Spending Time

Spending time: The impact of hours worked on workfamily conict

Cheryl L. Adkins a,, Sonya F. Premeaux b

a Longwood University, College of Business and Economics, 201 High Street, Farmville, VA 23909, USAb University of Arkansas at Little Rock, College of Business, 2801 S. University Avenue, Little Rock, AR 72204, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 16 July 2011Available online 1 October 2011

Scholars have long assumed that as workers spend more time at work fewer hours are avail-able for their non-work lives leading to negative effects in both domains, and most studiesexamining the impact of work hours on work and life domains have supported this viewpoint.However, the majority of these studies have used one-dimensional measures of workfamilyconflict (WFC) and familywork conflict (FWC) on homogenous samples which included pri-marily married managers and professionals with children. Further, despite calls to examinenon-linear relationships between work hours and WFC and FWC, few studies have done so.This study uses multi-dimensional measures to examine the linear and non-linear (quadratic)effects of work hours on WFC and FWC in a heterogeneous sample and examines the moderat-ing effects of several work and family characteristics on these relationships. The findings indi-cate that whereas work hours have a linear relationship with WFC, the relationship betweenwork hours and FWC is curvilinear. Managerial support was found tomoderate the relationshipbetween work hours and one dimension of FWC. Number of children moderated the relation-ships between work hours and WFC and another dimension of FWC.

2011 Elsevier Inc. All rights reserved.

Keywords:Workfamily conflictFamilywork conflictWork hours

Time is a fixed income and, as with any income, the real problem facing most of us is how to live successfully within ourdaily allotment.

Margaret B. Johnstone (1960)

As the opening epigraph illustrates, time is a limited quantity and the more time that is expended on work, the less that isavailable for home and family. For the most part, many Americans are thought to suffer from a time bind, whereby the hoursthat workers spend at work have a negative impact on their non-work lives (Hochschild, 1997). Many researchers have longheld that as the number of hours an individual spends on work increases, conflict between the individual's work and homelives increases as well, and that these negative effects extend to other work and life outcomes. This is in keeping with the scarcityhypothesis which suggests that the more roles one occupies, and the more one's time is divided between those roles, the less thatwill be available for any specific role (Barnett & Gareis, 2000). Logically, as time spent on work increases and time available forhome and family decreases, workfamily conflict (WFC) will increase. Further, as DiRenzo, Greenhaus, and Weer (2011) noted,increased working hours may also result in increasing interference between work and family responsibilities, thus resulting infamilywork conflict (FWC).

Previous empirical investigations into the relationship between hours worked and outcomes, including WFC and FWC, havegenerally found a positive relationship between hours worked and WFC (e.g. Barnett, 1998; Byron, 2005; Eby, Casper, Lockwood,

Journal of Vocational Behavior 80 (2012) 380389

Corresponding author.E-mail addresses: [email protected] (C.L. Adkins), [email protected] (S.F. Premeaux).

0001-8791/$ see front matter 2011 Elsevier Inc. All rights reserved.doi:10.1016/j.jvb.2011.09.003

Contents lists available at SciVerse ScienceDirect

Journal of Vocational Behavior

j ourna l homepage: www.e lsev ie r .com/ locate / jvb

Bordeaux, & Brinley, 2005; Ganster & Bates, 2003). For example, Major, Klein, and Ehrhart (2002) found that work time was sig-nificantly and positively related toWFC and that WFCmediated the relationship between the number of hours spent on work andpsychological distress. Fu and Shaffer (2001) also found that hours spent on paid work amplified WFC. Although previous studieshave examined the relationship between hours worked and WFC, many studies of WFC in general, and by extension, studies ofthe relationship between work hours and WFC have often used samples of managerial and professional employees (Byron,2005; DiRenzo et al., 2011; Eby et al., 2005), thus not examining the effect of long work hours (and other correlates of WFCand FWC) in samples of hourly workers. Earlier samples also tended to focus primarily on married employees with children,thus overlooking single employees and single parents. Similarly, a variety of measures of WFC have been used, but many havebeen one-dimensional, even including single-item measures. While some studies have also examined the relationship betweenhours and FWC, again, one-dimensional measures of FWC have been used. Despite Barnett's (1998) call to examine non-linearrelationships between work hours and WFC/FWC, few studies have done so. Nor have they examined work and non-work vari-ables that are potential moderators of the relationship between hours of work and WFC and FWC.

The purpose of the present paper is to examine the relationship between hours worked and workfamily and dimensions offamilywork conflict, using a heterogeneous sample of employees in seven work organizations. The employees in our sampleworked in jobs ranging from production and clerical to managerial and professional jobs. We examine both linear and nonlinearrelationships between hours worked and WFC and FWC in conjunction with family and job variables which may affect theserelationships. We do this using a sample of workers doing a wide variety of jobs across multiple organizations, who represent avariety of family structures. We use a more molecular measure of workfamily and familywork conflict so that we can examinethe relationship between hours worked and specific family roles including the roles associated with being a spouse or partner, theparental role, homecare role and leisure role. This approach to tappingWFC answers the call to examineWFC at a more molecularlevel, including the various roles that individuals play in their lives (Kossek & Ozeki, 1998; Lambert, 1990). This method shouldhelp us more fully understand not only if the number of hours worked contributes to WFC and FWC, but also which rolesare affected and how, thus examining predictions consistent with role theory.

