1 Introduction

Over recent decades, transformations have occurred in various fields, e.g., demographic, technological and organizational, that have increased interest in the relationship between work and family domains (Aryee et al. 2005; Greenhaus and Parasuraman 1999). Despite research has been dominated by the analyze of conflict between work and family (Greenhaus et al. 2006), a positive perspective has emerged in line with positive psychology (Snyder and Lopez 2002). Scholars have increased their interest in examining how involvement in one role positively influences the other role. In the specific case of enrichment, this describes the process by which experiences in one role improve individual performance in the other role (Greenhaus and Powell 2006). In work-to-family enrichment (WFE), the resource (e.g., material, skills and perspectives, flexibility, psychological and physical social-capital) gains in work improved the individual performance in family role.

Studies have demonstrated the relationship between WFE and employees’ well-being, such as life and family satisfaction (Van Steenbergen et al. 2007), physical health (McNall et al. 2010) and mental health (McNall et al. 2010). Thus, it is in the best interest of the organizations to execute practices that give rise to WFE. In this way, researchers have identified some organizational and job characteristics as WFE antecedents. For example, Bhargava and Baral (2009) demonstrated that job characteristics, i.e., autonomy, skill variety, task identity, task significance, feedback from one’s job, feedback from others, dealing with others, and supervisor support were predictors of WFE. Wayne et al. (2006) demonstrated that informal or emotional support within the work domain, rather than formal or instrumental support, was associated with greater WFE. Siu et al. (2010) revealed that WFE was not only positively related to organizational characteristics, namely family-friendly organizational policies and supervisor support, but also to job autonomy. Carlson et al. (2011) also suggested that job resources, i.e., skill discretion and job security are directly and positively related to WFE.

However, very few attempts have been developed to integrate these job characteristics into a specific job design to explain WFE and its relationship with employees’ well-being. Our study is designed to contribute to overcoming this limitation. First, we propose that job characteristics included in Job Demands-Control-Support (JDCS) Model (Karasek and Theorell 1990) are important to explain WFE, which, in turn, is related to employees’ well-being. Concerning the well-being, as Ryan and Deci (2001) suggested well-being is a complex construct that concerns to optimal experience and functioning. Moreover, they affirm that opportunities for positive affect and life satisfaction can be detracted by functional limitations that illness may provoke. Other authors also found empirical results that supported the association between health status and well-being (Okun et al. 1984; Ryan & Frederick 1997; Ryff and Singer 2000). Based on this assumption, we decided to measure general well-being through satisfaction of life and health perceptions.

Secondly, we propose that a high-performance work system (HPWS) is an important organization characteristic that is related to these job characteristics. HPWS refers to a group of separate but interconnected human resource (HR) practices that involve flexible job assignments, rigorous and selective staffing, extensive training and development, developmental and merit-based performance appraisal, competitive compensation, and extensive benefits (Takeuchi et al. 2007). The literature asserts that these practices increase employees’ knowledge, skills, and abilities (Becker and Huselid 1998; Delery and Shaw 2001) and the result is greater job satisfaction, lower employee turnover, higher productivity and better decision making (Becker et al. 1997). However, various authors have questioned these assumptions because they postulated that there are differences between intended HPWS policies and employees’ experience (Wright and Boswell 2002; Paul and Anantharaman 2003; Kinnie et al. 2005). Taking this into account we opted to consider the employees’ perceptions of HPWS. Concretely, we propose that employees’ HPWS perceptions are related to the employees’ perceptions of job demands, job autonomy and supervisor support (Holman 2005). By looking at the antecedents of WFE, this study goes beyond previous studies, exploring not only job characteristics but also a system of organizational practices related to these same job characteristics.

2 Literature Review and Hypotheses

2.1 Job Characteristics and Work-to-Family Enrichment

Concerning the job characteristics, job demands, job autonomy and supervisor support are key variables in the recognized job demands control-support (JCDS) model (Karasek and Theorell 1990) to explain employees’ well-being. In a general explanation, in JDCS model authors argued that high levels of strain inhibit learning, whereas high levels of learning inhibit stress.

