The Relationship of Job and Elder Caregiving Involvement to Work-Caregiving Conflict and Work Costs

Abstract

This article examines the relationship between job and elder caregiving involvement, bi-directional work-caregiving conflict (work interfering with caregiving [WIC] and caregiving interfering with work [CIW]), and work costs (job interruptions and job changes). Specifically, we consider the effects of both behavioral and psychological involvement and external and internal work caregiving conflict in a sample of 583 women between the ages of 50 and 64 who work full-time and have significant elder caregiving responsibilities. A telephone survey was administered using random-digit-dial procedures. Structural equation model analyses confirmed that behavioral job involvement was associated with external CIW, psychological job involvement was associated with internal WIC and internal CIW, behavioral caregiving involvement was associated with external CIW and work costs, and psychological caregiving involvement was associated with internal CIW. Internal WIC, external CIW, and internal CIW were associated with job costs. Some mediation, particularly through external and internal CIW, occurred.

Keywords

older workers elder caregiving job involvement caregiving involvement work-family conflict work costs

Many women care for an elder at the same time as they remain actively engaged in the full-time work force. The extent to which older women are actively involved in their roles of caregivers and employees likely has consequences for their work performance and their personal well-being. Research has shown that role involvement, or the extent to which a person is invested in a particular role, in both the work and family domains, has an important impact on an individual’s attitudes and behavior, and this impact can be positive or negative or both. For example, involvement in work and family roles has been associated with role conflict (Greenhaus & Beutell, 1985) and psychological well-being (Reitzes, Mutran, & Fernandez, 1994; Stephens, Townsend, Martire, & Druley, 2001). Additional studies considered the effect of job involvement on role conflict and work-related outcomes (e.g., Adams, King, & King, 1996; Frone, Russell, & Cooper, 1992). Yet only a few studies examined the relationship of family involvement to role conflict (e.g., Aryee, Srinivas, & Tan, 2005; Graves, Ohlott, & Ruderman, 2007), and almost no studies considered the impact of family involvement on job-related outcomes (e.g., Adams et al., 1996; see ten Brummenhuis, van der Lippe, Kluwer, & Flap, 2008, for exceptions).

Attempts to understand the relationship among the role involvement of working elder caregivers, the role conflict they experience, and the work costs they incur as a result of their involvement and role conflict have been inconsistent and inconclusive (Graves et al., 2007). Our research suggests that inconsistencies might be resolved if scholars take a more fine-grained approach to understanding role involvement. In this article, we answer the following research question: What is the relationship of various types of role involvement in the work and elder caregiving domains to work-caregiving conflict, and what are the work costs of the involvement and conflict? Answers to this question have important behavioral and financial implications for organizations and their employees.

Conceptual Background and Hypotheses

Role Involvement

We look at behavioral and psychological role involvement in two domains: job involvement in the work domain and elder caregiving involvement in the family domain. Behavioral involvement refers to the amount of time devoted to a specific role (Hepburn & Barling, 1996). An individual with a high level of behavioral involvement in a given role will spend more time performing the role than a person with a low level of involvement. Psychological role involvement describes the extent to which an individual identifies with a particular role and its responsibilities and considers performance of this role as related to his or her self-concept (Carlson & Frone, 2003). Individuals with high levels of psychological involvement will be more mentally preoccupied with and identify more with the role than individuals with low levels.

More research has focused on job involvement than on family involvement, and even less has examined the nature and impact of involvement in providing care for elders. Role involvement in elder caregiving may differ from role involvement in other aspects of family life. Because the demands of elder caregiving tend to become more burdensome over time (Azarnoff & Scharlach, 1988), behavioral involvement may become more extensive. However, because elder caregiving is not as consistent a responsibility as child care, behavioral involvement may be more intermittent. Psychological involvement, in contrast, may be less extensive than involvement associated with child care because typically caregivers invest more time, and likely energy, in a relationship with children than they do with elders (McCarthy, 1997).

An area of research has focused on the “sandwiched” generation (Hammer, Neal, Newsom, Brockwood, & Colton, 2005; Ingersoll-Dayton, Neal, & Hammer, 2001; Neal & Hammer, 2007). This work, for instance, has considered the role of workplace supports on work-family conflict in people who are simultaneously managing the demands of child and parent care, but has not examined several types of involvement and role conflict simultaneously. Although Barling, MacEwen, Kelloway, and Higginbottom (1994) studied elder care involvement and the interrole conflict related to it, their study did not include a detailed assessment of the relationship between elder care involvement and various types of role conflict. They did not look at the full range of work and caregiving involvement and work-caregiving conflict in the work domain, nor did they examine the work costs of such involvement and role conflict. In this article, we propose a detailed set of relationships among involvement, role conflict, and work costs, and test them within the realm of elder caregiving. This article uses conservation of resources theory to predict the consequences of behavioral role involvement and identity theory to predict the consequences of psychological role involvement.

