Introduction
In recent decades, many social changes have reshaped the society in which we live. The "knowledge work", which requires "flexible work schedules and spaces", has blurred the boundaries of work (Burke & Cooper, 2008). In addition, workers feel pressured by several job demands, such as physical workload, poor environmental conditions, demanding clients, individualized workplaces, time pressure, unfavorable schedules, and job insecurity, among others (Bauman, 2001; Beck, 1992, 2000; Castells, 1996).
The organizational world constantly evolves and so does the family world. Traditional gender roles have been renegotiated as a result of the increase in dual-career couples and single-parent households. The traditional single-earner family (i.e., with a male breadwinner and a female caregiver) is on decline, and when women entered the labor market much concern arose over the successful combination of work and family life (Pitt-Catsouphes & Christensen, 2004; Van der Lippe, Jager & Kops, 2006).
Work and family research has grown rapidly, although some gaps haven't been solved or addressed at all. The goal of this article is to provide a comprehensive review of the relevant theoretical and empirical contributions in the work and family research field, especially since the review contributions by Eby, Casper, Lockwood, Bordeaux and Brinley (2005) and Casper, Eby, Bordeaux, Lockwood and Lambert (2007) to the present. Considering that work and family (WF) literature remains highly fragmented, this review offers synthesis, explanation and critique on the existing WF research, demonstrating the diverse and multidisciplinary nature of the literature around the topic. This work hints at the development of the WF field, highlighting the gaps and limitations in research and serving as an introduction for new researchers in the field, Ph.D. students, human resources (HR) practitioners and policy-makers that might feel lost in the overwhelming explosion of information that has taken place during the last decade. The main contributions of this review are future research opportunities derived from each subject addressed in this review.
After presenting the methodology section, this paper explores i) work-home interface definitions and dominant perspectives and models, affording an overview of the conceptual and theoretical approaches; ii) the antecedents and outcomes of WF conflict, enrichment, and balance; and iii) the methodological gaps within the field. Research opportunities are presented for each of these three areas and an overall conclusion of the findings is presented as well.
Methodology
There is already an extensive and growing body of research on this subject, therefore this article is not intended to be an exhaustive review. In particular, the review aims to provide a qualitative and selective compilation of organized information that allows for the inclusion of more variables in future studies. However, the suggestions for future research stem from an analysis considering the selection of leading sources from 2004 to 2014 (peer-reviewed academic papers, book chapters, books, and conferences' presentations), chosen to cover the advances since the review contributions by Eby et al. (2005) and Casper et al. (2007) to the present.
The proliferation of WF research made it necessary for the researcher to develop specific criteria to limit the scope of this review. To begin with, in cases where a relatively large number of studies share a common theme, a sampling of those that are indicative of the overall thrust and general conclusions have been included. In addition, WF literature spans diverse academic disciplines and not all of the journals have the same value for management scholars.
The researcher accessed databases (Web of Science, PsycINFO, Sociological Abstracts, Wiley, ScienceDirect, SpringerLink, Oxford, JSTOR, Sage Journals Online, Scopus, Redalyc, and Google Scholar) containing articles related to three social sciences disciplines: economics, management, and psychology (social, organizational, occupational and/or behavioral). Many databases overlapped. MEDUNE/PubMed, Nature, American Chemical Society and other health databases were excluded since they covered articles and conference presentations related to WF and physical health, a topic out of the scope of this review. Welfare topics and social policies were excluded as well.
The literature search and criteria for inclusion were inductively oriented, first, by conducting a keyword search of terms in English (work or non-work, life or family or leisure, balance or conflict or integration or interference, work schedules, dual-earner, dual-career, childcare, dependent care, and coping or cope) in titles, abstracts, subjects or keywords. From there, new keywords emerged and the scope became broader. The second filter was the discipline: economics, management, and psychology. The analyst also took into consideration presentations from the main conferences on the WF field and management (biannual International Conference of Work and Family at the IESE Business School in Barcelona; biannual International Community, Work and Family Conference; biannual Work and Family Researchers Network Conference; the American Sociological Association meetings; annual Conference of the Society for Industrial Organizational Psychology (siop); and the Academy of Management meetings). Table 1 summarizes the resulting sample of 83 publications matching the criteria and supporting the findings reported in this comprehensive review.
Seventeen publications were included in the meta-analysis and the reviews, 14 were conceptual works proposing new theory or measure, and finally 52 were empirical studies -including 3 longitudinal studies. As usual in the WF field, the main peer-reviewed journals resulted in: Journal of Occupational Health Psychology, Journal of Vocational Behavior, Journal of Management, Journal of Organizational Behavior, Work & Stress, Journal of Marriage and Family, Journal of Managerial Psychology, Career Development International, and Community, Work & Family.
Findings
Definitions and Theories
In most individuals' lives the two perceived central domains are work and family. For this reason research efforts are primarily focused on the work-family interface. The relationship between work and other domain(s) has been addressed through different approaches. Work-family researchers in any discipline have frequently failed to base their predictions on strong conceptual frameworks (Hobfoll, 1989). This section (summarized in figure 1) strives to cover the main theoretical influences and their construct development.
