The Effects of Work-Family Conflict and Community Embeddedness on Employee Satisfaction : The Role of Leader-Member Exchange

Starting from the Social Exchange Theory, this study tested a theoretical model that examined the mediating role of Leader-Member Exchange in the relationship among two extra-work variables (Work-Family Conflict and Community Embeddedness) and Job Satisfaction. Using structural equation modeling, the relationships among the variables were explored in a sample of 561 health care workers from Italian Hospitals. Although not all the hypotheses were confirmed, findings supported that leaders can be considered an important factor in the mediation between nonwork dimension and job attitudes. Specifically, findings showed that extra-work dimension had a direct effect on the quality of relationship with a leader, which in turn influenced the employee’ job satisfaction. The implications of these findings on future research and organizational practice are discussed.


Introduction
In the last two decades, we are witnessing on one hand to a progressive physical separation between the home and the workplace, and on the other hand to the request for a better balance between work and nonwork-life.In a period of globalization, rapid technological and economic change, both workers and organizations need flexibility.The society has increasingly moved away from the Kanter's "myth of separate worlds" (Kanter, 1977) and now is embracing the view of Lewin's field theory (1951), where human behavior exists within a "life space" that in turn concerns several facets of human live (Lewin, 1951).Therefore, persons participate in many life domains, including work, family and community environment, and the resulting interface that exists between an employee's work demands and non-work needs has been an important topic for policy makers, researchers and practitioners interested in improving both employee's outcomes and organizational performance (Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005;Frone, 2003;Greenhaus & Parasuraman, 1999;McNall, Nicklin, & Masuda, 2010).
According to the spillover approach, nonwork and work dimensions should be positively related.The quality of an individual's work experiences is considered to affect the quality of experiences away from work.If work is boring and unexciting, nonwork experiences will reflect this quality (Wilensky, 1960).At the same time, if nonwork experiences are sources of conflicts, problems and preoccupations, work experiences will be influenced due to the fact that people usually "does not leave his/her own private life out of the workplace".As many studies have shown, the imbalance in work-life and nonwork-life interface has serious consequences for employees, their families, employers, and for society as a whole (Poelmans & Beham, 2008).
In the last 10 years, has become apparent to researchers and managers that the analysis of work and family should be expanded to include community.Also, results from many researches shown that work and family are embedded in the context of the community in which people are engaged (Voydanoff, 2007).Particularly, friendship, relationships and participation in community activities provides important resources in managing work and family duties, whereas the lack of adequate community engagement encumbers work-family activities (Bookman, 2004).
In this regard, scholars have progressively focused on the role of two off-the job variables, work-family balance (WFC) and community embeddedness (off-the job embeddedness, CE) and their influences on job attitudes, performance and turnover intention.
Results showed that both imbalance in the relationship between work and family and absence of social attachments (low community embeddedness) had influences in the everyday work experience and may create contextual forces that influence directly both employees' job attitudes (Griffeth, Hom & Gaertner, 2000;Portoghese, Battistelli, Saiani and Galletta, 2009).
As people experience more conflict between their roles in both spheres (work and private life), their level of (job and life) satisfaction falls.To control negative effects on employees' well-being and to help employees to better manage their work and non-work duties, several organizations have recently adopted and implemented work-family or work-life policies.Although these policies are important, regarding workplace relationships, little is known about the effects of extra-work variables on the relationship with leader.Contributes from literature do not provide unequivocal results concerning the effects of work-nonwork interface on the relationship with leader (Karatepe, 2009).
Traditionally, organizational support is one of the fundamental components involved in reducing the imbalance between work and private life (Thomas & Ganster, 1995), due to the role of leaders as a part of a social system.Also, leadership is one of the operative variables of social support, and has been conceived as one of the most important work-related resource aimed to reduce work-life imbalance.
However, actually there are not studies that consider leadership as an outcome of nonwork dimension.In fact, considering Social Exchange Theory, it is possible that leadership can be influenced by an employee's nonwork issues.If workers experience both imbalance between work and nonwork dimension and/or reduced social attachments, what are the effects in the everyday job?What are the effects on the relationship with own leader/supervisor?
The purpose of this study is to develop a conceptual model of the processes through which non-work variables influence job attitudes (see Figure 1), with a particular investigation of the role of Leader-Member Exchange (LMX) on the relationship among WFC and CE and employees Job Satisfaction (JS).Prior studies had focused exclusively on linear relationships between nonwork factors and JS (Kossek & Ozeki, 1998).Moreover, the role of LMX as a mediator of the influences of WFC and CE on employee satisfaction has not been addressed adequately.
It is expected that this article will be valuable to both scholars and practitioners.For scholars, it may improve the understanding of the potential mediating role LMX might play between extra-wok dimension (WFC and CE) and job satisfaction.For practitioners, it may draw attention to extra-wok demands as sources of stress, as well as their effects on job attitudes.This information may be used to reduce these effects in order to improve job attitudes and productivity.

