors, which are used to define a higher-order factor reflecting the variance shared among the first-order factors. Yet, the relation between an item and the higher-order factor is indirect (i.e., mediated by the first-order factor). This indirect effect is reflected as the product of (x) the item’s first-order factor loading by (y) the loading of this first-order factor on the higher-order factor. This second term (y) is thus a constant for all items associated with a specific first-order factor. Similarly, the relation between an item and the disturbance of the first-order factor to which it is associated is also reflected as the product of this item’s loading on the first-order factor (x) by another constant representing the link between the first-order factor and its disturbance (z). As a result, the ratio of item variance explained by the global (the higher-order factor; e.g., global burnout) and specific (the first-order factors; e.g., emotional exhaustion and disengagement) factors (xy/xz) is assumed to be identical for each first-order factor (y/z), and unlikely to hold in real life (Morin, Arens, & Marsh, 2016a).
Bifactor models provide an alternative to hierarchical models (Chen et al., 2006) and are not submitted to this unrealistic restriction. Bifactor models thus provide a more flexible way to address the same questions. According to a bifactor operationalization, each item is used to define both a Global (G) factor and one Specific (S) factor. This approach thus provides a way to obtain a direct estimate of the commonalities shared across all items (the G-factor, e.g., global engagement or global burnout), and an equally explicit estimate of the specificity associated with each component (specified as independent from one another)
A Person-Centered Perspective on the Complementary Role of Job Engagement and Burnout
Researchers relying on variable-centered analyses assume, often explicitly but sometimes implicitly, that their results would equally apply to all members of the population under study. Although it is possible to verify how the effects of one variable differ as a function of another one, such tests of interactions are virtually impossible to decode when more than three predictors interact together, especially in the presence of nonlinearity. Importantly, adopting a bifactor representation of job engagement and burnout would result in seven interacting predictors, making it impossible to rely on variable-centered analyses to achieve an integrated representation of the combined role played by these two G-factors (burnout and engagement) and five S-factors (emotional exhaustion, disengagement, cognitive engagement, emotional engagement, and physical engagement). Person-centered analyses do not rely on similar assumptions and are specifically designed to identify profiles of employees differing from one another on more or less extensive a set of variables (Meyer & Morin, 2016). Thus, rather than focusing on the additive or interactive effects of these variables, the person-centered approach rather focuses on the categorization of employees into discrete profiles differing in their unique experiences of job engagement and burnout dimensions, the outcome implications of these profiles, and the impact of various predictors on employees’ likelihood of corresponding to each of these profiles (Meyer & Morin, 2016).
Person-Centered Studies: A Summary
In Appendix 1 of the online supplements, Table S1 provides a summary of the results from previous person-centered research seeking to identify profiles of burnout and or engagement. Despite some variations possibly related to methodological differences (e.g., type of employees, measures), a high level of similarity is apparent across studies (Mäkikangas & Kinnunen, 2016). However, very few of these previous studies have adopted a comprehensive approach simultaneously incorporating multiple facets of engagement and burnout, while relying on a proper disaggregation of global levels of job engagement and burnout from the specificities associated with each job engagement and burnout facet. The estimation of latent profiles based on indicators capturing the bifactor structure of job engagement and burnout ratings (i.e., resulting in a proper disaggregation of global and specific levels of job engagement and burnout across facets) would make it possible to identify clearer, and more easily interpretable, profiles differing from one another in relation to both the global (i.e., global job engagement and burnout) and specific (i.e., the different dimensions of job engagement and burnout dimensions) components of these constructs (Morin et al., 2016b, 2017). This approach would thus help us to isolate the unique contribution of each specific dimension associated with both constructs (e.g., Gillet, Morin, Choisay, & Fouquereau, 2019b). Ignoring this dual global/specific structure carries the risk of inaccurately identifying profiles characterized by job engagement and burnout levels solely capturing the global components of these constructs (Morin & Marsh, 2015; Morin et al., 2016b, 2017). When we consider the results from previous person-centered studies, we can indeed note that a number of of these studies revealed a number of profiles mainly characterized by global types of differences. The present investigation, relying on representative sample of Defence employees, adopts an approach developed by Morin et al. (2016b, 2017) to identify profiles of burnout and job engagement while accounting for the global and specific components of these constructs.
Despite our difference in approach, it remains possible for us to expect the identification of profiles dominated by job engagement (i.e., an
However, in accordance with the subset of
Given the novelty of our approach, it would be possible to speculate regarding the possible identification of a rather large number of qualitatively distinct configurations of job engagement and burnout. However, despite their interest, these possibilities would remain largely speculative. Thus, and in a way that is aligned with the methodologically inductive nature of person-centered analyses, we leave as an open research question the nature of the profiles to be identified.
Research Question 1: Which profiles of job engagement and burnout will be identified among the current sample of Defence employees?
Research Question 2: Will these profiles differ quantitatively (based on employees’ global levels of job engagement and burnout), qualitatively (based on their specific configuration of burnout and job engagement components), or both?
Profiles of Burnout and Engagement: Implications for Turnover Intentions
Turnover intentions are the main predictor of voluntary turnover (Heavy, Holwerda, & Hausknecht, 2013), a relation that is particularly marked among military personnel (Lytell & Drasgow, 2009). Turnover itself has always been a ubiquitous outcome for organizations given its costs in terms of performance reduction, recruitment, and training (Heavey et al., 2013). Turnover intentions are also negatively related to job engagement and positively related to burnout (Alfes et al., 2013; Cheng et al., 2016), indicating that highly involved employees are more likely to want to stay in their job, whereas worn-out ones are more likely to seek alternative employment.