Conflict and Role Theory
Overall, the WF field has been dominated by role theory (Kahn, Wolfe, Quinn, Snoek & Rosenthal, 1964; Barnett & Gareis, 2006), with its origin lying on studies of inter-role conflict (Kanter, 1977; Sarbin, 1954). The guiding assumption is that work and family are in basic conflict because human energy is limited. Inter-role conflict denotes the conflict experienced when pressures arising in one role are incompatible with those emerging in another role (Greenhaus & Beutell, 1985). Greenhaus and Beutell (1985) suggested that work-family conflict (WFC) exists when there is role pressure incompatibility in terms of time, strain and/or behavior.
Where are the boundaries of the work domain? The pressures of work have intensified in recent decades (Burke & Cooper, 2008). Advances in information technology and information load, the need for quick responses, the importance attached to customer service quality and its implications for constant availability as well as the pace of change, can be sources of pressure (Guest, 2002). With such pressures, work can begin to dominate life, resulting in a sense of work-life imbalance. In addition, there are some work-related activities, such as overtime, commute time (Geurts, Beckers, Taris, Kompier & Smulders, 2009), organizing or networking, that may equally create conflict with life.
Simultaneously, interference theories began to flourish. First, the segmentation model (Piotrkowski, 1979) hypothesizes that work and non-work are two distinct domains of life experienced quite separately and exert no influence on one another. This appears to be offered as a theoretical possibility rather than a model with empirical support. In contrast, the spillover model (Andreassi & Thompson, 2008; Hanson, Hammer & Colton, 2006; Zedeck & Mosier: 1990) hypothesizes that one domain can influence the other in either a positive or a negative manner. This model operates by means of cross-domain processes, such as resource drain, resource generation and positive and negative spillover (Edwards & Rothbard, 2000). The compensation model (Staines, 1980) proposes that what may be lacking in one sphere in terms of demands or satisfactions can be made up in the other.
Directionality
Campbell-Clark (2000) stated that work and family life influence each other and introduced the work-family border theory. According to this approach, people are daily border-crossers between work and family spheres. The theory addresses how domain integration and segmentation, border creation and management, border-crosser participation and relationships between border-crossers (and others at work and home) influence work-family balance. This theory opens up the idea of the permeability of each domain, although the idea of work interfering with family (WIF) and family interfering with work (FIW) already existed, according to Greenhaus and Beutell (1985).
Testing Pleck's (1977) hypothesis concerning gender differences in the relative permeability of work and family boundaries, Frone, Russell and Cooper (1992a) reported that work and family boundaries are indeed asymmetrically permeable with family boundaries being more permeable than work boundaries, finding no gender differences in the pattern of asymmetry; which was also confirmed by Eagle, Miles and Icenogle (1997). Individuals typically report more WIF than FIW. These are distinct constructs with discriminant validity (Gutek, Searle & Klepa, 1991) and are strongly correlated with each other (Frone et al., 1992a, 1992b; Gutek et al., 1991; Kinnunen & Mauno, 1998). According to Gutek et al. (1991), this may have two explanations: i) work demands can generally be assessed better, and ii) employees' assessments are influenced by expectations of being a "good" employee. However, there is no evidence of gender differences in the pattern of asymmetry, indicating that the dynamics of work and family boundaries may operate similarly among men and women.
The border theory might help reveal the extent to which individuals are in control of issues determining balance. It also allows physical and psychological control analysis, thus permitting the accommodation of human agency. Nevertheless, a particularly interesting perspective by which to understand WFC is not as a one-construct perspective but rather as a two-direction perspective. While other theories regarding the intersection of work and family exist, the WIF and FIW constructs are rooted in conflict theory, which asserts that the work and family domains are incompatible due to their different norms and responsibilities (Greenhaus & Beutell, 1985) that cause intrusion and negative spillover from one domain to the other. Kossek and Ozeki (1998) showed that regardless the type of measure used (e.g., bidirectional WFC, work to family conflict or WIF, family to work conflict or FIW), a consistent negative relationship exists among all forms of WFC and job-life satisfaction, and this relationship may be stronger for women than men.
A great deal of research proposes to measure how work interferes with family or vice versa. Here, some examples that illustrates the former statement. Bronneberg (1996) showed that FIW conflict was negatively related to self-rated job performance and WIF conflict was positively related to self-rated job performance. Carr, Boyar and Gregory (2008) reported a statistically significant relationship between WIF and employee retention among employees from a manufacturing plant. Moreover, Bruck, Allen and Spector (2002) found a significant relationship between WFC and job satisfaction; when considering the three forms (described by Greenhaus & Beutell, 1985) of time-based, strain-based, and behavior-based conflict simultaneously, regression results revealed that behavior-based conflict was the only form that significantly related to job satisfaction, demonstrating why the study of coping strategies and self-efficacy is so important. Bruck et al. (2002) highlighted the importance of considering both the form and direction of WFC because WIF is more closely related to employee job satisfaction than FIW (Boles, Howards & Donofrio, 2001). This finding may have an explanation. Hochschild (1997) argued that although every mother and almost every father claimed "family comes first", few of these working parents questioned their long hours or took advantage of their companies' opportunities for flextime, paternity leave or other family-friendly policies. She concluded that the roles of home and work had reversed: work was offering stimulation, guidance, and a sense of belonging, while home had become the stressful place in which there was too much to do in too little time.