Theory and hypotheses

Two theoretical perspectives which may be used to address the relationship between hours worked and outcomes are roletheory and the conservation of resources model. Role theory would suggest that individuals play multiple roles, and since timeis a finite resource, when hours spent in one role increase, there is potential for inter-role conflict due to fewer hours left to al-locate to other roles (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964). Fulfilling the demands associated with one role necessarilyreduces the number of hours available to allocate to other roles; thus, according to role theory, fulfillment of one role inevitablyleads to fewer available hours for others, and plausibly results in greater conflict. Even if multiple roles are fulfilled simultaneous-ly, such as by bringing work home to be with one's family, the often inconsistent expectations of the work and family roles wouldlead to less energy and attention for each role. Time may in such instances be maximized to accommodate multiple roles; how-ever, conflict would still be likely.

The conservation of resources (COR) model (Hobfoll, 1989) is another theoretical perspective that is useful in studies of WFC(e.g. Grandey & Cropanzano, 1999). The COR model theorizes that individuals attempt to get and keep resources. Resources aredefined as those objects, personal characteristics, conditions, or energies that are valued by the individual or that serve as ameans for attainment of these objects (Hobfoll, 1989, p. 516). Resources, including social support, autonomy, establishedbehavior-outcome (i.e., reward) contingencies, and so on (Hobfoll, 1989; Lee & Ashforth, 1996; Wright & Hobfoll, 2004), maybe used in problem solving and coping (Greenhaus & Powell, 2006). Stress may result when an individual perceives a potentialloss of resources, when there is an actual loss of resources, or when an expected resource gain does not materialize.

The COR model suggests that conflict between employees' home and work lives may deplete resources, or fuel perceptions ofdepleted resources, and thus lead to negative outcomes which may include workfamily and familywork conflict. Anything thatserves to replenish these resources should lead to lower WFC/FWC and positive outcomes, while factors that serve to depletethese resources should lead to higher WFC/FWC and negative outcomes.

Valcour (2007), in a study of telephone call center representatives, found that the relationship between work hours andsatisfaction with workfamily balance was moderated by control over work time. There was no relationship between workhours and satisfaction with workfamily balance for workers with a high degree of control over their work time, and a negativerelationship betweenwork hours and satisfactionwithworkfamily balance for workers with little control over their work time.Valcour's measure of satisfaction with workfamily balance incorporates what she refers to as the two most critical personalresources for meeting work and family demands, time and attention (energy) (Valcour, 2007, p. 1517). She suggests that anindividual's satisfaction with worklife balance is due, in part, to how he/she parcels his/her time and other resources to variouslife roles. In other words, affective responses to worklife issues depend on how well the individual utilizes resources. Herconceptualization of satisfaction with worklife balance is consistent with both role and COR theories, i.e., work and familyrole demands compete for an individual's time and other resources leading to conflict potential. How well the individual al-locates resources impacts their affective responses to worklife issues. Whereas effective resource use leads to satisfactoryworklife balance, when these resources are depleted or not well allocated, WFC and/or FWC may result. Greenhaus, Bedeian,and Mossholder (1987), in a study of accountants, found that work hours were associated with WFC. Similarly, Grzywacz andMarks (2000), using data from the 1995 National Survey of Midlife Development in the United States found a relationshipbetween hours of work and workfamily spillover. In her meta-analysis of WFC and its antecedents, Byron (2005) found

381C.L. Adkins, S.F. Premeaux / Journal of Vocational Behavior 80 (2012) 380389

that workers who worked longer hours had higher levels of WFC than FWC. Eby et al. (2005) in their literature review of WFCfound a positive relationship between hours worked andWFC. Brett and Stroh (2003) in a study examining whymanagers workextreme hours (defined as 61 or more hours per week), found that increased work hours were positively associated with WFCfor male managers, but not for female managers. It should be noted that their sample of male managers was restricted to thosewho were currently married with children living at home. Their sample of female managers was not restricted and 75% weremarried and 54% had children living at home.

Although multiple studies have found a significant relationship between hours of work and WFC, it is important that we onceagain examine this relationship. Thus, consistent with role theory (Kahn et al., 1964), the COR model (Hobfoll, 1989), and previ-ous research (e.g. Byron, 2005; DiRenzo et al., 2011; Eby et al., 2005; Greenhaus et al., 1987; Major et al., 2002) we hypothesizethat:

Hypothesis 1a. Hours of work will be positively associated with WFC.

