Instead of being two opposite poles, burnout and engagement are viewed as independent, yet negatively correlated states of mind (Schaufeli & Bakker, 2004). Burnout was defined as ‘a syndrome of chronic exhaustion, a cynical, negative attitude regarding work, and reduced professional efficacy that could occur in any job (Maslach, Schaufeli, & Leiter, 2001)’ (Bakker & Demerouti, 2017, p. 273). Burnout is a tripartite syndrome (Maslach et al., 2001) consisting of emotional exhaustion, feelings of depersonalisation (also disengagement) and a diminished sense of personal accomplishment. Emotional exhaustion refers to ‘the depletion or draining of emotional resources caused by interpersonal demands’ (Schaufeli & Salanova, 2014, p. 296). Depersonalisation refers to ‘the development of negative, callous and cynical attitudes towards the recipients of one’s services’ (Schaufeli & Salanova, 2014, p. 296). Reduced personal accomplishment is ‘the tendency to evaluate one’s work with recipients negatively’ (Schaufeli & Salanova, 2014, p. 296). Although the multidimensional nature of this construct has been vigorously debated (e.g. Maslach et al., 2001), many researchers advocate a two-dimensional concept. For example:

Therefore, researchers argued that the specificity of the burnout syndrome lies in the combination of general reactions linked to stress (i.e. exhaustion dimension) and specific attitudinal manifestations that signal a crisis in the individual–work relationship (i.e. depersonalisation dimension) (Schaufeli & Salanova, 2014). Therefore, throughout the present study, we focused on these two principal components of burnout (i.e. exhaustion and depersonalisation).

Since the beginning of the 21st century and with the influence of positive psychology (Seligman & Csikszentmihalyi, 2000), research has begun to focus on a phenomenon more positive in nature, that is, work engagement. Work engagement refers to ‘a positive, fulfilling, work-related state of mind that is characterized by vigor, dedication, and absorption’ (Schaufeli, Salanova, Gonzalez-Roma, & Bakker, 2002, p. 74). Vigour is ‘characterized by high levels of energy and mental resilience while working, the willingness to invest effort in one’s work and persistence even in the face of difficulties’ (Schaufeli & Salanova, 2014, p. 300). Dedication refers to ‘being strongly involved in one’s work and experiencing a sense of significance, enthusiasm, inspiration, pride and challenge’ (Schaufeli & Salanova, 2014, p. 300). Absorption is ‘characterized by focused attention, being fully concentrated and happily engrossed in one’s work, whereby time passes quickly and one has difficulties detaching oneself from work’ (Schaufeli & Salanova, 2014, p. 300).

The JD-R model is mainly used to explain burnout (e.g. Bakker, Demerouti, & Euwema, 2005) and work engagement (e.g. Bakker, Hakanen, Demerouti, & Xanthopoulou, 2007). This model assumes that all types of job characteristics can be modelled using two different categories, namely, job demands and job resources (Proposition 1; Bakker & Demerouti, 2017). Job demands refer to ‘physical, psychological, social, or organizational aspects of the job that require sustained physical and/or psychological effort and are therefore associated with certain physiological and/or psychological costs’ (Bakker & Demerouti, 2017, p. 274). Job resources describe:

The central proposition of this model is the assumption of dual processes to well-being (Proposition 2; Bakker & Demerouti, 2017). Within the first process of the JD-R model, the motivational process, job resources are assumed to activate a motivational process whereby resources may:

As mentioned by Schaufeli and Salanova (2014), such resources may be located notably at the interpersonal level (i.e. social support from supervisor, recognition and rewards). The availability of job resources stimulates employee’s motivation notably in the form of work engagement (e.g. Bakker et al., 2007). This motivational process is in line with the conservation of resources (COR) theory (Hobfoll, 2002), arguing that people are motivated to obtain, retain, foster and protect things that they value most, known as resources. Therefore, individuals will try to expand these resources if possible and accumulate them. People invest resources in order to enrich their pool of resources. The idea is that initial resources generate more resources, ultimately creating a spiral gain of resources. The accumulation and gain of resources may result in positive outcomes such as better well-being (Hobfoll, 2002). According to Chen, Westman and Eden (2009), work engagement can be conceptualised as a result of continuous resource gain.

