Eldercare Demands and Time Theft: Integrating Family-to-Work Conflict and Spillover–Crossover Perspectives


Adapting the spillover–crossover model to the context of eldercare, we employed a 5-week weekly diary method in a sample of 82 Chinese dual-earner heterosexual couples to examine the relationships between family eldercare demands, family-to-work conflict, and time theft. Results from multilevel path modeling analyses found that family eldercare demands (as a shared/common stressor) were positively related to each partner’s family-to-work conflict and that the weekly family-to-work conflict of each partner was positively related. Furthermore, family eldercare demands were positively related to each partner’s time theft at the week level, and this relationship was mediated by weekly family-to-work conflict. Thus, this is one of the first studies to show that family eldercare demands can explain the interindividual crossover of family-to-work conflict between partners at the week level and this can spill over intraindividually to impact time theft at work. These findings enhance the field’s understanding of the process by which eldercare demands relate to time theft among dual-earner couples within a focused temporal framework.

The aging of the world’s population brings numerous challenges to the worldwide workforce (World Health Organization, 2015), and one of the most significant challenges is the potential demands and burdens that eldercare responsibilities place on workers. Eldercare generally refers to care or assistance provided to an elder family member or friend (Gorey, Rice, & Brice, 1992). Specific to the USA, nearly one in five workers acts as a caregiver for an elderly spouse, family member, or friend (AARP & The National Alliance for Caregiving, 2015), and almost 60% of working caregivers reported that work is impacted in some way by their eldercare responsibilities (Fortinsky, 2011) because of the excessive demands generally associated with the provision of eldercare (Barrah, Shultz, Baltes, & Stolz, 2004). Given the time-consuming nature of eldercare, the consequent time loss becomes a major concern for organizations (AARP & The National Alliance for Caregiving, 2015). Time theft refers to “time that employees waste or spend not working during their scheduled work hours” (p. 53; Henle, Reeve, & Pitts, 2010). Estimates of time theft among U.S. workers suggest that it is a major cost for organizations (McGee & Fillon, 1995).

In light of the increasing prevalence of eldercare responsibilities, and the costs to organizations related to time theft (e.g., Gorey et al., 1992; Martin, Brock, Buckley, & Ketchen, 2010), we suggest that the field would significantly benefit by developing a more coherent understanding of potential mechanisms linking the two concepts within a theoretically derived temporally grounded framework. To be clear, we are certainly not equating time theft due to eldercare responsibilities with more deviant forms of time theft (e.g., falsifying time reports; Martin et al., 2010). Rather, we are suggesting that, unfortunately, to be able to manage the excessive demands often associated with eldercare (Barrah et al., 2004), workers may find it necessary to “borrow” time that might otherwise normally be allocated to the work domain (Matthews, Winkel, & Wayne, 2014).

Because home is the shared life domain for husbands and wives, life events at home, such as caregiving for an elderly family member, are likely to simultaneously influence both partners (Westman, 2001). Arguably then, researchers should therefore examine the mutual influence and experience of both partners to understand how the family, as a system, reacts to eldercare demands. That is, for example, do family eldercare demands cross over between partners? If yes, will this further impact partner’s work outcomes and how? It has been argued that examining the broader family system by using the couple as the unit of analysis can enhance our understanding of the reciprocal influences and experiences of both members of dual-earner couples in the family and work domains (Hammer, Neal, Newsom, Brockwood, & Colton, 2005; Hammer, Bauer, & Grandey, 2003). As such, we apply the spillover–crossover model (SCM; Bakker & Demerouti, 2013) and use the couple as the unit of analysis (Hammer, Allen, & Grigsby, 1997; Hammer et al., 2003) to examine how family eldercare demands crossover within dual-earner couples and relate to employees’ time theft (Fig. 1).

Fig. 1

Theoretical model of eldercare and time theft among dual-earner couples

Additionally, given that most existing studies have been conducted using Western samples, scholars have called for research examining eldercare demands in countries with different cultural values and social policies (e.g., Zacher & Winter, 2011). Thus, we examine our model in a group of dual-earner couples from China, an East Asian country with Confucian values where eldercare rests primarily with the family. Because of its highest percentage (more than 90%) of dual-earner couples (Lu, Lu, Du, & Brough, 2016) and largest aging population in the world (Chen & Liu, 2009), China is an ideal setting to test the spillover and crossover of eldercare demands among dual-earner couples especially in light of the fact that at a societal level there is little or no employer or public support for eldercare (Pei, Luo, Lin, Keating, & Fast, 2017).

Our primary contributions to the field are twofold. First, we apply a within-person approach to eldercare. Past studies have generally applied cross-sectional between-person designs (e.g., Shoptaugh, Phelps, & Visio, 2004; Zacher, Jimmieson, & Winter, 2012) which are well suited to understanding how individuals differ from one another on their general levels of eldercare demands. Yet, such designs are limited in their ability to capture the fluctuations in eldercare demands, and how these fluctuations impact family and home experiences within a focused temporal framework in daily life. While existing research typically adopts a between-person view of eldercare demands, highlighting its average (relatively stable) influence, we adopt the view that eldercare demands are not static and may fluctuate over time. In the current study, we apply a diary methodology to study eldercare demands at the week level, an appropriate temporal frame (Matthews, Wayne, & Ford, 2014) that can enable researchers to understand psychological phenomenon at the appropriate time and level they are manifested (e.g., Ilies, Schwind, & Heller, 2007). Considering the factors (e.g., frequency, severity, exposure) that are related to predictors (i.e., caregiving; Matthews et al., 2014), studying family eldercare demands in a shorter time frame (i.e., week) can help explain within-person differences (i.e., why a person’s time theft may differ from one week to another). This design also allows us to test the impact of eldercare demands on each partner’s work outcomes with a short timeframe, supplementing findings from previous cross-sectional studies (Laurenceau & Bolger, 2005). The weekly diary method has several other advantages, including decreasing retrospective reporting bias (Ohly, Sonnentag, Niessen, & Zapf, 2010; Song et al., 2011) and social desirability by asking participants to focus on discrete behaviors and experiences in a weekly context instead of answering questions about typical behaviors and experiences (Bolger, Davis, & Rafaeli, 2003).

Another contribution of our study lies in furthering our understanding of eldercare by examining how family eldercare demands relate to the stress crossover within dual-earner couples, and how it relates to both partners’ time theft at work. Adapting the SCM framework, we explore the mutual effects between partners and test how family eldercare demands affect not just a given individual but also partner’s home and work experiences. More specifically, we test family eldercare demands as a stressor that is common to marital partners (i.e., simultaneously experienced by partners) and that relates to the crossover of family-to-work conflict between partners. We also test the mediating role of family-to-work conflict (as an explanatory variable) in the relationship between family eldercare demands and time theft. Recognizing and testing such explanatory mechanisms is a key issue of concern within the larger work–family literature (Matthews et al., 2014). Interestingly though, family-to-work conflict has received little attention in eldercare literature despite its obvious relevance (e.g., Barrah et al., 2004). Thus, we shed light on the role of family-to-work conflict in the context of eldercare. Findings from this study not only deepen our understanding of the process by which eldercare demands affect work outcomes but may also be used to guide interventions designed to help employees balance work and eldercare responsibilities, an issue of increasing societal concern (AARP & The National Alliance for Caregiving, 2015).