Results from previous person-centered research (see Table S1; e.g., Abós et al., 2019) suggest that profiles characterized by high levels of job engagement and low levels of burnout (e.g.,
Research Question 3: How do employees’ levels of turnover intentions differ as a function of their job engagement and burnout profiles?
A Multilevel Person-Centered Perspective on the Role of Job Demands and Resources
Job Demands
The job demands-resources (JD-R) model (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001) highlights the role of two categories of work conditions, job demands, and resources, in the prediction of employees’ engagement and burnout. Job characteristics requiring employees to expand psychological and/or physical efforts in an ongoing manner are referred to as job demands and tend to carry a toll for employees feeling exposed to such a work environment (Schaufeli & Bakker, 2004). Moreover, employees tend to perceive job demands as challenging or hindering (Crawford, LePine, & Rich, 2010). Challenging job demands have the potential to support mastery, personal growth, or future gains (i.e., demands to be overcome to learn and achieve), whereas hindering job demands have the potential to thwart growth, learning, and goal attainment. We focus on the effects of two types of hindering job demands with a known influence on job engagement and burnout (e.g., Ghorpade, Lackritz, & Singh, 2011; Reinke & Chamorro-Premuzic, 2014): Role ambiguity and work overload. Hindering demands are expected to interfere with employees’ functioning by impeding their self-actualization and the satisfaction of their psychological needs for autonomy, competence, and relatedness (Ryan & Deci, 2017). Hindering demands are thus likely to lead to a persistent psychophysiological and cognitive activation (Sonnentag & Fritz, 2015) as a result of being unable to attain personal goals (e.g., Kinnunen et al., 2017). This persistent activation is likely to interfere with the work recovery process (Sonnentag & Bayer, 2005). Not surprisingly, the effects of hindering job demands are well documented in the prediction of a range of outcomes likely to emerge from the quality of the work recovery process, such as higher levels of burnout and lower levels of job engagement (Gillet et al., 2020a, 2021). More specifically, when coping with role ambiguity, employees lack clear and consistent information about work expectations (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964). They are thus more likely to report higher levels of job anxiety and strain, subsequently leading to lower job engagement and higher burnout (Bakker & Demerouti, 2017). Similarly, attempts to cope with work overload may lead employees to exhaust their energetic resources, in turn increasing their likelihood of experiencing lower levels of job engagement and higher levels of burnout (Schaufeli & Bakker, 2004).
Job Resources
Contrasting with job demands, job resources refer to those aspects of a job that contribute to supporting employees in achieving their goals, to reducing the costs associated with job demands, and to stimulating personal development (Demerouti et al., 2001; Xanthopoulou, Bakker, Demerouti, & Schaufeli, 2009). Job resources are expected to help enhance employees’ psychological functioning, both by increasing job engagement and by decreasing burnout (Schaufeli & Bakker, 2004). Nielsen et al. (2017) proposed a multilevel framework focusing on whether job resources originate from the employees, their workgroup, their supervisors, or the organization. They also reported meta-analytic evidence supporting the complementary role of each type of resources for employees’ psychological health and behaviors.
Individual Resources
Individual resources are personal characteristics that help employees cope with job demands and achieve satisfactory levels of performance while remaining psychologically healthy (Xanthopoulou et al., 2009). In this study, we focus on psychological empowerment, which encompasses employees’ feelings of competence, autonomy, impact, and meaning in relation to their work (Spreitzer, 1995, 2008). Competence refers to feelings of having the abilities required for a successful execution of their work, a cognition close to the concept of self-efficacy. Autonomy refers to feelings of being in control when initiating and regulating work behaviors. Impact refers to feelings of being able to influence operational, strategic, or administrative outcomes at work. Finally, meaning refers to feelings that there is a good fit between work requirements and employees’ personal beliefs, standards, and values. Despite their distinct nature, these four cognitions have been systematically shown to converge on a global psychological empowerment construct (Morin et al., 2016c; Seibert, Wang, & Courtright, 2011). Psychological empowerment is positioned as a core psychological resource allowing employees to play a volitional role at work while feeling in control of their actions, and thus as an important mechanism allowing them to handle the stressfulness of their work (e.g., Spreitzer, 1995, 2008). Meta-analyses support the role of psychological empowerment as a driver of a variety of organizationally relevant outcomes and psychological health indicators (e.g., Seibert et al., 2011), including lower levels of burnout and higher levels of job engagement (e.g., Calvo & García, 2018; Livne & Rashkovits, 2018). Importantly, psychological empowerment is conceptually distinct from self-determined work motivation (Gagné, Senécal, & Koestner, 1997).
Workgroup Resources
Workgroup resources relate to the social and interpersonal context of the workplace; that is, to relationships among group members that foster efficient communications, positive interactions, and trust (Nielsen et al., 2017). In this study, we focus on interpersonal justice, which refers to workgroups in which employees interact respectfully with one another (Colquitt, 2001). Evolving in a workgroup in which employees feel respected across a range of situations is likely to improve the pleasantness of the work, to help employees feel supported when facing adversity, and to protect them against feelings of exhaustion, isolation, and disconnection (Colquitt et al., 2013). The role of interpersonal justice as a driver of positive functioning at work, including higher levels of job engagement and reduced levels of burnout, have been well established in research (Colquitt et al., 2013; Gillet, Fouquereau, Bonnaud-Antignac, Mokounkolo, & Colombat, 2013).