Work-Caregiving Conflict

The research on work-family conflict has widely accepted and conceptualized such role conflict as bidirectional (Carlson & Frone, 2003; Frone, Yardley, & Markel, 1997). In this article, we focus specifically on conflict between the responsibilities associated with work roles and the role of providing care to an elder and conceptualize this conflict as bidirectional—that is, work interfering with caregiving (WIC) and caregiving interfering with work (CIW). Research has suggested that role involvement likely affects the two directions of work-family interference differently. For example, job involvement was linked to work-family conflict and family involvement to family-work conflict (Adams et al., 1996; Frone et al., 1992). Aryee et al. (2005) found that job involvement was positively associated with family-work conflict, but that family involvement was not associated with work-family conflict. Graves et al. (2007), in contrast, found that neither parental nor marital role commitment was associated with increased conflict.

Another explanation for the inconsistent results related to role involvement and role conflict may be a need to move beyond global measures of work-family conflict. Carlson and Frone (2003) offered a four-factor conceptualization of work-family conflict, which distinguishes between external (behavioral) and internal (psychological) conflict in both directions—that is, work interfering with family and family interfering with work. We used their four-factor framework and applied it to the conflict between work and elder caregiving. External conflict refers to behavioral conflict between work and elder caregiving, such as when a person’s work schedule prevents her from attending a doctor’s appointment with the elder receiving care (external WIC) or when the requirements of elder caregiving prevent the employee from attending an important work meeting (external CIW). Internal conflict between work and elder caregiving refers to a psychological preoccupation with work or elder caregiving, such as when an individual thinks about the next day’s demands at work and cannot provide full attention to the care recipient (internal WIC) or when the caregiver is preoccupied with the next day’s caregiving activities and cannot provide full attention to work demands (internal CIW).

Inconsistent results in the literature may also result from the need to track the different consequences of behavioral and psychological involvement. Hepburn and Barling (1996) showed that behavioral caregiving involvement was associated with work-caregiving conflict—that is, the number of hours providing care to an elder and the number of hours interacting with the elder were associated with measures of external role conflict; however, they did not study other possible relationships. As noted earlier, Carlson and Frone (2003) expanded the components of work-family conflict and role involvement studied, but did not look at the relationships within the context of elder care, nor did they test any cross-domain relationships. Barling et al. (1994) showed that elder care involvement is positively associated with conflict between the roles of employee, caregiver for an elderly relative, spouse, and parent, but they did not separate the effects of particular types of involvement on particular types of role conflict, nor did they look at cross-domain relationships.

Involvement and Conflict

We build on Hobfoll’s conservation of resources theory and identity theory to develop hypotheses for the relationship between involvement and conflict. Hobfoll’s (1989, 2002) conservation of resources theory posits that excessive demands in the workplace create a resource drain that has negative consequences and results in role conflict and strain. This theory suggests that the time demands of behavioral work and caregiving involvement create demands in the workplace or at home, respectively, that cause a resource drain which ultimately results in conflict in performing the specific responsibilities of each role. Hence older working women with high involvement in their jobs, as well as those with high involvement in elder caregiving, should experience more external WIC and CIW. Involvement in one role may cause the individual to become less involved in a second role because of the scarcity of resources and time and ultimately result in role strain (Greenhaus & Parasuraman, 1994). Similarly, where different roles create competing demands, role conflict may result (Stephens & Franks, 1999). Individuals need to make tradeoffs to reduce demands, increase resources, and ultimately reduce the resulting role strain, which may ultimately result in non-financial costs to the individual and her employer, including interruptions or changes in their jobs. Thus, we hypothesize that behavioral role involvement, which creates competing demands, is positively associated with external WIC and external CIW (see Hypotheses 1a and 1c below).

Identity theory complements the conservation of resources theory and helps explain the relationship of psychological involvement to work-elder caregiving conflict. Psychological involvement reflects an individual’s identification with a particular role and the person’s perception that this identification is important to her self-concept (Carlson & Frone, 2003). The psychological preoccupation with one role, as well as the cognitive and emotional investment in that role, makes it challenging to engage psychologically in the other role. Hence, the more psychologically involved a person is with work or elder caregiving, the more likely that person considers work or elder caregiving as important to her self-concept, the more difficult it is to engage in the other role, and the more likely that internal (or psychological) conflict between the two roles will result (see Hypotheses 1b and 1d below).

Considering together, then, the conservation of resources and identity theories, we hypothesize the following:

Hypothesis 1a: Behavioral job involvement will be positively associated with external WIC.

Hypothesis 1b: Psychological job involvement will be positively associated with internal WIC.

Hypothesis 1c: Behavioral caregiving involvement will be positively associated with external CIW.

Hypothesis 1d: Psychological caregiving involvement will be positively associated with internal CIW.

Recent research suggests that cross-domain relationships exist (Ford, Heinen, & Langkamer, 2007; Huang, Hammer, Neal, & Perrin, 2004)—that is, aspects of the work domain may influence the caregiving domain, and vice versa. The same logic applies to cross-domain relationships as within-domain relationships. The conservation of resources theory posits that excessive demands in the workplace create a resource drain that results in role conflict and strain. This theory suggests that the time demands of work and caregiving involvement create demands in the workplace or at home, respectively, that cause a resource drain which ultimately results in conflict in performing the behaviors associated with each role, in this case across domains. Hence, behavioral work involvement is predicted to be positively associated with external CIW and behavioral caregiving involvement positively associated with external WIC (see Hypotheses 2a and 2c below).