Job autonomy is one of the most studied job resources and is related to workers’ possibilities of making decisions about their work (Karasek et al. 1998). Job design literature (Humphrey et al. 2007) has suggested that job autonomy allows employees to determine in which order and with whom they accomplish tasks, and that they should have the opportunity to experiment with different forms of task completion and to take responsibility for their results. Thus, job autonomy is a job characteristic that enriches the job domain and develops employees’ competences in terms of problem resolution and creativity (Volmer et al. 2012). Therefore, we expect employees who have job autonomy to develop problem resolution, creativity and flexibility skills that may be seen as a resource for the family domain and consequently, enrich the family. Moreover, with more job autonomy, employees manage their job demands, e.g., workload, and, in turn, this experience can also be profitable to the family domain, e.g., managing priorities in domestic tasks within the family domain (Friedman and Greenhaus 2000; Grzywacz and Butler 2005; Grzywacz and Marks 2000).

Supervisor support is defined as the extent to which employees feel they can count on their supervisors. The relationship with supervisors is sustained by both mutual trust and openness and the richness of verbal communication and bi-directional feedback between leaders and members (House and Shamir 1993). Hence, supervisor support acts as an efficient behavior model for employees which can be transferred to the family domain, i.e., transferring the support attitude to the family; understanding family problems and needs, and in this way, giving privilege to their benefits and stimulating the skills of family members. Voydanoff (2004) underlines supervisor support as an important resource for work family facilitation. The study of Wayne et al. (2006) observed that employees’ perception of supervisor support was critical for them to successfully integrate work and family. Bhargava and Baral (2009) suggested that supervisor support predicted WFE and represented a work-domain resource to assist employees’ effort to integrate work and family roles by providing instrumental support. In the same way, Siu et al. (2010) found a positive relationship between supervisor support and WFE.

Finally, job demands are defined here as workload and time-based are a threat to the work-home harmony (Beham et al. 2010; Brummelhuis and Bakker 2012). In fact, studies have shown that these job demands are associated with greater work–family conflict (Grzywacz and Butler 2005; Valcour 2007). The findings point to the prediction that the more time or energy one spends in a role, or the more involved one is in a role, the more interference in the secondary role (Byron 2005). According to the scarcity argument (Goode 1960), within-domain work demands limit the ability of employees to successfully manage their non-work domain responsibilities and are a negative work-to-family interference. Thus, this adverse characteristic contributes to a waste of time and energy and, as such, does not promote employees’ capacity to enrich the family domain.

Lower strain and high levels of learning (or outcomes of learning, such as feelings of efficacy and mastery; Parker and Sprigg 1999) occur in jobs in which there is low demand, high job autonomy and high support, whereas workers in high demand, low autonomy and low support jobs report high strain and low levels of learning (Luchman and González-Morales 2013; Van der Doef and Maes 1999). WFE is a model that privileges the presence of resources at work and define that through the instrumental path, employees acquire skills that are useful in the family context. In this sense, the combination of job characteristics in JCDS model is crucial for resources and new learning skills that WFE model focused. In this vein, we consider that these job characteristics are fundamental for WFE model (Greenhaus and Powell 2006) since one of the assumptions of this model is that WFE generate resources. In this sense, we considerer that these resources are more likely to arise in a work context with prevalence of job characteristics that lead to learning, as Karasek and Theorell (1990) proposed.

In this way, our first hypothesis establishes this relationship as follows:

H1

Job characteristics are related to WFE in a way that job demands, i.e., workload and time pressure, are negatively related to WFE and autonomy and supervisor support are positively related to WFE.

2.2 Work-to-Family Enrichment as a Mediator Between Job Characteristics and General Well-Being

The underlying theoretical argument of the JDCS model (Karasek and Theorell 1990) is that individual physiological strain results from the effects of job demands and the amount of job control (job autonomy) and available job support. As predicted in the above hypothesis, the additive effect of these job characteristics seems to be relevant to WFE.

Furthermore, research also points to the relationship between WFE and employees’ well-being: job satisfaction (Bhargava and Baral 2009; McNall et al. 2010), life satisfaction (Van Steenbergen et al. 2007), quality of life (Grzywacz and Marks 2000) and better mental and physical health (Hammer et al. 2005).

