Journal of Business and Psychology

, Volume 27, Issue 3, pp 255–269

Altering the Effects of Work and Family Conflict on Exhaustion: Telework During Traditional and Nontraditional Work Hours

Authors

    • Lally School of Management & TechnologyRensselaer Polytechnic Institute (RPI)
Article

DOI: 10.1007/s10869-011-9247-0

Cite this article as:
Golden, T.D. J Bus Psychol (2012) 27: 255. doi:10.1007/s10869-011-9247-0

Abstract

Purpose

The current study investigates the impact of time and strain-based work-to-family conflict (WFC) and family-to-work conflict (FWC) on exhaustion, by considering the moderating effect of telework conducted during traditional and non-traditional work hours.

Design/Methodology/Approach

Data were obtained from professionals in a large computer company using survey methodology (N = 316).

Findings

Results from this study suggest that time and strain-based WFC and FWC were associated with more exhaustion, and that exhaustion associated with high WFC was worse for individuals with more extensive telework during traditional and non-traditional work hours.

Implications

This study provides managers with findings to more carefully design telework programs, showing evidence that the adverse impact of WFC/FWC on exhaustion may depend on the type of telework and level of conflict experienced. This suggests that managers may need to be more aware of the full range of characteristics which encapsulate the teleworker’s work practices before making decisions about how telework is implemented.

Originality/Value

By differentiating the timing of telework and its role on the WFC/FWC—exhaustion relationship, this study delves deeper into the contingent nature of telework and suggests that the extent of telework conducted during traditional and nontraditional work hours may play an influential role. In addition, these considerations are investigated in light of the bi-directional time-based and strain-based nature of WFC and FWC, helping to unravel some of telework’s complexities.

Keywords

TeleworkFlexible workTelecommutingWork–family conflictExhaustion

Telework generally involves using computer technology to work from home away from the main worksite for a portion of the work week (Bailey and Kurland 2002). With growth of nearly 30% per year in the U.S. and other industrialized countries (Office of National Statistics 2005; WorldatWork 2007), the dramatic trends in telework may in part be spurred by the view that mixing work and home helps relieve work exhaustion and conflict between work and family (Bailey and Kurland 2002; Golden 2006). Exhaustion has been linked to a host of other detrimental work outcomes such as lower performance, burnout, increased turnover, and adverse health effects (e.g., Carson et al. 2010; Cropanzano et al. 2003; Demerouti and Bakker 2006; Maslach et al. 2001; Moore 2000a; Shirom 2011; Swider and Zimmerman 2010; Taris 2006). Given that estimates report upwards of 80% of employees struggle on a daily basis to meet work and family demands (Galinski et al. 1993), identifying factors such as telework that might alter the negative consequences of work–family conflict on exhaustion is therefore becoming a top concern (Barling et al. 2004). To date, however, research on work–family conflict has tended to focus on traditional work modes where work and family domains are highly segmented (Eby et al. 2005), and research on work exhaustion has been similarly focused (Moore 2000a, b).

Although the work–family conflict and exhaustion literatures are beginning to consider telework (e.g., Golden 2006; Standen et al. 1999), these literatures have treated telework as an omnibus factor, examining it as a unidimensional construct that only considers quantity/frequency of telework (e.g., Golden 2006; Wiesenfeld et al. 2001) or by generalized comparisons (e.g., Bailey and Kurland 2002; Igbaria and Guimaraes 1999). In part, this gap in the literature may be due to definitional ambiguity. Telework has historically been viewed as a substitution of work done in at the main worksite with work done at the home location (Madsen 2003; Nilles 1994; Pratt 1999). Telework, however, is not exclusively conducted during traditional business hours as it has generally been studied (e.g., Bailey and Kurland 2002; Igbaria and Guimaraes 1999), since it enables work to be easily accomplished during the evenings or weekends from home (Ammons and Markham 2004; Kuglemass 1995). Telework during nontraditional hours (hereafter referred to as nontraditional telework) involves the substitution of work done at the main worksite with work done at home during non-traditional work hours using technology. In contrast to augmenting work done in the office as with overtime or supplemental work (e.g., Duxbury et al. 1996; Venkatesh and Vitalari 1992), nontraditional telework involves the flexibility to conduct work at home instead of the typical “9 to 5” business hours. Hence in this study, I distinguish between telework conducted during traditional work hours (hereafter referred to as traditional telework), and nontraditional telework. From a researcher and practitioner perspective, failure to differentiate between these forms of telework makes it difficult to know why telework may influence work outcomes, since such treatments otherwise allow for only more general comparisons.

This study undertakes a more fine-grained analysis of telework, as part of an investigation to understand when work–family conflict impacts work exhaustion. Using a sample of 316 teleworkers, this study expands the telework literature to encompass both traditional telework and nontraditional telework. Moreover, this study considers these in light of the bi-directional time-based and strain-based nature of work–family conflict (WFC) and family–work conflict (FWC). Although some researchers have considered additional dimensions of conflict (e.g., Carlson et al.2000), given that the emphasis has been primarily on time- and strain-based forms (Adams et al. 1996; Rotondo et al. 2003), these dimensions are considered here. Drawing from the conservation of resources model (Hobfoll 1988, 1989), it is suggested that the collocation of work and family domains inherent in traditional and nontraditional telework may alter an individual’s resource recovery opportunities, thereby influencing energy and exhaustion. In addition to investigating WFC’s and FWC’s impact on work exhaustion and the moderating role of telework on this relationship, a more fine-grained analysis of telework may be necessary if we are to shed insights into its poorly understood complexities (Gajendran and Harrison 2007; Golden and Raghuram 2010).