Theoretical Background
People engage in work-related activity while are at home, and they carry on private purposes while are at the workplace.Organizational researchers have been giving increased attention to the difficulty people have in successfully to meet work and private-life demands.According with Voydanoff (2007), workplace, family, and community represent major domains in which people live out their lives.
Role theory (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964;Katz & Kahn, 1978) has provided the wide theoretical framework for much of the work and family literature.In line with role theory, work and family life dimensions entail multiple roles often resulting in conflict (Michel, Mitchelson, Pichler, & Cullen, 2010).It has also been suggested that domain boundaries are asymmetrically permeable from work to non-work (Frone, Russel, & Cooper, 1992).
Based on the work of Kahn and colleagues (1964), Greenhaus and Beutell (1985) defined work-family conflict as "a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect" (p.77).According to Greenhaus and Singh (2003), when the time, energy, and behavioral demands of a role in one domain (work or family) makes it difficult to meet the demands of the other domain (work or family), WFC arises.Moreover, Goode (1960) proposed that people have limited resources, and, as a result, Rode, Rehg, Near, & Underhill (2007) sustained that higher levels of resource commitment to one role can be in conflict with commitments to other roles, and this in turn leads to lower satisfaction with that role (Lazarus and Folkman, 1984).
Regarding to job attitudes, managers and researchers realize that dissatisfaction is the first factor that influences absenteeism and decision to quit of employees (Steel, 2002), two important issues concerned in the organizational efficacy.Therefore, organizations seek to control these kinds of outcomes as they seek to retain quality employees at work.Job satisfaction has been defined as a positive emotional state that represents the result of a perceptive appraisal of working conditions (Locke, 1969).It is one of the most important and widely researched variables in industrial-organizational psychology.Although the results have been mixed, a wide body of evidence from several researches indicates that high levels of work-family conflict have negative consequences, including high levels of turnover intentions, reduced well-being, greater general psychological strain, higher depression, and greater burnout (Allen et al. 2000;Byron 2005;Ford et al. 2007;Kossek and Ozeki 1998), the majority of studies have found that as WFC increases, job satisfaction decreases.H1: WFC is negatively related to JS Maertz (2004) argued that the community dimension influences individual choice about decisions concerning the job dimension.Particularly, in the Field Theory aspects of individuals' lives are symbolized and connected by a perceptual life space.Lewin (1951) purposed that individuals view themselves as entangled/embedded in a net of various forces and connections.Then, the more embedded the individual becomes the greater the likelihood that they leave their job is marginalized (Mitchell & Lee, 2001).
To illustrate the reasons why people do not quit organizations even when alternatives exist, Mitchell, Holtom, Lee, Sablynski, & Erez (2001) proposed a new construct called job embeddedness.This construct is assumed to be constituted of contextual and perceptual forces that connect individuals to the location, people, and issues at work (Yao, Lee, Mitchell, Burton, & Sablynski, 2004).Indeed, JE has been operationalized as a composited of two major factors: on-the-job embeddedness and off-the-job embeddedness (Mitchell et al., 2001).While on-the-job embeddedness refers to how enmeshed a person is in the organization, off-the-job embeddedness refers to how entrenched a person is the community (Mitchell et al., 2001).
In this sense, the embeddedness construct concerns the notion that workers are placed in a social, contextual web over time due to their fit within the situation, their links to other objects in the web, and the cost associated with quitting the situation (Mitchell et al., 2001;Reitz and Anderson, 2010).Lee, Mitchell, Sablynski, Burton, & Holtom (2004) found that individuals who are embedded within a local community, (e.g.family duties and leisure participation) show lower work absenteeism and less intent to quit their jobs.Many authors sustained (Lee, et al., 2004;Stroth, 2010) that employees develop attachment towards a specific social and physical context, regardless of their attachment to their organization.Thus, employees who develop a significant embeddedness in their community will show positive attitudes to their job.H2: CE is positively related to JS.
Organizational leadership is a matter of importance in the study of the occupational health of workers.Leadership is one of the extrinsic factors that have a significant impact on workers attitudes and because he/she is one of the most important agents for role definition, it is conceivable to expect that leadership can reduce the effect of off-the job dimension on job attitudes (Kelloway, Mullen, & Francis, 2004).
Adopting Social Exchange Theory's view, good relationships between employees and supervisors in workplaces is the essential condition to develop a mutual adjustment and that has positive influences on both the workers and the organization (Cole, Schaninger & Harris 2002;Wayne, Shore & Linden 1997).As such leaders can shape the work environment and contribute significantly to the atmosphere at work.Grounded in Social Exchange Theory (Blau, 1964;Cropanzano & Mitchell, 2005), LMX theory describes the development of a relationship between a leader and a member (Dienesch & Liden, 1986), recognizing how this relationship develops within a social context (Wayne et al., 1997).
A review of the literature (Gerstner & Day, 1997) shows that LMX significantly influence job satisfaction, organizational commitment, and intention to stay in the organization.H3: LMX is positively related to JS.
A central tenet of LMX theory is its focus on the working relationship between a leader and the various members of a work unit or organization.The LMX theory considers leader and workers as active actors in a developing process of transaction (Hollander, 1980).Particularly, Hollander (1980) sustained the potential for counter-influence on the part of the team-member.This facet represents a core assumption in LMX theory and emphasizes the necessity to consider the two-ways leader-member influence.In this regard, most researchers have focused primarily on work outcomes, ignoring the effect that nonwork dimension can have on leader-member relationships.
Starting from the theoretical rationale provided LMX theory, we predicted a direct effect of extra-work dimension on the quality of relationship with the leader.Moreover, we hypothesized the mediating role of LMX in the relationship between off-the job demands and job satisfaction.