Theoretical Background and Development of Hypotheses

The SCM (Bakker & Demerouti, 2013) suggests that there are two critical processes through which experiences in one domain (i.e., work) spill over to another domain (i.e., family) and cross over to another person. Spillover refers to an intraindividual transmission of stressors or strain where one’s experience in one domain is transferred into another domain and affects his/her experience in that domain (Bolger, DeLongis, Kessler, & Wethington, 1989). For instance, stress at work could lead to lower marital satisfaction at home. Crossover represents an interindividual transmission where a person’s experience in one domain could transfer to another person (Bolger et al., 1989; Westman, 2001). For instance, stress experienced by one person may increase the stress experienced by his or her marital partner. The SCM suggests a process whereby one’s experience in one domain (e.g., work) spill over to his/her experiences in the other domain (e.g., home) and cross over to other role partners within that domain (e.g., partner at home) (Bakker & Demerouti, 2013). However, in the present study, we first examine the crossover process between partners due to the major focus on eldercare demands at home, and as such, it represents a meaningful extension to the SCM.

Westman (2001) summarized three types of crossover models: shared factors through common stressors, direct crossover, and crossover through a mediator (e.g., social support and communication). The literature suggests that these mechanisms are not mutually exclusive and can operate simultaneously (e.g., Song et al., 2011; Westman & Etzion, 2005). Considering the purpose of the present study, we mainly focus on shared factors through a common stressor (i.e., weekly family eldercare demands) and direct crossover (i.e., weekly family-to-work conflict).

Shared Family Eldercare Demands and Direct Crossover of Family-to-Work Conflict

The common stressor mechanism of crossover suggests concurrent demands in a shared domain (Westman & Vinokur, 1998). For example, problems with other family members (e.g., a disabled parent) can contribute to the synchronization of stressful experiences between spouses. Given that family is a shared life domain for married couples, family eldercare represents a critical source of concurrent demands for both partners. Westman and Vinokur (1998) found that common life events for both partners influenced the crossover process by increasing each partner’s depression. As a common stressor at home, the shared family eldercare demands are likely to simultaneously impact both partners (Hobfoll & London, 1986; Westman, 2001; Westman & Etzion, 1995). Adapting the SCM framework, it is likely that demands at home (i.e., family eldercare) may transmit to the work domain in the form of family-to-work conflict. For instance, married couples who are caring for a parent who recently became disabled may feel exhausted at home and may continue to be influenced by these family demands when at work (e.g., family-to-work conflict).

This is also consistent with the conservation of resources (COR) theory (Hobfoll, 1989); as individual’s resources are fixed, couples with high family eldercare demands could perceive fewer resources (e.g., time, energy) to devote to their work due to the caregiving activities and thus perceive increased family-to-work conflict (Dugan, Matthews, & Barnes-Farrell, 2012). Conflict between home demands (e.g., eldercare) and work demands usually makes meeting the demands of the second role more difficult (Frone, Russell, & Cooper, 1992). Family eldercare demands, as a shared stressor at home, may deplete both partners’ resources and thus contribute to the experience of family-to-work conflict for both partners. Prior research has indeed found that family-related characteristics (or couple-level factors), such as family cohesion (Huffman, Matthews, & Irving, 2017) and family hassles (Song et al., 2011), could have a significant impact on both partners.

Furthermore, since the competing demands from home and work domains are stressful (e.g., Hammer et al., 2003; Matthews et al., 2014), these stressful experiences (i.e., family-to-work conflict) are likely to cross over between partners because they are likely to be aware of and influenced by each other due to the considerable time spent together (Westman, 2001). Given the physical and relational proximity between partners, crossover may occur via the communication and interaction between partners (e.g., coping, lack of social support; Bakker & Demerouti, 2013). Additionally, since both partners are living in the same home environment and are therefore exposed to a number of common stressors (e.g., the aforementioned family eldercare), they tend to experience similar stressful feelings. For instance, when the family elder caregiving requires a substantial amount of time and energy, both partners are likely to experience family-to-work conflict.

Empirically, crossover can be indicated by a positive correlation between the stress reported by two individuals in a dyadic relationship (Song et al., 2011). There is ample evidence showing the crossover of stress or strain within the home domain (e.g., Bakker, Demerouti, & Dollard, 2008; Westman & Etzion, 1995; Sanz-Vergel, Rodríguez-Muñoz, & Nielsen, 2015). For instance, Westman and Etzion (2005) found reports of work-to-family conflict of both partners to be significantly correlated. However, it is surprising that research has not yet examined the crossover of family-to-work conflict between partners that are facing family eldercare demands. On the basis of theoretical reasoning and the findings described above, we propose the following hypotheses for the crossover effect among couples with eldercare responsibilities:

Hypothesis 1: Weekly family eldercare demands are positively related to each partner’s weekly family-to-work conflict.

Hypothesis 2: Weekly family-to-work conflict of each partner is positively related.

Weekly Family Eldercare Demands and Time Theft

Because time is an organizational asset, the misuse of time can be as problematic as the misuse of any other asset (Martin et al., 2010). Examples of time theft include taking extra or longer than acceptable breaks, or daydreaming. These activities all represent instances where time is being misused or stolen from the organization, which violates organizational norms and may threaten the organization (Bennett & Robinson, 2000). Some scholars (e.g., Ketchen, Craighead, & Buckley, 2008; Martin et al., 2010) have discussed time theft as a subdimension of counterproductive work behavior (or deviant behavior). Consistent with this, researchers have assessed time theft using relevant items from established workplace deviance measures (Henle et al., 2010; Lorinkova & Perry, 2017). Brock, Martin, and Buckley (2013) have discussed three types of time theft behaviors: general or classic time banditry, technology-related time banditry, and socially oriented time banditry. The first type of time theft behaviors includes taking excessively lengthy breaks or leaving early; the second type involves behaviors that are related to computer abuse or misuse of technology during the work hours; the third type consists of social behaviors that misused work time such as lengthy conversations with coworkers about nonwork matters during work hours. Given that the current study examines the relationship between eldercare demands and employees’ general use of time at work, the time theft behaviors that we focus on fall largely under the general or classic category. Again, though, it should be noted that unlike production deviance (a subtype of deviance), time theft behavior may be unintentional and in many cases is not done to harm the organization (Brock et al., 2013).

Research on counterproductive work behavior suggests that employees typically engage in deviance either as a reaction to a stressful experience (i.e., hostile motives) or as a means to an end (i.e., instrumental motives). Studies focusing on the hostile motives for performing deviant behaviors usually conceptualize deviance as revenge or anger-induced retaliation (e.g., Lorinkova & Perry, 2017). Research on time theft usually does not focus on negative motivations for engaging in the misuse of time at work (Martin et al., 2010). Instead, scholars have argued that time theft could be simply driven by self-interest (Ketchen et al., 2008). This is consistent with past research showing that people may engage in deviance as a way to retain existing resources and reduce further resource depletion in response to stress (i.e., the instrumental motives; Krischer, Penney, & Hunter, 2010). For instance, leaving early from work allows individuals to escape from stress and to protect remaining resources or potentially to rebuild resources (Matthews et al., 2014).

Individuals facing family eldercare demands may engage in time theft at work because doing so allows them to have more time and energy to meet their eldercare demands. This proposition is consistent with the SCM, which suggests that one’s experience of stressors or strain at home (e.g., eldercare demands) could impact his/her work experiences. Similarly, COR theory suggests that because individuals’ personal resources (e.g., time, energy) are limited, they strive to protect those limited resources, are careful not to use up their remaining resources, and often take steps to replenish resources for the future. Within the COR theory (Hobfoll, 1989), and at the weekly level, family eldercare demands require employees to devote more resources (e.g., time, energy) to the home domain, leaving them with fewer resources to devote to the work domain. Employees may thus engage in time theft to conserve their remaining resources and recover their own personal resources.

Hypothesis 3: Weekly family eldercare demands are positively related to each partner’s weekly time theft.