Leader Resources
Leader resources refer to vertical interactions between employees and supervisors who may, by virtue of their position and leadership style, provide them with support, guidance, and security (Nielsen et al., 2017). We focus on transformational leadership, which refers to the ability of the supervisor to inspire and motivate employees’ loyalty and involvement (Bass & Avolio, 1994). Transformational leaders focus on employees’ individual needs, and provide them with a sense of mission and purpose which helps to protect them from adversity while maintaining their positive drive (Hildenbrand, Sacramento, & Binnewies, 2018). Similar to interpersonal justice, the role of transformational leadership as a mechanism able to support employees’ psychological health, including increases in job engagement and protection against burnout, has been supported by extensive research evidence (e.g., Hildenbrand et al., 2018; Montano, Reeske, Franke, & Hüffmeier, 2017).
Organizational Resources
Organizational resources refer to the broader work environment context, and the way it is organized and managed to support, motivate, and encourage positive functioning and growth (Nielsen et al. 2017). In this study, we focus on organizational support, which refers to the extent to which the organization values and supports employees’ contributions and well-being (Eisenberger, Huntington, Hutchinson, & Sowa, 1986). Organizational support contributes to fulfilling employees’ basic socioemotional needs at work and is expected to convey the idea that support (material or emotional) will be available to help them maintain adequate levels of performance under stressful conditions (Eisenberger & Stinglhamber, 2011). Not surprisingly, the beneficent role of employees’ perceptions of organizational support in relation to a wide range of outcome variables, including job engagement and burnout, has also been well established in research (e.g., Eisenberger & Stinglhamber, 2011; Gillet et al., 2018a).
A Person-Centered Perspective
Despite their importance, no research has examined the effects of these job demands and resources on job engagement and burnout profiles. Indeed, whereas variable-centered predictions simply highlight the role of job demands and resources in the prediction of each burnout or job engagement component considered in isolation, the person-centered perspective makes it possible to consider this role more broadly in the prediction of distinctive multidimensional configurations of job engagement and burnout. In other words, it makes it possible to directly account for the role of these job demands and resources in the prediction of the complete reality of employees’ engagement and burnout.
Despite the novelty of our approach, the variable-centered evidence presented thus far suggests that transformational leadership, interpersonal justice, organizational support, and psychological empowerment, as well as lower levels of role ambiguity and work overload, should predict a higher likelihood of membership into the
Building on the JD-R model, we tested these possibilities by considering an individual resource (psychological empowerment) in addition to a series of job demands (work overload and role ambiguity) and workgroup (interpersonal justice), supervisor (transformational leadership), and organizational (organizational support) resources in the prediction of profile membership.
Research Question 4: How will job demands (work overload and role ambiguity) as well as individual (psychological empowerment), workgroup (interpersonal justice), supervisor (transformational leadership), and organizational (organizational support) resources relate to employees’ likelihood of belonging to the profiles of job engagement and burnout identified in this study?
A Multilevel Perspective
JD-R research has, for the most part, focused on job demands and resources assessed at the individual level via employees’ report, without often considering how the effects of these work-related characteristics may differ at the work unit level. Yet, Bakker and Demerouti (2017) remarked that it would be critical to adopt a more systematic multilevel approach to the study of these multilevel phenomena. Indeed, employees evolve in complex multilayered workplaces in which at least a part of their work experiences are likely to be shared by all members of their workgroups (i.e., reflecting their exposure to more objective work characteristics), thus conflating two sources of influence in a single estimate when relying on single-level analyses (Gonzàlez-Romà & Hernàndez, 2017; Morin et al., 2021). We adopt a multilevel perspective to achieve a clearer understanding of the role played by employees’ shared perceptions of the job demands and resources present in their work unit (i.e., a more objective, or at least consensual, picture of their work unit environment) properly disaggregated from their individual exposure to job demands and resources (i.e., inter-individual differences in their perceptions of their work unit).
More precisely, as part of the instructions provided to them in the questionnaires, employees were explicitly asked to report on their individual perceptions of the job demands, as well as the workgroup, supervisor, and organization resources present in their work unit. Using these ratings, our multilevel perspective allowed us to disaggregate their shared perceptions (from the group level aggregation of their individual perceptions) from their unique individual experiences. In organizational research, this type of rating makes it possible to assess “climate” or “consensus” constructs at the work unit level (Bliese, Chan, & Ployhart, 2007; Quigley, Tekleab, & Tesluk, 2007; Morin et al., 2021). In fact, many have argued that when the referent of the ratings is the work unit, then it is unreasonable to assume that the unique reality of the individual employee who provided the rating is the only cause of that rating (thus committing the fundamental attribution error of ignoring the work unit reality as an equally important source of influence on the rating; e.g., Ross, 1977). In this case, the proper level at which these predictors should be considered is the work unit (the object of the rating), allowing researchers to separately consider the role played by inter-individual deviations in these ratings. These deviations, however, are more likely to reflect social comparison processes or inter-individual differences in exposure to specific work characteristics than the whole reality of individual levels of exposure to these work characteristics (Marsh et al., 2012; Morin et al., 2014, 2021). This perspective highlights the risk of failing to separate these two layers of influences, especially when focusing on job demands and resources explicitly conceptualized, and measured, as characteristics of the work unit.
In contrast, being explicitly defined as an individual resource, psychological empowerment needs to be studied as such. Although meaningful individual variables can sometimes create a specific work context (such as sex, which is an meaningful individual variable and yet can create a male- or female-dominated work context), previous research has shown that did not happen when psychological empowerment was considered (Morin, Blais, & Chénard-Poirier, 2021); that is, that the construct of psychological empowerment (located at the individual level) was qualitatively distinct from the construct of team empowerment (located at the work unit level; Maynard et al., 2013), which is not considered in the present study.