Carlson et al. (2011) presented a study were they examined the associations of three job resources (job security, skill discretion, and schedule control) with WFE and they considered subsequent impact of WFE on women’s health (physical and mental health). They verified that skill discretion and job security, both job resources, directly and positively related to enrichment and that WFE positively predicted health. In a similar vein, we intend to examine the association of three job characteristics with WFE, which in turn, related to employees’ well-being. Concretely, we analyze the relationship of work-to-family enrichment on health perceptions and satisfaction with life. Thus, in this study we expect WFE to be a crucial mechanism enabling an explanation of the relationship between job characteristics and employees’ general well-being, i.e., health perceptions and satisfaction with life.

As such, hypothesis 2 has been established:

H2

The relationship between job characteristics and employees’ general well-being is mediated by WFE.

H2a

The relationship between job characteristics and employees’ health perceptions is mediated by WFE.

H2b

The relationship between job characteristics and employees’ satisfaction with life is mediated by WFE.

2.3 High-Performance Work System and Job Characteristics

The human resources management system is designed to ensure that employees’ behavior is consistent with the requirements of functions according to the strategy chosen by the organization (Wood et al. 2006). More specifically, the high-performance work system (HPWS) refers to a group of separate but interconnected HR practices designed to enhance employees’ skills and effort (Datta et al. 2005). HPWS practices contribute to improving employee performance in three interrelated parts: (1) people flow, including selective staffing, training, employee mobility such as broad career paths and promotion within the firm, and guarantee of job security; (2) appraisal and rewards, including long-term and results-orientated performance appraisal, and compensation and other benefits, such as extensive, open-ended rewards; (3) employment relations, including job design, such as broad job descriptions and flexible job assignments, and encouragement of participation (Zhang and 2008).

In fact, in the presence of HPWS, alternative job redesigns are adopted, including the promotion of job autonomy (Armstrong 2009). According to White et al. (2003) HPWS was designed to provide, among other things, greater participation in decision making. Accordingly, Wood et al. (2006) found a positive relationship between both initial and continuous training and workers’ autonomy. Correspondingly, empirical evidence has demonstrated that HR practices: training, participation, reward systems, and performance-related pay, directly influence the nature of the jobs, either reducing or increasing workload and time pressure and autonomy (Holman 2005). In the same way, Castanheira and Chambel (2010) found that HR involvement systems, which include training, participation, and performance-related pay, are significantly related to lower job demands and higher job autonomy.

On the other hand, the supervisor’s action is a key to human resource practices since it is he/her who operationalizes them. Supervisors are crucial as they make decisions about systems for recruiting, motivating, and developing people that will ensure that the organization has the necessary talent both to develop and execute an effective strategy (Pfeffer 1998). Therefore, the supervisor aids the successful implementation of the HR system. In fact, in high performing cultures supervisors are indispensable to making decisions and solving problems to meet goals that are congruent with the organization’s vision, mission, values and strategies (Bohlander and Snell 2007; Owen et al. 2001) Thus, in a HPWS, through his/her behaviors a supervisor ensures clear expectations, promotes membership, fosters employee involvement in decision making and solving problems, places an emphasis on the importance of quality, promotes a consistent focus on meeting customers’ needs and requirements, and encourages and rewards learning and skill development. Thus, with HPWS the presence of job autonomy and supervisor support are ensured and, in accordance with the JDCS model (Karasek and Theorell 1990), job demands are faced. So, we consider that employees’ perception of HPWS is a key antecedent to these job characteristics (Fig. 1).

Fig. 1
figure 1

The research model

Full size image

Thus, we formulate our hypothesis:

H3

Employees’ HPWS perceptions are related to job characteristics in a way that presents a negative relationship with workload and time pressure and a positive relationship with autonomy and supervisor support.

3 Method

3.1 Participants and Procedure

Participants were employees (N = 1,390) from different departments of a Portuguese bank. The participants had different functions: manager (N = 527, 37.9 %), administrative (N = 222, 16 %), commercial (N = 228, 16.4 %), technician (N = 279, 20.1 %) and other (N = 133, 9.6 %). Most respondents were males (797 = 57.4 %) compared to the number of female responses (592 = 42.6 %). Respondents answered the questionnaire online and were assured of the anonymity of their responses and the opportunity to receive feedback.