Consistent with role theory and the COR model, DiRenzo et al. (2011) found that longer work hours were associated withincreased WFC and suggested that longer work hours may also increase FWC simply by making less non-work time availableto employees. Less non-work time may result in employees needing to take care of non-work matters during the extra hoursof work time. Barnett (1998) notes that negative home to work spillover is less likely for those who work part-time, suggestingthat working fewer hours allows more time to care for home and family. Again, the number of hours available for taking care offamily issues potentially impacts affective responses. Thus, we also examine the relationship between hours of work and FWC.

Hypothesis 1b. Hours of work will be positively associated with FWC.

Barnett (1998) noted that there have been weak and inconsistent findings for linear effects of work hours on WFC and otheroutcomes and suggested that these relationships may actually be non-linear. Rather than negative effects for the number of hoursworked regardless of the level, work hours may begin to have damaging effects only when an individual spends an inordinateamount of time on the job. It is also possible that the intensity of the relationship between hours of work and WFC/FWC maychange as the hours of work increases, thus resulting in a non-linear relationship. Barnett (1998) noted that although nonlinearrelationships are possible, few researchers have examined nonlinear relationships between work hours and outcomes. The excep-tions are Ganster and Bates (2003) and Ng and Feldman (2008).

Ganster and Bates, using data from the 1997 National Study of the Changing Workforce, examined curvilinear effects of hoursworked onWFC. Although Ganster and Bates (2003) did find curvilinear effects for hours worked on both job stress and WFC, theeffects actually diminished at higher levels. This suggests, contrary to popular belief, that working long hours leads to less, notmore, job stress andWFC. Given that theirs was an early study testing for non-linear effects of hours worked, and that their resultsrun counter to commonly held views concerning the impact of longer work hours on job outcomes, we elected to investigate non-linear effects in the current study.

Ng and Feldman (2008) in a meta-analysis using a social identity perspective to examine long work hours conducted an ex-ploratory analysis and found a curvilinear relationship between hours squared and the relationship between work hours andWFC variables (both WFC and FWC). They found a stronger relationship between hours worked and WFC and FWC respectivelyas hours increased; however, they are silent on the direction of that relationship. Thus, consistent with Ng and Feldman (2008)and Ganster and Bates (2003), we hypothesize that:

Hypothesis 2. Hours worked will have a curvilinear relationship with WFC and FWC.

Potential moderators of the number of hours workedoutcome relationships

Conditions in the employee's work setting and family life may moderate the relationship between hours worked andWFC andFWC. Barnett (1998) posits that there is no a priori reason to expect long work hours to have the same effects across workers andconditions, and that instead long work hours should be construed as a risk factor that when combined with other factors may leadto negative outcomes. She recommended examining the indirect effects of hours worked on various outcomes by incorporatingassorted moderators that address the nature of the job, tradeoffs and benefits associated with working longer hours, and the im-pact of longer hours on family members. Alternatively, Ganster and Bates (2003) tested a number of moderators and concludedthat the effects of work hours are universally experienced in the same way by different people. Our multi-dimensional measure ofWFC will allow us to better test Barnett's (1998) hypothesis that work hours interact with other factors to influence outcomes andGanster and Bates (2003) opposing conclusion. We examined the work variable managerial support, and the family variablesmarital status, number of children, and family support as moderators of the relationship between hours worked and WFC andFWC.

Work variables

We examined managerial support as a workplace moderator of the relationship between hours of work and WFC and FWC.Managerial support is a dimension of workfamily culture which addresses the extent to which an employee perceives thathis/her manager is supportive of his/her needs to balance their work- and non-work lives (Thompson, Beauvais, & Lyness,

382 C.L. Adkins, S.F. Premeaux / Journal of Vocational Behavior 80 (2012) 380389

1999). Consistent with COR theory (Hobfoll, 1989) managerial support for worklife balance is a resource that employees maydraw upon to reduce WFC/FWC as work hours increase. Prior research has shown that managerial support negatively influencesWFC (Anderson, Coffey, & Byerly, 2002; Premeaux, Adkins, & Mossholder, 2007). Having a manager who understands the de-mands of an employees' personal time may facilitate the employees' adjustment to working longer hours. Similarly, managerialsupport may lessen perceived FWC by creating a work environment in which the employees' family roles are affirmed, thus re-ducing the perception that family and home are an intrusion on work. Accordingly, consistent with both COR and role theories,supportive managers may be a resource who supports the multiple roles of the employee.

Hypothesis 3a. Managerial support will moderate the relationship between hours worked and WFC.

Hypothesis 3b. Managerial support will moderate the relationship between hours worked and FWC.

Family variables

We also examined the family variables presence of a spouse or partner, number of children, and family support as potentialmoderators. The presence of a spouse or partner, family support, and children may be conceptualized as resources that allowthe worker to better cope with long work hours. On the other hand, while children may increase psychological resources, theyalso are less likely to be understanding and supportive of a parent's long work hours and may be viewed as demanding more re-sources thus exacerbating the effects of longer work hours.