Theory and Hypotheses

Work exhaustion is a growing concern for many professional-level employees, who must not only be responsive to increasing demands and expectations from colleagues but also to family demands spurred by changed family structures and expectations for high involvement in all aspects of family activities. Work exhaustion is the depletion of energy needed to meet job demands (Moore 2000a, b). It occurs when employees feel unable to meet the demands placed upon them (Hobfoll 1989; Lee and Ashforth 1996; Wright and Cropanzano 1998). As noted by researchers in the literature on exhaustion and burnout (e.g., Cropanzano et al. 2003; Densten 2001; Wright and Cropanzano 1998), exhaustion is central to the impacts sustained by individuals struggling to cope with feelings of being overwhelmed. Researchers have previously linked exhaustion to an array of negative outcomes, including increased absenteeism, turnover, physical illness, reduced satisfaction, and lower job performance (e.g., Carson et al. 2010; Cropanzano et al. 2003; Demerouti and Bakker 2006; Maslach et al. 2001; Moore 2000a; Parker and Kulik 1995; Shirom 2011). Meta-analyses have also supported the adverse consequences of exhaustion on outcomes (Lee and Ashforth 1996; Swider and Zimmerman 2010).

Individuals who experience conflict between work and family are subject to exhaustion as a result of tension between work and family domains, which is likely to be draining on their emotional and physical energy. According to the conservation of resources perspective (Hobfoll 1988, 1989), individuals strive to retain, protect, and build resources such as energy and time. Moreover, during recovery periods when not confronted with an immediate need for resources, individuals attempt to recover and stockpile resources to counter future losses (Lee and Ashforth 1996; Wright and Cropanzano 1998). To the extent that conflict between work and family act to consume an individual’s energy and other resources, and prevent ‘resource stockpiling’ that may be necessary to ward off other current and future resource needs (Hobfoll 1989), individuals are apt to experience higher levels of exhaustion.

Individuals with a high level of WFC are prone to exhaustion since conflicts associated with demands in the work domain that interfere with family obligations are apt to be psychologically and physically draining, hindering or preventing an individual’s ability to cope and leading to feelings of being overwhelmed (Hobfoll 1989). Time-based WFC entails work interfering with time needed for family activities (Carlson et al. 2000; Kossek and Ozeki 1998), initiating personal deficits that wear away emotional resources and lead to exhaustion. With higher time-based WFC, individuals are unable to accomplish demands in the family domain due to time needed by work activities, creating time-based shortfalls that sap their emotional stamina, pulling away energy that might otherwise have been available to cope with their resource demands (Hobfoll 1989). Moreover strain-based WFC, or strain from work interfering with family (Carlson et al. 2000; Kossek and Ozeki 1998), entails the consumption of additional energy to cope with anxiety and pressure that depletes personal resources and leads to exhaustion. In this way, strain from work interfering with family creates anxiety and drains energy that is likely to result in higher levels of exhaustion.

Similarly, because time-based FWC involves family demands that hinder or prevent time being spent on work tasks due to one’s family/personal role (Netemeyer et al. 1996), higher time-based FWC is apt to wear away emotional resources and lead to greater exhaustion (Hobfoll 1989). Whereas time spent in the family domain is not available to spend on work, high levels of time-based FWC are apt to create time-induced shortfalls that deplete personal resources and exhaust individuals trying to juggle the opposing time demands (Hobfoll 1989). Furthermore, strain-based FWC is likely to evoke exhaustion from dealing with the burdens and anxiety generated by family obligations that are incompatible with work (Kossek and Ozeki 1999). With high strain-based FWC, the greater anxiety and pressures are likely to tap emotional and physical energy reserves and deplete resistance to stress (Geurts and Demerouti 2003; Meijman and Mulder 1998), draining energy and leading to greater exhaustion.

In this way, individuals with higher levels of time and strain-based WFC and FWC are likely to have higher levels of exhaustion and burnout (Allen et al. 2000; Demerouti et al. 2004, 2005). Therefore, consistent with outcomes reported in the work and family literature for those in traditional work modes (Haar 2006; Hall et al. 2010; Kossek and Ozeki 1999; Wittmer and Martin 2010), it is expected that teleworkers with higher time and strain-based WFC and FWC will experience higher levels of exhaustion than those with less WFC and FWC. Stated formally,

H1

(a) Time-based WFC, (b) strain-based WFC, (c) time-based FWC, and (d) strain-based FWC are positively related to exhaustion.

Unpacking the Moderating Role of Telework

In contrast to more traditional work arrangements, telework’s inherent flexibility in the location and timing of work is likely to affect the ability to recover from WFC and FWC and influence work exhaustion. Through shifts in “where” and “when” work is accomplished, telework may influence the ability to recover from resource drains (Hobfoll 1988, 1989) brought about by tension between work and family domains and an individual’s ability to cope with the conflict, with corresponding changes in energy and exhaustion. Whereas a number of reviews of the telework literature have noted the varying conceptualizations of telework and the narrow treatment of its effects (e.g., Bailey and Kurland 2002; Baruch 2000; Gajendran and Harrison 2007), and telework researchers are beginning to consider telework’s more complex moderating effects (e.g., Golden and Raghuram 2010; Wiesenfeld et al. 1999), this study differentiates two critical forms of telework that have up to this point been undifferentiated. As noted earlier, these are the amount of telework conducted during typical work hours (traditional telework), and the amount of telework conducted outside of typical work hours (nontraditional telework).

These two forms of telework help characterize the changed nature of work experienced by teleworkers (Bailey and Kurland 2002; Kugelmass 1995; Valcour and Hunter 2005), and the manner or degree to which individuals may be able to recuperate from the draining effect of conflict (Hobfoll 1989; Nippert-Eng 1996a, b). Since telework generally involves the collocation of workspace within the home, the additional physical reminders of conflict and lack of role demarcation may prevent psychological disengagement (Kossek et al. 2006; Standen et al. 1999), hindering recovery from conflict and depleting additional energy and resources that lead to exhaustion. By acting to influence the degree to which time- and strain-based WFC and FWC depletes resources that lead to energy loss (Hobfoll 1988, 1989), telework may alter the ‘resistance capacity’ that leads to work exhaustion. Considering that existing research examines more generalized conceptualizations of telework rather than considering them separately as done here (e.g., Bailey and Kurland 2002; Wiesenfeld et al. 2001), it is hoped this more realistic and nuanced approach will shed additional insights.