Family-to-Work Conflict as a Mediator

Demands at home (i.e., eldercare) may impact the work domain in the form of family-to-work conflict, which in turn, may influence employees’ functioning at work. Conflict between caregiver and work roles usually makes meeting the demands of the second role more difficult (Frone et al., 1992). Such conflict could further impact employed caregivers’ well-being and work outcomes. For instance, employees who are confronted with high eldercare demands may feel exhausted at home and may continue to be influenced by these family demands while at work. Research has shown that family-related demands lead to family-to-work conflict, which often influences work-related outcomes (e.g., Bakker et al., 2008; Dugan et al., 2012). This is also consistent with the COR theory. Because an individual’s resources are fixed, employees with high eldercare demands could perceive fewer resources (e.g., time, energy) to devote to their work due to the caregiving at home and family-to-work conflict (Dugan et al., 2012). Furthermore, people facing family-to-work conflict strive to protect their remaining resources and to rebuild resources for the future. Employees may engage in various forms of time theft such as arriving late or leaving early from work, or even daydreaming at work, for the purpose of conserving their remaining resources and preventing resources being further consumed by their work (Krischer et al., 2010). Thus, time theft may provide employees with more resources that can help in managing the stress of family-to-work conflict.

Indirect evidence for the relationship between family-to-work conflict and time theft has been found in between-person level survey studies. For example, Barling, MacEwen, Kelloway, and Higginbottom (1994) found that work and eldercare conflict was positively related to partial absenteeism. Additionally, family-to-work conflict has been found to be related to both work withdrawal (Hammer et al., 2003) and production deviance (Ferguson, Carlson, Hunter, & Whitten, 2012). However, prior research has yet to examine the relationship at the within-person level. Furthermore, they did not examine the possibility that family-to-work conflict may be a consequence of demands at home such as eldercare demands. To address this gap, we examine whether the relationship between family-to-work conflict and time theft may be the result of a spillover from home to work. We propose that weekly eldercare demands at home will positively relate to time theft at work through the mediating effect of weekly family-to-work conflict.

Hypothesis 4: Weekly family-to-work conflict mediates the relationship between weekly family eldercare demands and each partner’s weekly time theft.


Participants and Procedure

Data collection involved two phases: first, an initial survey assessing demographics and job information and second, a weekly diary study (across 5 weeks—beginning 1 week after the initial survey). To be eligible to participate, respondents were required to be employed, have an eldercare recipient in their family (i.e., parents, relatives), and be married or cohabitating with a partner. Ten undergraduate students majoring in business administration at a university in Southeast China volunteered to be trained as research assistants to distribute surveys to potential participants. Research assistants first approached and introduced the study to potential respondents. If potential participants expressed an interest in this study and met the inclusion criteria, they were either handed paper surveys or sent an online survey link through a smartphone application (i.e., WeChat, a commonly used social media) by the research assistant. After instructing respondents how to participate, each couple was assigned a code which was used to match each couple’s responses. Partners were asked to independently complete the surveys. Each completed survey was rewarded with $1.5 (for each week’s survey). Research assistants directly reported to the third author, who met with them and discussed potential problems (if any) every week.

Of the 188 participants who completed the initial survey, a total of 90 couples were matched and participated in the weekly diary survey. Eight couples (n = 16) with incomplete responses were removed. The final sample consisted of 82 couples (total n = 164) who provided complete responses for at least 3 of the 5 weeks, which resulted in a final sample of 390 wife–husband dyads. All couples were married except one that was cohabitating. The average age of participants was 45.82 years old (SD = 5.59) and their average tenure in the current job was 16.32 years (SD = 10.50). Additionally, 61% of participants had a college degree or above, and 60.4% worked in nonsupervisory positions. Among the participants, 22.6% were government employees, 31.7% worked in government-owned companies and organizations, and 30.5% worked in private companies. The majority were caring for either their own parent(s) or parent(s)-in-law (94.5%), did not hire professional staff to provide eldercare service (93.9%), have spent at least 100 hours on elder caregiving in the past year (57.5%), half had no help from other relatives (50%), and 72.6% were not living with eldercare recipients in the same house/apartment. The average age of eldercare recipients was 74.6 years (SD = 5.9). Compared with other studies in the field of eldercare that are based on Chinese samples, the characteristics of our sample are similar to those studies in terms of caregivers’ age, marital status, and relationships with care recipients (e.g., Zhan, 2002).


Because measures used in the current study were originally developed in English, the first and third authors followed a commonly used translation–back translations procedure to ensure that all survey items were accurately translated into Chinese (Brislin, 1980). For all measures, respondents were instructed to consider the past week when responding.

Eldercare Demands

The original scale contained 22 items developed by Zarit, Orr, and Zarit (1985). The current study used the shortened 4-item version validated by Bédard et al. (2001), who reported a strong correlation (r = .83) between this 4-item measure and the full scale. Furthermore, compared to the full scale, correlations between the 4-item measure and activities of daily living and frequency of problem behaviors among care recipients were very similar (Bédard et al., 2001). This 4-item version has been widely used in aging and gerontology research (e.g., Trukeschitz, Schneider, Mühlmann, & Ponocny, 2012). Participants were instructed to report the extent to which they experienced each of the feelings in the past week on a scale from 1 (never) to 5 (nearly always). An example item is “Do you feel strained when you are around your eldercare recipient?”. Across the five measurement points, the average Cronbach’s alpha was .80 (range from .77 to .83) for all participants. As the home is the shared life domain for both partners, eldercare demands are likely to simultaneously influence both partners (Westman, 2001). Thus, we analyzed eldercare demands at the couple level and termed it as family eldercare demands. Furthermore, at the couple level, the average rwg(j) value was .87, and the ICC1 and ICC2 values were .68 and .81, respectively. Based on previous studies (e.g., Lance, Butts, & Michels, 2006; LeBreton & Senter, 2008), it is justifiable to aggregate each partner’s rating into a couple-level construct. In accordance with past research (e.g., Song et al., 2011), we averaged the scores of each partner’s rating to create the weekly family eldercare demands.

Family-to-Work Conflict

The 3-item subscale of family-to-work conflict by Grzywacz, Frone, Brewer, and Kovner (2006) was used. Participants were instructed to report how often they had experienced each of the situations in the past week on a scale from 1 (never) to 5 (nearly always). An example item is, “How often did your home-life interfere with your job or career?”. The average Cronbach’s alpha across weekly observations was .83 (range from .76 to .87) for all participants.

Time Theft

In accordance with previous studies (Henle et al., 2010; Lorinkova & Perry, 2017), we used 3 items from Bennett and Robinson (2000) and 3 items from the literature on workplace deviance (e.g., Dalal, Lam, Weiss, Welch, & Hulin, 2009). The content of these items is very similar to the content of the general or classic type of time theft behaviors discussed by Brock et al. (2013). Participants were instructed to report how often they had engaged in each of the behaviors during the last week on a scale ranging from 1 (never) to 5 (nearly always). Items include “Worked on a personal matter instead of working for your employer,” “Spent too much time fantasizing or daydreaming at the job,” “Took an additional or a longer break than is acceptable at your workplace,” “Left work earlier than is allowed at your workplace”, “Spent time on tasks unrelated to work”, and “Surfed the Internet to deal with personal matters at your workplace”. The average Cronbach’s alpha across 5 weeks was .83 (range from .76 to .88). The result of multilevel confirmatory factor analysis confirmed the one-factor model of this measure, χ2 = 60.02 (p < .001), df = 27, CFI = .96, TFI = .93, RMSEA = .04.