More specifically:
H4a: WFC is negatively related to LMX.
H4b: CE is positively related to LMX.H5a: LMX mediates the relationship between WFC and JS.
H5b: LMX mediates the relationship between CE and JS.

Sample
The research setting was the Public Health System in Italy and the study focused principally on nurses in three public hospitals.The researchers assured employees that their participation in the survey was voluntary.Surveys were administered to 1091 nurses and 719 (66%) completed the survey; of those, 78% provided usable responses for purposes of this study.In total, the sample is composed by 561 nurses with an average age of 37.80 (SD = 7.38) varying between 24 and 64 years.22.1% are males with a 38.25 average age (SD = 7.90) and the remaining 77.9% are females with a 37.68 average age (SD = 7.23).Altogether, participants have been working as nurse for an average period of 12.89 years (SD = 7.48).For what concern family duties, 43.70% has no children, 23.20% has one child, 28.30% has two children and 4.80% has more than three children.Regarding the type of employment contract, 82.50% has a "Full-time permanent contract", 13.40% has a "Full-time fixed-term contract", and 4.1% has a "Part-time permanent contract".

Data collection/measures
Consistent with past research in this area, data were collected using an existing survey instrument.For all survey items, respondents were asked to indicate the extent to which they agreed or disagreed with statements using a five-point response scale ranging from 1 (strongly disagree) to5 (strongly agree).

Work-Family conflict.
WFC was measured using five-item scale developed by Netemeyer, Boles, & McMurrian (1996).A sample item is "The demands of my work interfere with my home and my family life".The estimated Cronbach's alpha coefficient was .91.