Consistent with the COR model (Hobfoll, 1989), evidence suggests that family support, either in the form of family memberssharing in duties and responsibilities or providing emotional support, counteracts WFC to some degree. (Adams, King, & King,1996; Carlson & Perrew, 1999). Similarly, role theory suggests that a spouse or partner may assist the worker by taking-onsome of his or her role-related duties at home (for example, taking on a household task that is normally the responsibility ofthe worker). Consistent with role theory, supportive family relationships are also plausibly less demanding resulting in less feltconflict than non-supportive family relationships. This may be especially true in terms of time spent at work, as family members'help and understanding should ameliorate to some degree the negative effects of working long hours. Thus we hypothesize thatthe presence of a spouse or partner and family support will moderate the relationship between hours of work andWFC and FWC.Alternatively, the presence of a spouse or partner may simply represent an additional role demand on the employee which de-pletes resources and thus strengthen the relationship between hours worked and WFC and FWC.

Hypothesis 4. The presence of a spouse or partner will moderate the relationship between hours of work and WFC and FWC.

Hypothesis 5. Family support will moderate the relationship between hours of work and WFC and FWC.

The number of children an employee has may also moderate the relationship between hours worked and WFC and FWC.Although in the context of the COR model (Hobfoll, 1989), children may serve as a source of support, thus reducing felt conflict,the COR model also suggests that children may be viewed as a drain on resources, and as the number of children increases, therelationship between hours worked andWFC and FWCmay be stronger. Similarly, role theory (Kahn et al., 1964) would suggestthat childrenwould place the demands of an additional role upon the employee and these demandsmay increase with the num-ber of children and thus increase WFC and FWC.

Hypothesis 6. The number of children an employee has will moderate the relationship between hours of work andWFC and FWCsuch that as the number of children increases the relationship between hours of work and WFC and FWC becomes stronger.

In the next sections we will present the methodology used in the present study, the study results, and a discussion of the studyresults.

Methods

Settings and participants

Data were gathered at seven organizations in the southern United States as part of a larger study. A total of 544 participantsprovided usable responses. The sites included an insurance company (21 participants), a distribution center (44 participants),a manufacturing organization (30 participants), a bank (18 participants), a utility company (128 participants), a not-for-profitservice organization (198 participants) and a hospital (105 participants). These sites thus represent both manufacturing andservice sectors of our economy as well as including a not-for-profit organization. Participants included 148 men (27.2%) and396women (70.8%). Approximately 52% of the respondents had at least one child under the age of 18, and 69.1% of the respondentshad a spouse or partner. The average age of the participants was 43.12. The jobs of the participants ranged from operations/production to top management. Participants were contacted through their employing organization and in some cases wereallowed to complete the questionnaire on work time. Participation was voluntary and all participants were assured thattheir individual responses would be totally confidential. To reward participation, participants had the opportunity to enterinto a random drawing for cash rewards with one or more rewards offered at each site.


Measures

Hours workedParticipants were asked to indicate how many hours per week they worked at their job and how many hours per week they

worked on company business at home. An open-ended response was used for both items. The variable hours worked was the sumof the responses to both items.

Work characteristicThe managerial support dimension of workfamily culture was measured using a scale developed by Thompson et al. (1999).

Eleven items measured managerial support (=.82) using a 5-point response format (1 = strongly disagree; 5 = stronglyagree).

Family characteristicsFamily characteristics measured included demographic variables and perceptions of family support. Participants were asked to

indicate the number of children (age 18 or younger) that they have. Participants were also asked to indicate whether they weresingle, married, or cohabiting with a significant other. Responses to this item were coded single = 0, married or cohabiting = 1.Family support was measured using a scale developed by Carlson and Perrew (1999). The seven-item scale of family support(=.88) asks participants to indicate the degree to which various features, such as appreciation, feedback from others,sharing of duties, and sharing of responsibilities are present in their family life. These items were measured on a four-pointscale (1 = not present; 4 = to a great extent).

OutcomesWe measured both WFC and FWC. Our measure of multi-dimensional WFC was adapted from a scale developed by Carlson,

Kacmar, and Williams (2000). The scale includes bi-directional (workfamily, and familywork) measures of time-, strain-, andbehavior-based conflict. We adapted the time- and strain-based scales for the roles of 1) spouse, 2) parent, 3) home care, and 4)leisure. Conflict between the role demands of work and spouse was measured using four sub-scales of three items each. Theyare: time-based workspouse conflict, time-based spousework conflict, strain-based workspouse conflict and strain-basedspousework conflict. Parallel scales measured conflict between the role demands of work and parent, work and home care, andwork and leisure. We conducted a factor analysis of these items and found that a three-factor solution provided the best fit. Thisresulted in one scale measuring WFC (24 items, =.97), a scale measuring spouse/parent FWC (12 items, =.93), and a scalemeasuring homecare/leisure FWC (12 items, =.95). A sample item for spouse/parent FWC is The time I spend on parentalresponsibilities often interferes with my work responsibilities. A sample item for home/leisure FWC is Tension and anxietyfrom my household responsibilities often weaken my ability to do my job. All items were measured using a 5-point responseformat (1 = strongly disagree; 5 = strongly agree). For factor loadings please see Premeaux et al. (2007).