Fortunately, some emerging variable-centered attempts have been made to study the effects of job demands and resources across the individual and group levels. For instance, Demerouti et al. (2001) found similar associations between job demands and resources and employees’ burnout and disengagement at the group and individual levels, showing job demands to be associated with higher levels of burnout, and job resources to be associated with lower levels of disengagement. Likewise, Bakker, Van Emmerik, & Van Riet (2008) found supervisor and workgroup resources to be associated with lower levels of cynicism, whereas job demands were positively related to emotional exhaustion at the individual level. Rather than focusing on global constructs reflecting job demands and resources, other studies established the multilevel role of specific work environment characteristics, such as leadership, organizational support, or justice perceptions (e.g., El Akremi, Colaianni, Portoghese, Galletta, & Battistelli, 2014; Gagné et al., 2020; Kiersch & Byrne, 2015) in the prediction of various indicators of psychological functioning, including job engagement and burnout. Despite similarities, these studies are inconsistent regarding the relative role of individual perceptions and group aggregates, making it hard to establish clear expectations and to transpose these expectations to the person-centered context.
Research Question 5: How will the associations between job demands and resources and employees’ likelihood of profile membership differ across levels of analyses (i.e., inter-individual differences in perceptions of work-related demands and resources and shared perceptions at the work unit level)?
Method
Participants
This study relies on a stratified random sample of Canadian Armed Forces/Department of National Defence (CAF/DND) non-deployed personnel, selected from a sampling frame of 100,018 military and civilian personnel covering a wide range of occupations. Random samples were drawn from 67 organizational strata with proportional allocation for the sector (i.e., Regular Force, Primary Reserve, and civilian personnel), sex, rank (i.e., non-commissioned members and officers) for military personnel, and years of service for civilian personnel. This random sampling scheme yielded a total sample of 41,387 personnel with a small expected margin of error (<1%). Of those, 13,088 respondents (31.6%), nested within 65 work units (including 46 to 576 employees,
Sampling weights were calculated to ensure that the sample was representative of the target population (i.e., to ensure that the results can be generalized to the whole CAF/DND population from which the sample has been recruited). Taking into account these weights, 55% of the population were members of the Regular Force, 20% were members of the Primary Reserve, and 25% were civilian employees. Nineteen percent of the military members were officers, whereas 26% of the civilian employees occupied a managerial or supervisory position. Seventy-five percent of the population was male, 37% was younger than 35, 50% was between 35 and 54 years of age, and 13% was older than 54. Thirty-eight percent of the population had served within the CAF/DND for fewer than 11 years, 33% between 11 and 20 years, and 29% served for 20 years or more.
Most respondents (81.7%) completed the English version of the DWWS, whereas the remaining completed the French version. For the few measures (role ambiguity and work overload) not already validated in both official languages of Canada, translators from the Government of Canada’s Translation Bureau translated the original English items into French. Bilingual experts from CAF/DND then back-translated these items into English. Discrepancies were resolved by consensus.
Measures
Burnout
Disengagement (four items; α = .81; e.g.,
Job engagement
Cognitive (six items; α = .93; e.g.,
Psychological empowerment (individual resource)
Feelings of meaning (three items; α = .96; e.g.,
Role ambiguity (job demand)
Employees’ perceptions of role ambiguity were assessed using the relevant six-item scale (α = .92; e.g.,
Work overload (job demand)
Employees’ perceptions of work overload were assessed with the six-item (α = .93) short version (Thiagarajan, Chakrabarty, & Taylor, 2006) of Reilly’ (1982) questionnaire. All items (e.g.,
Transformational leadership (supervisor resource)
Perceptions of the supervisors’ transformational leadership practices were assessed using the seven-item (α = .96; e.g.,
Interpersonal justice (work group resource)
Interpersonal justice perceptions were measured with the four-item subscale (α = .93; e.g.,
Organizational support (organizational resource)
Respondents described the level of support received from their organization with the French adaptation (Gillet et al., 2015a) of a questionnaire originally developed by Eisenberger, Huntington, Hutchinson, and Sowa (1986). This questionnaire includes eight items (e.g.,
Turnover intentions (outcome)
Turnover intentions were assessed with a measure developed by Colarelli (1984). The four items from this scale (α = .86; e.g.,
Analyses
Mplus 8.3 (Muthén & Muthén, 2019) was used to conduct analyses via the Maximum Likelihood-Robust (MLR) estimator, which is robust to multilevel nesting and non-normality. Missing data was handled with Full Information Maximum Likelihood (Enders, 2010). Preliminary measurement models estimated for the individual-level measures of job engagement, burnout, empowerment, and turnover intentions, as well as unconditional LPA based on the indicators of job engagement and burnout were estimated at the individual level. For these models, we relied on Mplus design-based correction procedures (Asparouhov, 2005) to obtain standard errors and tests of model fit that accounted for participants’ nesting within work units. Preliminary measurement models for the multilevel constructs of job demands and workgroup, supervisor, and organization resources were specified as multilevel with employees (L1) nested under work units (L2). These latent variable models make it possible to assess constructs corrected for measurement errors at both levels of analyses (via the estimation of latent factors), together with L2 ratings reflecting aggregated individual perceptions corrected for inter-rater reliability, and L1 ratings reflecting inter-individual differences in perceptions of the L2 reality (Marsh et al., 2012; Morin et al., 2014, 2021). Conditional multilevel LPA were then used to allow L1 predictors to influence the likelihood of profile membership at the individual level (L1) and L2 predictors allowed to influence the frequency of occurrence of each profile at the work unit level (L2) (Finch & French, 2014; Mäkikangas, Tolvanen, Aunola, Feldt, Mauno, & Kinnunen, 2018). All models were estimated while incorporating stratified sampling weights using Mplus complex survey design functionalities (Asparouhov, 2005).
