How Work–Family Research Can Finally Have an Impact in Organizations

Abstract

Although work–family research has mushroomed over the past several decades, an implementation gap persists in putting work–family research into practice. Because of this, work–family researchers have not made a significant impact in improving the lives of employees relative to the amount of research that has been conducted. The goal of this article is to clarify areas where implementation gaps between work–family research and practice are prevalent, discuss the importance of reducing these gaps, and make the case that both better and different research should be conducted. We recommend several alternative but complementary actions for the work–family researcher: (a) work with organizations to study their policy and practice implementation efforts, (b) focus on the impact of rapid technological advances that are blurring work–family boundaries, (c) conduct research to empower the individual to self-manage the work–family interface, and (d) engage in advocacy and collaborative policy research to change institutional contexts and break down silos. Increased partnerships between industrial–organizational (I–O) psychology practitioners and researchers from many industries and disciplines could break down silos that we see as limiting development of the field.

According to a 2011 American Psychological Association study, only about a third (36%) of U.S. workers are satisfied with the manner in which their employers assist them in balancing work and family and other personal life demands—a drop from 42% in 2009 (Clay, 2011). And growing numbers of employers are not very satisfied with work–family benefits either. A recent Society for Human Resource Management (SHRM) survey (2010) reports many firms are reducing or eliminating some family friendly benefits, with the biggest drops in flextime, elder care referral, and adoption assistance. Another survey shows a significant drop in employer provision of full-time pay and time off for maternity leave, down to 16% from 27% in the prior decade (Shellenbarger, 2008). This decline in satisfaction with employer work–life supports is paradoxical as work–family research has mushroomed over the past several decades.

Work–family research—the study of positive and negative processes, antecedents, and outcomes related to work and family roles—has finally moved from the margins to the mainstream of industrial–organizational (I–O) psychology, management, and organizational behavior. Yet, research evaluating the effective implementation of work–family initiatives seems to have declined in the past decade. It is our contention that, to date, work–family researchers have not made a significant impact in improving the lives of employees relative to the amount of research that has been conducted. Evidence of having an impact would include the reduction of workplace stress, more positive work–family relationships in employment settings, and effectively implemented work–family policies.

We believe a big part of the problem is that an implementation gap remains between work–family research and practical impact. Even before the economic downturn, the implementation gap was evident (Kossek, 2005, 2006; Kossek & Lambert, 2005). However, the current economic climate has put into sharpened focus the disconnect between what research suggests are key ingredients about balancing work and family (e.g., positive organizational culture and supervisory support, employee control over work hours and workload, and opportunities for recovery outside the workplace), what organizations offer and allow for, and what options employees feel they can genuinely access that will be truly effective in helping them excel and thrive on and off the job.

The purpose of this article is to make the argument that better and different work–family research should be conducted to improve its impact in employing organizations. Before turning to the main body of the article, we give a brief overview of work–family conflict origins and define concepts related to organizational support of work and family roles. We discuss why there has been an implementation gap and argue why employers should (still) put organizational support of work and family roles higher on the corporate agenda than ever, even in light of current economic challenges. We will then discuss positive developments related to work–family research and suggest areas for improvement to enhance impact. Our discussion is not intended to be exhaustive, as the topic has benefited from numerous reviews (see Allen, Herst, Bruck, & Sutton, 2000; Casper, Eby, Bordeaux, Lockwood & Lambert, 2007; Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005; Ford, Heinen, & Langkamer, 2007; Kossek & Distelberg, 2008; Kossek & Ozeki, 1998, 1999; Michel, Mitchelson, Kotrba, LeBreton, & Baltes, 2009). We close with alternative yet complementary strategies to help bridge research–practice gaps.

Work–Family Research Origins

Most I–O psychologists study how the work side of the work–family relationship affects individual performance at work more predominantly than employee effectiveness in other life roles or how the work affects the family or the community (Thompson, Beauvais, & Allen, 2006). The main theoretical foundation that I–O psychologists use to study work and family relationships emanates from role theory (Katz & Kahn, 1978), which focuses on how individuals in social contexts enact expectations of their roles or positions. The notion of work–family conflict derives from the belief that when individuals have to enact multiple roles, psychological distress increases from the increased conflicts that are likely to occur when the expectations of one role, such as work, interfere with the expectations of the family role (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964).

The assumption that conflicts are the predominant dynamic between work and family roles emanates from stress theory and theories of role accumulation suggesting that human energy and time is a fixed commodity (Thompson et al., 2006). The more roles individuals add on to their life demands, the greater the probability of role strain, role overload, and conflict (Marks, 1977). The construct of work–family conflict has been central to the work–family field’s development (Eby et al., 2005; MacDermid & Harvey, 2006). It has been well studied as an antecedent (e.g., of job satisfaction and life satisfaction; Kossek and Ozeki, 1998), interrelated processes (Lambert, 1990), and an outcome (MacDermid & Harvey, 2006). Allen, Herst, Bruck, and Sutton’s (2000) meta-analysis demonstrates that work-to-family conflict has clear negative effects on individual well-being as it is strongly related to job burnout, depressive symptoms, general psychological stress, physical health symptoms, and family strain.

Why the focus on conflict between the family role and the work role? Scholars examining the historical roots of work–family conflict (e.g., MacDermid & Harvey, 2006; Near, Rice, & Hunt, 1980) note that at the turn of the 20th century as economic development occurred and large social institutions grew (such as organizations employing a larger number of individuals), classical sociologist and organizational theorists such as Karl Marx and Max Weber became concerned with the potential negative impacts of power dynamics of large institutions, namely employers, on worker well-being and alienation.