Community Embeddedness
CE was measured using 10-item scale developed by Lee, Mitchell, and colleagues (Lee et al., 2004;Mitchell et al., 2001).The three community-focused composites (Links, Fit and Sacrifice) were aggregated to form an off-the-job embeddedness construct (Lee et al., 2004;Hom, Tsui, Wu, Lee, Zhang, Fu, & Li, 2009).Sample items are "I really love the place where I live" and "Leaving this community would be very hard".The estimated Cronbach's alpha coefficient for the composite measure was .85.

Leader-member Exchange
LMX quality was measured using the seven-item scale from Scandura and Graen (1984).A sample item is "How would you characterize your working relationship with your manager?".The estimated Cronbach's alpha coefficient was .93

Job Satisfaction
To measure nurse's satisfaction, a reduced version of the Organizational Satisfaction Questionnaire by Cortese (2001) has been used.Job satisfaction was measured by three widely recognized job satisfaction facets: degree of satisfaction with the work itself (α = .93),sample item is "Autonomy level of my job"; degree of satisfaction with coworkers (α = .90),sample item is "Dialogue with colleagues"; degree of satisfaction with the supervisor, sample item is "Support and indications my supervisor gives me".The three job satisfaction facets were aggregated (α = .93)to form a composite measure.The estimated Cronbach's alpha coefficient for the composite measure was .93

Results
As a preliminary analysis, several steps were conducted to ensure data validity and reliability.Anderson and Gerbing's (1988) two-step approach was used to examine the conceptual model in this study.The evaluation of the measurement model was carried out using exploratory analysis (EFA), complemented by confirmatory factor analysis (CFA) which was conducted using the principles of structural equations modeling (SEM).The second step tests linkages in the proposed structural equation model using structural equation modeling (SEM).We conducted our SEM analysis from a partial disaggregation model (Bagozzi & Edwards, 1998) by creating parcels of items as recommended by Hall, Snell, and Singer Foust (1999).Mediation as tested by inspecting statistical significance of structural coefficients making up a mediational pathway (Kenny, Kashy, & Bolger, 1998).
In order to examine model fit, the χ², Root Mean Square Approximation Error (RMSEA), goodness-of-fit index (GFI), and comparative fit index (CFI) were used.For these two fit-indices, as a rule of thumb, values greater than .95are considered as indicating a good fit, whereas values smaller than .06for RMSEA are indicating an acceptable fit.These analyses were performed with covariance matrices and the maximum likelihood (ML) estimation procedures of AMOS 7 software package (Arbuckle, 2006).

Measurement Model
We examined responses to the survey items using confirmatory factor analysis.A measurement model (where the four multi-item scales were constrained to load on their respective factors) fits the data: χ² (28, N = 561) = 50.8,p < .05,GFI = .98;CFI = .99;RMSEA = .04.Convergent validity is observed when the path coefficients from the latent constructs to their corresponding manifest indicators are statistically significant (i.e., t > 2.0, Anderson & Gerbing, 1988).All factor loadings were significant and sizeable (mean estimated standardized loading = .80)suggesting convergent validity (Anderson & Gerbing, 1988).
Results from the test for discriminant validity among the hypothesized constructs indicated that they were empirically distinct.The comparison of the measurement model with that obtained from an alternative one-factor model [χ² (34, N = 561) = 1120.4,p < .05,GFI = 76; CFI = .65;RMSEA = .24]showed that the hypothesized structure was a better fit to the data at the .001level of significance.Specifically, χ² different test was conducted: Δχ² (6) = 1069.6,p < .001.