Conservation of resources model (COR) states that individuals seek to acquire and maintain resources, including objects (e.g., homes, clothes, food), personal characteristics (e.g., self-esteem), conditions (e.g., marital status, tenure, financial security), and energies (e.g., time, money, and knowledge) (Hobfoll, 1989). COR proposes that stress arises when there is a loss of resources, or a threat of loss. On this topic, Grandley and Cropanzano (1999) applied and supported the COR model to the WF field. The positive side of the model, diminishing conflict by acquiring resources, has been supported by the enrichment perspective. In other words, when individuals generate resources they seem better able to cope with stressful situations, resulting in increased psychological and physical well-being (McNall, Niclin & Masuda, 2010). Putting together two models rooted in the stress literature - the COR model and the job demands and resources model - a new model arose.
The resulting work-home resources (W-HR) model describes work-home conflict as a process whereby demands in one domain deplete personal resources and impede accomplishments in the other domain. Enrichment is described as a process of resource accumulation: work and home resources increase personal resources. Those personal resources, in turn, can be utilized to improve home and work outcomes (Ten-Brummelhuis & Bakker, 2012, p. 545).
Balance
There are many definitions of work-life balance. Though a consensus does not exist regarding a term to designate the concept, it is used as a convention in the field. Work-life balance is a metaphor that conveys the equilibrium between two primary spheres of life (yet there can be more). First, there is an individual's work, or his/her source of paid income. Then there is the term "life" (often replaced by "non-work"), which refers to everything besides work, including leisure, housekeeping and family care - or simply "family" - which has been given the greatest amount of attention (see Eby et al., 2005). The weight of the spheres is not supposed to be equally distributed, because it depends on the centrality work and non-work have along the life course (Ahmad, 2007; Demerouti, Peeters, & van der Heijden, 2012; Erickson, Martinengo & Hill, 2010; Lu, Gilmour, Kao & Huang, 2006).
Historically, work-family balance was narrowly conceptualized as the absence of work-family conflict (i.e., incompatibility of roles in two domains) (Greenhaus & Beutell, 1985). Then, it was proposed that people experience work-family balance as "the extent to which individuals are equally involved in - and equally satisfied with - their work role and their family role" (Greenhaus & Singh, 2003, p. 2). Kalliath and Brough (2008) provide a review of six conceptualizations of balance found in the literature. Greenhaus and Allen (2006, 2011) redefined work-life balance as "the extent to which an individual's effectiveness and satisfaction in work and family roles are compatible with the individuals' life role priorities at a given point in time". In an attempt to shed light on the different notions and terms in the field, Clarke, Koch and Hill (2004) indicate that fit and balance are two separate constructs and suggest that fit is based more on the structural aspects of wf interactions whereas balance appears to be based more upon the psychological factors.
Positive Relationships between Work and Family
Along with the influence of positive psychology in the earlier 2000, two recent positive spillover approaches arose: expansion and enrichment/facilitation. The work-family expansion refers to the notion that simultaneously engaging in multiple work and family roles is beneficial for an individual's physical, mental and relationship health (Barnett & Hyde, 2001). The positive side of the WF interface (Parasuraman & Greenhaus, 2002) has taken a variety of names: enhancement (Ruderman, Ohlott, Panzer & King, 2002), positive spillover (Hanson, Hammer & Colton, 2006) , facilitation (Wayne, Grzywacz, Carlson & Kagmar, 2007) , and enrichment (Carlson, Kagmar, Wayne & Grzy-wacz, 2006; Greenhaus & Powell, 2006). Greenhaus and Powell (2006) proposed a theoretical model of work-family enrichment as the extent to which experiences in one role improve the quality of life in the other role and they offered a series of research proposals that reflect two paths to enrichment: an instrumental path and an affective path. As well as conflict, enrichment is thought to be bidirectional, in which positive benefits may flow from work to family (work enrichment of family, WEF) and from family to work (family enrichment of work, FEW). At the resource-level analysis, Powell and Greenhaus (2006) suggested that work-family enrichment is either unrelated or negatively related to WFC, depending on the specific process under consideration by which experiences in one role may affect experiences in the other role.
Frone presented a fourfold taxonomy of work-family balance, which comprises direction of influence (work to family or family to work) and types of effect (conflict or facilitation). According to him, "low levels of interrole conflict and high levels of interrole facilitation represent work-family balance" (Frone, 2003, p. 145). Role balance reflects more than the presence of enrichment and absence of conflict (Marks & MacDermid, 1997). Specifically, while conflict and enrichment act as linking mechanisms between work and family, work-family balance reflects a summa-tive characterization of an individual's engagement in and enjoyment of a multitude of roles across work and family domains (Marks & MacDermid, 1997; Valcour, 2007); in addition, this statement was empirically tested. Carlson, Grzywacz and Zivnuska (2009) demonstrated by a discriminant validity and confirmatory factor analysis that constructs they created of work-family balance, work-family conflict and work-family enrichment are conceptually distinct, while Wayne (2009) provided conceptual definitions of each construct (enhancement positive spillover, enrichment, and facilitation).