The historical development of U.S. societal interest in work–family policies is also relevant to understanding the current research–practice relationship. Work–family policies have some ties to equal employment opportunity and social resistance to changing the workplace to be inclusive of new labor market entrants. With the passage of equal employment opportunity legislation (e.g., the Civil Rights Act of 1964 and the Pregnancy Discrimination Act of 1971), work–family policies such as child care centers and part-time and flexible working arose to integrate women, who were participating in the labor force in unprecedented numbers, into the work place. The term work–family policies became popular in the business press, referring to employers’ support of child care and flexible work hours for employees with the most salient family demands—often women with young children. Yet, some employees without families or with stay-at-home spouses felt that work–family policy users were not carrying their weight at work and that their lower “face time” made them less committed. To address this stigmatization, large employers began to switch the names of their work–family policies to be labeled work–life policies to create the narrative that all workers need to be supported for personal life roles. Yet many researchers still prefer to study work and family policies, perhaps given a special interest in understanding organizational response to work–family conflict. These historical roots help explain why we see it as critical that I–O psychological research results in improved organizational support for work and family. It should be noted that we define family broadly and do not use the term simply to refer to traditional nuclear families, but to the nonwork and personal roles of all employees. In light of the shifting rhetoric on the evolution of work–family policies, it is not surprising that social conflict and ambiguity regarding their implementation remain.

Defining Organizational Support of Work and Family

Organizational support of work and family roles pertains to the degree to which the workplace is designed to reduce work–family conflicts and enhance work–family interactions. Employer work–family supports include three workplace characteristics that influence work–family relationships: (a) job conditions and the structure of work, such as work hours and job designs that give workers control over when, where, or how they do their job; (b) organizational culture and norms about the hegemony of work and nonwork relationships; and (c) human resource policies supporting the juggling of work and family roles (Kossek, 2006). Of these three areas, the business press has paid most attention to employer work–family policies, setting up a cultural expectation that “progressive” employers offer these policies as a means of being nominated to working mother or other best employer lists.

Work–family policies are organizational programs, policies, and practices that are designed to assist employees with the joint management of a paid work role with nonwork roles such as parenting, elder care, leisure, education, volunteering, and self-care (e.g., exercise or medical needs; Ryan & Kossek, 2008). Examples include flexibility in the scheduling (flextime), location (telework, mobile virtual office), or amount of work (job sharing, part time, vacation, and leaves); benefits such as health-, and child-, and elder care and domestic partner; and information such as referral programs. A key problem is that work–family policies have often been implemented in silos and not well linked to the other workplace characteristics such as the conditions of employment (i.e., expected work schedules and the ability to control workload and work hours) or workplace cultural support (e.g., can the policies be used without backlash?).

Even more importantly, managers are unsure how to implement and manage new work–family policies such as flexibility where they cannot see the employees at work. Consequently, there is considerable anecdotal evidence suggesting that employers have had problems implementing policies in ways that create an inclusive workplace and increase productivity, yet research clarifying exactly how to address these problems is limited (Kelly et al., 2008; Ryan & Kossek, 2008). Thus, although work–family researchers have demonstrated that employees who are able to effectively take advantage of work–family policies do benefit psychologically (e.g., have higher job satisfaction), we have failed to fully clarify how use of the policies link to productivity, how to avoid cultural stigmatization from use of the policies (e.g., being viewed as less committed or performing less effectively), and what challenges managers and organizations may face in implementing these policies to enhance competitiveness and, importantly, how to address these challenges to increase organizational effectiveness.

Why has work–family research been somewhat disconnected from improving practice? For one, how to translate and implement work–family research findings is not necessarily simple. Employers face many complexities in the development of work–family practice as workplace innovation. Many employers have far less experience in the work–family arena than in other traditional I–O and Human Resources (HR) areas such as selection or training or performance appraisal. Employers may also be unsure how to implement work–family policies effectively into existing HR systems and management philosophies of labor management. For example, increasing organizational support of employees’ ability to devote more time and energy to nonwork demands by giving workers more control over their schedules challenges traditional labor economic and capitalistic principles that employers should strive to get as much working time out of people as possible in the short run in exchange for the least amount of money.

Scalability is also an issue. For example, should policies like flextime be mandated for all employees or just those in tight labor markets, in conducive jobs, and who have an established work record? What about the tensions of standardization and customization? Should employers adopt a one-size-fits-all approach to ensure equity and ease of administration or customize support to individuals such as single parents compared to older workers? Given growing workforce diversity, it is increasingly difficult for employers to determine which strategies work best for which workers, how to ensure equity and coordinate diverse schedules and benefits, and how much work–family conflict is too much while trying to run a business and make a profit. These are the types of research questions that would help employers, but not much research has focused on these issues.

Another reason for a growing research–policy gap is that researchers and practitioners often have different agendas and fail to collaborate as much as they could. They also sometimes simply fail to trust each other and come from different perspectives. For example, they often use a different language. More specifically, practitioners talk about work–family balance, whereas most work–family scholars (barring some exceptions such as Greenhaus & Powell, 2006) emphasize the notion of work–family conflict. And evaluation of effectiveness could be problematic. For some firms, offering work–family policies serves as a public relations tool to increase worker recruitment and attraction. What if researchers show a failure to implement and that these policies are really more window dressing than real support?