Preliminary Analyses

To assess the distinctiveness of the study variables, we performed a three-level confirmatory factor analysis (i.e., weekly data are nested within individuals, and individuals are nested within couples). The hypothesized three-factor model (i.e., eldercare demands, family-to-work conflict, and time theft) demonstrated good fit, χ2 = 362.82 (p < .001), df = 183, CFI = .92, TFI = .90, RMSEA = .04, and was better than either an alternative two-factor model (i.e., combine family-to-work conflict and time theft into one factor), χ2 = 843.97 (p < .001), df = 189, CFI = .72, TFI = .66, RMSEA = .07, Satorra–Bentler ∆χ2(6) = 1377.55 (p < .001), or an alternative one-factor model, χ2 = 1121.23 (p < .001), df = 192, CFI = .61, TFI = .52, RMSEA = .08, Satorra–Bentler ∆χ2(6) = 1065.41 (p < .001). Thus, the results supported construct distinctiveness.

Furthermore, based on a separate heterogeneous sample of 243 working U.S. employees with eldercare responsibilities from TurkPrime (Litman, Robinson, & Abberbock, 2017), we sought to better establish the construct validity of the abbreviated measures used in the current study. Our results found that the shortened measures were all significantly and positively related with the relevant and longer versions of the scales, suggesting that our short measures are valid measures of their respective constructs (more detailed information is reported in the Appendix).

Analytical Approach

Because our data points are nested in nature (weekly ratings nested within each couple), we used multilevel path modeling (Kaplan, 1998) with Mplus 7.2 to test the model presented in Fig. 1. This multilevel modeling method accommodates the hierarchical data structure with multiple observations nested within each couple (Raudenbush, Brennan, & Barnett, 1995). Furthermore, this method allowed us to simultaneously test multiple path coefficients explaining the crossover and spillover effects (Jöreskog, 1977).


Table 1 provides means, standard deviations, and correlations among all study variables. As reported, the pattern of within-couple relationships was in the expected direction. Furthermore, we calculated the ICC1 of each study variable to examine the amount of total variance explained by couple membership and if these variables had sufficient within-couple variations. The ICC1 was .68 for family eldercare demands; .43 and .45 for family-to-work conflict of husbands and wives, respectively; and .60 and .61 for time theft of husbands and wives, respectively. The results indicate that 62% of variance in family eldercare demands can be attributed to couple membership, and 32% of variance can be attributed to within-couple variation, that is, the weekly fluctuations. Similarly, a substantial amount of variance (39~57%) in family-to-work conflict and time theft of husbands and wives was found to fluctuate from week to week. According to LeBreton and Senter (2008), the value of ICC1 larger than .05 justifies the use of multilevel modeling. We thus conduct multilevel modeling analyses.

Table 1 Means, standard deviations, and correlations between the study variables (N = 82 dyads, N = 390 observations)

In the current study, family eldercare demands were shared by partners, and the crossover effect of family-to-work conflict between husbands and wives was also tested. Furthermore, given that our primary goal was to examine the within-couple relationships among family eldercare demands, family-to-work conflict, and time theft at the week level, we integrated the variables from husbands and wives into the same model and tested the path coefficients of both partners simultaneously. Specifically, we specified a two-level path model (i.e., weekly level and couple level) in which the within-couple and between-couple variances were estimated and partitioned.

Test of Hypotheses

The two-level path model including all the hypothesized parameters (Fig. 2) fits the data well, χ2 = 1.27 (p = .25), df = 2, CFI = .99, TFI = .99, RMSEA = .03. In support of Hypothesis 1, family eldercare demands were positively related to family-to-work conflict of husbands (γ = .39, SE = .06, p < .001) and the wives (γ = .47, SE = .06, p < .001). In support of Hypothesis 2, one’s own family-to-work conflict was positively related to his/her partner’s family-to-work conflict (γ = .17, SE = .03, p < .001).

Fig. 2

Estimated model of eldercare and time theft among dual-earner couples based on both original and post hoc analyses. Coefficients are unstandardized. Estimates using the original number of items are reported first, and estimates, after removing potentially problematic items, are reported second (i.e., after the hash line). All shown estimates are significant at .001

Consistent with Hypothesis 3, family eldercare demands were positively related to the husband’s (γ = .22, SE = .04, p < .001) and the wife’s time theft at work (γ = .16, SE = .05, p < .001). To test Hypothesis 4, we adopted a Monte Carlo resampling approach (Zhang, Zyphur, & Preacher, 2009) to test the indirect effect using the RMediation package in R. With 95% CI of 10,000 repetitions, we found that significant indirect effects of family eldercare demands on weekly time theft via each partner’s family-to-work conflict were significant for both husbands (effect = .09, CI 95% = [.04, .14]) and wives (effect = .08, CI 95% = [.03, .12]), supporting Hypothesis 4.

Post Hoc Analyses

Gender Differences

To delve deeper into our results, we examined potential gender differences by comparing the path coefficients between husbands and wives via the model constraint function implemented in Mplus 7.2. We did not find significant differences between husbands and wives in the paths from family eldercare demands to family-to-work conflict (difference = .08, SE = .08, p = .32), from family-to-work conflict to time theft (difference = .06, SE = .04, p = .19), nor from family eldercare demands to time theft (difference = .04, SE = .06, p = .48). There were no significant gender differences in the indirect effect of family eldercare demands on time theft via family-to-work conflict (difference = .01, SE = .02, p = .70). We further tested the model without constraining the crossover paths to be equal between partners, and we found no gender differences in the crossover effect (difference = .04, SE = .05, p = .49).

Test of Lagged Relations

We also sought to develop a more fine-grained understanding of the time-lagged relations. We structured lagged variables such that family eldercare demands of the week k had a lagged effect on family-to-work conflict of the week k + 1, and family-to-work conflict of the week k + 1 had a lagged effect on time theft of the week k + 2. This approach allowed us to test the lagged effects with stronger statistical power by using all available data points without losing valuable information. We performed multilevel path modeling to test all the parameters simultaneously. Family eldercare demands of the week k were positively related to family-to-work conflict of the husband (γ = .35, SE = .06, p < .001) and the wife (γ = .32, SE = .07, p < .001) in the week k + 1. In addition, family-to-work conflict of the week k + 1 was positively related to time theft for both the husband (γ = .11, SE = .06, p = .04) and the wife (γ = .10, SE = .05, p = .03) in the week k + 2. Finally, results supported the mediating role of family-to-work conflict between family eldercare demands and time theft for both the husband (effect = .04, CI 95% = [.001, .08]) and the wife (effect = .03, CI 95% = [.001, .06]). Taken together, results of the lagged analyses provided support for our hypotheses.

Supplemental Analyses Without Conceptually Overlapping Items

It should be noted that while our measures demonstrated systematic discriminant validity, conceptually, two of the eldercare demands items (i.e., “because of the time you spend on elder caregiving that you don’t have enough time for yourself”, “stressed between caring for elderly family member(s) and trying to meet other responsibilities”) and two of time theft items (i.e., “worked on a personal matter instead of working for your employer”, “surfed the Internet to deal with personal matters at your workplace”) overlap with the family-to-work conflict items. Thus, while our preliminary analyses suggest that the constructs are distinct, we sought to verify the robustness of our findings. Specifically, we performed additional analyses wherein we eliminated the overlapping items in order to examine if the conceptual overlap unduly influenced our interpretation of the empirical data. The analyses with the reduced items yield similar results showing that all our hypotheses were supported. In Fig. 2, unstandardized coefficients are reported wherein the original measures are reported first, and estimates, after removing potentially problematic items, are reported second (i.e., after the hash line).


Adapting the SCM framework (Bakker & Demerouti, 2013) to the context of eldercare, we examined family eldercare demands at the week level with an emphasis on understanding its effects on family-to-work conflict and time theft within dual-earner couples. Our results suggest that eldercare demands, like other family domain–specific constructs (e.g., Huffman et al., 2017; Song et al., 2011; Westman et al., 2008), act as a common stressor that relates to the crossover of weekly family-to-work conflict between partners and, consequently, time theft at work. Furthermore, we demonstrate that scholars would do well to take a more temporally nuanced approach when studying eldercare demands. Collectively then, and as will be discussed in more detail shortly, this study also extends the application of the SCM into the field of eldercare.