Critical perspectives
In line with feminist critical, postmodernist and poststructuralist definitions of language, the current WF scholarship does not take processes of the social construction shaping our work and family lives as given, neutral or apolitical (Weedon, 1987). Caproni (2004) argued that the "work-life discourse" reflects individualism, goal focus, achievement orientation and instrumental rationality. She stated that WF research fails to consider the emotional side of human beings and researchers must transcend the language of the compartmentalization of life that actually traps people in perceived work-life imbalance. Ransome (2007) claimed that by staying in the dominant sexual/gender division of the labor paradigm, much of the current academic work-life balance discourse has become rather narrow. The current discourse is particularly difficult to apply because in most households, because primary care, particularly of infants, does not actually account for a major share of the total responsibility burden. In the same vein, Pitt-Catsouphes and Christensen challenged generally accepted assumptions in wf research (2004). They encouraged scholars to think outside the box and focus on calling into question the relationship between WF, gender and employment with both an equity and differential perspective. Lewis, Gambles and Rapoport (2007) stated that the work-life balance metaphor is a social construct located within a particular period of time and originating in the Western context. This, in addition to creating enormous difficulties when comparing the past and other cultures, is related to globalization and reorganization, mainly because people are experiencing changes at work (e.g., overload and time pressure) that engender feelings of lack of time and pressure, which are then translated into the personal/social/family domains. They argued that there are assumptions surrounding work-life balance terminology as gender neutral, individual agency and culture free. The challenge now is to move on to a new discourse paradigm.
Opportunities for Future Research
The difficulty in work-life research begins with an inaccurate definition of the subject. Besides, that definition will be charged with perspective. This renders the comparison of studies very difficult. Empirical studies generally give a functional definition of family (e.g., nuclear and extended) rather than a legal or traditional definition. Therefore, researchers and policy-makers should delineate "work", "family" and "life" accurately and focus on specific roles of non-work, such as "caregiving", "household", "studying", rather than "family" or "life". Furthermore, there is a call to further test the distinctness of other measures of work-family conflict, balance, and enrichment (Carlson, Grzywacz & Zivnuska, 2009).
Scholars still need to develop critical perspectives and move to a new discourse paradigm. Moreover, researchers need to respond to the following questions from an applied perspective: Does the dynamics of work and family boundaries operate similarly among men and women? Is enrichment determined by the accumulation of resources described by COR model, WF enrichment model, and/or W-HR model?
Background and Outcomes
The relationship between the WFC and its antecedents and consequences has been examined by reviews and meta-analyses (see Byron, 2005; Eby et al., 2005; Greenhaus & Allen, 2011; Kossek & Ozeki, 1998, 1999). This section will summarize the results found in those studies and in the collection of new studies since then, whose findings are still not reflected in meta-analysis (figure 2).
Three decades ago, Greenhaus and Beutell (1985) reviewed research studies on the antecedents of WFC. Later on, when Kossek and Ozeki (1999) reported consistent support for the bidirectionality of WFC, theory and evidence-based research started to suggest that WIF and FIW may have different causes and effects (Frone, Russell & Cooper, 1992a, 1992b). Parasuraman and Greenhaus (2002) argued that a disproportionate emphasis was placed on environmental and situational factors (the context) rather than individual differences and psychological characteristics as antecedents of WFC.
WIF and FIW Antecedents
WIF and FIW antecedents were reviewed by Frone (2003) and Byron (2005) and Michel, Kotrba, Michelson, Clark and Baltes (2011), and all concluded that WIF and FIW are related and the antecedents of WIF reside primarily in the work domain, whereas the antecedents of FIW reside primarily in the family domain, although a meta-analysis conducted by Mesmer-Magnus and Viswesvaran (2005) found that the measures of WIF and FIW converge. Byron (2005) and Mesmer-Magnus and Viswesvaran (2005) found asymmetry in the domain effects, meaning that work variables consistently showed stronger relationships with WIF than with FIW.
Consistent with Eby et al. (2005) and Byron (2005), the previously researched antecedents can be classified into three categories: work domain variables (schedule flexibility, job stress, etc.), non-work domain variables (marital conflict, number of hours spent on housework and care, age of the youngest child, etc.), and individual and demographic variables (personality, gender, income, coping style, etc.). Meta-analysis results (Byron, 2005; Michel et al., 2010) showed that job stress and family conflict are the major factors related to higher levels of WIF and FIW, respectively. Furthermore, those who have children at home are likely to experience higher WFC (Carlson, 1999) as are those who are worried or anxious about childcare (Buffardi & Erdwins, 1997), or experience tension with their family and have little family support (Carlson & Perrewe, 1999). In addition, greater family involvement translates into greater general FIW (Carlson & Perrewe, 1999; Parasuraman & Simmers, 2001) while greater job involvement is related to wif (Gutek et al., 1991; Kossek & Ozeki, 1999).