Within the second process, the health impairment process, job demands:

This energetic process is also in line with the COR theory (Hobfoll, 2002). Over time, increasing demands threaten one’s available resources. People try to react to resource losses by investing resources. However, when resource losses are not counterbalanced by resource gains in situations of resources investment, a progressing stressful process appears. This stressful process also appears when no action is taken facing situations of resource losses. When losses continue, there is an accelerated negative effect of ongoing loss spirals (i.e. loss spirals of resources) which can lead to burnout.

Consequently, job demands drain more and more individuals’ resources, leading them to perceived burnout.

Even if job demands are known to play a crucial role in the health impairment process, they could also have a motivational role. In line with this idea, LePine, Podsakoff and LePine (2005) distinguished two types of job demands. The first one is the hindrance job demands that refer to ‘job demands or work circumstances’ (e.g. role conflict, role ambiguity, etc.) ‘that involve excessive or undesirable constraints that interfere with or inhibit an individual’s ability to achieve valued goals (Cavanaugh, Boswell, Roehling, & Boudreau, 2000)’ (Bakker & Demerouti, 2017, p. 277). The second one is the challenge job demands (e.g. workload, time pressure, responsibility, etc.) that are defined as ‘demands that cost effort but that potentially promote personal growth and achievement of the employees (Podsakoff, LePine, & LePine, 2007)’ (Bakker & Demerouti, 2017, p. 277). The latter type of demands could be considered as ‘rewarding work experiences well worth the discomfort involved, and are therefore considered as “good” stressors’ (Bakker & Demerouti, 2017, p. 278). However, depending on the context, a challenge demand may be experienced as a hindrance demand, and vice versa.

Many empirical studies have tested these dual pathways, which have an influence on workers’ well-being.

Concerning the motivational process, through their study conducted on teachers, Bakker et al. (2007) demonstrated that supervisor support enhanced work engagement. In the same vein, Laschinger and Finegan (2005) showed that recognition from supervisors was related to employees’ work engagement. Even if job resources are related to such positive outcomes, they may also be related to negative outcomes. Indeed, according to the COR theory (Hobfoll, 2002), job resources should be negatively related to burnout because individuals with greater resources are more capable to meet demands and to protect themselves from the strains of further resource depletion. Empirical studies highlighted that perceived supervisor support was negatively related to burnout (e.g. Rich et al., 2010). The lack of recognition has negative consequences on variables such as burnout, turnover and a low sense of well-being (e.g. Amutio et al., 2009). Indeed, a failed reciprocity between effort and rewards is likely to elicit negative emotions, and conversely, appropriate rewards (e.g. recognition) evoke positive emotion and well-being (Siegrist, 2000).

Concerning the energetic process, meta-analyses have found that workload is amongst the key determinants of emotional exhaustion (e.g. Bowling, Alarcon, Bragg, & Hartman, 2015). Emotional demands contributed positively to emotional exhaustion and disengagement (e.g. Verweij et al., 2017). Even if job demands are related to negative outcomes such as burnout, they may also be related to positive outcomes like work engagement. Job demands, such as workload, drain employees’ energy, exhausting them mentally (Schaufeli & Bakker, 2004). Workers perceiving that their resources are drained because of workload may engage in their work into a lesser extent in order to protect their remaining resources (Hobfoll, 2002).

Throughout their life, people have to perform multiple social roles (i.e. parent, husband or spouse, employee, etc.). The question of compatibility of these roles is important in the prediction of workers’ health and well-being. Several models or theories have been advanced to explain the relationship between work and family roles.

According to the scarcity hypothesis (Sieber, 1974), people have limited resources (i.e. time and energy) and are therefore unable to satisfy the demands of all their roles. When too many demands are placed on an individual’s limited resources, he or she may experience work–family conflict generally defined as ‘a form of inter-role conflict in which the role pressures from the work and family domains are mutually incompatible in some respect’ (Greenhaus & Beutell, 1985, p. 77). In situations where work-role demands impede the performance of family responsibilities, workers may perceive WFC between their work and family lives (Netemeyer, Boles, & McMurrian, 1996).

According to the expansion theory (Sieber, 1974), by having multiple roles, individuals have access to more resources, leading them to expand their level of energy and increase their psychological well-being.