Proposed Model
The results of the hypothesized model suggest that the overall fit of the model with the data is good, χ² (31, N = 561) = 56.4,p < .01,GFI = .98;CFI = .99;RMSEA = .04.With respect to the predicted paths, 3 of 5 were significant.The completely standardized path loadings are presented in Figure 2.
With respect to the non-significant paths, the paths between WFC and JS (β = -.02,ns), and between CE and JS (β = .08,ns) were not significant.These results did not support both Hypothesis 1 and Hypothesis 2.
To test the first hypothesized mediation (H5a), we considered the relation between WFC and LMX.It was significant, β = -.17,p < .001supporting Hypothesis 4a.Then we considered the path between LMX and job satisfaction β = .87,p < .001.Also this relation was significant supporting Hypothesis 3. Finally then LMX mediated the relation between WFC and job satisfaction.This result support Hypothesis 5a.
To test the second hypothesized mediation (H3b), we inspected the relation between CE and LMX (β = .19,p < .001)and between LMX and job satisfaction (β = .87,p < .001).These relations were all significant and LMX, therefore, mediated the relation between CE and job satisfaction, thus supporting Hypotheses 4b and 5b.88% of the variance in job satisfaction was explained by the variables in the model (see Figure 2).

Discussion
The aim of this study was to integrate results from the nonwork dimension and LMX theory into a theoretical model that would further comprehending of the role of leadership in the relationship between non-work variables and job attitudes.Most researches on leadership have recognized the existence of the link between support/leadership and extrawork factors but have largely ignored the influence of these factors on leadership quality construct in favor of a focus on the role of leadership as antecedent.
According with LMX theory, time and resource constraints and personal impressions of their subordinates drive leaders in their daily actions (Dansereau, 1995;Wim van Breukelen, Schyns, Blanc, 2006).).The relation between workers and leader can influence directly job satisfaction but also can be influenced from different factors (Wayne et al., 1997).One of these can be the private life of worker, and, therefore, it can be plausible that when subordinates feel some kind of imbalance between on-the job and off-the job dimensions, they probably will perceive differently the relationship with the supervisor.
In this regard, the findings of this study support our hypotheses and extend the notion that the quality of the relationship with leader can be directly influenced from off-the job dimension.Also, results show that LMX can act as a full mediator in the relationship between private life of subordinates (work-family conflict and community embeddedness) and their job satisfaction.In fact, contrary to what hypothesized and to the results from literature, WFC and CE do not show direct influence on JS, suggesting that these linkages are more complex than previously thought.Therefore, as hypothesized, when conflict between work and family is present, the quality of the relationship with the supervisor-leader can be an important mediator factor in reducing the influences of this variable on job satisfaction.
The results also demonstrate the importance of considering community embeddedness into leadership theories.Results of this investigation demonstrate that CE, although originally theorized to predict job retention, also has important implications for employee job satisfaction.Our findings show that although CE has not a direct impact on job satisfaction, it can play an important influence on LMX quality and then on job satisfaction.
Our results have numerous implications for both research and practice.
First, our findings show the importance to consider LMX as a mediating factor.In fact our research extends LMX theory beyond the on-the job dimension, within which LMX research has been traditionally constrained and also contributes to the off-the job literature by providing explanations for relationships between WFC, Community Embeddedness and LMX quality.Specifically, our results suggest that ignoring the influence of WFC and CE on LMX can have indirect negative effects on job satisfaction by decreased quality of relationship with supervisor.Unfortunately, researchers know little about this issue and, thus, additional research needs to be conducted.
Even more importantly, our findings confirm that Community embeddedness may increase the quality of the relationship with leader.Then, managers can increase off-the-job embeddedness by providing people with information about the community surrounding their workplace and by providing social support for local activities and events (Mitchell, et al., 2001).

Limitations
Despite the contributions of this study, some limitations must be noted.
First of all, this study was a cross-sectional study, thus no hard conclusions can be drawn with regard to causation.A longitudinal study may reduce this issue, but even a longitudinal design may not be fully able to find causation.Thus, future research could employ longitudinal design to explore how the changing of WFC, community embeddedness, or LMX influence employee job satisfaction.
Second, all data were provided via self-reports raising the possibility of response set tendencies and the possible role of individual dispositions or negative affect in accounting for some of the findings.Third, data were collected from only one occupation.It remains to be seen whether the obtained findings can be generalized to other professions.

Figure 1 .
Figure 1.Proposed conceptual model Hypothesis are shown in the given figure to be tested for the study