Extent of Telework During Traditional Hours

The extent of traditional telework refers to the degree to which a teleworker’s work activities are conducted at home during traditional work hours using technology, instead of at the main worksite (Kraut 1989; McCloskey and Igbaria 1998; Olson 1989). Similar to telework “intensity” (Gajendran and Harrison 2007), individuals with extensive telework spend a larger proportion of their week teleworking from home rather than the corporate office, thereby collocating their work and family and altering their ability to recover from the time and strain-based conflict between domains. Extensive telework is likely to have a detrimental effect on an individual’s ability to cope with the energy and resource drains associated with conflict between work and family, leading to higher levels of exhaustion. Individuals who telework more extensively experience more constant physical reminders of the conflict between work and family due to their greater presence in the home (Kossek et al. 2006; Standen et al. 1999). Examples include laundry, bills, and the routines of family members such as school buses. Unlike counterparts who spend little time at home, those who telework extensively therefore have more constant and salient reminders of the conflict (Fiske and Taylor 1991), which adds anxiety and drains energy. Moreover, these serve as a continual irritant preventing psychological detachment and subsequent recovery (Demerouti et al. 2007; Fritz et al. 2010; Taris et al. 2006), leading to higher levels of exhaustion (Bakker et al. 2008).

For individuals who telework extensively, this lack of reprieve and recovery from the time and strain-based conflict between work and family is apt to be draining on energy levels, yet this is likely to be especially draining among individuals with high levels of WFC/FWC, since a reprieve under these conditions may be particularly necessary. With high WFC/FWC and a large proportion of the workweek spent working at home, the inability to physically and mentally distance oneself from conflict is particularly problematic (Sonnentag and Fritz 2007) and is likely to prevent recovery and self-regulatory resources from being effective (Baumeister et al. 1998). While greater presence in the home is apt to be of benefit in terms of saved commute time and strain (Baruch 2000; Igbaria and Guimaraes 1999), these are likely to be off-set by the additional resource drains stemming from mental and physical preoccupation with high levels of conflict and the associated lack of recovery time (Demerouti et al. 2007; Fritz et al. 2010; Taris et al. 2006). Consequently, the increase in exhaustion brought about by rising levels of WFC/FWC should be exacerbated by extensive telework, because working extensively at home in conjunction with high levels of conflict deceases the individual’s ability to recover and depletes additional energy and resources.

H2

The positive relationship between WFC/FWC (H2a: time-based WFC; H2b: strain-based WFC; H2c: time-based FWC; H2d: strain-based FWC) and exhaustion will be moderated by telework, such that exhaustion will be higher for individuals with more extensive traditional telework.

Extent of Non-Traditional Telework

The extent of nontraditional telework refers to the amount of time spent teleworking at home during non-traditional work hours. Though receiving far less attention in the literature than telework conducted during traditional work hours (Bailey and Kurland 2002; Towers et al. 2006), nontraditional telework replaces traditional work hour time with nontraditional work hour time (Fenner and Renn 2004; King 1998; Towers et al. 2006). Often discussed in the popular press (e.g., Zeytinoglu et al. 2009), individuals engaged in nontraditional telework generally work some of the day during regular work hours, but also spend evenings or weekends teleworking from home to make up for hours not worked during the regular work day. More extensive nontraditional telework is likely to exacerbate WFC’s impact on exhaustion, since spending a greater proportion of work periods during evening hours is likely to interfere with rest and recovery that typically takes place during these periods (Fritz et al. 2010). From a COR theory perspective (Hobfoll 1988, 1989), nontraditional telework represents a decrease in recovery periods from the tension created by conflict between work and family, given that evenings and other extended periods away from conflicting demands tend to bolster psychological detachment and recovery (Sonnentag and Fritz 2007). Because working during the evening hours in the midst of other family routines and rhythms may prevent mentally distancing oneself from work (Etzion et al. 1998), extensive nontraditional telework is likely to hinder more complete recuperation from the conflict and lead to greater resource drain and exhaustion.

Moreover, because individuals who engage in extensive nontraditional telework tend to work a significant proportion of their time during evening or non-standard hours (e.g., Kugelmass 1995; Nilles 1994), the neighbors, friends, and family members of these teleworkers may not fully understand or adhere to traditional social customs that respect boundaries around work time (Standen et al. 1999). Instead, they may possess expectations for participation in activities that demand additional time and energy during these work periods (Stephens and Szajna 1998), thereby further hindering resource recovery and adding to exhaustion. For individuals with high WFC/FWC, this lack of recovery and draining effect is likely to be especially apparent, as the salience and intensity of WFC/FWC in the home domain after regular work hours is likely to be higher than for individuals with less conflict. With more constant preoccupation with the conflict and a lack of recovery periods in the evening hours due to extensive nontraditional telework, those with high WFC/FWC are likely to experience even more resource drain and exhaustion (Bakker et al. 2008). In addition, while nontraditional telework may permit working less during regular work hours and may therefore provide some resource recovery opportunities, these are likely more than off-set by the intrusion into traditional family time that depletes further resources and hinders time spent with family to truly facilitate rest and recovery (Hobfoll 1988, 1989; Fritz et al. 2010). The draining effect of high work family conflict and extensive work outside of regular work hours, although unique to each family setting, therefore serves to intensify the exhaustion experienced. Thus, the increase in exhaustion brought about by rising levels of WFC/FWC is likely exacerbated by extensive nontraditional telework, since working extensively during periods outside of regular work hours in conjunction with high levels of conflict decreases the individual’s ability to recovery and depletes additional energy and resources.

H3

The positive relationship between WFC/FWC (H3a: time-based WFC; H3b: strain-based WFC; H3c: time-based FWC; H3d: strain-based FWC) and exhaustion will be moderated by nontraditional telework, such that exhaustion will be higher for individuals with more extensive nontraditional telework.