Why Work–Family Research Can Really Help Organizations

A recent review of research notes that it is clear that employers benefit from offering and supporting the use of work–family policies in several ways: more positive employee attitudes, better talent, and cost savings (Kossek & Michel, 2010). Recruitment and retention, job satisfaction and commitment, and workforce quality are improved from offering work–life policies and implementing them. Employers who implement these policies are more likely viewed as an employer of choice and reap the benefits of a larger applicant pool (Kelly et al., 2008). Cost savings comes from a decrease in problematic employee behaviors, such as turnover, absenteeism, or accidents, as well as lower labor costs as some workers are willing to trade off flexibility for wages (Kossek, 2006; Kossek & Hammer, 2008; Kossek & Michel, 2010). Yet, these trends have been documented for several decades and still have not resulted in a major tipping point that creates a ground swell of employer excitement for work–family policies. As noted previously, we have actually seen a decline in support for some work–family policies (e.g., Shellenbarger, 2008; SHRM, 2010).

What we would like to see is a research agenda focused on how work–life policies and positive role relationships can help organizations solve pressing critical organizational and societal problems. These include the facilitation of workforce engagement and creativity, more time for learning, and better communication and collaboration in a global world. Clear statistical evidence of the heightened work–family issues we are facing is provided from the National Study of the Changing Workforce (Aumann & Galinsky, 2009; Galinsky, Aumann, & Bond, 2009). Although granting that there are currently many U.S. workers who lack jobs or are underemployed, the nationally representative survey results from Aumann and Galinsky (2009) suggest that many employees perceive overwork stress. Over a third report they “often or very often” felt overwhelmed by the amount of work they had to do on their jobs in the last several months. In the same survey, one third of respondents report signs of clinical depression, and one in five report high blood pressure. Only a little more than one quarter (28%) of workers rate their health excellent, down from 34% in 2002. Blurring work–family boundaries and interruptions are other issues: Nearly one third of workers report being contacted by a work colleague or client outside of normal work hours at least once a week, and over half of employees surveyed report they often or very often worked on too many tasks and multitasked too much. What if work–family research actually addressed these trends by providing tangible solutions for organizational practice or individuals?

Before discussing four possible paths for work–family research to improve its impact, we discuss what we see as positive developments in work–family research and areas where improvement is still needed, looking in turn at the framing and language used in the work–family field, the clarity of our constructs, issues of measurement and research design, and research samples.

Framing and Language

A positive development in work–family research has been the movement to focus not only on the negative aspects of work–family relationships, such as work–family conflict, but to include the study of positive work–family relationships as well. The work–family enrichment perspective advanced by Greenhaus and Powell (2006) and others argues that resources such as psychological assets (e.g., resilience and mood), social capital (e.g., knowledge and influence), or behaviors and skills can be transferred back and forth between work and family domains to enrich each. This spotlight on positive resource transfer among multiple roles has its roots in role expansion theories, which contrary to stress theories assume that greater involvement in meaningful roles involving positive experiences can enhance well-being and positive functioning (Barnett & Hyde, 2001; Sieber, 1974).

Including positive perspectives advances work–family research for a number of reasons. First, more and more families and working individuals today have no choice but to juggle work and caregiving roles. Given these trends, research might as well examine how to make the joint enactment of work–family roles as positive as possible to create resilience and positive outcomes. Second, given the recent global economic slowdown and the aging population world-wide, it is clear that a growing proportion of the population will be involved in employment for longer periods of the life span and caring for their own health needs. And the high unemployment rates have highlighted the need to shift mindsets to the positive benefits of having a job and an income for individuals’ identities, families’ household economic security, and communities at large. Finally, a positive focus may increase understanding that can help individuals better optimize how they manage work–family conflict (Baltes & Heydens-Gahir, 2003) or help organizations learn how to increase social support in the workplace to create more positive work–family climates (Allen 2001; Kossek, Colquitt, & Noe, 2001; Kossek, Pichler, Bodner, Hammer, 2011). This could provide a new theoretical home for work–family research in the growing positive organizational scholarship (POS; Cameron & Caza, 2004) and positive organization psychology (POP) fields. Donaldson and Ko (2010) define POP as the study of “positive subjective experiences and traits in the workplace and positive organizations, and its application to improve the effectiveness and quality of life in organizations,” which seems very apropos for application of work–family research into a core I–O psychology and human resource policy domain.

At the same time, we argue that the work–family field needs to broaden and update its language. Language and cultural framing are very powerful. As Kegan and Lahey (2001) note, language, as the main way we communicate issues, does more than merely reflect attitudes and emotions, but rather the mere choice of one word over others can shape beliefs and attitudes and, ultimately and most critically, our actions. Language regarding work–family issues often implies deviancy or negative connotations regarding commitment to work. For example, those who use flexible work schedules are viewed as working “alternative” or “nontraditional” schedules. Teleworkers are “distant workers.” Those who do not work this way are “traditional” workers. We talk about the “nonwork” role as if it is juxtaposed against the work role, where work is seen as the norm, that is, the primary role or the standard with which all other life roles should be compared. Careers that are “full time” are “normal.” Any work arrangement that deviates from this, such as “part time” or “reduced load” or “customized work,” uses language implying that the worker is not giving his or her full effort and is getting special treatment. Yet in the Netherlands, nearly 39% of the workforce works part time and nearly 75% of women who work do so part time. This suggests that part-time work is the norm for some employee groups, and they are valued members of the workforce. Furthermore, research shows part-time workers can often bring higher intensity and productivity than full-time workers (Kossek & Michel, 2010). We need new ways to talk about careers and jobs that move away from the notion that working flexibly is somehow cheating the employer, particularly as our companies become more global.