Generally speaking, a supportive work and organizational culture (Carlson & Perrewe, 1999; Greenhaus et al., 1987; Thompson, Beauvais & Lyness, 1999), the availability of work-family benefits (Thompson et al., 1999), and having a mentor (Nielson, Carlson & Lankau, 2001) are related to experience less WFC. Moreover, placing social value on one's work and having access to promotional opportunities are related to lower WFC (Wallace, 1997). Demographic variables tended to be poor predictors of WIF and FIW, however, individuals with better coping styles and skills (e.g., time management, emotional intelligence, resilience), an individual variable, had less WIF and FIW (Braunstein-Bercovitz, Frish-Burstein & Benjamin, 2012; Byron, 2005, Yanchus et al., 2010).
WIF AND FIW OUTCOMES
WIF and FIW outcomes have been reviewed by Amstad, Meier, Fasel, Elfering and Semmer (2011) as domain-specific effects. Before that study, other works explored WFC general outcomes on psychological and physical health (Greenhaus et al., 2006), or only WIF outcomes, but not FIW outcomes (Allen, et al., 2000). Additionally, some studies examined links between WFC and healthy behaviors and safety behaviors. Food choices are used as a way to cope with conflicting demands between work and family and also exacerbated feelings of stress (Devine et al., 2006). FIW, but not WIF, was associated with decreased compliance with safety rules and less willingness to participate in discretionary safety meetings (Cullen & Hammer, 2007). In the few existing longitudinal studies, results seem mixed. A study conducted by Kinnunen, Geurts and Maumo (2004) showed that WIF perceived by women at Time 1 significantly predicted job dissatisfaction, parental distress, and psychological symptoms at Time 2 (one year later), while WIF assessed for men in Time 1 was not related to outcomes in Time 2. In addition, the experience of work-to-family conflict turned out to be relatively stable for both genders between both assessments. Hammer, Cullen, Neal, Sinclair and Shafiro (2005) found bivariate relationships between WIF and FIW at Time 1 and depression at Time 2 (one year later), but when controls were included in regression analyses the results were not significant.
Amstad et al. (2011), based on Bellavia and Frone (2005), developed a classification that distinguished three categories of "potential" (causality is not yet tested) consequences of WFC: work-related, family-related, and domain-unspecific outcomes. Both directions of WFC have been found to be associated with work-related outcomes, such as job satisfaction, organizational commitment (Aryee, Srinivas & Tan, 2005; Odle-Dusseau, Britt & Bobko, 2012), engagement, job performance, absenteeism, tardiness, turnover/intentions (Ahmad, 2008; Andres, Moelker & Soeters, 2012; Boyar, Maertz & Pearson, 2005; Greenhaus, Parasuraman & Collins, 2001; Kossek & Ozeki, 1998; Netemeyer, Boles & McMurrian, 1996), "opting-out" - or resigning from one's job - (Stone , 2007), perceived career success (Van Eck Peluchette, 1993), employee physical health - e.g., hypertension - (Frone et al., 1997b), mental health (e.g., anxiety and mood disorders), job burnout (Kinnunen & Mauno, 1998; Netemeyer et al., 1996; Peeters, Montgomery, Bakker & Schaufeli, 2005), work-related strain (Netemeyer, Brashear-Alejandro & Boles, 2004), and organizational citizenship behavior (Netemeyer, Maxham & Pullig, 2005), as well as family-related outcomes, such as marital satisfaction (Mihelic , 2014; Voydanoff, 2005a), family satisfaction (Bedeian, Burke & Moffett, 1988; Cardenas, Major & Bernas, 2004), and family-related strain.
Finally, the third category -domain-unspecific outcomes-has also been found to be related to both directions of WFC; these outcomes include life satisfaction (Greenhaus, Collins & Shaw, 2003), psychological strain (Parasuraman & Simmers, 2001), somatic complaints (Peeters, de Jonge, Janssen & van der Linden, 2004), depressive symptoms and health behaviors, such as sleep and exercise or cardiovascular disease (Allen, Herst, Bruck & Sutton, 2000; Allen & Armstrong, 2006; Bianchi, Casper & King, 2005), and substance use or abuse (Grzywacz & Bass, 2003).
But not only employees benefit from diminishing their WFC and enhancing their work-life balance. Employees' families benefit when WFC is ameliorated as well (Crouter & McHale, 2005; Hammer et al., 2005; Schneider & Waite, 2005) in terms of higher levels of well-being of the family members. According to studies conducted in the United States, communities benefit if employed individuals have more time and energy to be involved in local volunteer work and civic engagement (Barnett & Gareis, 2008; Bookman, 2004; Swisher, Sweet & Moen, 2004). At the societal level, concerns relate to family disruption and community disconnection as well as reduced social citizenship and community engagement (Voydanoff, 2005b). Also, WFCS have been linked to labor market decisions and fertility decisions (Stone, 2007). On the other hand, the business outcomes of WFC or initiatives aimed at reducing WFC refer to the aggregate impact of many of these individual-level work outcomes on overall organizational-level performance. The business outcomes of WFC are productivity measures and financial performance information, including stock market performance and return on investment (roi) that can directly portray both the costs of WF initiatives and the financial benefits associated with those organizational changes (e.g., aggregated absenteeism or turnover rates at the department or organizational level as well as health-care costs) (Kelly et al., 2008).