Preliminary Analyses
Preliminary analyses were conducted to verify the psychometric properties of all measures. These analyses were also used to obtain factor scores (estimated in standardized units with
A first measurement model was estimated for the profile indicators. In this model, participants’ ratings of burnout and job engagement were represented together by bifactor confirmatory factor analytic (bifactor-CFA) models including one global factor per construct (G-factor: Global burnout and Global engagement) and a series of orthogonal specific factors (S-factors; for burnout: Disengagement and emotional exhaustion; for job engagement: Cognitive, physical, and emotional engagement; Morin et al., 2016b, 2017). Bifactor models make it possible to explicitly isolate one global component underlying participants’ responses to all burnout or engagement items from specific components associated with responses to items forming each subscale left unexplained by the global components and reflecting imbalanced levels of burnout or engagement across dimensions. This approach is consistent with the high correlations typically observed among burnout (e.g., Demerouti et al., 2003) and job engagement (Rich et al., 2010) components, and with research supporting a similar operationalization of burnout (Barcza-Renner et al., 2016; Isoard-Gautheur et al., 2018; Sinval et al., 2019) and engagement (Gillet et al., 2019a; 2020c). Importantly, this approach has been recommended to identify clearer profiles in situations where a global construct is assumed to co-exist with specificities assessed from the same indicators (Morin et al., 2016b, 2017). Bifactor factor scores result in cleaner differentiations between the profiles as the indicators are uncorrelated (their “overlap” is rather explicitly represented by the global factor). Then, the indicators are free to vary independently of one another to provide a clearer representation of the distinct configurations (or profiles) observed in the sample. This has been extensively discussed in statistical (Morin et al., 2016b) and statistically oriented (Morin et al., 2017) publications.
A second model was estimated for the individual covariates. In this model, one higher-order factor was used to define participants’ global levels of psychological empowerment from four first-order factors reflecting autonomy, meaning, impact, and competence matching the well-established higher-order structure of this construct (Morin et al., 2016c; Seibert et al., 2011). One additional factor was included to reflect turnover intentions. Three a priori correlated uniquenesses were incorporated to this model to reflect the negative wording of three items from the autonomy subscale (Marsh, Scalas, & Nagengast, 2010).
A third model was estimated for the multilevel constructs. In this model, participants’ ratings of role ambiguity, work overload, transformational leadership, interpersonal justice, and organizational support were used to estimate five a priori CFA factors at the individual (L1) and work unit (L2) levels. These multilevel CFA models were estimated using doubly latent procedures to estimate latent constructs corrected for measurement errors at both levels, while also relying on a latent aggregation procedure to correct for agreement among work unit members in the assessment of the L2 constructs (Marsh et al., 2012; Morin et al., 2014, 2021). These models included six a priori correlated uniquenesses at the individual level (L1) to control for the negative wording of three items from the role ambiguity subscale and three items from the organizational support subscale (Marsh et al., 2010). Doubly latent models rely on an automatic group-mean centering procedure, so that L1 ratings can be directly interpreted as inter-individual deviations from the average rating of the L2 group reality, which has been shown to be the appropriate centering procedure for the type of constructs considered in the present study (Morin et al., 2014, 2021). This multilevel model was also used to assess the measurement
Once the optimal models were identified, we combined all three solutions into a global single-level (L1) measurement model to assess the measurement invariance (Millsap, 2011) of participants’ responses as a function of their language (English vs. French), sex (males vs. females), and status (military vs. civilian). Goodness-of-fit was estimated using the Root Mean Square Error of Approximation (RMSEA), the Tucker-Lewis Index (TLI), the Comparative Fit Index (CFI), and a visual examination of parameter estimates. The robust χ2 will also be reported. According to common guidelines, RMSEA values under .06 and .08, and TLI/CFI values above .95 and .90, respectively, support excellent and acceptable fit (Hu & Bentler, 1999; Marsh, Hau, & Grayson, 2005). The results from these analyses are reported in Appendix B (Tables S2 to S6) of the online supplements and support the adequacy of all measurement models, their isomorphism, and their measurement invariance.
Weighted Correlations and Reliability for All Variables Used in the Present Study.
More precisely, although the first-order model was able to achieve an acceptable level of fit to the data, the bifactor model with two G-factors (burnout and engagement) and five S-factors (emotional exhaustion, disengagement, cognitive engagement, emotional engagement, and physical engagement) was able to achieve a better level of fit across all indicators. This bifactor solution revealed two G-factors that were both well-defined by strong positive loadings from all items (λ = .538 to .797 for burnout and λ = .587 to .791 or job engagement). Over and above this G-factor, four S-factors retained a satisfactory level of specificity: Physical engagement (λ = .279 to .519), emotional engagement (λ = .481 to .695), cognitive engagement (λ = .209 to .505), and exhaustion (λ = .265 to .482). In contrast, the S-disengagement factor (|λ| = .041 to .455) appeared to be weakly defined, suggesting that disengagement ratings mainly served to define G-levels of burnout, and only retained a limited amount of specificity when these G-levels were taken into account. The fact that this S-factor retained less specificity does not mean that it has no meaning, especially when modeled using an approach that explicitly controls for both measurement errors and associations with the G-burnout construct, such as the approach taken in the present study. It should also be noted that, despite this low level of specificity, the factor scores used as input to our main analyses remain corrected for measurement errors (e.g., Skrondal & Laake, 2001; Morin et al., 2020).