3.2 Measures

3.2.1 High-performance Work System

We measured HR practices using a seventeen item scale developed by Takeuchi et al. (2007). A sample items was “Employees are involved in decision making”, “The performance evaluation is based on clear and objective criteria”. Items were scored on a seven-point rating scale from (1) “totally disagree” to (7) “totally agree” (α = 0.94).

3.2.2 Job Characteristics

We measured job characteristics using the Job Content Questionnaire (Karasek et al. 1998), which was used in a previous Portuguese study (Castanheira and Chambel 2010). Items were: workload and time pressure (7 items)—“I have too much to do”; autonomy (4 items)—“I have the opportunity to decide how to organize my work”. Items were scored on a five-point rating scale from (1) “totally disagree” to (5) “totally agree”. (α = 0.88 for workload and time pressure and 0.79 for autonomy).

3.2.3 Supervisor Support

We measured supervisor support using five item scale developed by Scandura and Graen (1984). A sample item was “My supervisor understands my problems and needs” and “My supervisor recognizes my potential”. Items were scored on a seven-point rating scale from (1) “totally disagree” to (5) “totally agree” (α = 0.94).

3.2.4 Work-to-Family Enrichment

We measured WFE using the nine items scale of Carlson et al. (2006) that was used in a previous Portuguese study (Chambel, in press). A sample of items is “My involvement with my work helps me to understand different viewpoints and this helps me be a better family member” and “My involvement in my work helps me to develop my abilities and this helps me be a better family member”. Items were scored on a seven-point rating scale from (1) “totally disagree” to (5) “totally agree” (α = 0.94).

3.2.5 General Well-Being

We measured employees’ general well-being with the assessment of Satisfaction with Life and Health Perceptions. Satisfaction with Life was assessed with five items scale of Diener et al. (1985), which had already been used in Portugal (Neto 1992). A sample of items is—“I am satisfied with my life” and “If I could live my life again I wouldn’t change hardly anything at all”. Items were scored on a seven-point rating scale from (1) “totally disagree” to (7) “totally agree”(α = 0.88). The Health Perceptions Questionnaire developed by Ware et al. (1978) was used to assess Health Perceptions. The scale was composed of four-items—“I am as healthy as others” and “My health is excellent”. Items were scored on a five-point rating scale from (1) “definitively false” to (5) “definitively true” (α = 0.90).

It is important to note that the High-performance work system (Takeuchi et al. 2007), Health Perceptions Questionnaire (Ware et al. 1978) and Leader-Member Exchange Scale have not been previously used in Portugal. They were translated into Portuguese, and then a translator was asked to provide a back translation of the Portuguese into English (Brislin 1980). The HR manager of the Bank then read the questionnaire and confirmed the clarity and familiarity of items.

3.2.6 Control Variables

To explore the possible main effects of participants’ demographics (gender) and tenure, we included in the structural models as an ordinal variable.

4 Results

4.1 Confirmatory Factor Analysis

Our theoretical Model covered seven factors (i.e., HPWS perceptions, job demands, job autonomy, supervisor support, WFE, satisfaction with life and health perceptions) and obtained an adjusting fit (cf. Table 1). We compared this Model with another two models, one single factor Model where all items loaded on a single latent variable as also a six-factor Model where Satisfaction with life and Health perceptions were grouped together under one factor. Our analysis showed a significantly lower fit on the single factor Model as also on the six-factor Model compared to our theoretical Model. In that sense, our theoretical Model presented the best fit to the data (Table 2).

Table 1 Results of confirmatory factor analysis
Full size table
Table 2 Means, standard deviations and correlation coefficients between variables
Full size table

4.2 Structural Models

To test hypothesized relationships, we first tested our theoretical Model (Model 1) where we considered a full mediation among the constructs of the study. Model 1 provided an adequate fit for the data. All the paths in the Models were significant. Then, we compared our theoretical Model with alternative Models in which we included different paths that were also theoretically plausible (Table 3).