Theoretical Implications

Overall, the current research has several meaningful theoretical implications. First, we contributed to the eldercare literature by using the couple as a unit in examining the impact of eldercare demands. This is one of the first studies to show that eldercare demands may act as a common family stressor that relates to time theft at work, suggesting that the common life events (i.e., eldercare) at home could affect both partners’ work outcomes (Westman, 2001). We suggest that research on eldercare could develop a more comprehensive understanding of the impact of eldercare demands by taking a broad family system view. That is, it is meaningful to examine the experiences of eldercare in both partners and explore how the family, as a system, is impacted by eldercare demands (Hammer et al., 2003, 2005).

Second, we contributed to research by using a weekly diary design to examine the effects of family eldercare demands at the week level. By investigating eldercare demands cross-sectionally, it is assumed that eldercare experiences are stable in nature. Yet, the demands of caring for an elderly family member are such that this may not always be the case. Instead, because of the more episodic nature caregiving, it is likely that eldercare demands are more consistent with an acute or episodic stressor model. This has important implications for future research in that the timing and duration of the effects of eldercare demands and potential coping resources required depend on the nature of the stressor, that is whether it is acute or chronic (Barling et al., 1994). Extending previous cross-sectional studies (e.g., Zacher et al., 2012; Zacher & Winter, 2011), our findings suggest that family eldercare demands fluctuate from week to week. In turn, as predicted, we found that family eldercare demands were significantly associated with each partner’s family-to-work conflict and time theft at the week level. These findings suggest that the negative outcomes of family eldercare demands could be observed over short time periods (i.e., week). The present weekly diary study provides tentative support for the conclusion that variations in eldercare demands, even over a short timeframe, have significant and measurable consequences for employed elder caregivers such as increased next-week family-to-work conflict and time theft at work. While this study focuses on the short-term effects of eldercare demands, our results may be interpreted as “early warning signs” for the potential development of more serious outcomes (e.g., depression, poor performance, or turnover).

Third, we contributed to research by showing the value of the SCM framework (Bakker & Demerouti, 2013) in exploring the effect of eldercare demands on both partners. Our findings suggest that the SCM framework could be helpful to guide future research to test other spillover and crossover processes of family eldercare demands. It would be helpful to investigate stressors and resources in the work domain (e.g., job insecurity) that might spill over to the family domain and impact both partner’s perceptions of eldercare demands, which could further influence other important work and well-being outcomes. We also suggest that it is helpful to link the eldercare research with the traditional work–family literature (e.g., Barrah et al., 2004) by explicitly testing eldercare as an antecedent of family-to-work conflict (rather than treat eldercare as a covariate).

Finally, unlike past research that has been conducted using Western samples, our results are based on a sample of Chinese dual-earner couples. Despite the different macroeconomic structure, China is experiencing an aging trend similar to Western countries (Chen & Liu, 2009). The “one-child” policy may also make eldercare a more prevalent phenomenon in China. Our findings are largely consistent with results from previous Western samples (Hammer et al., 2003) regarding the effects of family-to-work conflict. This is among the first studies that have replicated and extended these findings in a Chinese sample. Looking forward, though beyond the scope of this study, it would be interesting to explore whether organizations in China have different reactions to time theft among employed caregivers given that eldercare mainly rests with the family with little or no employer or public support for eldercare available (Pei et al., 2017). It is possible that caregivers may perceive a norm that they can occasionally take time away from work, or perhaps they make up for this time at other times (e.g., weekends).

Limitations and Future Research

Although this study advances knowledge on the spillover and crossover of eldercare demands, the results should be interpreted in light of some limitations. First, given that our study relied on self-report measures, there may be concerns about the common method variance (CMV). However, repeated measures completed by both members of the couple (i.e., multisource data) were collected on a weekly basis, thereby reducing the likelihood of our results being biased by transient mood effects (cf. Podsakoff, MacKenzie, & Podsakoff, 2012). Moreover, given the nature of perceived eldercare demands and family-to-work conflict, self-report measures seem to be more appropriate than other forms of measurement. Although measuring time theft via self-report might be a limitation and objective (e.g., attendance record) or other-reported measures (e.g., peers, managers) could be used, prior meta-analytic research has found a moderate to strong correlation between self- and other-reported deviance (Berry, Carpenter, & Barratt, 2012).

A second limitation is that our predictor, mediator, and outcome data points were collected simultaneously across weeks, limiting causal inferences. This might raise concerns about reverse causality. However, our hypotheses were based on a sound theoretical framework and prior empirical findings. Both the SCM framework and empirical research suggest that eldercare demands at home would spill over to experiences at work through family-to-work conflict. Furthermore, the present study mainly focused on examining whether and how family eldercare demands (i.e., when the demands deviate from the “normal” levels) relate to within-person differences (i.e., why a person’s time theft may differ from 1 week to another) in family-to-work conflict and time theft at the week level. The results of our lagged analyses also provided preliminary support for the mediation effect. Relating to this, the relatively small number of couples constrained us from using more sophisticated approach (e.g., cross-lagged SEM analysis, latent growth modeling, and latent difference score modeling) to analyze the theoretical model (Selig & Preacher, 2009). Future research may consider using advanced longitudinal mediation analysis which may allow more rigorous tests of the hypotheses and better elucidation of the spillover and crossover processes of eldercare demands.

Another possible limitation of this study is the sampling strategy. Our participants were not randomly sampled, implying that they may not represent the general population with respect to eldercare demands. Because the study design is complex (including dual-earner couples who are facing eldercare demands, 5 weeks of data collection) and thus demanding for the participants, we recruited participants with the assistance of several trained research assistants. Doing this allowed us to recruit a sufficient number of participants. These research assistants were trained before they started to recruit participants. During the recruiting process and throughout data collection, the researcher met regularly with them and checked the progress.

Finally, the present study failed to consider the amount of time spent in other forms of caregiving (e.g., childcare). For employees who are facing multiple forms of caregiving, the effects of eldercare demands might be even stronger. It would be important for future research to pay attention to the population of employees who are dealing with more than one form of caregiving demands (e.g., “sandwiched” couples caring for both children and aging parents). Research may also investigate strategies or resources that can aid these caregiving employees.

Practical Implications

Our findings have implications for organizational practitioners striving to help employees successfully balance eldercare and work responsibilities (Greenhaus & Powell, 2006). Based on our results, practitioners should be aware that employees facing high levels of family eldercare demands may occasionally need to take time away from their jobs. This is a hidden cost of eldercare that both employees with eldercare responsibilities and organizational practitioners should recognize. Furthermore, the mediating role of family-to-work conflict suggests that organizational practitioners may help to reduce employed caregivers’ family-to-work conflict. Considering that increasing supervisors’ family-supportive supervision has been found to benefit physical health, job satisfaction, and turnover intentions among employees with high levels of family interference with work (Hammer, Kossek, Anger, Bodner, & Zimmerman, 2011; Odle-Dusseau, Hammer, Crain, & Bodner, 2016), interventions involving family supportive supervision may particularly benefit employees with eldercare responsibilities (Kossek et al., 2017). Additionally, organizational practitioners may consider other policies and/or practices such as scheduling flexibility. Allowing employees to have a say in scheduling their work (Paullin & Whetzel, 2012) could provide employees the flexibility and/or time needed to meet their eldercare needs (Peng, Jex, & Wang, 2018). Finally, given that most employees with eldercare responsibilities are aged 40 and older, future interventions may also employ selection, optimization, and compensation strategies that could help older workers cope with work–family conflict (e.g., Thrasher, Zabel, Wynne, & Baltes, 2016). For instance, when caring for a parent with a chronic disease, an employee may choose to focus on a limited number of family goals (e.g., focus on caring older parents and not teenager children), decide which skills (e.g., caregiving skills) need to be improved to insure successful goal accomplishment, and seek out resources to reduce their caregiving load (e.g., adult day care use). Use of selection, optimization, and compensation strategies is related to lower amounts of family and job stressors and subsequently lower levels of family interference with work (Baltes & Heydens-Gahir, 2003).