Latent Profile Analyses (LPA)
The procedures used to select the optimal number of latent profiles present in our data is fully disclosed in Appendix C of the online supplements and led to the selection of a five-profile solution in which the means of the profile indicators were allowed to differ across profiles. Multilevel relations (Finch & French, 2014; Mäkikangas et al., 2018) between the L1 predictors and participants’ likelihood of membership in the various profiles, as well as between L2 predictors and the relative frequency of each profile occurring at the work unit level were assessed with a multilevel multinomial logistic regression link function based on the direct inclusion of the predictors into the final LPA solution (Diallo, Morin, & Lu, 2017). The profiles were also contrasted in relation to participants’ turnover intentions, which were directly included to the final solution, using the multivariate delta method (Raykov & Marcoulides, 2004). Annotated Mplus inputs, used to estimate our main analytic models, and are reported in Appendix D of the online supplements.
Results
Latent Profiles
The results from the five-profile solution are illustrated in Figure 1 (see Appendix C of the online supplements for details). The first profile was characterized by high global levels of burnout (1.25 SD above the sample mean) and low global levels of job engagement (2 SD under the average), coupled with close to average to moderately low specific levels of job engagement across dimensions (between .2 and .6 SD under the average), moderately low specific levels emotional exhaustion (.5 SD under the average), and high specific levels of disengagement (.6 SD above the average). This means that employees’ levels of emotional, physical, and cognitive job engagement are moderately lower than their global levels of job engagement across dimensions, suggesting that these employees do not feel the need to invest any specific resource beyond their already low levels of job engagement. Similarly, employees’ levels of emotional exhaustion are moderately lower than their global levels of burnout across dimensions, indicating that these employees do not feel exertion going beyond their levels of burnout. In contrast, employees’ levels of disengagement were higher than their levels of global burnout across dimensions, suggesting feelings of disengagement or demoralization going beyond their global levels of burnout. This Final Five-Profile Solution. 
The second profile was also characterized by high global levels of burnout (+1.4 SD), but only by close to average levels (−.2 SD) of job engagement. In addition, employees’ corresponding to this profile presented moderately high to high specific levels of physical (+.7 SD) and cognitive (+.4 SD) engagement, coupled with specific levels of emotional exhaustion corresponding to the sample average, moderately low specific levels of disengagement (−.5 SD), and low specific levels of emotional engagement (−1.5 SD). This
The fourth profile was characterized by slightly above average global levels of burnout (+.2 SD) and job engagement (+.5 SD), coupled with low specific levels of disengagement (−.3 SD), and moderately high (+.2 SD) to high (+1 SD) specific levels of physical engagement, emotional engagement, cognitive engagement, and emotional exhaustion. This
Turnover Intentions
Turnover intentions differed in a statistically significant (
Job Demands and Resources
Results from the Predictive Analyses.
In terms of job demands, inter-individual deviations in perceptions of work overload at the individual level were systematically associated with the likelihood of membership into all of the profiles in a manner that was the direct opposite of psychological empowerment. More precisely, higher work overload perceptions were linked to an increased likelihood of membership into the
For job resources, inter-individual deviations in perceptions of interpersonal justice at the individual level were linked to an increased likelihood of membership into the
Results were not as numerous at the work unit level. In terms of job demands, work unit levels of work overload were associated a higher frequency of occurrence of the
Discussion
The dual global and specific multidimensional nature of job engagement and burnout is well established in research. Job engagement can be seen as a global construct, which also encompasses physical, cognitive, and emotional facets (Rich et al., 2010), just like burnout can be viewed as a global construct minimally encompassing emotional exhaustion and disengagement (Demerouti et al., 2010). However, despite the widely acknowledged recognition of the complementary role played by these two multidimensional constructs in shaping employees’ psychological functioning (Salmela-Aro et al., 2019), the most typical configurations taken by the combination of the global and specific facets of job engagement and burnout among distinct profiles of employees remain essentially unknown. The present study sought to address this limitation while building on recent person-centered research conducted on burnout (Berjot, Altintas, Grebot, & Lesage, 2017; Guidetti, Viotti, Gil-Monte, & Converso, 2018; Laverdière, Kealy, Ogrodniczuk, & Morin, 2018; Leiter & Maslach, 2016; Portoghese et al., 2018), job engagement (Gillet et al., 2019a, 2020c; Simbula, Guglielmi, Schaufeli, & Depolo, 2013), and both constructs (see Table S1 in the online supplements) without relying on a comprehensive operationalization of their multidimensionality (Abós et al., 2019; Mäkikangas et al., 2012; 2014; 2017; Moeller et al., 2018; Salmela-Aro et al., 2019). To document the practical relevance and construct validity of these profiles, we also considered their implications for turnover intentions and adopted a multilevel perspective to investigate the role of job demands and resources in the prediction of profile membership.