Therefore, individuals may experiment work–family enrichment when, because of participation and experience in one domain, the quality of their life in the other domain is enhanced through the transfer of resources from one role to the other (Greenhaus & Powell, 2006). When work provides resources that enhance the individual’s functioning in the family domain, workers may perceive WFE (Greenhaus & Powell, 2006).

Job demands generally imply that employees devote such time and efforts (i.e. resources) to their work that they are left with insufficient time and energy for their personal and familial activities (i.e. depletion of resources; Bakker & Demerouti, 2007). Amongst job demands, workload and emotional load are highlighted by numerous studies as importantly associated with WFC and/or WFE. According to the COR theory (Hobfoll, 2002), the negative effect of these job demands on employees’ functioning in their family domain is because of the depletion of resources resulting from the important efforts required by workload and emotional load.

This resource depletion may hamper the amount and the effectiveness of resource transferred from the work to family domain, causing WFC and preventing WFE. Indeed, high workload reduced the perception of WFE (e.g. Grzywacz & Marks, 2000) and increased the perception of WFC (e.g. Molino, Cortese, Bakker, & Ghislieri, 2015), whereas emotional job demands increased the perception of WFC (e.g. Peeters et al., 2005).

Concerning job resources, a resourceful work environment creates a sense of personal control, self-esteem, self-efficacy, self-confidence, psychological resources and positive emotions (Greenhaus & Powell, 2006), which have positive impacts on workers’ ability to manage their work and family lives. Job resources may also stimulate the motivational process (Bakker & Demerouti, 2017) facilitating enrichment between work and family. According to the enrichment process (Greenhaus & Powell, 2006), resources gained from work may be transferred to the family in a direct way (i.e. transfer of skills, behaviours and rewards; referring to the instrumental path) or an indirect way (i.e. transfer of mood and emotions; referring to the affective path), helping individuals to perform better in the family domain. Support at work indicates that emotional and material resources or help are available when needed to allow workers to carry out their job more effectively (Rhoades & Eisenberger, 2002). Support at work also enhances workers’ perceived competence (i.e. self-efficacy) and a sense of being able to cope effectively with the demands of various roles (Rhoades & Eisenberger, 2002). Social support generates feelings of love, care and value (Hobfoll, 2002) that enable individuals to more effectively transfer resources gained in one domain to the other domain. The more the employees feel supported, the more they know that help is available, perceiving WFC less and WFE more. Indeed, supervisor support has been associated with lower WFC (e.g. Bowling et al., 2015) and higher WFE (e.g. Nicklin & McNall, 2013). We are not aware of any study investigating the relationship between recognition from supervisors and WFE or WFC. However, we reasonably assume that positive emotion resulting in a reward such as recognition from a supervisor (Siegrist, 2000) and the positive psychological functioning resulting from recognition at work (i.e. autonomy, resilience, self-esteem and optimism; Merino & Privado, 2015) can be transferred to the family domain (Greenhaus & Powell, 2006).

Workers who feel recognised and valued by supervisors will be better able to manage their work and family lives effectively (i.e. perception of lower WFC), and enhance the quality of life in their family sphere (i.e. perception of higher WFE).

Concerning WFC’s outcomes, it is well established that conflict has a negative influence on health and well-being (e.g. Amstad, Meier, Fasel, Elfering, & Semmer, 2011). From a COR perspective (Hobfoll, 2002), resources (such as time, money, mental and physical energy, etc.) are lost in the process of managing work and family demands. When this resource depletion is not offset by resource gain, burnout occurs. Indeed, several studies found that WFC can foster emotional exhaustion (e.g. Demerouti, Bakker, & Schaufeli, 2005) and disengagement (e.g. Verweij et al., 2017). In order to protect their remaining resources in situations of resource depletion, workers perceiving WFC are more likely to reduce their level of work engagement, which is also in line with the COR theory (Hobfoll, 2002). Indeed, WFC was found to have a negative impact on work engagement (e.g. Babic et al., 2017).