The scales arising from the factor analysis, spouse/parent FWC and homecare/leisure FWC, are based on the amount ofdiscretion the worker has in performing the role. The expectations and responsibilities accompanying the familial roles of spouseand parent are often inflexible. These cannot be postponed for very long if positive relationships with family members are to bemaintained and there is no substitute for a spouse or parent. Although non-family members may provide child care, workers arebound to day-care center closing times, and must attend immediately to a child who is ill. On the other hand, homecare andleisure demands often are flexible and can often be delayed indefinitely or can often be omitted altogether. Some householdtasks can be completely outsourced (for example, by hiring a housekeeper or eating meals out).1

Control variablesPrevious research has found a relationship between age and sex and hours worked such that younger workers and males tend

to work longer hours (e.g. Brett & Stroh, 2003); thus, respondents were asked to indicate their age and sex. Age was coded inyears. Sex was coded as male = 0, female = 1.

Results

Table 1 presents means, standard deviations, and correlation coefficients for all study variables. All hypotheses were testedusing hierarchical multiple regression. To maximize usable sample size, mean replacements were used for missing variables.No mean replacements were used for the variable hours worked.

As Table 1 shows, hours worked were negatively related to gender and age. Men and younger workers tended to work morehours. Work hours were positively related to marital status participants with a spouse or partner tended to work longer hours.Work hours were also positively associated with WFC, home/leisure FWC, and spouse/parent FWC, thus lending preliminarysupport for our hypotheses.

The first two hypotheses were tested using hierarchical regression analysis. WFC and the two dimensions of FWC wereregressed on the control variables in the first step. The variable hours worked was added in the second step, and hours worked

1 Thanks to an anonymous reviewer for highlighting the distinction between these scales.


squared was added in the third. As Table 2 shows, hours worked was a significant predictor of WFC above and beyond the controlvariables. Hours worked squared was a significant predictor of homeleisure FWC and of spouse/parent FWC. The pattern of thenonlinear effects for both outcomes was such that the relationship between work hours and FWC was initially positive, and thenbecame negative. Although the nonlinear relationship is significant, increased work hours are associated with lower levels ofFWC. Thus, Hypotheses 1 is supported with respect to WFC and Hypothesis 2 is partially supported with respect to FWC, inthat there is a significant curvilinear relationship.

Hypotheses 36 were also examined using hierarchical regression analysis. WFC, homeleisure FWC, and spouse/parent FWCwere regressed on the control variables. The appropriate main effects were added in the second step, and the interaction term foreach hypothesis was added in the third. As Table 3 shows there were three significant moderated relationships. Because of spacelimitations, only the results in which the interaction was significant are reported.

The relationship between hours worked and homeleisure FWC was moderated by managerial support, thus lending supportto Hypothesis 3b. The relationships between hours worked and WFC and spouse/parent FWC were moderated by the numberof children, thus supporting Hypothesis 6. The patterns of the interactions were examined by plotting the interaction effects onthe dependent variable one standard deviation below and above the mean of the independent variable. As Fig. 1 shows, the levelsof homeleisure FWC were higher for the group with lower managerial support than for those with higher levels of managerialsupport at lower levels of hours worked. As hours of work increased, the level of homeleisure FWC declined for those with lowerlevels of managerial support, but increased for those with higher levels of managerial support. Perhaps those with higher levelsof managerial support were more able to respond to home care issues or leisure pursuits as hours of work increased, and thus,perceived some level of conflict.

As Figs. 2 and 3 show, the overall levels of WFC and spouse/parent FWC were higher for employees with more children.The slope of the relationship for those with fewer children was steeper than for those with more. Neither the presence of a spouseor partner, nor family support moderated the relationship between hours worked and WFC/FWC. Thus, there is support forHypothesis 3b and partial support for Hypothesis 6, but no support for Hypotheses 4 and 5.

Table 2Results of regressing WFC and WFC on control variables, hours worked, and hours worked squared.

Workfamily conicta Homeleisure familyworkconicta

Spouseparent familyworkconicta

Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Step 1 Step 2 Step 3

Sex .094 .127 .127 .012 .000 .000 .065 .054 .053Age .218 .187 .188 .065 .054 .058 .230 .219 .223Equation F 14.61 1.23 17.65

R2 .051 .005 .061F for R2 14.61 1.23 17.65

Hours worked .194 .13 .065 .254 .07 .243Equation F 17.10 1.55 12.70

R2 .036 .004 .005F for R2 21.00 2.18 2.69

Hours worked squared .068 .333 .328

Equation F 12.86 2.51 10.98

R2 .000 .010 .009F for R2 .238 5.36 5.50

a Standardized Betas. pb .05. pb .01.

Table 1Means, standard deviations, and correlations among study variables.