Employees’ Profiles of Job Engagement and Burnout (Research Questions 1 and 2)
Our results revealed that five distinct profiles best represented the job engagement and burnout configurations observed among a nationally representative sample of Canadian Defence employees: (1)
More specifically, our results also showed that a more imbalanced configuration of specific facet scores seemed to be associated with profiles displaying high global levels of burnout (
A key take-home message from the present study is that the similarity between the current results and those obtained in the context of previous studies (see Table S1) relying on different measures and methodological approaches reinforces the robustness of our findings, and the idea that the current profiles might be generalizable enough to support interventions seeking to maximize employees’ likelihood of experiencing more desirable profiles. Beyond similarity, however, the differences and specificities between our results and previous ones supports the need to rely on a precise operationalization of the multidimensional nature of job engagement and burnout. By providing the first direct source of evidence of job engagement and burnout profiles defined according to their recently recommended bifactor operationalization (e.g., Gillet et al., 2020c; Isoard-Gautheur et al., 2018), the present study represents an important step forward in job engagement and burnout research. Indeed, the reliance on a more traditional approach (ignoring the dual global and specific nature of job engagement and burnout) would have simply resulted in the estimation of profiles suggesting that there was little value in considering the unique nature of each dimension over and above these global levels. In contrast, our results show that both components play a key role in the definition of job engagement and burnout profiles, and thus bring valuable information to our understanding of job engagement and burnout.
The Implications of the Profiles for Turnover Intentions (Research Question 3)
Supporting the meaningfulness of these profiles, our results revealed that they shared well-differentiated associations with turnover intentions in a way that matched our expectations and previous results see Table S1). Indeed, employees presenting the lowest levels of global job engagement coupled with high levels of global burnout (
On the one hand, these results reinforce the idea that more aligned levels of job engagement and burnout yield higher benefits in terms of turnover intentions. The idea that alignment among these components could be, in some situations, more important than overall levels of psychological functioning has been previously documented in self-determination theory (e.g., Gillet et al., 2019b) and job engagement (e.g., Gillet et al., 2019a, 2020c) research. Our results demonstrate that these observations extend to a more comprehensive consideration of psychological functioning, encompassing burnout and job engagement. This observation suggests that this form of balance could stem from a more adequate allocation of one’s psychological resources at work, which is known to help reduce stress and recovery. On the other hand, the
A Multilevel Perspective on the Impact of Job Demands and Resources (Research Questions 4 and 5)
Individual-level predictions
Our results supported the role of interpersonal justice, transformational leadership, organizational support, and psychological empowerment as key drivers of psychological functioning at work (e.g., Colquitt et al., 2013; Eisenberger & Stinglhamber, 2011; Montano et al., 2017; Seibert et al., 2011). More specifically, individual levels of psychological empowerment, perceptions of interpersonal justice, and perceptions of organizational support were associated with membership into the
In contrast, and unexpectedly, perceptions of organizational support were associated with an increased likelihood of membership into the
Nevertheless, Caesens et al.’s (2020) recent findings suggest that high levels of social support perceptions might be detrimental in some situations. This “too much of a good thing” interpretation is aligned with prior variable-centered results revealing curvilinear relations between employees’ perceptions of organizational support and their levels of affective organizational commitment, trust, in-role performance, taking charge behaviors, extra-role performance, and deviance (Burnett, Chiaburu, Shapiro, & Li, 2015; Harris & Kacmar, 2018). Just like here, these studies reveal that higher levels of perceived organizational support are not always associated with more desirable outcomes. In line with this, Gillet et al. (2019b) also found that perceived organizational support was negatively related to specific levels of imbalance in the satisfaction of employees’ need for competence. They interpreted this result by suggesting that higher levels of perceived organizational support could lead employees to believe that their organization has doubts regarding their competence, ultimately leading to negative consequences (e.g., lower global levels of job engagement). What the present results suggest is that these undesirable effects of organizational support perceptions might be particularly marked among burned-out employees. Clearly, additional studies are needed to replicate the present results and to identify the mechanisms underlying these unexpected relations.
In terms of job demands, inter-individual differences in perceptions of work overload and role ambiguity were related to membership into the arguably least desirable
Work unit-level predictions
To answer repeated calls for increases in multilevel research focusing on the effects of job demands and resources (e.g., Bakker & Demerouti, 2007), we examined the role of work overload, role ambiguity, interpersonal justice, transformational leadership, and organizational support at the work unit level in the prediction of the relative frequency of occurrence of the profiles at the work unit level. Supporting the documented role of work overload in the emergence of burnout (e.g., Reinke & Chamorro-Premuzic, 2014), our results showed that work overload was associated with membership into the
Surprisingly, work overload was also related to an increased likelihood of membership into the
In contrast, role ambiguity and transformational leadership were unrelated to the frequency of profile occurrence at the work unit level. This result differs from that of previous variable-centered research (e.g., De Clercq, 2019; Montano et al., 2017), which could be explained by our adoption of a multivariate perspective in which various job demands and resources are simultaneously considered. Adopting a multivariate perspective means that all of the variance that is shared among the various predictors is controlled for (once the moderate correlations among them, as shown in Table 2, are accounted for), allowing for a more precise identification of the unique contribution of the most potent predictors. More precisely, what the present results suggest is that these specific job demands and resources do not seem to further contribute to the prediction of the relative frequency of occurrence of the profiles at the work unit level once the effects of other, arguably more potent, types of job demands and resources are considered. These findings encourage researchers to look at how various job demands and resources uniquely contribute to employees’ job engagement and burnout profiles.