Practitioners should also realize that eldercare demands can relate to stress crossover within dual-earner couples. From the organizational perspective, organizational interventions designed to help employees with caregiver responsibilities should consider the roles of the partner when designing and delivering the intervention program. To maximize the effectiveness of these interventions, both employees and partners could be involved. Many family supportive practices, such as scheduling flexibility and working from home, might be more effective when both partners participate. For instance, a caregiving employee working from home on Monday may need more time to catch up with colleagues on Wednesday. In this situation, it would be particularly helpful if his or her spouse could take advantage of the flexible work scheduling during the days when the employee needs to stay longer at work.


This study contributes to the literature by showing for the first time that family eldercare demands act as a common stressor at home and can relate to family-to-work conflict between partners and spill over intraindividually to relate to time theft at work. Our findings enhance the field’s understanding of family eldercare demands and time theft among dual-earner couples at the within-person level. Organizations should recognize the hidden cost of time loss and consider the roles of the partner when designing and delivering the intervention program.


  1. AARP & The National Alliance for Caregiving (2015). Caregiving in the U.S. Retrieved from: http://www.aarp.org/content/dam/aarp/ppi/2015/caregiving-in-the-united-states-2015-report-revised.pdf. Accessed 31 Aug 2017.

  2. Bakker, A. B., & Demerouti, E. (2013). The spillover-crossover model. In J. G. Grzywacz & E. Demerouti (Eds.), New frontiers in work and family research (pp. 54–69). Hove, UK: Psychology Press.

    Google Scholar 

  3. Bakker, A. B., Demerouti, E., & Dollard, M. F. (2008). How job demands affect partners’ experience of exhaustion: Integrating work-family conflict and crossover theory. Journal of Applied Psychology, 93, 901–911.

    PubMed  Article  PubMed Central  Google Scholar 

  4. Baltes, B. B., & Heydens-Gahir, H. A. (2003). Reduction of work-family conflict through the use of selection, optimization, and compensation behaviors. Journal of Applied Psychology, 88, 1005–1018.

    PubMed  Article  PubMed Central  Google Scholar 

  5. Barling, J., MacEwen, K. E., Kelloway, E. K., & Higginbottom, S. F. (1994). Predictors and outcomes of elder-care-based interrole conflict. Psychology and Aging, 9, 391–397.

    PubMed  Article  Google Scholar 

  6. Barrah, J. L., Shultz, K. S., Baltes, B., & Stolz, H. E. (2004). Men’s and women’s eldercare-based work–family conflict: Antecedents and work-related outcomes. Fathering, 2, 305–330.

    Article  Google Scholar 

  7. Bédard, M., Molloy, D. W., Squire, L., Dubois, S., Lever, J. A., & O’Donnell, M. (2001). The Zarit Burden Interview: A new short version and screening version. The Gerontologist, 41, 652–657.

    PubMed  Article  PubMed Central  Google Scholar 

  8. Bennett, R. J., & Robinson, S. L. (2000). Development of a measure of workplace deviance. Journal of Applied Psychology, 85, 349–360.

    PubMed  Article  PubMed Central  Google Scholar 

  9. Berry, C. M., Carpenter, N. C., & Barratt, C. L. (2012). Do other-reports of counterproductive work behavior provide an incremental contribution over self-reports? A meta-analytic comparison. Journal of Applied Psychology, 97, 613–636.

    PubMed  Article  PubMed Central  Google Scholar 

  10. Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54, 579–616.

    PubMed  Article  PubMed Central  Google Scholar 

  11. Bolger, N., DeLongis, A., Kessler, R. C., & Wethington, E. (1989). The contagion of stress across multiple roles. Journal of Marriage and Family, 51, 175–183.

    Article  Google Scholar 

  12. Brislin, R. W. (1980). Translation and content analysis of oral and written materials. In H. C. Triandis & J. W. Berry (Eds.), Handbook of cross-cultural psychology (pp. 137–164). Boston, MA: Allyn and Bacon.

    Google Scholar 

  13. Brock, M. E., Martin, L. E., & Buckley, M. R. (2013). Time theft in organizations: The development of the time banditry questionnaire. International Journal of Selection and Assessment, 21, 309–321.

    Article  Google Scholar 

  14. Buhrmester, M. D., Talaifar, S., & Gosling, S. D. (2018). An evaluation of Amazon’s mechanical turk, its rapid rise, and its effective use. Perspectives on Psychological Science, 13, 149–154.

    PubMed  Article  PubMed Central  Google Scholar 

  15. Chen, F., & Liu, G. (2009). Population aging in China. In P. Uhlenberg (Ed.), International handbook of population aging (pp. 157–172). New York, NY: Springer.

    Google Scholar 

  16. Dalal, R. S., Lam, H., Weiss, H. M., Welch, E. R., & Hulin, C. L. (2009). A within-person approach to work behavior and performance: Concurrent and lagged citizenship-counterproductivity associations, and dynamic relationships with affect and overall job performance. Academy of Management Journal, 52, 1051–1066.

    Article  Google Scholar 

  17. Dugan, A., Matthews, R. A., & Barnes-Farrell, J. L. (2012). Understanding the role of subjective and objective experiences of time in the work-family interface. Community, Work & Family, 15, 149–172.

    Article  Google Scholar 

  18. Ferguson, M., Carlson, D., Hunter, E. M., & Whitten, D. (2012). A two-study examination of work–family conflict, production deviance and gender. Journal of Vocational Behavior, 81, 245–258.

    Article  Google Scholar 

  19. Fortinsky, R. (2011), Juggling work and eldercare responsibilities. The National Institute for Occupational Safety and Health, CPH News Views (20):1–2.

  20. Frone, M. R., Russell, M., & Cooper, M. L. (1992). Antecedents and outcomes of work-family conflict: Testing a model of the work-family interface. Journal of Applied Psychology, 77, 65–78.

    PubMed  Article  Google Scholar 

  21. Gorey, K. M., Rice, R. W., & Brice, G. C. (1992). The prevalence of elder care responsibilities among the work force population: Response bias among a group of cross-sectional surveys. Research on Aging, 14, 399–418.

    Article  Google Scholar 

  22. Greenhaus, J. H., & Powell, G. N. (2006). When work and family are allies: A theory of work-family enrichment. Academy of Management Review, 31, 72–92.

    Article  Google Scholar 

  23. Grzywacz, J. G., Frone, M. R., Brewer, C. S., & Kovner, C. T. (2006). Quantifying work-family conflict among registered nurses. Research in Nursing & Health, 29, 414–426.

    Article  Google Scholar 

  24. Hammer, L. B., Allen, E., & Grigsby, T. D. (1997). Work–family conflict in dual-earner couples: Within-individual and crossover effects of work and family. Journal of Vocational Behavior, 50, 185–203.

    Article  Google Scholar 

  25. Hammer, L. B., Bauer, T. N., & Grandey, A. A. (2003). Work-family conflict and work-related withdrawal behaviors. Journal of Business and Psychology, 17, 419–436.

    Article  Google Scholar 

  26. Hammer, L. B., Kossek, E. E., Anger, W. K., Bodner, T., & Zimmerman, K. L. (2011). Clarifying work-family intervention processes: The roles of work-family conflict and family-supportive supervisor behaviors. Journal of Applied Psychology, 96, 134–150.