Table 3 Results of structural equation models
Full size table

Model 2, to which the connection between job characteristics and well-being (i.e., health perceptions; satisfaction with life) was added to test partial mediation, presented also an acceptable fit. When we compared this Model with the theoretical Model 1, this Model presented a better fit (Δχ2 (6) = 32.58, p < .001). Model 3 to which a direct connection between HPWS and WFE was added to test partial mediation also presented a good fit. Thus, we compared this Model 3 with Model 2 and this Model 3 presented a better fit. Finally, we tested Model 4 by adding the relationships established in Model 2 to the relationships established in Model 3. This Model also presented an adequate fit and when compared to Model 3, we observed significant differences. In this sense, Model 4 displayed the best fit to the data and became our final Model.

As expected, observing the standardized coefficients for each of the significant paths in Model 4 (see Fig. 2) we found a significant and negative relationship between job demands and WFE, a significant and positive relationship between job autonomy and WFE and a significant and positive relationship between supervisor support and WFE that support our Hypothesis 1.

Fig. 2
figure 2

Model 4

Full size image

To analyze our mediation hypothesis, with regard to health perceptions (Hypothesis 2a), we observed a significant and negative relationship between job demands and health perceptions and a significant and positive relationship between supervisor support and health perceptions. However, regarding job autonomy and health perceptions we did not observe a significant relationship. Through the Sobel Test, the relationship between job demands and health perceptions (Z = −2.77; p < .001), between job autonomy and health perceptions (Z = 3.97; p < .001) and between supervisor support and health perceptions (Z = 5; p < .001) were found to be mediated by WFE. Thus, we observed that the relationship between job autonomy and health perceptions was fully mediated by WFE, but the relationships between job demands and supervisor support with health perceptions were only partially mediated by WFE. These results partially support H2a. With regard to satisfaction with life (Hypothesis 2b), we observed a significant and positive relationship between job autonomy and satisfaction with life. However, regarding job demands and satisfaction with life and supervisor support and satisfaction with life we did not observe a significant relationship. Through the Sobel Test, the relationship between job autonomy and satisfaction with life (Z = 4.20; p < .001), between job demands and satisfaction with life (Z = −2.85; p < .001) and between supervisor support and satisfaction with life (Z = 5.50; p < .001) were found to be mediated by WFE. Thus, we observed that the relationships between job demands and satisfaction with life and between supervisor support and satisfaction with life were fully mediated by WFE, but the relationships between job autonomy with satisfaction with life were only partially mediated by WFE. These results partially support H2b.

We also found a significant and negative relationship between HPWS and job demands, a significant and positive relationship between HPWS and job autonomy and a significant and positive relationship between HPWS and supervisor support. Thus, hypothesis 3 was supported. However, we also observed a significant and positive relationship between HPWS and WFE, suggesting that this system of practices had a direct relationship with WFE that was not hypothesized by us.

The control variables contributed in explaining variance. Gender was positively related to WFE (β = .09, p < .001) and tenure was positively related to HPWS perceptions (β = −.07, p < .01), supervisor support (β = −.05, p < .05), job demands (β = −.07, p < .01), job autonomy (β = .06, p < .05) and health perceptions (β = −.16, p < .001).

5 Discussion

The present study has demonstrated that job characteristics, i.e., job demands, job autonomy and supervisor support, are related to WFE. More interestingly, our results suggest that WFE is an important mechanism to explain the relationship between these job characteristics and employees’ well being. Furthermore, as we expected, the employees’ perceptions of high performance work systems is relevant for explaining their perceptions of job characteristics. We can also emphasize that, beyond our original expectations, HPWS perceptions has a direct relationship with WFE and not only through job characteristics.

The results confirmed previous findings on the relationship between job characteristics and WFE (Brummelhuis and Bakker 2012; Friedman and Greenhaus 2000; Grzywacz and Butler 2005). In accordance with the principles of the JDCS model (Karasek and Theorell 1990), the results suggest that when employees experience high job autonomy and supervisor support they have resources to deal with job demands which will most likely result in the construction of positive resource transference from the work to family domain.