Interaction Between Traditional and Nontraditional Telework

Whereas extensive telework during traditional or nontraditional hours is likely to exacerbate the effects of high WFC and FWC on exhaustion, the interactive effects of telework during both time periods may also influence exhaustion. Although prior research has not empirically considered the interaction of these two variables with WFC/FWC, the scant available evidence suggests that their combined effects will alter work exhaustion. Following earlier arguments, with more extensive telework during traditional and nontraditional times, time and strain-based WFC and FWC are apt to be salient to the individual a much greater proportion of time, compared to those individuals who telework little. Whereas individuals who work predominantly at the main worksite may benefit from recovery opportunities due to a greater separation of work and family domains (Demerouti et al. 2007; Fritz et al. 2010), individuals with extensive traditional and nontraditional telework are not afforded such buffering effects. Instead, these individuals have a higher comingling of work and family cues (Standen et al. 1999), which precipitate more constant and continual exposure to the existing conflict between work and family (Demerouti et al. 2007; Fritz et al. 2010; Taris et al. 2006). Indeed, for those with extensive traditional and nontraditional telework accompanied by high levels of conflict between work and family, the ever-present tension between demands in each domain is likely to be especially draining on their physical and emotional energy (Bakker et al. 2008). With greater tension and accompanying anxiety, these individuals are apt to miss out on sufficient periods of time to recover from such resource drains, and such constant exposure is likely to further sap energy levels and increase exhaustion (Hobfoll 1988, 1989). While there is likely to be some saved energy and resources due to less commuting and extra time during regular work hours (Stephens and Szajna 1998), these are also likely off-set by the greater tension brought about by collocation of work and family domains and the corresponding lack of sufficient recovery time (Hobfoll 1989). Compared to those who engage in limited traditional and nontraditional telework and have little conflict between work and family, extensive teleworkers are likely to experience less adequate recovery time and more exposure to the tension between work and family domains, eroding resources which leads to higher levels of exhaustion.

H4

There will be a three-way interaction between traditional telework, nontraditional telework, and WFC/FWC, such that the relationship between WFC/FWC (H4a: time-based WFC; H4b: strain-based WFC; H4c: time-based FWC; H4d: strain-based FWC) and exhaustion will be stronger for individuals with extensive traditional and nontraditional telework.

Method

Sample and Procedure

The collection of data was undertaken in a large computer company. The company’s highly technical workforce developed computer solutions for clients, which often necessitated work–family trade-offs due to work demands, and management desired to implement programs so as to better understand what could be done to curtail exhaustion among its employees. While norms existed for working traditional business hours (e.g., 8:30 am–5:30 pm), employees desired greater freedom in where and when they worked. The telework program was established at the company as a means to address these employee quality of life concerns ~15 months before the survey was administered. Technology professionals offer an ideal sample since the location of their work is often flexible enabling them to participate in telework programs, and this group of workers is often prone to exhaustion since they typically face excessive job demands which force coping with significant work–family conflict (Moore 2000a). The sample was constructed based on records provided by the company for individuals who teleworked. A senior manager sent participants an email inviting them to take part in the study. The email contained a link to a web-based survey. Participation was voluntary and respondents were assured confidentiality. Requests for participation were sent to 800 salaried technical professionals in the firm, with 316 completed surveys received (39% response rate). Typical of many technology professionals, respondents were mostly male (71%), nearly 30 years of age on average, and had an average tenure of 28 months (SD = 16 months). On average there were 2.45 children living at home with the respondent, and most of the respondents had children (88%).

Measures

Work–Family Conflict (WFC) and Family–Work Conflict (FWC)

Conflict between work and family was assessed using the bidirectional measure developed by Carlson et al. (2000) with time-based and strain-based subcomponents. This measure has been shown to capture the time and psychological strain aspects of conflict and has been widely used with sound psychometric properties (e.g., O’Driscoll et al. 2003). The measure contains twelve items using a 5-point likert scale; six items which assess WFC (three for time-based WFC, and three for strain-based WFC), and six items which assess FWC (three time-based FWC, and three strain-based FWC). Example WFC items include “My work keeps me from my family activities more than I would like” (time), and “Due to all the pressures at work, sometimes when I am home I am too stressed to do the things I enjoy”. Example FWC items include “The time I spend on family responsibilities often interfere with my work responsibilities” (time), and “Because I am often stressed from family responsibilities, I have a hard time concentrating on my work” (strain). Items were averaged to form overall scores (time-based WFC, α = .94; strain-based WFC, α = .93; time-based FWC, α = .89; strain-based FWC, α = .95).

In order to ensure this measure of conflict between work and family had sound properties in the sample, additional evidence was sought of discriminant validity by conducting a series of confirmatory factor analyses (CFAs). The proposed four-factor model was examined, in which each factor corresponded to items assessing either time or strain components of WFC and FWC, and compared to one and two factor models representing the more generalized forms of conflict. Results indicated that the proposed four-factor model [χ2 (48, N = 316) = 132.2 p < .000, Comparative Fit Index (CFI) = .98, Incremental Fit Index (IFI) = .98, Goodness of Fit Index (GFI) = .94, Tucker–Lewis Index (TLI) = .97, Root Mean Square Error of Approximation (RMSEA) = .07, Akaike Information Criterion (AIC) = 192, Expected Cross Validation Index (ECVI) = 1.2, Normed Fit Index (NFI) = .97] was superior to a two-factor model representing the more generalized form of work–family and family–work conflict [χ2 (53, N = 316) = 808.6 p < .000, CFI = .82, IFI = .83, GFI = .66, TLI = .78, RMSEA = .21, AIC = 858, ECVI = 2.7, NFI = .81], and to a general one-factor work and family conflict model [χ2 (54, N = 316) = 1501.1 p < .000, CFI = .66, IFI = .66, GFI = .49, TLI = .59, RMSEA = .29, AIC = 1549, ECVI = 4.9, NFI = .65]. It was also superior to a two factor model comparing time-based and strain-based dimensions collapsing WFC and FWC [χ2 (53, N = 316) = 1274.88 p < .000, CFI = .72, IFI = .72, GFI = .53, TLI = .65, RMSEA = .27, AIC = 1324, ECVI = 4.2, NFI = .72]. Thus, given the significant χ2 differences (p < .001) and the better fit of the proposed four-factor model to the other models, it was concluded that proposed WFC and FWC (time- and strain-based) are best captured as separate constructs.