We also need to drop the use of “work–life” as the politically correct way to talk about work–family matters. As noted in the historical review of work–family policy development, a number of years ago “work–life” replaced the term “work–family” in many major companies and in the field as a well-meaning effort to minimize backlash against giving increased support and attention to people with caregiving demands. Yet we need to drop the term “work–life” as “work” is a part of “life.” We simply, in our research, need to be clearer on what personal life roles we are juxtaposing vis-à-vis the work role. Why don’t we just use the terms “work role” and “caregiving role” and “leisure role” for example, and actually name the role being compared to give greater clarity?

Finally, some authors use terms such as “work–family balance” and “work–family integration” as positive outcomes, assuming all workers want the same thing. Yet not all workers necessarily want work–family balance as some may have higher work identities. Others may have higher family identities. And, yes, still some others are truly “dual-centric,” meaning their identities truly are high for both work and family roles (Galinsky et al., 2003; Lobel, 1991). Put another way, one person working 80 hours a week may perceive that they have achieved work–family balance, but another employee working a “traditional” 40-hour week may feel imbalanced. This is a function of the relative importance these two individuals place on work and family; balance does not mean the same thing to everyone. But the persistence of research on work–family balance normalizes this as the outcome that all workers want and the norm.

Similarly, not all workers want “work–family integration” as the goal. Some really do desire the ability to separate or control work–family boundaries more. As Kossek and Lautsch (2008) and Bulger, Matthews, and Hoffman (2007) note, people vary in terms of their preferences and their abilities to segment work and family and to detach from work. In fact, recent research suggests that some individuals are very unwilling to integrate work and family, and when they do, they experience more work–family conflict (Matthews, Barnes-Farrell, & Bulger, 2010). We need to be more careful in our language to not assume that all workers want the same thing or normalize some outcomes as more desirable. Doing so suggests that variation from balance, integration, or other outcomes implies deviancy or being “less normal.”

Clarity of Constructs

We see a large challenge remaining in the work–family literature with respect to construct overlap. For example, on the conflict side, commonly measured constructs are work–family conflict, work–family interference, and work–family spillover. Work–family conflict is said to arise from simultaneous pressures from the work and family domains that are incompatible in some respect—participation in one role is made more difficult by virtue of participation in the other role (Greenhaus & Beutell, 1985). Work–family interference describes the extent to which employees’ work (or family) demands interfere with their family (or work) responsibilities (Gutek, Searle, & Klepa, 1991). Work–family spillover is defined as emotions, attitudes, skills, and behaviors established at work (or home) that employees carry into their family (or work) life (Lambert, 1990). All three terms are often used interchangeably in empirical studies (e.g., Leiter & Durup, 1996; Major, Klein, & Ehrhart, 2002). Unfortunately, few authors have clarified if they are the same constructs with different names that could be used interchangeably, distinct but related constructs, or separate constructs altogether.

This lack of construct clarity is also reflected in overlapping items in scales used to measure these constructs. For example, an item from the work interference with family scale (Gutek et al., 1991) reads “My work takes up time that I would like to spend with my family/friends.” An item from the work–family conflict scale (Netemeyer, Boles, & McMurrian, 1996) reads “Things I want to do at home do not get done because of the demands my job puts on me.” An item from the negative work-to-nonwork spillover scale (Bond, Thompson, Galinsky, & Prottas, 2003) reads “How often has your job kept you from concentrating on important things in your family or personal life?” All three of these items illustrate some construct overlap.

Similar issues exist on the positive side of the work–family interface where constructs like work–family enrichment (the degree to which the work role enhances the family role), work–family facilitation (the degree to which the work role makes it easier to perform the family role), work–family positive spillover (the degree to which positive emotions, skills, or other resources carryover over to the family role), and work–family integration (the degree to which work and family roles are synthesized or combined) are often used interchangeably, yet they all have different meanings (cf., Thompson et al., 2006). The field will be held back if we continue to use several overlapping constructs interchangeably to measure phenomena.

The lack of construct clarity also manifests itself in our measurement tools. Even when scholars and practitioners are claiming to be measuring the same thing (i.e., using the same construct name such as work–family conflict with conceptually similar items), two issues often arise. First, we have multiple measures of the same construct that potentially differ in many functional ways. For example, Matthews, Bulger, and Barnes-Farrell (2010) point out that measures of conceptually similar constructs can differ in terms of number of items, response scale, and scale orientation. These differences can affect the relationships found between those measures and other variables. Second, we too often create our own short measures or trim items from existing measures in a way that is often not specified in our publications. These abbreviated measures typically have not been subjected to much psychometric or validity evaluation (Matthews, Barnes-Farrell, & Bulger, 2010). In summary, we think much more needs to be done to help with construct clarity, both theoretically and from a standardized measurement perspective.

Measurement and Research Design

The trend toward collecting multisource data from not just the worker but family members, coworkers, supervisors, and even clients represents an improvement in work–family research measures. This development moves the field away from its long and not so rigorous history of over reliance on same source self-report data (Casper, Eby, et al., 2007). It also helps us better tap into the notion of cross-over, the idea that increasingly work–family issues impact not only the individual employee, but families and coworkers as well (Westman & Etzion, 2005).