Despite the proliferation of work-family policies, the growing media attention on them and the literature on WF issues, little is known about whether and which organizational initiatives actually reduce WFC and how these changes are likely to impact employees and the organization as a whole (Kelly et al., 2008). Evidence of the benefits of WF policies is generally anecdotal, so in many cases it is impossible to achieve a precise analysis of their effects. Success indicators must be developed for these policies that can provide tangible proof of whether a company is benefiting from adopting work-life balance arrangements. According to the evidence up till know, there is no general or "one fits all" solution to work and family conflict. Each individual is influenced by a different context, which changes during his/her lifetime as well. According to this, each individual requires different resources; albeit this literature review demonstrates that many work resources such as a meaningful job, informal flexible work arrangements, a supportive supervisor (Valcour, Olli-er-Malaterre, Matz-Costa, Pitt-Catsouphes, & Brown, 2011), a family-friendly culture, have the potential to reduce conflict (Mauno, Kinnunen, & Ruokolainen, 2006) and enhance work-family facilitation.
WEF and FEW Antecedents and Outcomes
wef and few antecedents and outcomes have been examined in this last decade. McNall, Nicklin and Masuda (2010) studied the WEF and FEW outcomes and found that WEF was more strongly related to work-related variables, while few was closer to non-work-related variables. Studies have found that people who report greater WEF were more likely to report higher job satisfaction (Aryee, Srinivas & Tan, 2005; Balmforth & Gardner, 2006; McNall, Nicklin & Masuda, 2010). Resources acquired at work (e.g., flexibility) may result in better performance at work, which has the effect of creating more positive effects at work, ultimately transferring to more positive effects in the family domain (i.e., WEF). In turn, individuals experiencing more positive emotions about their work should experience higher job satisfaction and lower turnover intentions (Greenhaus & Powell, 2006). On the other hand, McNall, Nicklin and Masuda (2010) found no positive relation of WEF or FEW to turnover intentions in their meta-analysis study. Mediation analyses conducted by Allis and O'Driscoll (2008) indicated that psychological involvement in family and personal activities was associated with increased positive spillover, which in turn enhanced well-being, supported by the meta-analysis of McNall, Nicklin and Masuda (2010). Moreover, Ilies, Schwind, Wilson and Wagner (2009) found a main effect of daily job satisfaction on daily marital satisfaction and affect at home (i.e., WEF).
Work-family Balance
Work-family balance seems to account for additional variance in several key work and family outcomes (such as job satisfaction, organizational commitment, family satisfaction, family functioning and family performance) above and beyond WF conflict and WF enrichment (Carlson, Grzywacz & Zivnuska, 2009) and has been found to be positively related to job satisfaction, organizational commitment, quality of life and career accomplishment (Aryee, Srinivas & Tan, 2005; Carlson, Grzywacz & Zivnuska, 2009; Greenhaus et al., 2006; Marcinkus, Whelan-Berry & Gordon, 2007), although there is not enough evidence of casualization.
Opportunities for Future Research
As previously mentioned, the WF field is full of complexities. Firstly, personal and business outcomes of conflict have been studied the most and only a little is known regarding the antecedents (Byron, 2005). Second, recent theory and empirical research on WIF and FIW suggests these concepts may have different causes and effects (Frone, Russell & Cooper, 1992a, 1992b), therefore, the continued use of bidirectional measures of conflict are supported (Byron, 2005). Causal direction of WF relationship and the temporal element of WFC, enrichment (both directions) or balance, need to be better understood. In spite of studies conducted by Carlson et al. (2009) the concept of work-family balance still remains underdeveloped and poorly understood in the WF literature, having no agreed upon definition nor a widely accepted measure (Maertz & Boyar, 2011; Valcour, 2007; Wayne, Grzywacz, Carlson & Kacmar, 2007).
The literature showed there are many ways to deal with work and home demands. There is a need for studies that examine workers' coping strategies and styles to analyze whether they are more likely to tackle demands or to relieve pressure. Job stress and family conflict are the major factors that affect WIF and FIW. How can one effectively cope with it? How can one enhance WF enrichment/facilitation? Demerouti, Bakker and Tetrick (2014) found that expectations and styles of the couple could influence family life, and therefore the work life, intertwining spillover of facilitation and spillover of conflict on opposite directions (Michel, Mitch-elson, Kotrba, LeBreton, & Baltes, 2009). Future research efforts should seek scenarios in which one style of coping is effective and other contexts in which it is not and should analyze the short and long-term implications for health and well-being (Bianchi & Milkie, 2010). The academic community is now gradually revealing the antecedents, outcomes and other linkages (e.g., moderators, mediators) of both work-to-family and family-to-work enrichment/facilitation. How to better increase WEF and FEW? There is a need to further explore new work arrangements (McNall, Masuda & Nicklin, 2010), communicational skills from the manager to better transmit perceived support from supervisor and organizational culture (Kelly et al., 2008), and the role of job design (Grzywacz & Butler, 2005).