    PubMed  Article  Google Scholar 

  27. Hammer, L. B., Neal, M. B., Newsom, J. T., Brockwood, K. J., & Colton, C. L. (2005). A longitudinal study of the effects of dual-earner couples’ utilization of family-friendly workplace supports on work and family outcomes. Journal of Applied Psychology, 90, 799–810.

    PubMed  Article  Google Scholar 

  28. Henle, C. A., Reeve, C. L., & Pitts, V. E. (2010). Stealing time at work: Attitudes, social pressure, and perceived control as predictors of time theft. Journal of Business Ethics, 94, 53–67.

    Article  Google Scholar 

  29. Hobfoll, S. E. (1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44, 513–524.

    PubMed  Article  Google Scholar 

  30. Hobfoll, S. E., & London, P. (1986). The relationship of self-concept and social support to emotional distress among women during war. Journal of Social and Clinical Psychology, 4, 189–203.

  31. Huffman, A. H., Matthews, R. A., & Irving, L. H. (2017). Family fairness and cohesion in marital dyads: Mediating processes between work–family conflict and couple psychological distress. Journal of Occupational and Organizational Psychology, 90, 95–116.

    Article  Google Scholar 

  32. Ilies, R., Schwind, K. M., & Heller, D. (2007). Employee well-being: A multilevel model linking work and nonwork domains. European Journal of Work and Organizational Psychology, 16, 326–341.

    Article  Google Scholar 

  33. Jöreskog, K. G. (1977). Factor analysis by least-squares and maximum-likelihood methods. In K. Enslein, A. Ralston, & H. S. Wilf (Eds.), Statistical methods for digital computers (pp. 125–153). New York: Wiley.

    Google Scholar 

  34. Kaplan, D. (1998). Methods for multilevel data analysis. In G. A. Marcoulides (Ed.), Modern methods for business research (pp. 337–357). Mahwah, NJ: Erlbaum.

    Google Scholar 

  35. Ketchen, D. J., Craighead, C. W., & Buckley, M. R. (2008). Time bandits: How they are created, why they are tolerated, and what can be done about them. Business Horizons, 51, 141–149.

    Article  Google Scholar 

  36. Kossek, E. E., Thompson, R. J., Lawson, K. M., Bodner, T., Perrigino, M. B., Hammer, L. B., ..., Bray, J. W. (2017). Caring for the elderly at work and home: Can a randomized organizational intervention improve psychological health? Journal of Occupational Health Psychology. Advance online publication.

  37. Krischer, M. M., Penney, L. M., & Hunter, E. M. (2010). Can counterproductive work behaviors be productive? CWB as emotion-focused coping. Journal of Occupational Health Psychology, 15, 154–166.

    PubMed  Article  Google Scholar 

  38. Lance, C. E., Butts, M. M., & Michels, L. C. (2006). The sources of four commonly reported cutoff criteria: What did they really say? Organizational Research Methods, 9, 202–220.

    Article  Google Scholar 

  39. Laurenceau, J., & Bolger, N. (2005). Using diary methods to study marital and family processes. Journal of Family Psychology, 19, 86–97.

    PubMed  Article  Google Scholar 

  40. LeBreton, J. M., & Senter, J. L. (2008). Answers to 20 questions about interrater reliability and interrater agreement. Organizational Research Methods, 11, 815–852.

    Article  Google Scholar 

  41. Litman, L., Robinson, J., & Abberbock, T. (2017). TurkPrime.com: A versatile crowdsourcing data acquisition platform for the behavioral sciences. Behavior Research Methods, 49, 433–442.

  42. Lorinkova, N. M., & Perry, S. J. (2017). When is empowerment effective? The role of leader-leader exchange in empowering leadership, cynicism, and time theft. Journal of Management, 43, 1631–1654.

    Article  Google Scholar 

  43. Lu, C. Q., Lu, J. J., Du, D. Y., & Brough, P. (2016). Crossover effects of work-family conflict among Chinese couples. Journal of Managerial Psychology, 31, 235–250.

    Article  Google Scholar 

  44. Martin, L. E., Brock, M. E., Buckley, M. R., & Ketchen, D. J. (2010). Time banditry: Examining the purloining of time in organizations. Human Resource Management Review, 20, 26–34.

    Article  Google Scholar 

  45. Matthews, R. A., Wayne, J. H., & Ford, M. T. (2014). A work–family conflict/subjective well-being process model: A test of competing theories of longitudinal effects. Journal of Applied Psychology, 99, 1173–1187.

    PubMed  Article  PubMed Central  Google Scholar 

  46. Matthews, R. A., Winkel, D. E., & Wayne, J. H. (2014). A longitudinal examination of role overload and work–family conflict: The mediating role of inter-domain transitions. Journal of Organizational Behavior, 35, 72–91.

    Article  Google Scholar 

  47. McGee, M. K., & Fillon, M. (1995). Honesty is still the best policy. Information Week, 519, 156.

    Google Scholar 

  48. Odle-Dusseau, H. N., Hammer, L. B., Crain, T. L., & Bodner, T. E. (2016). The influence of family-supportive supervisor training on employee job performance and attitudes: An organizational work–family intervention. Journal of Occupational Health Psychology, 21, 296–308.

    PubMed  Article  PubMed Central  Google Scholar 

  49. Ohly, S., Sonnentag, S., Niessen, C., & Zapf, D. (2010). Diary studies in organizational research: An introduction and some practical recommendations. Journal of Personnel Psychology, 9, 79–93.

    Article  Google Scholar 

  50. Paullin, C., & Whetzel, D. L. (2012). Retention strategies and older workers. In J. W. Hedge & W. C. Borman (Eds.), The Oxford handbook of work and aging (pp. 392–418). New York, NY: Oxford University Press.

    Google Scholar 

  51. Pei, X., Luo, H., Lin, Z., Keating, N., & Fast, J. (2017). The impact of eldercare on adult children’s health and employment in transitional China. Journal of Cross-Cultural Gerontology, 32, 357–372.

    PubMed  Article  Google Scholar 

  52. Peng, Y., Jex, S. M., & Wang, M. (2018). Aging and occupational health. In K. Shultz, & G. Adams (Eds.), Aging and work in the 21st century (2nd Ed; pp. 213–233). Routledge.

  53. Podsakoff, P. M., MacKenzie, S. B., & Podsakoff, N. P. (2012). Sources of method bias in social science research and recommendations on how to control it. Annual Review of Psychology, 63, 539–569.

    PubMed  Article  PubMed Central  Google Scholar 

  54. Raudenbush, S. W., Brennan, R. T., & Barnett, R. C. (1995). A multivariate hierarchical model for studying psychological change in married couples. Journal of Family Psychology, 9, 161–174.

    Article  Google Scholar 

  55. Sanz-Vergel, A. I., Rodríguez-Muñoz, A., & Nielsen, K. (2015). The thin line between work and home: The spillover and crossover of daily conflicts. Journal of Occupational and Organizational Psychology, 88, 1–18.

    Article  Google Scholar 

  56. Selig, J. P., & Preacher, K. J. (2009). Mediation models for longitudinal data in developmental research. Research in Human Development, 6, 144–164.

    Article  Google Scholar 

  57. Shoptaugh, C. F., Phelps, J. A., & Visio, M. E. (2004). Employee eldercare responsibilities: Should organizations care? Journal of Business and Psychology, 19, 179–196.

    Article  Google Scholar 

  58. Spector, P. E., Fox, S., Penney, L. M., Bruursema, K., Goh, A., & Kessler, S. (2006). The dimensionality of counterproductivity: Are all counterproductive behaviors created equal? Journal of Vocational Behavior, 68, 446–460.

    Article  Google Scholar 

  59. Song, Z., Foo, M. D., Uy, M. A., & Sun, S. (2011). Unraveling the daily stress crossover between unemployed individuals and their employed spouses. Journal of Applied Psychology, 96, 151–168.