We are also seeing more triangulation of qualitative and quantitative methods as well as more longitudinal work (cf., Hammer et al., 2011; Kossek, Huber, & Lerner, 2003), although this trend is still somewhat limited (Casper, Eby, et al., 2007). Building on the movement toward more multisource data, many studies are examining dyadic (e.g., Hoobler, Wayne, & Lemmon, 2009; Ilies, Schwind, & Wagner, 2009) and multilevel relationships (Major, Fletcher, Davis, & Germano, 2009) to more accurately reflect the fact that work–family attitudes and outcomes are not independent but rather reflect nested relationships at home and at work.

However, our current measures of work–family conflict may not be up to the task when the goal is to understand dyadic and multilevel relationships. Although measuring perceptions of work–family conflict is valuable as a psychological measure and should be continued when this issue is central to the research question, it is most probably not the correct measurement technique when one is attempting cross-person prediction. For example, in dual career couples, if one wants to predict a husband’s level of marital satisfaction from his wife’s level of work–family conflict, we would argue that both objective and subjective data are needed. Specifically, it would be useful to predict a husband’s perceived level of satisfaction with his marriage from his wife’s level of work-to-family conflict. It may not be his wife’s perceived level of work-to-family conflict that is the most important in this scenario but indicators of how work is affecting family members. These include, for example, how often she is late because of work, how often she misses family functions because of work, and how often she has to complete work after hours at home. Some may argue that a spouse’s perception of his or her partner’s work impinging on family is still subjective, although it is better than having the employee’s self-report on the family’s views of their work–family conflict. Truly objective indicators of interference would be the number of family occasions missed because of work and the number of hours worked at home after hours before the children are asleep. Or health of family might be measured objectively to show the objective impact of work–family conflict on family. In summary, for many of the cross-person predictions we would like to see explored, objective data is needed. As objective scales will not meet many of the requirements made of perceptual measures (e.g., internal consistency), it will require some “relearning” for researchers, practitioners, readers, and reviewers in the work–family arena to be more open when it comes to objective, nonsame source scale development.

Samples

Unlike early research that focused on samples with a very restricted range of work–family issues, we are seeing a much greater diversity of job and family demographics represented in research samples. Examples of formerly understudied populations include singles (Casper, Weltman, & Kwesiga, 2007), blue-collar workers, (Grandey, Cordeiro, & Michael, 2007), low-income retail workers (Hammer, Kossek, Bodner, Anger, & Zimmerman, 2011; Lambert, 2008), immigrant families (Chien-Juh, 2009), unionized workers (Berg, Kossek, & Misra, 2010), welfare-to-work recipients (Kossek et al., 2003), and kin that live together but are not comprised of a married couple or a parent and children under 18 (Rothausen-Vange, 2005).

At the same time, there is limited international sampling in work–family studies. U.S. cultural norms about the primacy of the work role and individualistic approach to work–family issues should not be viewed as the worldwide norm. The United States is also unique among industrialized countries as having a more limited public safety net for work–family issues because of the lack of nationalized paid family leave for newborn children. And the United States has a very individualist employer-driven approach to work–family support, where employers have far more latitude than in some other nations to determine how to manage work–family matters. We increasingly need to include measures of cultural values in work–family studies, particularly on collectivism and femininity (Hofstede, 2001), which may deeply shape perceptions of the appropriateness of employer support of work and family demands.

A final issue is sampling heterogeneity. Numerous inconsistencies in antecedents and outcomes of work–family conflict have been noted in the past research. Some of these include flexible work arrangements (Shockley & Allen, 2007), family friendly policies (Brough, O’Driscoll, & Kalliath, 2005), job performance (Cullen & Hammer, 2007), job and life satisfaction (Kossek & Ozeki, 1998), social support (Kossek et al., 2011), and perceived stress (Voydanoff, 2005). For example, two meta-analyses conducted only a year apart provided widely different estimates of the relationship between flexible work arrangements and work-to-family conflict—from .30 in Byron’s (2005) study to −.01 in Mesmer-Magnus and Viswesvaran’s (2005) study. We argue that the lack of systematic management of sample heterogeneity may be a driving factor in some of the inconsistent findings. Sample heterogeneity is defined as the degree to which participants within a particular sample differ in terms of a variety of individual difference variables.

In a single study where sample heterogeneity is so high that two or more distinct subpopulations exist, incorrect inferences can be generalized to all subpopulations (Matthews, Bulger, & Barnes-Farrell, 2010). Alternatively, when multiple subpopulations exist, a null effect might be observed for the entire sample if the process works in opposite directions for the groups (e.g., as in a traditional interaction effect). As an illustration of this issue, Matthews, Bulger, and Barnes-Farrell (2010) found in their overall sample that family social support had no relationship with work-to-family conflict. However, when conceptually distinct age groups were considered, they note that the null effect in their overall model was driven by a canceling out of the effect across the three age groups.

Across studies examining the same issues, sample heterogeneity may also play a role in the reporting of inconsistent results. For example, it is possible that within two independent studies, sample heterogeneity may be relatively low (all participants are relatively similar) within each sample. However, each study might demonstrate contradictory results for the constructs of interest if the samples are systematically different from one another. For example, if one study examines flexible work arrangements for a sample of nurses and a second study examines the same issue in a sample of blue-collar workers, different results may be observed as a function of the nature of the job, types of policies available, issues of organizational culture, and other distinguishing characteristics of individuals who comprise these occupations. Thus, the issue is not so much an inconsistency in the observed results but variation in results generalization. We must systematically examine the degree to which these sample idiosyncrasies may bias study results or how our findings must be understood as valid only for a very specified context.