Methodological Gaps
There are common criticisms of WF research that concern to methodological deficiencies (figure 3).
Level of Analysis, Approaches, Techniques and Instruments
Some studies argue that an overemphasis is placed on the individual level of analysis (Eby et al., 2005; Zedeck & Mosier, 1990) rather than on the group, dyad, or organizational level. Methods and research designs mainly comprise quantitative approaches, however, researchers have requested qualitative approaches (Greenhaus & Parasuraman, 1999). Qualitative approaches and theory testing techniques (i.e., path analysis and structural equation modeling) have been used infrequently (Casper et al., 2007). To the detriment of a better understanding of changes over time, there is an over-reliance on cross-sectional designs (Barnett, 1998; Greenhaus & Parasuraman, 1999) and a lack of longitudinal or experimental studies, which reverberate in poorly understood causal relations (Casper et al., 2007; Greenhaus & Parasuraman, 1999; Kelly et al., 2008) and a lack of triangulation or corroborating evidence (Barnett, 1998; Greenhaus & Parasuraman, 1999; Zedeck & Mosier, 1990).
With regard to data analysis techniques, most studies in industrial-organizational psychology and organizational behavior WF research use simple inferential statistics and techniques examining one dependent variable with a single relation (Casper et al., 2007). Allen et al. (2000) suggested that it would be valuable to include possible moderators such as organizational and personal characteristics in future studies. Multiple efforts have been made across the world to develop and test new instruments to measure the work-life interface, such as WFC, WIF and FIW measures (Carlson, Kacmar & Williams, 2000; Frone et al., 1992a, 1992b; Greenhaus & Beutell, 1985; Gutek et al., 1991; Kopelman, Greenhaus & Connolly, 1983; Netemeyer et al., 1996). There has also been an over-reliance on single-source self-report survey data (Barnett, 1998; Greenhaus & Parasuraman, 1999) and poor measures (Green-haus & Parasuraman, 1999).
According to the variables used, genetics, personality type and coping styles are variables understudied as antecedents of WFC. The mere characteristics of the sample (or the "context") point to the multilevel aspects of WF studies. Contemporary work-life theory often incorporates concepts of work and family role demands, work and family role resources, and role salience. Each of these concepts is embedded in family life course location. The concepts of family life stages are useful for exploring differences in the experience of work and family life throughout the course of life. The size and range of family life stages will provide the needed contrast and context in which to theorize more specifically with regard to changes in the WF experienced across the life course (Grzywacz, Almeida & McDonald, 2002). Exploring differences across family life stages grounds the importance of the family life course approach and invites the refinement and qualification of theoretical concepts. Moreover, Kossek et al. (1999) and Byron (2005) noted that scholars often overlook how caring for parents differs from caring for children, leaving an interesting gap to fill.
Methodological Problems in Social Sciences
The following elements generate frequent methodological problems but are very common in applied models and proper of social sciences disciplines (e.g., management, applied psychology, sociology, etc.). These elements are causality, endogeneity, simultaneity, effect size and self-selection, as explained below.
Causality: Does work-life balance cause job or family satisfaction? Does job or family satisfaction cause well-being?
Does work-family conflict cause burnout? Or is it the other way around? Based on these questions, longitudinal and randomized experimental studies are needed to strengthen theories based on evidence (Casper et al., 2007; Greenhaus & Parasuraman, 1999; Kelly et al., 2008).
Endogeneity: In a statistical model, a parameter or variable is said to be endogenous when there is a correlation between the parameter or variable and the error term (Wooldridge, 2013). When the independent variable has not been exogenously manipulated there is a strong possibility that its relationship to a dependent variable will not be correctly estimated, leading to spurious findings. Fortunately, there are at least two ways of addressing endogeneity (Antonakis et al., 2014): first, a well-designed randomized experimental study may eliminate the "need to correct endogeneity", and second, using procedures borrowed from econometrics (i.e., two-stage least square regression estimator).
Simultaneity: There is evidence in the WF field that some variables are antecedents and outcomes at the same time (e.g., stress). Work and family characteristics that reflect role involvement and quality are considered as independent variables in some studies and as outcomes in others.
Effect Size: It is evident that some variables have an impact on the other. Those variables are related. But is this important? How important? What is the effect size? The meta-analysis by Byron (2005) and Mesmer-Magnus and Viswesvaran (2005) examined effect sizes of work-family policies (i.e., flexibility) on WIF and FIW, but results were radically different and not conclusive.
Self-selection: Researchers suspect that participants of their studies are more sensitive to WF issues, and therefore they are willing to participate. If this is true, the studies may not be capturing representative evidence.
Sampling in the WF Field
Not only have theories and models varied, but also samples. Some relevant examples are shown here. The knowledge of WF issues is based on studies of heterosexual, caucasian, managerial, and professional employees in traditional family arrangements (Casper et al., 2007).