Concerning WFE’s outcomes, it is established that enrichment has a positive influence on health and well-being (e.g. Zhang, Xu, Jin, & Ford, 2018). According to the COR theory (Hobfoll, 2002), being involved in several roles has advantages for people because of greater access to resources. Work-to-family enrichment can be viewed as a resource-generating process which could protect from burnout (Tement & Korunka, 2015). Moreover, when employees perceive that their work provides something (i.e. resources) beneficial to themselves or their family (i.e. WFE), they want to obtain or gain more resources (COR theory; Hobfoll, 2002) and consequently they get more involved in their job. Indeed, empirical research found that when people experience WFE, they report putting greater efforts into their work and being better engaged in their job (e.g. Wayne et al., 2004).

Even if research on WFE and burnout is limited (e.g. Dishon-Berkovits, 2014), the few studies examining these relationships found that WFE was negatively related to burnout (e.g. Verweij et al., 2017).

Therefore, based on these elements, we postulated that:

In addition to the main effects of job demands and job resources on well-being (i.e. Proposition 2), the JD-R model also postulates that job resources can buffer the impact of job demands on strain (Proposition 3; Bakker & Demerouti, 2017). In other words, job resources may offset the negative impacts of job demands on employee well-being. Several studies have provided evidence for such an interaction effect. For example, Xanthopoulou et al. (2007) conducted a study amongst home care professionals and found that job resources (i.e. autonomy, social support, performance feedback and opportunities for professional development) buffered the relationship between job demands (i.e. emotional demands, patient harassment, workload and physical demands) and burnout.

Based on these elements, we postulated that:

Figure 1 depicts the hypothesised theoretical model.

Analyses were performed using Mplus 6 (Muthén & Muthén, 2010). We assessed the measurement models through a series of confirmatory factor analyses to evaluate the independence of constructs examined in our study. Concerning hypotheses 1 and 2, we proceeded with the assessment of the hypothesised structural relationships amongst latent variables. We used the bootstrapping technique to estimate indirect effects (Preacher & Hayes, 2008). Concerning Hypothesis 3, we tested the moderation hypotheses using the two-step estimation procedure proposed by Klein and Moosbrugger (2000) for estimating latent moderated structural (LMS) equations in Mplus. By using the balancing technique (Little, Cunningham, Shahar, & Widaman, 2002), we reduced the number of items to three for each factor (except for work engagement because the three sub-dimensions were already evaluated using three items).

We compared several nested models in order to test the distinctiveness of the constructs included in our study. Firstly, we examined the fit of our hypothesised 11-factor model (that comprises emotional load, workload, recognition from superiors, perceived supervisors support, exhaustion, disengagement, vigour, dedication and absorption) as well as a series of more constrained measurement models. Chi-square difference tests were used to compare the fit of these nested models with that of the 11-factor model. Results indicated that the 11-factor model was significantly superior to all alternative models.

Consequently, we treated these 11 constructs as independent from each other in subsequent analyses. Table 1 presents fit indices of some of these alternative models.

Means, standard deviations, Cronbach’s alphas and correlations amongst variables are presented in Table 2.

Internal consistency reliabilities ranged from 0.72 to 0.93. Table 3 presents fit indices for the hypothesised structural model (Model 0) and several alternative models (models 1–20). In all models, disturbance terms of our outcomes were allowed to correlate. Model 0 fit the data reasonably well, as indicated by the following indices: χ²(df) = 1060.85(525), root-mean-square error of approximation (RMSEA) = 0.07, comparative fit index (CFI) = 0.90. We compared this Model 0 with 20 alternative nested models containing additional paths that were theoretically plausible. The Chi-square difference test was used to compare the fit of these alternative nested models with that of Model 0. Results of the comparisons indicated that Model 19 fit the data reasonably well (χ²[df] = 1023.90(520), RMSEA = 0.07, CFI = 0.91) and was significantly superior to all models. We thus retained Model 19 as the best fitting model. In Model 19 we added, starting from Model 0: (1) a path from emotional load to exhaustion, (2) paths from recognition from supervisors to disengagement, vigour and absorption and (3) a path from perceived supervisor support to dedication.