Work Exhaustion

Work exhaustion was measured via the 5-item scale used by Moore (2000a) based on the work exhaustion subscale of the General Burnout Questionnaire (Schaufeli et al. 1995). This measure was derived from the Maslach and Jackson (1981) exhaustion scale adopted to reflect general exhaustion in the work context. Items were rated on a 7-point Likert scale (1 = strongly disagree; 7 = strongly agree). An example item includes “I feel emotionally drained from my work”. Items were averaged to form an overall score (α = .91).

Telework During Traditional Hours

The extent to which employees engaged in traditional telework was assessed by asking participants to indicate the proportion of hours in an average workweek they spent working from home during typical work hours. This measurement approach has been used in a number of other studies (e.g., Golden 2006; Raghuram et al. 2003) and captures the proportion of time employees spend during typical hours teleworking instead of working from their main worksite (e.g., corporate office). Responses ranged from 3–100%, representing the full range of time spent available to telework.

Telework During Nontraditional Hours

The extent to which employees engaged in telework was measured by asking participants to indicate the proportion of hours in an average workweek they spent working from home during the evenings or weekends rather than during the typical business day. Similar methods have been used by others (e.g., Duxbury et al. 1996; Venkatesh and Vitalari 1992), which capture the proportion of time spent working during periods of the day that do not fall within the organization’s typical working hours. Responses ranged from 2 to 60%, indicating the varying proportion of time they spent during the week working outside of typical work hours.

Control Variables

Based on prior research and to mitigate spurious effects, age (years), biological sex (1 = male, 2 = female), number of children, tenure (months), and total work hours were controlled in all analyses. While the sample is typical of the largely male technology-based workforce, biological sex, age, and number of children were controlled since they may influence levels of work family conflict and the experiences of teleworkers (Bailey and Kurland 2002; Belanger 1999; Parasuraman and Greenhaus 2002). Total work hours were controlled in order to eliminate its potential influence on work exhaustion. Similarly, tenure was controlled since there is some indication that the length of time spent working in the organization may change telework outcomes (Ramsower 1983).

Confirmatory Factor Analysis

To examine the distinctiveness of the scales a CFA was also conducted on the measures of work exhaustion, time-based WFC, strain-based WFC, time-based FWC, and strain-based FWC. The fit of the proposed five factor model was empirically tested to examine whether it fit the data better than did competing models (Kelloway 1998). As anticipated the five factor model indicated an adequate fit with the data, [χ2 (109, N = 316) = 267.3, p < .001, CFI = .97, IFI = .97, GFI = .91, TLI = .96, RMSEA = .06, AIC = 355, ECVI = 1.1, NFI = .95]. In contrast, other conceivable models with fewer factors did not exhibit adequate fit statistics. For instance, a two factor model that included work exhaustion and combined time-based WFC, strain-based WFC, time-based FWC, and strain-based FWC into one factor exhibited a poor fit, [χ2 (118, N = 316) = 1664.6, p < .001, CFI = .72, IFI = .72, GFI = .55, TLI = .67, RMSEA = .20, AIC = 1734, ECVI = 5.5, NFI = .70], as did a two factor model where exhaustion and strain-based WFC/FWC was a single factor [χ2 (118, N = 316) = 2523.0 p < .000, CFI = .56, IFI = .56, GFI = .42, TLI = .49, RMSEA = .25, AIC = 2593, ECVI = 5.2, NFI = .55]. A one-factor model exhibited a similarly poor fit, [χ2 (119, N = 316) = 2718.6, p < .001, CFI = .52, IFI = .53, GFI = .41, TLI = .46, RMSEA = .26, AIC = 2786.6, ECVI = 8.8, NFI = .52]. These analyses provided further support that the measures are empirically distinguishable.

Results

Descriptive statistics including the means, standard deviations, and correlations for variables are shown in Table 1. Hierarchical regression analyses were used to test the hypotheses, and following accepted procedures for calculating interaction effects, variables were centered (Cohen et al. 2003). Consistent with scholars studying flexible work (e.g., Shockley and Allen 2007) regression were analyzed for the forms of work–family conflict, with control variables entered in the first step followed by the main effects of time- and strain-based WFC and FWC in step 2. In step 3, direct effects were entered. Then in step 4, the focal 2-way interaction terms were entered into the equation. In step 5, the two-way interaction (traditional telework x nontraditional telework) was entered. Finally, the three-way interaction terms were entered in step 6 of the regression equation. The incremental variance explained in the final step represents an estimate of the size of the three-way interaction when predicting work exhaustion.
Table 1

Descriptive statistics and correlations among variables

Variables

Mean

SD

1

2

3

4

5

6

7

8

9

10

11

1. Age

29.67

7.61

          

2. Sex

1.29

0.45

.09

         

3. Children living at home

2.45

2.08

−.01

−.03

        

4. Tenure

28.29

16.70

.33**

−.08

−.04

       

5. Hours worked per week

54.71

9.12

.04

.14*

.12*

.04

      

6. Work exhaustion

3.49

1.57

.02

.04

−.07

−.04

−.02

(.91)

     

7. Time-based WFC

3.50

1.30

.08

.13*

.36**

.06

.32**

.14*

(.94)

    

8. Strain-based WFC

3.15

1.46

.08

.15**

.48**

−.01

.19**

.16**

.56**

(.93)

   