Little appears to be known about WF issues among employees from different racial and ethnic groups, different cultures and non-traditional families (Casper et al., 2007). Gay and lesbian employees may face different WF issues because of discrimination and stigmatization (Button, 2001). Generation (i.e., Baby Boomers, Gen X and Gen Y, etc.) is a sociological taxonomy that reflects several shared characteristics (e.g., values, age, and work attitudes) and is used mainly in the United States, although the taxonomy may not be fully applicable to other cultures (Powell, Francesco & Ling, 2009). Byron (2005) found that demographic factors (e.g. gender and marital status) were weak predictors of both wif and FIW, but when samples included a greater number of parents women tended to report more FIW and WIF than men (Stier, Lewin-Epstein & Braun, 2012), and those who were single reported greater WIF and FIW than married people. The differences are tangible in topics such as the use of time; the gender role attitudes and the sexual division of labor (Crompton, Brockmann & Lyonette, 2005); discrimination for being a mother reflected in pay and rank (Bacik & Drew, 2006; Bhave, Kramer & Glomb, 2013); work and family cen-trality (Cinamon & Rich, 2002); take-up of WF resources at the organization, such as teleworking or homeworking (Sullivan & Smithson, 2007); and skills development and coping strategies (Somech & Drach-Zahavy, 2007).
In addition, it appears that the redistribution of roles within the family to match increased role responsibilities outside the home has not yet occurred (Duxbury & Higgins, 1991). Women tend to emphasize their family roles to a greater extent than men do (Gutek et al., 1991). Furthermore, compared to men, women are still more likely to have the primary responsibility of balancing family obligations with obligations to their employer (Greenhaus & Parasuraman, 1993). Often, if not always, the sample is disaggregated by sex. In any case, some works have shown less and less influence of the sex/gender variable, such as a merely limited moderating influence on the relationships between the antecedents and the components of WF balance (Aryee, Srinivas & Tan, 2005) or even null gender moderation of the influence of any of the family-responsive variables on retention-relevant outcomes (Aryee, Luk & Stone, 1998). Some studies have confirmed while other have rejected evidence of gender differences in WF experiences and outcomes (Parasuraman & Greenhaus, 2002). This leads to the hypothesis that gender has an indirect mediation effect or moderating effect on family responsibilities and status. On the other hand, working fathers are conceptually and empirically underrepresented in the WF field (Hill, 2005; Hill, Hawkins, Martinson & Ferris, 2003).
Opportunities for Future Research
There are common criticisms of WF research such as the level of analysis, approaches, techniques and instruments concerning to methodological deficiencies in terms of conclusions about causality, endogeneity, simultaneity, effect size, and self-selection. There are many instrumental and socio-demographic variables that have been generally ignored in the field. Researchers still have to explore other work arrangements (e.g., freelancer, self-employed, entrepreneur, copreneurial couple) impacted by the emerging digital economy, other occupations, type of companies, industries, countries, cultures and new family configurations (e.g., single-parent families, same-sex couples, successive marriage), although Byron (2005) called to homogenize samples in order to make comparisons possible.
Future research will explore how sampling strategies and time frames (in longitudinal studies) may explain differences in results. As warned by Field (2001, p. 179) "estimates and significance test from meta-analytic studies containing less than 30 samples should be interpreted very cautiously"; therefore, more empirical studies and meta-analysis are needed. As results vary from one empirical study to the other, Byron (2005) suggested that the difference might have its origin on the sample composition.
Conclusions
Ambitiously, this article intended to cover multiple concepts in order to relate a fragmented work-family literature and the terminology that might be confusing for HR practitioners, Ph.D. students and newcomer academics in the field. The suggestion for researchers is to be consistent in their use of terms. This review offers a synthesis and integration of past literature with advances in the field during the decade of 2004-2014.
While there are new reviews and meta-analysis on the WF field (e.g., Amstad et al., 2011; McNall, Nicklin & Masuda, 2010; Michel et al,, 2011; Rabinowitz, 2007), the field has plenty of gaps and missing links, as shown in this study. Each subtheme presents a section that gathers the opportunity for future research and proposes specific research questions. There is a need to take further stock of the evidence. For instance, Byron (2005) stated that there is a lack of knowledge of WIF antecedents, and as shown by
Heraty, Morley and Cleveland (2008, p. 209) "the literature on the work-family interface is complex, and theory in the field is uncertain and under-developed". Work-family enrichment has been understudied, although it has developed during that decade (Nicklin & Masuda, 2010).
At the same time, many of the methodological problems faced by this field are common in Social Sciences. This article suggested alternatives to mitigate those methodological issues. New work demands, new work arrangements (e.g., freelancers, self-employed, entrepreneurs) and new occupations and contexts add new complexities to the field, as new forms of coping with WF conflict might arise. Researchers from different countries put emphasis on different realities. As a way to illustrate this idea, researchers from the United States (country that does not have a paid maternity leave after the birth or adoption of a child) are highlighting the benefits of WF arrangements on performance (Aryee et al., 2012) and policy design on women's employment (Hegewisch & Gornick, 2011), while European and Northern countries (with an aging population) is doing research on the elderly both as family burden as well as caregivers from their grandchildren (Moya-Martínez, Escribano-Sotos & Pardo-García, 2012). A future paradigm change and new methodologies (e.g., qualitative, longitudinal and experimental studies) are required to help fill these gaps.