Standardised parameter estimates for this model are shown in Figure 2. For ease of presentation, we show the structural model rather than the full measurement model. Amongst the four independent variables included in the model, only two were significantly related to WFC. Workload and perceived supervisor support were found to be, respectively, positively and negatively associated with WFC. In turn, WFC was found to be positively associated with exhaustion, disengagement and absorption and was negatively associated with vigour. As specified in Table 4, the results of the bootstrap analyses indicated that the indirect effects of workload on burnout and work engagement (except dedication) through WFC were significant, with WFC playing the role of a total mediator. These findings partially supported Hypothesis 1a (i.e. WFC will mediate the relationships between job demands and well-being). The indirect effects of perceived supervisor support on burnout and work engagement (except dedication) through WFC were significant (see Table 4), with WFC playing the role of a total mediator. These findings partially supported Hypothesis 1b (i.e. WFC will mediate the relationships between job resources and well-being).

Only one of the four independent variables studied was significantly related to WFE. This finding did not support Hypothesis 2a (i.e. WFE will mediate the relationships between job demands and well-being).

Recognition from supervisors was found to be positively associated with WFE. In turn, WFE was found to be negatively associated with exhaustion and disengagement and was positively associated with vigour and dedication. Moreover, recognition from supervisors was also directly and negatively related to disengagement and positively related to vigour. The indirect effects of recognition from supervisors on burnout and work engagement (except absorption) through WFE were significant (see Table 4). Thus, WFE totally mediated the effects of recognition from supervisors on exhaustion and dedication. Moreover, WFE partially mediated the effects of recognition from supervisors on disengagement and vigour. These findings partially supported Hypothesis 2b (i.e. WFE will mediate the relationships between job resources and well-being).

The results also highlighted three direct effects. Emotional load was found to be positively related to exhaustion. Perceived supervisor support was found to be positively related to dedication. Recognition from supervisors was found to be positively related to absorption.

To investigate moderation, Hypothesis 3, we tested structural models using latent interactions separately. The relative fit of LMS models versus the model without interaction terms was determined using log-likelihood ratio tests. Only two log-likelihood ratio tests were significant, that is, Mod 3: D = −2 [(−8202.999)–(−8200.662)] = 4.67(1) (p < 0.05); Mod 4: D = −2 [(−8202.999)–(−8197.915)] = 10.17(1) (p < 0.01). Within Mod 3, the recognition from supervisors × workload interaction on WFC was significant (−0.16, SE = 0.08, p < 0.05). High workload when combined with high recognition from supervisors resulted in lower perception of WFC (see Figure 3a for a graphical representation of this moderation). Within Mod 4, the perceived supervisor support × workload interaction on WFC was significant (−0.21, SE = 0.07, p < 0.01). High workload when combined with high perceived supervisor support resulted in lower perception of WFC (see Figure 3b for a graphical representation of this moderation). These findings partially supported Hypothesis 3 (i.e. job resources will moderate the relationships between job demands and WFC).

Based on recent recommendations (Baeriswyl et al., 2016), the present study had three objectives. The first two objectives consisted of investigating the mediating role of WFC (i.e. hypotheses 1a and 1b) and WFE (i.e. hypotheses 2a and 2b) in the relationships between work environment (i.e. emotional load and workload as job demands; supervisor support and recognition from supervisors as job resources) and well-being outcomes (i.e. work engagement and burnout). The third one consisted of investigating the buffering effect of job resources in the job demands–WFC relationships (i.e. Hypothesis 3). These hypotheses were tested on employees working in the medical division of a Belgian Federal Public Service active in the field of security.

On one hand, our findings indicate that WFC mediates the effects of workload and perceived supervisor support on burnout and work engagement (except the dedication sub-dimension). These results are consistent with some previous research investigating the relationships between workload and WFC (e.g. Molino et al., 2015), between supervisor support and WFC (e.g. Bowling et al., 2015), between WFC and burnout (e.g. Verweij et al., 2017) and between WFC and work engagement (e.g. Babic et al., 2017). By depleting individual’s resources (Bakker & Demerouti, 2007), workload increases workers’ perception of WFC, whereas supervisor support through energising the motivational process (Bakker et al., 2005) reduces the perception of WFC. The resource depletion because of the process of juggling work and family demands (Hobfoll, 2002) leads workers to experience burnout and to reduce their level of vigour in order to protect their remaining resources (Hobfoll, 2002). Surprisingly, WFC is found to be positively related to absorption. This unexpected result could be explained by the reactive compensation theory (Edwards & Rothbard, 2000) postulating that unpleasant experiences in one domain are remedied by pleasant experiences in another domain. By perceiving unpleasant experiences in their family domain, workers confronted with WFC could make some efforts at work (by being fully concentrated and engrossed in their work) to increase positive experiences within their professional sphere and therefore counter the negative experiences in the family sphere.