9. Time-based FWC

2.67

1.31

.04

−.01

.39**

−.03

.00

−.02

.51**

.54**

(.89)

  

10. Strain-based FWC

2.11

1.37

.04

.03

.42**

−.04

.06

−.03

.47**

.60**

.58**

(.95)

 

11. Traditional telework

0.24

0.16

−.06

.09

.01

−.05

.17**

−.07

.02

−.13*

−.19**

−.15**

12. Nontraditional telework

0.17

0.09

−.05

.14*

−.04

−.04

.10

.01

.06

−.06

−.13*

−.14*

.33**

Notes: Age is in years; sex (1 = male; 2 = female); tenure is in years; traditional telework denotes the extent of telework conducted during traditional work hours; nontraditional telework denotes the extent of telework conducted during nontraditional work hours

p < .05, ** p < .01

Hypothesis 1 predicted positive relationships between time-based and strain-based WFC and FWC with exhaustion. As shown in Table 2, this hypothesis was supported for time-based WFC (H1a: β = .30, p < .001) and strain-based WFC (H1b: β = .36, p < .001), but not for time-based FWC (H1c: β = .04, p > .05) or strain-based FWC (H1d: β = .07, p > .05). Higher work–family conflict was associated with higher work exhaustion (time and strain-based), whereas family–work conflict was not, providing partial support for the hypothesis.
Table 2

Hierarchical regression analysis

Hypotheses

H1a, 2a, 3a, 4a

H1b, 2b, 3b, 4b

H1c, 2c, 3c, 4c

H1d, 2d, 3d, 4d

β

R2

β

R2

β

R2

β

R2

Step 1

 

.01

 

.01

 

.01

 

.01

Age

.02

 

.03

 

.04

 

.04

 

Sex

.02

 

−.01

 

.05

 

.05

 

Children

−.11

 

−.13*

 

−.04

 

−.04

 

Tenure

−.06

 

−.07

 

−.06

 

−.06

 

Hours worked

−.06

 

−.03

 

−.01

 

.00

 

Step 2

 

.04***

 

.05***

 

.00

 

.00

Time-based WFC

.30***

       

Strain-based WFC

  

.36***

     

Time-based FWC

    

.04

   

Strain-based FWC

      

.07

 

Step 3

 

.01

 

.01

 

.01

 

.01

Extent of traditional telework

−.07

 

−.02

 

−.09

 

−.08

 

Extent of nontraditional telework

−.02

 

−.01

 

.03

 

.00

 

Step 4

 

.04***

 

.06***

 

.01

 

.01

Traditional telework × time-based WFC

.16*

       

Nontraditional telework × time-based WFC

.13*

       

Traditional telework × strain-based WFC

  

.24***

     

Nontraditional telework × strain-based WFC

  

.14*

     

Traditional telework × time-based FWC

    

.07

   

Nontraditional telework × time-based FWC

    

.18*

   

Traditional telework × strain-based FWC

      

.15

 

Nontraditional telework × strain-based FWC

      

.01

 

Step 5

 

.01

 

.01

 

.01

 

.01

Traditional telework × nontraditional telework

−.02

 

−.01

 

.04

 

−.01

 

Step 6

 

.00

 

.00

 

.01

 

.00

Time WFC × traditional × nontraditional tele

.03

       

Strain WFC × traditional × nontraditional tele

  

.07

     

Time FWC × traditional × nontraditional tele

    

.13

   

Strain FWC × traditional × nontraditional tele

      

.01

 

Final F

2.51***

 

3.53***

 

.88

 

.68

 

Values represent standardized regression coefficients from the final equation

p < .05, **  p < .01, *** p < .001

To assess moderation of the relationships between work–family and family–work conflict with exhaustion, variables were entered into the regression models following procedures outlined by Baron and Kenny (1986). Moderators were entered as a block into the regression analysis because doing so takes into account their simultaneous effects, and therefore represents a more conservative and robust approach (Kohler and Mathieu 1993). As shown in Table 2, Hypothesis 2, which predicted that the extent of traditional telework would moderate the relationship between time-based and strain-based WFC and FWC with exhaustion, was partially supported. Telework during traditional hours moderated the relationship between time-based WFC (H2a: β = .16, p < .05) and strain-based WFC (H2b: β = .24, p < .001), but not time-based FWC (H2c: β = .07, p > .05) or strain-based FWC (H2d: β = .15, p > .05). To further interpret the interaction effects, I followed the procedures recommended by Cohen et al. (2003) by creating two simple regressions of WFC and FWC on exhaustion, given conditional values of the telework (mean ±1 SD). As shown in Fig. 1 (time-based WFC) and Fig. 2 (strain-based WFC), when telework is extensive, exhaustion increases at a faster rate, as expected. Individuals who had extensive telework tended to have lower levels of work exhaustion at low and moderate levels of time and strain-based WFC compared to those with more limited telework, whereas for those with high WFC higher work exhaustion was experienced.
https://static-content.springer.com/image/art%3A10.1007%2Fs10869-011-9247-0/MediaObjects/10869_2011_9247_Fig1_HTML.gif
Fig. 1

The moderating role of telework during traditional work hours on the time-based work-to-family conflict–exhaustion relationship

https://static-content.springer.com/image/art%3A10.1007%2Fs10869-011-9247-0/MediaObjects/10869_2011_9247_Fig2_HTML.gif
Fig. 2

The moderating role of telework during traditional work hours on the strain-based work-to-family conflict–exhaustion relationship