Mean Std. deviation 1 2 3 4 5 6 7 8 9

1. Total hours 40.40 8.47 2. Sex .72 .44 .19 3. Age 43.12 11.94 .18 .14 4. Managerial support .29 .45 .01 .06 .13 5. Marital status .69 .46 .10 .13 .04 .01 6. Number of children .85 1.01 .07 .03 .30 .11 .20 7. Family support 19.66 4.95 .03 .11 .03 .15 .13 .01 8. Workfamily conflict 58.29 30.72 .20 .06 .20 .32 .28 .33 .16 9. Homeleisure FWC 17.61 8.05 .08 .03 .06 .13 .07 .03 .14 .41 10. Spouseparent FWC 13.32 9.85 .12 .09 .22 .16 .32 .43 .04 .62 .47

pb .05. pb .01.


Discussion

In this study we examined the relationship between hours worked, WFC and FWC using a heterogeneous sample of workersin jobs ranging from operations and clerical to professional and top management, and using a multidimensional measure ofWFC and FWC. We found evidence of a linear relationship between hours worked and WFC, and a curvilinear relationship be-tween hours worked and both homeleisure FWC and spouse/parent FWC. Our findings are consistent with role theory (Kahnet al., 1964) and with the COR model (Hobfoll, 1989). The finding that hours worked has a linear relationship with WFC is alsoconsistent with previous research (e.g. Byron, 2005; DiRenzo et al., 2011; Eby et al., 2005; Greenhaus et al., 1987; Major et al.,2002).

Our findings for a non-linear relationship between hours worked and both dimensions of FWC are consistent with those ofGanster and Bates (2003). The form of this relationship was such that as hours worked increased, homeleisure FWC andspouse/parent FWC increased to a point and then declined with additional hours of work indicating an inverted U-shaped

Table 3Results of regressing WFC and WFC on control variables, hours worked, and moderators.

Homeleisure familywork conicta

(moderator: managerial support)Workfamily conicta

(moderator: number of children)Spouseparent familywork conicta

(moderator: number of children)

Step 1 Step 2 Step 3 Step 1 Step 2 Step 3 Step 1 Step 2 Step 3

Sex .216 .232 .308 6.51 8.73 8.87 1.44 1.20 1.16Age .044 .026 .021 .56 2.57 .22 .19 .081 .07Equation F 1.23 14.61 17.65

R2 .005 .051 .061F for R2 1.23 14.61 17.65

Hours worked .061 .058 .691 .78 .08 .10

Moderator .124 .127 8.70 8.85 3.82 3.87Equation F 3.11 25.90 35.36

R2 .018 .11 .147F for R2 4.97 35.55 49.87

Interaction term .015 .24 .08Equation F 3.64 21.77 29.52

R2 .01 .007 .007F for R2 5.65 4.57 5.12

a Unstandardized Betas; Independent variables centered. pb .05. pb .01.

Fig. 1. Managerial support as a moderator of the relationship between hours worked and home/leisure familywork conflict.


curve. This suggests that once hours of work reach a certain level, employees may make accommodations, such as seeking addi-tional childcare or other assistance to reduce WFC and FWC. Alternatively, those working longer hours may be workers who aresimply less concerned with workfamily balance.

As a post-hoc analysis we examined the relationship between hours worked on company business at home andWFC and FWC,controlling for total hours worked. Hours worked at home was a significant predictor of spouse/parent FWC above and beyondthe control variables (age and sex) and total hours worked; however, it must be noted that the overall equation regressingspouse/parent FWC on age, sex, and total hours worked was not statistically significant. Because some respondents reportedworking over 40 h a week at the workplace, we also examined the relationship between hours worked over 40 a week (regardlessof location) andWFC/FWC. Hours over 40 a week was significantly related toWFC and both dimensions of FWC above and beyond

Fig. 2. Number of children as a moderator of the relationship between hours worked and workfamily conflict.

Fig. 3. Number of children as a moderator of the relationship between hours worked and spouseparent family work conflict.


total hours worked. This finding suggests that working hours over what is considered a standard weekmay lead to higher levels offelt conflict.2

We also examined potential moderators of the relationship between hours worked and WFC and FWC consistent withBarnett's (1998) suggestion that moderators should be examined. The finding that managerial support moderated the relation-ship between hours worked and homeleisure FWC is consistent with DiRenzo et al. (2011) suggestion that work variables mayneed to be examined as antecedents of FWC. The fact that home and leisure FWC is the outcome of interest is of special note.There is a call for the workfamily literature to be expanded to include the worklife balance of all employees, not just thosewith spouses and/or children. This finding suggests that a positive worklife culture may have a positive impact beyondmarriedemployees with children and that caring for home and leisure activities also warrant managers' support. Even single individualswithout children have lives beyond work that need to be accommodated.