Finally, work unit levels of interpersonal justice were related to an increased likelihood of membership into the
Limitations and Directions for Future Research
Even though this study represents the first systematic attempt to investigate the nature, predictors, and outcomes of employees’ job engagement and burnout profiles while relying on a methodological approach allowing us to properly disaggregate the global and specific components of these multidimensional constructs, limitations remain. First, the present study relied entirely on self-reports, raising possible concerns regarding the possible impact of various forms of self-report biases and social desirability. It would thus seem desirable for researchers to incorporate more objective, or multi-informant, measures to future investigations of similar issues. Second, although we have no reason to expect that our results would differ among other samples of employee (which is supported by the similarity between the nature of the profiles observed in this study relative to previous studies relying on different measures and methods), this study was conducted within a Canadian military organization which still serves to limit the generalizability of our findings. As a result, it would be important for future research to systematically verify the replicability of our results among more diversified samples of workers from different types of organizations (e.g., less hierarchical or authoritarian, or without the same level of job security and benefits) and cultures. The ability to demonstrate generalizability is important to support the value of interventions inspired by person-centered solutions. Third, we relied on a cross-sectional research design which made it impossible to verify the directionality of the observed associations, or even the possibility of changes. Although predictors or outcomes were selected based on their theoretical relevance (Bakker & Demerouti, 2017), it remains important for future research to extend our results longitudinally. Fourth, the novelty of our inductive approach made it impossible to rely on an a priori selection of predictors or outcomes that would allow us to precisely tease apart the qualitative differences observed between our profiles. Based on our results, we suggest that future research might benefit from a consideration of a wider set of outcomes (e.g., job performance, work–family conflicts) and from predictors likely to explain the specificity of the identified profiles (e.g., factors likely to explain involvement among otherwise burned-out employees or exertion among otherwise engaged employees). Finally, we considered the role of work characteristics at the individual and work unit levels in the prediction of employees’ likelihood of membership into the various profiles identified here. Alternative, and complementary, approaches would include the investigation of work unit profiles characterized by different frequencies of occurrence of individual profiles, as well as investigating work unit profiles characterized by distinctive sets of job demands and resources (e.g., Collie, Malmberg, Martin, Sammons, & Morin, 2020; Mäkikangas et al., 2018).
Practical Implications for Assessment, Research, and Intervention
In terms of research, the present investigation highlights the importance of accounting for the dual global and specific nature of employees’ multidimensional ratings of burnout and job engagement. Ignoring this dual nature might lead to the erroneous conclusion that each specific component of these two constructs are relatively independent from one another and result in similar effects caused in fact by employees’ global levels of burnout and job engagement. More concerning is the fact that these apparently comparable effects could mask the unique role played by each specific component beyond this global level. For applied researchers, this observation is particularly worrisome, given that biased results may serve as guides for the development of incomplete, or improper, interventions tailored at distinct profiles of employees defined by their global levels of burnout and job engagement while completely ignoring the specificities related to their unique manifestations of burnout and engagement.
In terms of measurement, our results pinpoint the value of adopting a bifactor operationalization of burnout and job engagement. Indeed, the failure to do so is likely to increase the risk of multicollinearity by the estimation of construct scores reflecting a confusing combination of global and specific components. Importantly, although bifactor models can separate the variance of both constructs shared across dimensions from the unique role of each specific dimension, the meaning of these global and specific dimensions remains the same as in more traditional approaches. Although it is reasonably simple to adopt this recommendation in research, practical applications of a bifactor operationalization for scoring purposes are not as straightforward. Indeed, the ability to score employees’ ratings of burnout and job engagement will require the development of online calculators, developed based on results from more representative
In terms of intervention and practice, this study reinforces the value of managerial practices seeking to reduce burnout and nurture engagement. Managers need to pay attention to employees feeling exposed to particularly high workloads, or who lack the ability or opportunity to adopt a volitional approach at work (i.e., low psychological empowerment). These individuals seem to be at risk of experiencing low global levels of job engagement coupled with high global levels of burnout, leading them to develop higher turnover intentions. Care should be taken to ensure that any unforeseen increase in workload be shared, in a reasonably equitable manner, among colleagues. Changes in the work organization designed to increase psychological empowerment might sustainably increase job engagement and decrease burnout levels in the long run. For instance, moving towards or enhancing high-involvement managerial systems (e.g., performance-related remuneration schemes) may help to improve employees’ psychological empowerment (Rehman, Ahmad, Allen, Raziq, & Riaz, 2019). Organizations should also allocate resources to enactive mastery experiences, promote self-directed decision-making, and create opportunities for personal growth. Efforts to promote justice perceptions in terms of workload allocations also seem promising (Emery, Booth, Michaelides, & Swaab, 2019).
Moreover, our findings suggest that initiatives seeking to increase employees’ perceptions of organizational support at work are likely to have widespread benefits when care is taken to ensure that this increase is perceived equivalently by all work unit members. Among possible ways to achieve this objective, top management might promote a supportive culture within their organization, for instance, by providing employees with the resources or materials they need to perform their job effectively, by providing useful training and developmental programs, by providing assurance of security during stressful times, and by promoting justice and fairness in the way policies are implemented and rewards distributed (Eisenberger & Stinglhamber, 2011). Importantly, care should be taken to maximally limit perceptions of inequity in the availability of these improved support mechanisms. Finally, programs designed to sensitize managers to the benefits of adopting a more transformational approach, and to provide them with tools on how to implement such an approach, might prove beneficial.
Supplemental Material
Supplemental Material - A Multilevel Person-Centered Perspective on the Role of Job Demands and Resources for Employees’ Job Engagement and Burnout Profiles
Supplemental Material for A Multilevel Person-Centered Perspective on the Role of Job Demands and Resources for Employees’ Job Engagement and Burnout Profiles by Nicolas Gillet, Alexandre J. S. Morin, and Ann-Renée Blais in Group & Organization Management
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: Preparation of this paper was supported by grants from the Canadian Institute for Military and Veteran Health Research (CIMVHR) and from the Social Science and Humanities Research Council of Canada (435-2018-0368).
Supplemental Material
Supplemental material for this article is available online.
Associate Editor: Simon Restubog
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