Hypothesis 3, which predicted the extent of nontraditional telework would moderate the relationship between time- and strain-based WFC and FWC with exhaustion, received partial support as well. Nontraditional telework moderated the relationship with time-based WFC (H3a: β = .13, p < .05) and strain-based WFC (H3b: β = .14, p < .05), but not time-based FWC (H3c: β = .18, p < .05) or strain-based FWC (H3d: β = .01, p > .05) given the lack of significant variance accounted for in the regressions. As shown in Fig. 3 (time-based WFC) and Fig. 4 (strain-based WFC), when nontraditional telework is extensive, exhaustion increases at a faster rate, as expected. Individuals who had extensive nontraditional telework tended to have lower work exhaustion at low levels of time and strain-based WFC compared to those with more limited nontraditional telework, whereas for those with high WFC higher work exhaustion was experienced.
https://static-content.springer.com/image/art%3A10.1007%2Fs10869-011-9247-0/MediaObjects/10869_2011_9247_Fig3_HTML.gif
Fig. 3

The moderating role of telework during nontraditional work hours on the time-based work-to-family conflict–exhaustion relationship

https://static-content.springer.com/image/art%3A10.1007%2Fs10869-011-9247-0/MediaObjects/10869_2011_9247_Fig4_HTML.gif
Fig. 4

The moderating role of telework during nontraditional work hours on the strain-based work-to-family conflict–exhaustion relationship

Hypothesis 4, which predicted three-way interactions between traditional telework, nontraditional telework, and time-based or strain-based WFC/FWC in predicting exhaustion, was not supported. Traditional telework’s interaction with nontraditional telework did not moderate the relationship between time-based WFC (H4a: β = .03, p > .05) or strain-based WFC (H4b: β = .07, p > .05), nor did it moderate the relationship between time-based FWC (H4c: β = .13, p > .05) or strain-based FWC (H4d: β = .01, p < .05) with exhaustion.

Discussion

This study examined the extent to which time-based and strain-based WFC and FWC impacts work exhaustion, and if the extent of telework during traditional and non-traditional work hours played an influential role. By building on earlier telework research, this investigation extends the study of work and family conflict’s consequences to teleworkers, by empirically examining if time- and strain-based WFC and FWC impacts exhaustion among professional workers who telework. As anticipated, time- and strain-based WFC was found to have a positive association with exhaustion, suggesting that individuals who have high time-based and strain-based WFC experience greater depletion of resources that lead to higher levels of exhaustion. Counter to expectations, FWC was not associated with higher levels of exhaustion, considering both time and strain-based dimensions. This suggests that conflict due to family interference with work may have different effects on exhaustion, perhaps reflecting the asymmetrically permeable boundaries which scholars have suggested between work and family (Ashforth et al. 2000; Nippert-Eng 1996b). These results reinforce the bidirectional nature of conflict between work and family found by other researchers (e.g., Frone 2000). Whereas individuals may adopt telework as a means to enhance their quality of life (Bailey and Kurland 2002), given higher levels of WFC, higher levels of exhaustion are likely to ensue.

More importantly, by encompassing two previously undifferentiated types of telework and their role influencing the WFC/FWC—exhaustion relationship, this study delves deeper into the contingent nature of telework and suggests that the extent of telework during traditional and non-traditional work hours are highly influential on work exhaustion. More extensive traditional telework appears to moderate the impact of time-based and strain-based WFC on work exhaustion, such that more exhaustion is associated with higher levels of WFC and this is made worse when individuals telework extensively. As shown in Fig. 1 (time-based WFC) and Fig. 2 (strain-based WFC), in comparison to individuals with limited telework, those with extensive telework who had high WFC tended to have higher levels of work exhaustion. For those with high WFC, it may be that while working extensively as a teleworker could alleviate strain due to reduced commutes (Guimaraes and Dallow 1999), other factors in these situations may become more dominant. For instance, increased home distractions, cognitive preoccupation with family-member needs, and a lack of reprieve due to the collocation of work and home may off-set any gains in resources (Ammons and Markham 2004; Kossek et al. 2006), thwarting any benefit in terms of exhaustion. Absent the physical and corresponding psychological separation to facilitate resource replenishment (Hobfoll 1989), teleworkers who have high WFC may suffer increased resource drain and exhaustion compared to those with less flexibility. Future research investigating detailed work and family domain segmentation (Ashforth et al. 2000; Nippert-Eng 1996b) among teleworkers with high WFC may shed additional insights into this possibility.

With respect to telework during nontraditional work hours and its contingent effects, results suggest that this form of telework moderates the impact of time-based and strain-based WFC on exhaustion, such that more exhaustion is associated with higher levels of WFC and this is made worse when individuals telework extensively during non-traditional hours. As shown in Fig. 3 (time-based WFC) and Fig. 4 (strain-based WFC), in comparison to individuals with limited nontraditional telework, those with extensive nontraditional telework who had high WFC tended to have higher levels of work exhaustion. As occurred with traditional telework, extensive nontraditional telework appears to exacerbate the rising impact of time and strain-based WFC on exhaustion. Perhaps the extensive nontraditional telework, as well as the cognitive preoccupation with WFC during this time (Kossek et al. 2006), serves to siphon off further personal resources and hinder energy replenishment (Hobfoll 1988, 1989). Since working extensively during the evening hours in the midst of other family routines and rhythms may prevent mentally distancing oneself from work (Etzion et al. 1998), extensive telework during nontraditional work hours appears to hinder more complete recuperation from the conflict that typically takes place during these hours (Fritz et al. 2010), thereby leading to greater resource drain and exhaustion. An additional factor not studied here may be the extent to which individuals maintain psychological distance from the continual irritants brought about by collocation of work and family domains, through means such as doors and other physical separation methods. Future research investigating such refinements may be especially insightful, and given the important role that flexibility plays in telework arrangements (e.g., Mokhtarian et al. 1995), the results discussed here highlight the need for further more detailed studies so as to understand nontraditional telework’s precise effects.