The number of children a worker had moderated the relationship between work hours and WFC and spouse/parent FWC.Workers with more children exhibited higher overall levels of WFC and spouse/parent FWC than did those with fewer children.For workers with fewer children, the steeper slope of the relationship indicates that additional hours of work had a strongerrelationship with WFC and FWC. This finding is consistent with the COR model (Hobfoll, 1989) which suggests that childrenmay serve as an additional drain on the resources of the worker and managers need to accommodate employees' childcareneeds as much as possible. The stronger relationship between work hours and WFC and FWC of individuals with fewer chil-dren may be explained several ways. One possibility is that those with fewer children had younger children so that childcarearrangements were more crucial and expensive than for those with more, but older children. School-age children normallyrequire less child care and beyond a certain age daily child care arrangements are no longer necessary so resource use de-clines as children get older. Another possibility is that although more children are a greater resource drain, they may, para-doxically, also serve as a resource source by helping care for younger siblings, assisting with household duties, and byproviding emotional support. Another explanation is the emotional and instrumental support that more children can provideto one another. Parents may feel less guilt and anxiety, and thus less conflict, about working longer hours when they knowthat their children have each other.

Consistent with the findings of Ganster and Bates (2003), the percent of variance in the dependent variables explained bywork hours (alone or in interactions with other predictors) was relatively small; however, we believe that these findings are im-portant. The relationship between work hours and work- and family-related outcomes is a phenomenon that, by nature, can onlybe studied in the field. As with any field study, there is a myriad of unmeasured variables that may potentially attenuate observedrelationships among study variables. Further, in the present study, the variance of the variable hours workedmay have attenuatedobserved relationships. The majority (55.9%) of the sample reported working 40 h a week. Of the remainder of the sample, 21.9%reported working less than 40 h per week (range 639.5), and the remaining 22.2% reported working more than 40 h per week(range 40.390). The relatively low variance in this critical variable may have attenuated statistical relationships; however,the sample was representative of a diverse group of workers. Thus, the observed relationships may well under represent thephenomenon.

To develop this argument, we draw an analogy from Locke's (1986) discussion of generalizing from laboratory to field settings.He cited Mook's (1983) argument that some lab experiments may be conducted simply to examine whether a phenomenon ex-ists; regardless of whether the findings could be expected to generalize. Locke and Mook's points about the value of the tightlycontrolled conditions of the lab are well-taken. We would argue that our results, combined with those of Ganster and Bates(2003), have demonstrated that, even outside of the tightly controlled conditions of a laboratory, the phenomenon exists. Thesmall effects in a diverse sample do not eliminate the possibility that for some groups of workers, in some situations, the effectsof long hours worked may be marked. The latter argument is consistent with the discussions of Barnett (1998). Further, we wouldposit that the observed effects may be further attenuated by self-selection into jobs with time demands compatible with other liferoles. For example, childless workers may choose jobs that demand long work hours, while workers with young children may se-lect jobs that offer reduced hours and/or flexible work schedules. An anecdotal example of this is the fact that the SAS Institute,which guarantees workers a 30 h work week (albeit with lower salaries than the industry average) is repeatedly cited as a bestcompany to work for largely because it allows workers to choose a job compatible with life roles (Levering & Moskowitz, 2003,January 20). Further, when balancing between work and non-work time demands, workers may choose to emphasize some rolesat the expense of others. For example, a parent facing increased work hours may choose to devote his/her non-work hours to theparenting role at the expense of the homecare role. Such juggling among non-work roles would also serve to attenuate observedrelationships between work hours and outcomes.

Limitations and future research

As with most studies, our study is not without limitations. The relatively low variance in hours worked, noted above, is a lim-itation to the present study. Clearly, the relationship between hours worked and the three dimensions of WFC should be exam-ined by comparing samples of workers with different levels of work hours. As noted, however, our sample extends the literaturebeyond studies of managerial and professional workers who tend to work longer hours on average. Future studies should also beexpanded to capture more of the complexities of worklife balance for all workers.

2 The authors thank an anonymous reviewer for the suggestion for these post-hoc analyses.


Another limitation is the cross-sectional nature of the study. Because cross-sectional studies do not allow for a true test ofcausality or rule out the possibility of reverse causality, future research should attempt to incorporate the effects of time onthe various types of WFC. It is plausible to expect the natural ebb and flow of workers' lives, including their career stage andtheir stage in the family life cycle to cause the different types of interrole conflict to vary over time and for the impact ofhours worked to vary accordingly.

Conclusions

In sum, we have demonstrated that hours worked, in conjunction with job and family variables are related to WFC and FWC.We argue that although the observed effects were small, they are important. Our findings, together with those of Ganster andBates (2003), have shown that the amount of time devoted to a life role, specifically work, does impact other life roles. Futureresearch should build on these findings to explore the scope and boundary conditions of this phenomenon.

References

Adams, G. A., King, L. A., & King, D. W. (1996). Relationships of job and family involvement, family social support, and WFC with job and life satisfaction. Journal ofApplied Psychology, 81, 411420.

Anderson, S. E., Coffey, B. S., & Byerly, R. T. (2002). Formal organizational initiatives and informal workplace practices: Links to workfamily conflict and job-relatedoutcomes. Journal of Management, 28, 787810.