  60. Thrasher, G. R., Zabel, K., Wynne, K., & Baltes, B. B. (2016). The importance of workplace motives in understanding work–family issues for older workers. Work, Aging and Retirement, 2, 1–11.

    Google Scholar 

  61. Trukeschitz, B., Schneider, U., Mühlmann, R., & Ponocny, I. (2012). Informal eldercare and work-related strain. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 68, 257–267.

    Article  Google Scholar 

  62. Westman, M. (2001). Stress and strain crossover. Human Relations, 54, 557–591.

    Article  Google Scholar 

  63. Westman, M., & Etzion, D. (1995). Crossover of stress, strain and resources from one spouse to another. Journal of Organizational Behavior, 16, 169–181.

    Article  Google Scholar 

  64. Westman, M., & Vinokur, A. D. (1998). Unraveling the relationship of distress levels within couples: Common stressors, empathic reactions, or crossover via social interaction?. Human Relations, 51, 137–156.

  65. Westman, M., & Etzion, D. L. (2005). The Crossover of Work‐Family Conflict From One Spouse to the Other. Journal of Applied Social Psychology, 35, 1936–1957.

  66. Westman, M., Keinan, G., Roziner, I., & Benyamini, Y. (2008). The crossover of perceived health between spouses. Journal of Occupational Health Psychology, 13, 168–180.

  67. World Health Organization. (2015). World report on ageing and health. Geneva: World Health Organization.

    Google Scholar 

  68. Zacher, H., Jimmieson, N. L., & Winter, G. (2012). Eldercare demands, mental health, and work performance: The moderating role of satisfaction with eldercare tasks. Journal of Occupational Health Psychology17, 52–64.

  69. Zacher, H., & Winter, G. (2011). Eldercare demands, strain, and work engagement: The moderating role of perceived organizational support. Journal of Vocational Behavior, 79, 667–680.

    Article  Google Scholar 

  70. Zarit, S. H., Orr, N. K., & Zarit, J. M. (1985). The hidden victims of Alzheimer’s disease: Families under stress. New York: New York University Press.

    Google Scholar 

  71. Zhan, H. J. (2002). Chinese caregiving burden and the future burden of elder care in life-course perspective. The International Journal of Aging and Human Development, 54, 267–290.

    PubMed  Article  PubMed Central  Google Scholar 

  72. Zhang, Z., Zyphur, M. J., & Preacher, K. J. (2009). Testing multilevel mediation using hierarchical linear models: Problems and solutions. Organizational Research Methods, 12, 695–719.

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Yisheng Peng.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.



Description of Supplemental Data Set and Measures Used to Evaluate Construct Validity of the Shortened Eldercare Demands and Time Theft Measures

Participants and Procedure

For the purposes of providing additional construct validity information for the eldercare demands and time theft measures used in this study, a separate heterogeneous sample of 243 working U.S. eldercare providers was recruited from TurkPrime (Litman et al., 2017) to complete an online survey for a financial compensation of $.50. Prior research argued that researchers can recruit diverse samples from such channels and can acquire very acceptable quality of data for academic research (Buhrmester, Talaifar, & Gosling, 2018). Respondents had to be organizationally employed at least 35 hours per week and be caring a loved one (i.e., family member, relative, friend, neighbor) who had to be 50 years and older. Initially, 510 individuals from a TurkPrime panel who are currently providing eldercare completed a prescreening survey, 301 of whom met the criteria and were invited to complete the full survey, and 292 responded to the full survey. Five validation questions were embedded to ensure effortful responding (e.g., “Select often if you are paying attention to this item”). Respondents who failed to correctly complete at least 4 of the 5 questions were excluded. After data cleaning, 243 individuals were retained. The majority of the sample was female (58%) and White/non-Hispanic (69%) and was caring for a parent/parent-in-law and grandparent/grandparent-in-law (85%). Sixty-two percent had at least a 4-year college degree and 62% worked in nonsupervisory positions. Hours worked per week ranged from 35 to 90 (M = 41.32 h, SD = 5.69). Hours devoted to elder caregiving per week ranged from 1 to 80 (M = 23.29 h, SD = 15.29). Participants’ ages ranged from 20 to 69 (M = 37.93 years, SD = 11.26). Eldercare recipients’ ages ranged from 50 to 99 (M = 75.09 years, SD = 11.29). Average job tenure was 6.52 years (SD = 6.10).


Eldercare demands were measured with all 22 items developed by Zarit et al. (1985). We followed the recommendation of Bédard et al. (2001) and computed the average score of the 4-item version, which has been used in the weekly diary study. Furthermore, following the editor’s and reviewers’ suggestion, we eliminated two conceptually overlapping items (i.e., do not have enough time for yourself, stressed between caring for elders and trying to meet other responsibilities) and computed the average score of the remaining two items. Family-to-work conflict was measured by using the same 3-item subscale of family-to-work conflict (Grzywacz et al., 2006). Time theft was measured by the same 6 items used in the weekly diary study (Bennett & Robinson, 2000; Dalal et al., 2009; Henle et al., 2010; Lorinkova & Perry, 2017). Following the editor’s and reviewers’ suggestion, we eliminated two conceptually overlapping items (i.e., worked on a personal matter instead of working for your employer, surfed the Internet to deal with personal matters at your workplace) and computed the average score of the remaining four items. Organizational deviance was measured by using the 12-item subscale of organizational deviance developed by Bennett and Robinson (2000). Time banditry was measured using the 31-item questionnaire developed by Brock et al. (2013). This measure contains three subscales measuring classic time banditry (18 items), technology-related time banditry (7 items), and socially oriented time banditry (5 items). Production deviance (3 items) and withdrawal (4 items) were measured by using the two subscales from the 32-item Counterproductive Work Behavior Checklist (CWB-C; Spector et al., 2006).


Appendix Table 2 provides Cronbach’s alphas and correlations among all measured variables. All other descriptive statistics are available from the first author. The 4-item eldercare demands measure (r = .87, p < .01) and the 2-item eldercare demands measure (r = .82, p < .01) were significantly correlated with the full 22-item eldercare demands measure. The 4-item eldercare demands measure and the 2-item eldercare demands measure were significantly correlated (r = .90, p < .01). Furthermore, a similar pattern of relationships with all other measured variables was observed, and the magnitudes of these correlations were quite similar.

The 6-item time theft measure and the 4-item time theft measure were significantly correlated (r = .98, p < .01). Moreover, the 6-item time theft measure and the 4-item time theft measure were similarly correlated with production deviance (r = .58, p < .01; r = .59, p < .01), withdrawal (r = .74, p < .01; r = .74, p < .01), overall time banditry (r = .78, p < .01; r = .75, p < .01), classic time banditry (r = .78, p < .01; r = .77, p < .01), technology-related time banditry (r = .51, p < .01; r = .44, p < .01), socially oriented time banditry (r = .64, p < .01; r = .60, p < .01), organizational deviance (r = .73, p < .01; r = .75, p < .01), and family-to-work conflict (r = .54, p < .01; r = .53, p < .01).


The short measures of eldercare demands were strongly correlated with the full measure of eldercare demands. The short measures of time theft were also strongly correlated with other established measures, including time banditry, production deviance, and organizational deviance. These results suggest that the short measures used in the weekly diary study are valid measures of their respective constructs.

Table 2 Bivariate correlations

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Peng, Y., Jex, S., Zhang, W. et al. Eldercare Demands and Time Theft: Integrating Family-to-Work Conflict and Spillover–Crossover Perspectives. J Bus Psychol 35, 45–58 (2020). https://doi.org/10.1007/s10869-019-09620-3

Download citation


  • Eldercare
  • Time theft
  • Family-to-work conflict
  • Spillover–crossover model