Regarding our mediation hypothesis, we observed different results with different dimensions of employees’ well-being, i.e., health perceptions and satisfaction with life. We verified that the relationship between autonomy and health perceptions occurs through WFE, but job demands and support maintain a direct relationship with this outcome. This result could be associated with the fact that job demands and supervisor support are job characteristics that have a direct relationship with health, as suggested by the literature (for example, Kirkcaldy et al. 2000; Shultz et al. 2010). Moreover, we verified that the relationship between job demands and supervisor support with satisfaction with life occur through WFE, but job autonomy maintains a direct relationship with this outcome. This result could be associated with the fact that job autonomy is a resource that has a direct relationship with satisfaction with life, as suggested by the literature (for example, Judge et al. 1998).

Our findings also demonstrated the importance of employees’ perceptions of High Performance Work System to job characteristics in line with the findings of previous studies (Appelbaum et al. 2000; Castanheira and Chambel 2010; Holman 2005; Wood et al. 2006). When employees’ consider that organization applies HPWS they consider have more job autonomy and support and less job demands. Consequently, in order to ensure the presence of these job characteristics, it is important for the organization to invest in selection policies, to enhance training opportunities, develop fair performance appraisals, competitive practices and equity rewards and to promote employees’ participation and empowerment.

Stemming from our prediction, we verified that HPWS perceptions were significantly and positively related to WFE. This result might be related to the affective and instrumental paths in which WFE occurs (Greenhaus and Powell 2006). Through the instrumental path, resources are directly transferred from one role to another, e.g., a skill acquired through work is transferred and applied at home, resulting in improved interaction with family members. In the affective path, authors suggested that employees who develop a positive job affect could transfer this positive affect from work to family and consider that, for this reason, work enriches the family domain. Previous studies related HPWS with affective job variables, for example, affective commitment (Takeuchi et al. 2007) and job satisfaction (e.g., Appelbaum et al. 2000). Therefore, it is possible that these positive affects may indirectly affect the family domain. On the other hand, in relation to the instrumental path and as postulated above, the HPWS enhances a broad range of employees’ skills and efforts. We may consider these skills and efforts to be directly transferred to the family domain and not only through job characteristics. As Edwards and Rothbard (2000) suggested, skills obtained in one domain may be abstracted into general knowledge structures of scripts that apply across life domains. Furthermore, the relationship between HPWS perceptions and WFE could be mediated by other job characteristics. For example skill variety, task identity, feedback from one’s job, feedback from others, dealing with others (Bhargava and Baral 2009), team resources (Hunter et al. 2010) and emotional labor (Yanchus et al. 2009).

Finally, we verified that control variables, i.e., gender and tenure, contributed to explaining variance. These results are in accordance to literature by demonstrating that gender is associated with WFE (Powell and Greenhaus 2010) and that tenure is predictive of negative job outcomes (Duffy et al. 1998)

5.1 Limitations and Future Studies

Our study contains several limitations. Firstly, data was cross-sectional and did not permit make causal interferences regarding variables relations. Secondly, data was self-reported that presented the common method variance problem. However, self-reported data seemed the most appropriate avenue to capture employees’ perceptions and appraisals of these variables (Fox and Spector 1999) and according to Spector (2006) concerns associated with relying heavily on self-reported data measurements may be overstated. Furthermore, to minimize this impact we followed several of Podsakoff et al.’s (2003) methodological and statistical recommendations. Another limitation is related to the sample since all participants were bank employees. One should therefore be careful in generalizing the results across sub-populations, different subjects and settings (Shadish et al. 2002).

Limitations notwithstanding, the results of this study underline several points for enhancing theory, practice and research in the WFE field. Firstly, the JDCS model was used to explore the relationship between job characteristics and WFE and showed that job demands, autonomy and supervisor support are essential to WFE. As nowadays it is common for employees to be constantly under job demands, organizations should put strategies in place to address work pressures and provide sufficient resources in order to promote WFE. Secondly, we observed that HPWS could have positive effects on these job characteristics and WFE. Once these HPWS practices have this important contribution, organizations should apply them. Moreover, White et al. (2003) suggested that certain HPWS practices are related to negative work-life spillover. In this way, it would be interesting in the future to analyze the simultaneous effects of HPWS perceptions on the work- family conflict and work-to-family enrichment to fully understand their real contribution to the work—family relationship. Finally, future studies should also explore different well-being dimensions since, as we demonstrated through health perceptions and satisfaction with life, these variables did not represent the same relationship with WFE.