Barnett, R. C. (1998). Toward a review and reconceptualization of the work/family literature. Genetic, Social, and General Psychology Monographs, 1242, 125153.Barnett, R. C., & Gareis, K. C. (2000). Reduced-hours job-role quality and life satisfaction among married women physicians with children. Psychology of Women

Quarterly, 24, 358364.Brett, J. M., & Stroh, L. K. (2003). Working 61 plus hours a week: Why do managers do it? Journal of Applied Psychology, 88, 6778.Byron, K. (2005). A meta-analytic review of workfamily conflict and its antecedents. Journal of Vocational Behavior, 67, 169198.Carlson, D. S., Kacmar, K. M., &Williams, L. J. (2000). Construction and initial validation of amultidimensionalmeasure of workfamily conflict. Journal of Vocational

Behavior, 56, 249276.Carlson, D. S., & Perrew, P. L. (1999). The role of social support in the stressorstrain relationship: An examination of workfamily conflict. Journal of Management,

25, 513540.DiRenzo, M. S., Greenhaus, J. H., & Weer, C. H. (2011). Job level, demands, and resources as antecedents of workfamily conflict. Journal of Vocational Behavior, 78,

305314.Eby, L. T., Casper,W. J., Lockwood, A., Bordeaux, C., & Brinley, A. (2005).Work and family research in IO/OB: Content analysis and review of the literature (19802002).

Journal of Vocational Behavior, 66, 124197.Fu, C., & Shaffer, M. A. (2001). The tug of work and family: Direct and indirect domain-specific determinants of workfamily conflict. Personnel Review, 30,

502522.Ganster, D. C., & Bates, C. (2003). Do long work hours decrease general well-being and increase workfamily conflict? A paper presented at the annual meeting of

the Academy of Management, Seattle, WA.Grandey, A. A., & Cropanzano, R. (1999). The conservation of resources model applied to workfamily conflict and strain. Journal of Vocational Behavior, 54,

350370.Greenhaus, J. H., Bedeian, A. G., & Mossholder, K. W. (1987). Work experiences, job performance, and feelings of personal and family well-being. Journal of Vocational

Behavior, 31, 200215.Greenhaus, J. H., & Powell, G. N. (2006). When work and family are allies: A theory of workfamily enrichment. Academy of Management Review, 31, 7292.Grzywacz, J. G., & Marks, N. F. (2000). Reconceptualizing the workfamily interface: An ecological perspective on the correlates of positive and negative spillover

between work and family. Journal of Occupational Health Psychology, 5, 111126.Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44, 513524.Hochschild, A. R. (1997). The time bind. New York: Holt and Company.Johnstone, M. B. (1960). How to live every day of your life. Indianapolis, Indiana: Bobbs-Merrill.Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Organizational stress: Studies in role conflict and ambiguity. Oxford, England: JohnWiley.Kossek, E. E., & Ozeki, C. (1998). Work-family conflict, policies, and the joblife satisfaction relationship: A review and directions for organizational behavior-

human resources research. Journal of Applied Psychology, 83, 139149.Lambert, S. J. (1990). Processes linking work and family: A critical review and research agenda. Human Relations, 43, 239257.Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the three dimensions of job burnout. Journal of Applied Psychology, 81, 123133.Levering, R., & Moskowitz, M. (2003, January 20). 100 best companies to work for. Fortune, 147(1), 127+.Locke, E. A. (1986). Generalizing from laboratory to field settings. Lexington, MA: D. C. Heath & Company.Major, V. S., Klein, K. J., & Ehrhart, M. G. (2002). Work time, work interference with family, and psychological distress. Journal of Applied Psychology, 87, 427436.Mook, D. G. (1983). In defense of external invalidity. American Psychologist, 38, 379387.Ng, T. W. H., & Feldman, D. C. (2008). Long work hours: a social identity perspective on meta-analysis data. Journal of Organizational Behavior, 29, 853880.Premeaux, S. F., Adkins, C. L., & Mossholder, K. W. (2007). Balancing work and family: A field study of multi-dimensional, multi-role workfamily conflict. Journal

of Organizational Behavior, 28, 705727.Thompson, C. A., Beauvais, L. L., & Lyness, K. S. (1999). When workfamily benefits are not enough: The influence of workfamily culture on benefit utilization,

organizational attachment, and workfamily conflict. Journal of Vocational Behavior, 54, 392415.Valcour, M. (2007). Work based resources as moderators of the relationship between work hours and satisfaction with workfamily balance. Journal of Applied

Psychology, 92, 15121523.Wright, T. A., & Hobfoll, S. E. (2004). Commitment, psychological well-being, and job performance: An examination of conservation of resources (COR) theory and

job burnout. Journal of Business and Management, 9, 389406.


Spending time: The impact of hours worked on workfamily conflictTheory and hypothesesPotential moderators of the number of hours workedoutcome relationshipsWork variablesFamily variables

MethodsSettings and participantsMeasuresHours workedWork characteristicFamily characteristicsOutcomesControl variables

ResultsDiscussionLimitations and future research

ConclusionsReferences

Documents

Spending Time