One way to address sampling issue is to include greater use of individual, group, and organizational differences as moderators. Overall, relative to the total number of work–family studies that have been conducted, the role that important individual, group, or organizational difference characteristics play as moderators of relevant processes is understudied (Casper, Eby, et al., 2007). Although we continue to develop a better overarching conceptual model of how the work–family interface functions (Eby et al., 2005), we persist in ignoring the role that surface-level demographics and deep-level characteristics such as values play in moderating important relationships. We also would like to see more organizational- and occupational-level moderators used, given wide variation in occupations and family supportive organizational cultures and access to and use of work–family policies.

Different Paths Forward in Work–Family Research

In addition to addressing some of the persistent problems in work–family research so that “better” research is conducted, resulting in richer and more valid knowledge for organizations to use in figuring out what to do, we advocate for a renewed emphasis on several substantive paths. We do not see these paths as an either/or crossroad but rather that each are worthy of pursuit.

Path 1: Researchers Actually Study Effective Policy and Practice Implementation

Overall, we need work–family research to improve the understanding of effective policy and practice implementation. Going down this path, researchers would work to help organizations better implement policies. This would require more research focused on program evaluation and on linking work–family policies to organizational development and change. Such research would also focus on contextual differences—one-size does not fit all organizations and cultures. What would comprise such a research agenda?

First on the agenda are integrative studies of policy availability, timing, extent of use, and subsequent outcomes. A shortcoming of research examining policy implementation is that it often simply examines whether a policy, such as a flextime or telework, is reported by organizations as being available but does not examine variation in use and effectiveness of the policy, or consequences (good and bad) of policy use (Kossek & Michel, 2010). We need to really look into organizations more carefully to understand what is happening within the firm and across workgroups. Many organizations have “lumpiness” or wide unevenness in policy implementation. That is, individuals are nested in workgroups that vary considerably in the degree to which they have jobs or bosses supporting work–family flexibility, which creates understudied work–family micro-climates. Unevenness in within-firm implementation and access also highlights issues of organizational stratification—the notion that within organizations there is workforce stratification or rigidity where only some workgroups, employees, or occupations are allowed to work flexibly while others are not. Yet at the firm level, a company can still publicize that it offers flextime or telework (Ryan & Kossek, 2008).

Researchers should also not assume that all work–family policies, practices, or interventions affect people the same way. For example, different family structures (e.g., single parents, families with children with disabilities, elder caregivers) are likely to view (and experience) these policies differently. The ability to telework several days a week may be a moderately attractive job policy for an unmarried recent college graduate entering the labor market but viewed very differently by someone with an on going lifelong critical work–family demand. More than simply “nice to have,” the psychological meaning of such a telework policy may be viewed as a godsend for a parent whose child has long-term learning disabilities or physical or mental health problems (Parish, 2006).

Scholars also need to more carefully measure and examine the conditions under which policy use leads to positive outcomes. Few studies include both measures of cultural support for work–family policies and actual policy use (from family members or work) in the same study or assess the negative consequence of policy use. Regarding the later, some studies have found that the heaviest users of work–family policies have higher work–family conflict than nonusers. For example, a study of teleworkers found that they had higher work–family conflict than nontelework users (Kossek, Lautsch, & Eaton, 2006). One explanation (given the cross-sectional design) is that most teleworkers had higher work–family conflict than nonusers a priori, that is, the presence of work–family conflict motivated them to make use of the telework option. Work–family conflict may increasingly need to be measured not just as a dependent variable but as an independent variable shaping the effects of work–family policy use (Hammer et al., 2011).

Hammer, Neal, Newsom, Brockwood, and Colton (2005) found that greater users of work–family flexibility had higher work–family conflict but surmised users were using policies to their detriment. Rather than ameliorating work–family conflict, sometimes work–family flexibility policies have the reverse effect of enabling more negative work–family spillover. This occurs because workers are trying to do both their family demands and their work demands at the same time without cutting back on either or having more open boundaries between work and home. For example, we should examine the possible negative consequences of flexibility policies leading to lower and more permeable boundaries between work and family, which in turn increases interruptions from family-to-work or work-to-family.

Another explanation for negative effects is that policy use may lead to backlash. That is, if policies do not fit with the company culture, using these policies can actually lead to greater work–family conflict, thereby hurting the users the policies were designed to help. We must increase our understanding on potential backlash of use from childless workers, clients, managers, and those in nondual-earner families. We also need to understand “family backlash” from use of flexibility policies enabling higher work-to-family conflict. Finally, we need to look at the stickiness of positive or negative effects from policy use. Some research has shown (Baltes et al., 1999) that the positive effects of new flexible schedules may decline over time, particularly in terms of how they relate to job satisfaction. This may be because people start getting used to policies, they just become part of the normal expectations people have of a job design.

It could also be that having access to flexibility, particularly for professionals, will not reduce work–family conflict unless workload is also reduced. This may also be because professionals often have the “flexibility” to work 24/7, and policies that are well intentioned such as flextime and telework may actually be producing overwork and workaholism, and people are having increasing difficulty shutting work off.

Path 2: Researchers Focus on the Impact of Rapid Technological Advances That Are Blurring Work–Family Boundaries

This path could create a whole new ballgame in terms of understanding the work–family interface and perhaps mainstream work–family issues as essential to implementation of workforce deployment as a global organizational strategy.