On the other hand, our findings indicate that WFE mediates the effects of recognition from supervisors on burnout and work engagement (except the absorption sub-dimension). These results are consistent with some previous research investigating the relationships between WFE and burnout (e.g. Verweij et al., 2017) and between WFE and work engagement (e.g. Wayne et al., 2004). Being recognised by supervisors generates happiness and a positive mood (e.g. Siegrist, 2000), which overflows into the family domain and improves the quality of workers’ life in their family sphere, leading them to perceived WFE. The resource-generating process resulting in the WFE (Tement & Korunka, 2015) reduces workers’ experience of burnout and increases their level of vigour and dedication to work in order to obtain or gain more resources (Hobfoll, 2002). We found no impact of job demands on WFE experiences. This is consistent with the fact that the relationships between job demands and positive outcomes are less pronounced than the relationships between job demands and negative outcomes (Schaufeli & Bakker, 2004). This is also in line with the fact that WFE originates mainly in a resource-rich work environment (Carlson et al., 2006).

One interesting finding of the present study concerns the fact that recognition from supervisors is related to WFE, whereas support from supervisors is related to WFC. These results assume that these two job resources operate in different ways regarding work-to-family interface. The fact that WFC is only influenced by supervisor support highlights the importance of social support. When individuals feel socially supported at work, they feel cared for by socially related others and feel that they have access to help (Hobfoll, 2002) and that emotional and tangible support is available when needed (George, Reed, Ballard, Colin, & Fielding, 1993).

By perceiving more social support, workers’ emotional and psychological supplies for coping with daily stressors increase, and perceptual appraisals of stressors decrease (Jex, 1998). By being supportive, supervisors indicate that they are sensitive to the workers’ needs (notably their needs related to family obligations). Employees feeling supported by supervisors would expect managers to show greater empathy and emotional support when they are confronted with difficult work-related problems. When individuals have more social support and content specifically for managing work–family issues, these positive dynamics may overflow into the family role, thereby reducing WFC (e.g. Frone, Russell, & Cooper, 1992).

Support received from supervisors is viewed by workers as a sign that managers take care of their well-being and are sensitive to their needs, thus reducing the resource depletion process induced by WFC.

Recognition from supervisors refers more to the resource generation and motivational processes, therefore increasing the perception of WFE. Indeed, incentives, rewards and recognitions are the prime factors that impact employee motivation (Danish & Usman, 2010). Implementing effective rewards and recognition within an organisation generates a favourable working environment, which motivates employees to perform better in their work (Freedman, 1978). Taking recognition as their feelings of value and appreciation boosts up the morale of workers, which increases their productivity. According to the instrumental path of the enrichment process (Greenhaus & Powell, 2006), rewards gained are directly transferred from one role to another, and help one to perform better in the receiving domain, increasing workers’ perception of WFE. In the same vein, the increased performance at work generated by recognition produces a positive environment at work which seeps in and positively impacts how one feels, acts and behaves in the receiving domain (i.e. affective path of the enrichment process) Greenhaus & Powell, 2006).

These results emphasise the fact that recognition from supervisors activates the resource generation and motivational processes more, whereas support from supervisors is more important in the resource depletion and health impairment processes. Recognition can motivate and help to build feelings of confidence (Keller, 1999), as well as encourage employees to extend their efforts (Freedman, 1978). Perceiving supervisors as supportive, who take care of worker’s well-being, is an important resource for workers. Social support notably reduces the likelihood that work situations will be perceived as stressful impacting the perception of work–family conflict (Carlson & Perrewe, 1999).

However, the fact that we found no significant relationship between supervisor support and WFE might be explained by the very low level of WFE experienced by the respondents of our study (i.e. mean of 0.80, range = 0–3). Nicklin and McNall (2013) found that the relationship occurred when WFE was experienced at a higher level. Possibly, within a sample perceiving more WFE, the support–WFE relationship would become significant. According to Greenhaus and Powell (2006), the social capital resources generated from work can improve the quality of life in the family. This issue should be investigated in future studies.