Given the apparently similar moderating role of traditional telework and nontraditional telework, it seems that individuals who work extensively in either of these forms of telework may be particularly susceptible to energy drain and exhaustion, since they lack the physical separation and associated cognitive reprieve offered by more separated work and family domains (Nippert-Eng 1996a, b; Olson-Buchanan and Boswell 2006). In other words, in these conditions it may be that the constant and salient physical reminders of work and home found with extensive presence at home (Ammons and Markham 2004; Goldsborough 2000; Mirchandani 1999) precludes individuals from greater psychological separation and the needed reprieves from the draining effects of high WFC and FWC (Greenhaus and Powell 2003), thereby depleting resources and leading to more work exhaustion. In this way a teleworker’s recovery from the draining effect of high time-based or strain-based WFC and FWC are apt to be of shorter duration and less fulfilling (Meijman and Mulder 1998), preventing adequate energy from being replenished and contributing to exhaustion (Hobfoll 1989). Future research investigating such lack of resource stockpiling or recovery opportunities that might otherwise counter future losses (Lee and Ashforth 1996; Wright and Cropanzano 1998) therefore seems especially promising.

Whereas the data in this study do not support the anticipated three-way interaction between traditional telework, nontraditional telework, and WFC/FWC, it may be that other more influential factors come into play when considering these interactions, or that they may simply cancel each other out (Kohler and Mathieu 1993). Given that this is an initial examination into more refined aspects of telework to include traditional telework and nontraditional telework, researchers need to understand if the pattern of results found here is typical of telework’s influential role. From a larger perspective, however, considering the widespread benefits of telework reported by researchers and practitioners such as job satisfaction, autonomy, and other quality of life benefits (Baruch 2000; Stephens and Szajna 1998), it may be that the exhaustion from WFC is off-set by other considerations in the mind of the teleworker. In other words, whereas for the highest levels of WFC extensive traditional telework and nontraditional telework may exacerbate exhaustion, this may be an acceptable trade-off for individuals seeking telework’s other benefits. For this sample, telework was a voluntary work mode, and as such these results may be a reflection of the experiences of those who choose to telework in response to WFC/FWC. Although these possibilities remain tentative pending further investigation, this study demonstrates the pressing need to understand more completely the many differential impacts of telework. Understanding such complex interactions and trade-offs may be key to revealing future insights into the nature of telework and its contingent effects.

Given that results presented here cannot infer causality and the influential role of telework is likely to remain elusive without inferential capability, future investigations into telework may benefit from studies whereby conditions can be experimentally manipulated so that causality can be more clearly determined. While the hypothesized interactions demonstrate the importance of understanding types of telework rather than considering telework as one omnibus variable, understanding interactions are clearly fundamental to future research and investigating them in a longitudinal manner may prove particularly insightful. Such longitudinal investigations would also help untangle telework’s potential association with work–family conflict. Although a recent meta-analysis did not find support for a relationship between flexible work arrangements like telework and WFC or FWC (Mesmer-Magnus and Viswesvaran 2006), other studies have (Gajendran and Harrison 2007; Golden et al. 2006). Results from supplementary analyses mirrored these equivocal findings. For example, telework was not associated with WFC whereas it was associated with FWC for traditional telework but not nontraditional telework. These analyses and prior findings in the literature highlight the need for future longitudinal studies and perhaps even experience sampling methods to explicitly test associations and resource drains suggested earlier.

In terms of other limitations, the data reported here is self-reported, so that future studies would benefit from additional data sources that may be compared to self-report data. Moreover, participants in this study were all technology professionals, and it may be that these results are not entirely applicable to employees in other fields or industries, so care should be taken when attempting to generalize these results to other groups of individuals. In addition, given that 88% of participants in this sample had children with an average of 2.45 children at home, results obtained here may not be applicable to other groups of individuals with less substantial child-rearing activities in the home. Finally, given that WFC and FWC are related to exhaustion, and exhausted employees are more likely to suffer performance decrements or leave the organization (e.g., Carson et al. 2010; Cropanzano et al. 2003; Demerouti and Bakker 2006; Maslach et al. 2001; Moore 2000a; Parker and Kulik 1995; Shirom 2011; Taris 2006), it is critical that future researchers investigate other aspects of work that could reduce or contain rising levels of exhaustion. Such findings would have direct implications for flexibility policies (e.g., Kossek et al. 2006) and other employee work-life programs, with far reaching consequences.

Managers and professionals seeking to make decisions about this work modality may find the results presented here particularly intriguing, as they seek out avenues with which to contain the impacts of WFC and FWC experienced by their employees. In particular, this research suggests that teleworkers with higher WFC and FWC tend to experience more exhaustion than those with less conflict, and the impact on work exhaustion depends on the extent of telework conducted during traditional and non-traditional work hours. Results suggest extensive traditional telework and nontraditional telework may exacerbate exhaustion for individuals with high time and strain-based WFC, whereas exhaustion may be lower for teleworkers with extensive telework who have low WFC. Given that the influence of traditional telework and nontraditional telework on exhaustion was similar among this group of voluntary teleworkers, managers may have some leeway in allowing employees to telework according to individual employee preferences, and potentially permit subordinates to flexibly employ telework to meet their own individual desires. Results also suggest that telework may offer a means with which to reduce exhaustion levels, at least for individuals without high levels of time- or strain-based WFC. Extensive traditional telework and nontraditional telework therefore seem to be beneficial in terms of exhaustion for those with moderate levels of WFC, suggesting that employees without high WFC may be able to reap some resource replenishment from extensive flexibility.

Finally, the results suggest that managers need to be especially aware of making individual telework decisions without considering the full range of characteristics which encapsulate the teleworker’s work practices. For those with high WFC, although extensive telework during traditional and non-traditional work hours appear to generate more exhaustion than lower levels of telework, this does not occur for individuals with more moderate levels of conflict between work and family. As these complex interactions illustrate, managers may need to be cognizant of the home situations of their employees, in terms of the degree of WFC and FWC, when making telework decisions if they are to maximize telework’s benefits. It may also be that telework’s flexibility reduces the impact of selected stress and strains but also introduces new ones. Although such interpretations remain tentative pending further research, the growth and continued importance of telework in many work organizations requires continued researcher-practitioner collaboration to understand this expanding form of virtual work.

Copyright information

© Springer Science+Business Media, LLC 2011