Technology is increasing the spread of work into people’s personal time and the spread of the private domain into working time. As a result, individuals will have more difficulty maintaining work–family boundaries, unless they actively seek to do so (Kossek & Lautsch, 2008). Research is needed to understand how work–family flexibility practices are being self-regulated by the employee, particularly for those who work on laptops or use cell phones. More and more employees are being given autonomy to enact personal differences in how they manage boundaries between work and home—either by choice (e.g., psychological preference to multitask) or constraints (e.g., global travel job, single parent).

Kossek and Lautsch (2008) have developed a typology of boundary-management styles: segmenters, integrators, and volleyers. Some employees prefer to segment boundaries and work in blocks where they separate work and family. Perhaps they only take phone calls from a doctor during their lunch hour, for example. Other employees like to integrate work and family boundaries and blend work and family all day long. Still other employees are volleyers who have periods of high segmentation and then high integration depending on how their jobs are structured (e.g. accountants who have peaks and valleys during tax season) or their family situation (e.g., divorced parent who gets full child custody every summer).

We need to continue to update our work–family measures to include preferences for balance and segmentation, integration, and how individuals and groups seek to actively structure temporal boundaries and work–family relationships. This will also enhance objective measurement issues discussed above, as now studies can measure actual behaviors such as the number of e-mails and texts individuals send and receive a day, how long they are on cell phones and computers for personal and work demands, and whether this is by individual choice. We can also measure the length of time someone is spending on work interruptions when at home and family interruptions when at work as e-mails and cell calls arrive.

Path 3: Researchers Work to Empower the Individual

Another avenue of research that could help employees achieve work–family balance is one that focuses on the individual and not the organization. Given the fact that many work–family policies still lag in impact in their current form, and the fact that the employer is powerful in fostering work-to-family boundary blurring via technology, one could argue that employees may often be on their own when it comes to dealing with work–family conflict. On this path, researchers assume organizations are unlikely to ever really make organizational support for work–family issues a top priority. Here researchers would focus studies to help individuals better help themselves.

Although the body of research that examines how individuals successfully deal with work–family conflict has grown considerably in recent years, relatively little is known about what individual factors may allow some employees to better cope with the antecedents of work–family conflict than others. It is reasonable to assume that both personality factors and coping strategies used by individuals in dealing with the limited resources resulting from work–family conflict may influence whether an employee can achieve some level of work–family balance, and recent research supports this hypothesis (e.g., Baltes & Heydens-Gahir, 2003; Rosenbaum & Cohen, 1999). Of special interest are coping strategies because individuals can use these and they would seem to be trainable (i.e., malleable), whereas personality factors are not. Recent research on coping strategies (e.g., Baltes & Heydens-Gahir, 2003) has shown that they are linked to lower levels of work–family conflict.

The most often tested model of coping in the work–family domain is the meta-model of selection, optimization, and compensation (SOC; Baltes, 1997; Baltes & Baltes, 1990; Freund & Baltes, 2002). SOC is based on the underlying assumption that “the coordinated use of behaviors involving SOC can (a) increase one’s resources in the sense of developmental enhancement, (b) help maintain functioning in the face of challenges, and (c) help regulate impending losses in resources” (Baltes & Heydens-Gahir, 2003, p. 1006).

Selection is concerned with the degree to which individuals identify and select goals, alternative contexts, outcomes, and goal structures. Behaviors associated with the selection strategy provide a direction to behavior. Optimization refers to the acquisition, refinement, and use of means to achieve goals. Compensation concerns the acquisition and use of alternative means to maintain a desired level of functioning in the face of decreases in resources.

Early studies examining the usage of SOC and how it relates to work–family conflict have demonstrated that individuals who use SOC behavioral strategies in the work or family domain tend to experience lower amounts of job and family stressors and in turn lower levels of work–family conflict (Baltes & Heydens-Gahir, 2003; Clark, Bal, Zhdanova, & Baltes, 2009) and that people in the most resource-stretched situations benefit the most from using SOC coping behaviors (Young, Baltes, & Pratt, 2007).

Although a handful of studies that have examined coping strategies suggest that they can be very effective in helping individuals reduce work–family conflict, much more research is required before a practical impact may be seen. We propose three interrelated themes of research: (a) research that investigates which strategies work and which do not (taking into consideration boundary conditions that might influence the effectiveness of different techniques), (b) research that examines how the use of coping strategies affect the employees’ immediate partner and family, (c) research on the degree to which effective coping strategies are indeed trainable, what training methods are the most efficacious, and how organizations can enhance transfer of training.

Path 4: Researchers Engage in Advocacy and Collaborative Policy Research to Change Institutional Contexts and Break Down Silos

The movement toward research to change public policy and to advocate for the effective implementation of work–family policies could be achieved if organizations like SIOP were more active on Capitol Hill or partnered more with the U.S. Department of Labor and Census Bureau to better measure policy availability and use. There is an increasing availability of NIH grants to look at intervention and policy effectiveness research. Recently, the SHRM published principles for a 21st-century flexibility policy. The principles advocate for flexibility policies that facilitate employees’ abilities to jointly meet growing personal and work demands in their daily lives while still ensuring productivity and stability for employers and not hurting job creation (SHRM, 2011). SHRM also argues for better coordination of state and federal laws and “safe harbors” standards where employers could voluntarily offer a certain number of paid leave days that workers could use for any need, as long as consistent with collective bargaining agreements or other organizational policies. I–O psychologists could play a bigger role here. Partnering with SHRM, employers, and other groups, we could help design and evaluate these policies and demonstrate the potential benefits of effective implementation.