The study findings also indicate some direct effects of work environment on well-being. Firstly, emotional load is positively related to exhaustion. In line with previous studies (e.g. Verweij et al., 2017), emotional load increases exhaustion by draining individuals’ resources. The fact that emotional load was found to be unrelated to the second sub-dimension of burnout (i.e. disengagement) is consistent with the fact that job demands relate more strongly with exhaustion than with the other sub-dimensions of burnout (Alarcon, 2011). Secondly, support from supervisors is directly related in a positive way to only one dimension of work engagement, namely, dedication. By fulfilling socio-emotional needs (e.g. Rhoades & Eisenberger, 2002), support leads workers to be more involved in their work. The effects of support on the other four investigated outcomes were found to be indirect through WFC. Thirdly, recognition from supervisors is directly positively related to vigour and absorption and directly negatively related to disengagement. These results are consistent with some previous research investigating the relationships between recognition from supervisors and work engagement (e.g. Laschinger & Finegan, 2005) and between recognition from supervisors and burnout (e.g. Amutio et al., 2009). By stimulating the resource generation and motivational processes (Freedman, 1978), recognition from supervisors increases the level of workers’ vigour and absorption and reduces workers’ level of disengagement. The effects of recognition on the other investigated outcomes (i.e. dedication and exhaustion) were found to be indirect through WFE.

Considering the buffering effects, our findings indicate that both support and recognition from supervisors moderate the positive effect of workload on WFC. These results are consistent with previous research investigating the moderating role of social support (e.g. Xanthopoulou et al., 2007). These results reveal that under demanding work conditions (i.e. perception of workload), employees receiving support and recognition from their supervisors dispose more supplies and, thus, are better able to deal with these demands, protecting them from a negative outpour from the work domain into the private domain (i.e. WFC). This finding is in line with the fact that resources become salient and gain their motivational potential when employees are confronted with demanding jobs (Hobfoll, 2002).

Our study has several limitations, leading us to interpret the findings reported here with caution. The first major limitation is the fact that the data are cross-sectional, precluding any inference of causality amongst the studied constructs. Secondly, the specificity of our data makes it difficult to generalise the results to other professional sectors. Thus, it would be helpful to replicate these results amongst other organisations. Thirdly, as we used self-reported data, common method variance may have biased our results (Podsakoff, MacKenzie, & Podsakoff, 2012). However, considering that the single-factor model showed a poor fit to the data, this common method bias was partially treated (i.e. Harman’s single-factor test; Podsakoff et al., 2012). Despite these limitations, our study has some strong points too. Indeed, this study extends previous research by integrating WFE in the JD-R process to explain well-being. Our study is also in line with theories or models largely acknowledged in the literature (i.e. COR theory; JD-R model; enrichment process). Moreover, by including recognition from supervisors, our study also contributes to the literature considering that, to the best of our knowledge, no previous research has investigated the relationship between recognition and work-to-family interface.

From a practical perspective, our findings suggest that workload, support and recognition from supervisors may play a crucial role in the emergence of WFC, WFE, burnout and work engagement. Promoting a supportive work environment and reducing the workload may contribute to reduce the perception of WFC, and therefore enhance workers’ well-being. To promote a positive social climate at work, managers have to provide assurance in forms such as empathy for the employee’s difficulties, confidence in the employee’s abilities and offers of aid as needed. To reduce workload, organisations should, for example, provide workers adequate opportunities to gain control of the amount of time available to spend on various job requirements. Providing training programmes focusing on time and stress management skills could help employees set realistic goals and prioritise them so as to better cope with high workload. However, workload is often difficult to reduce at short notice. But, as suggested by our study, support and recognition from supervisors buffer the impact of workload. Therefore, to inhibit the health impairment process induced by workload, it is important to invest in social support and in social recognition processes. Moreover, if recognition from supervisors buffers the workload–WFC relationship, such job resources also have motivational potential which fosters WFE. To promote recognition, managers could spontaneously praise people, thank the person by name, specifically state what workers did that is being recognised or highlight workers’ competencies and value.