The Need for Researcher – Practitioner and Interdisciplinary Partnerships

Providing technical assistance to help employers better implement work–family policies and develop supportive work–family cultures as organizational change interventions is needed. Scholars, policy-makers, and practitioners increasingly need to build bridges to help ensure continuous improvement in the implementation of work–family initiatives. Until this is done, practitioners may continue to be more comfortable with consulting firms who are not necessarily selling evidenced-based work–family policies.

Rather than seeing work–family consulting as a way to implement a quick fix to the work–family “problem,” consulting firms and companies should embed scholars to help with confidential evaluation as a neutral third party. This would help ensure that the work–life policies being sold by work–family vendors actually work. Researchers could be instrumental in measuring a pre- and post snapshot of current work–family stress. They could then do a follow-up evaluation to ensure that policies and practices being implemented actually work and are being uniformly implemented across the company. This would also lead to the development of multiple stakeholder measures of effectiveness as the views of employees, employee subgroups, and companies may differ. We need not spend more money on new policies, but we do need to better implement the policies we already have, as noted above. Certainly conducting needs assessments of unique workforce needs to prioritize work–family issues also would be helpful to ensure that employers don’t just buy what vendors are selling carte blanch but are customizing policies to unique workforce needs.

We would also like to see companies move toward adopting certified or validated interventions. Just like the total quality movement moved toward certification of certain quality procedures, we would like to see this done for work–family policies. We’ve noted that consultants rarely validate the interventions they are promoting. Yet in the employee selection realm, the International Journal of Selection and Assessment encourages practitioners to provide research briefs using “real” data as part of their Information Exchange manuscript submission process (Viswesvaran, 2007). The write-ups are short and theory is not needed or warranted. This serves as fertile ground and empirical evidence for theorists to work from and supports inductive theory generation. We hope developments for validated knowledge sharing are something we can see in the work–family arena either as a recurring section in at least one of our top journals, in a series of special issues, or a SIOP online community.

Work–family research also needs to become more interdisciplinary, and the next generation of work–family scholars need to be more cross-disciplinary. Work–family phenomena, problems, and solutions indeed span many disciplines. Although there are many lenses to understand work–family issues, the field has been fragmented and studies do not talk to each other well. Deciding what research questions to ask and the perspectives used to “see” and interpret work–family phenomenon are very disciplinary specific. Take this example from a chapter in the Handbook of Work and Family (Kossek, Sweet, & Pitt-Catsouphes, 2006).

A young woman who has been employed at the local factory for several years with an excellent attendance and performance record decides to quit work shortly after the birth of her and her husband’s second child. She had returned to work for a few months after her 6-week maternity leave and it seemed as if things were working fine, but then she gave her 2 weeks’ notice to her supervisor. Senior management noted that the young woman’s behavior was similar to that of many other working parents at the plant and they were becoming worried about the lost productivity. They decided to consult a variety of researchers from the local university. The psychologist stated that the working parents were probably experiencing role conflict between the demands of work and those of the family. The sociologist added that traditional societal and marital gender expectations were causing women to work a second domestic shift when they got home from their jobs. The economist surmised that the wages the plant was offering were not sufficiently generous to offset the cost of paying for quality child care. The demographer observed that workforce data suggest that if a working mother has more than one child under 3 years old and is also part of the sandwich generation providing elder care, she is likely to temporarily leave the labor force. The historian stated that the factory has retained employment approaches not much different from those first adopted a century ago, and these traditional production methods have made it difficult to alter workplace structures to provide flexible work hours. This perspective was echoed by the anthropologist, who noted the strong factory cultural norms reinforcing segmentation of the workplace for at least 10 hours a day from personal life. (p. 67)

As this example illustrates, researchers from different disciplines tend to frame research questions in very unique ways, using different measures, replicating and culturally reproducing knowledge in ways reflective of how they were socialized as doctoral students. I–O scholars, practitioners, students, or policymakers have probably overfocused on the employer and employee psychological perspectives and consequently have limited the utility of our studies. We need to look at work–family relationships more broadly to better understand the structural root causes perpetuatingwork– family issues—perspectives more often investigated by broader social and behavioral sciences.

An exciting university and company partnership that is interdisciplinary is the National Work–Family Health Network, which is supported by the U.S. National Institutes of Health and many other foundations and agencies. This team of leading scholars from a broad range of disciplines, including public and occupational health, I–O psychology, human resources, and sociology, has teamed up to design workplace interventions to reduce work–family conflict, increase social support for work–family issues, increase schedule control and change the culture to be more supportive, and ultimately improve health and effectiveness of workers, companies, and families. When the studies are done, interdisciplinary research teams will have collaborated over a decade to conduct studies with workers, families, and managers spanning retail, corporate headquarter professional, IT, healthcare, and hotel environments. Some of the studies are using clinical trial methodology where workgroups and organizations are randomized to receive work–family interventions as a treatment and then compared to control workplace groups. As supportive supervisors and work–family cultures are so critical to the effective implementation of work–family policies, this interdisciplinary practitioner, scholarly, and government partnership is an important innovation and hopefully one that can be replicated (http://www.kpchr.org/workfamilyhealthnet work/public/default.aspx).