Identity theory suggests that the more psychologically involved a person is with work or elder caregiving, the more likely that person considers work or elder caregiving as important to her self-concept, the more difficult it is to engage in the other role, and the more likely that psychological or internal conflict between the two roles will result, in this case across domains. Thus, we hypothesize that psychological work involvement is positively associated with internal CIW and psychological caregiving involvement is positively associated with internal WIC (see Hypotheses 2b and 2d below). Together, the conservation of resources and identity theories predict that role involvement will also have cross-domain associations with conflict—that is, that work involvement will be positively associated with CIW and caregiving involvement will be positively associated with WIC. More specifically, we hypothesize the following:

Hypothesis 2a: Behavioral job involvement will be positively associated with external CIW.

Hypothesis 2b: Psychological job involvement will be positively associated with internal CIW.

Hypothesis 2c: Behavioral caregiving involvement will be positively associated with external WIC.

Hypothesis 2d: Psychological caregiving involvement will be positively associated with internal CIW.

Work Costs

Elder caregiving has been shown to have significant short-term and long-term costs at work for employees and employers. Caregivers experience daily disruptions, such as arriving late to work, leaving early, missing work, and experiencing frequent interruptions at work because of their caregiving responsibilities (Barling et al., 1994; Barnett, 2005; Wakabayashi & Donato, 2005). In the longer term, caregivers may switch to part-time employment or change jobs to allow them more time for caregiving responsibilities (Stone & Short, 1990). Thus, costs at work may include but are not limited to financial or economic costs; daily interruptions, lack of attentiveness, lowered performance, and other performance-related costs may occur. Most research related to work interruptions and job withdrawal has focused on either absenteeism or turnover (see meta-analysis by Allen, Herst, Bruck, & Sutton, 2000), although Hammer, Bauer, and Grandey (2003) also included family interruptions at work and lateness to work in the withdrawal behaviors they examined. They found, for example, that work-to-family conflict was positively associated with job interruptions for both the wife and husband. Wegge, Schmidt, Parkes, and van Dick (2007) noted that the interaction of job involvement and job satisfaction was positively associated with both the frequency of absences and time lost on the job. Yet limited, if any, research has examined the relationship of behavioral and psychological work and caregiving involvement to these work costs.

We can again use conservation of resources and identity theories to predict how role involvement is associated with work costs. Excessive demands in the workplace create a resource drain that has negative consequences. Increasing involvement with work responsibilities may increase job interruptions and motivate job changes because of a lack of available time during the work day to handle caregiving matters. Involvement with caregiving responsibilities, in contrast, may spill into the work domain, also increasing the likelihood of work costs. Hence, we hypothesize that behavioral work and caregiving involvement are positively associated with work costs (see Hypotheses 3a and 3c below).

Second, identity theory suggests that the less an individual identifies with a particular role, the less involved she will be with that role and the more likely she will interrupt or leave that role. We predict that the more a person identifies with his or her work roles, the more psychologically involved he or she will be involved with work roles and consequently the more likely to remain in the work role and therefore in their present job situation. Conversely, the more they identify with their caregiving role, the more psychologically involved they will be with the caregiving role and the more likely they are to interrupt or leave their work roles and hence incur the work costs of job interruptions and job changes (see Hypotheses 3b and 3d below). Considering conservation of resources and identity theories together, we hypothesize the following:

Hypothesis 3a: Behavioral job involvement is positively associated with work costs.

Hypothesis 3b: Psychological job involvement is negatively associated with work costs.

Hypothesis 3c: Behavioral caregiving involvement is positively associated with work costs.

Hypothesis 3d: Psychological caregiving involvement is positively associated with work costs.

Work-caregiving conflict is also related to job interruptions and job changes. Role theory suggested that work-family conflict was associated with role strain (Cooper, Dewe, & O’Driscoll, 2001; Kahn & Byosiere, 1992). Because of the strain, role conflict is stressful (Barling, 1990; MacEwen & Barling, 1991) and can result in undesirable outcomes, including the work costs of job interruptions and job changes. Hence, we hypothesize the following:

Hypothesis 4a: External CIW is positively associated with work costs.

Hypothesis 4b: Internal CIW is positively associated with work costs.

Hypothesis 4c: External WIC is positively associated with work costs.

Hypothesis 4d: Internal WIC is positively associated with work costs.

Figure 1 shows the hypothesized relationships. The hypotheses we have presented so far suggest that work-caregiving conflict may mediate the relationship between role involvement and work costs. Although Michel, Mitchelson, Kotrba, LeBreton, and Baltes (2009) did not find such mediating relationships for job satisfaction, no test of such mediation has occurred for work costs, which includes job interruptions and job changes. Hence, we test for mediating relationships in our analyses.

Figure 1.

Model With Hypothesized Relationships

Method

Sample

Women eligible to participate in the study were between the ages of 50 and 64, were engaged in the labor force for at least 35 hours per week, and provided significant levels of care to an older person. They assisted a person over age 50 who lived in the community rather than in a long-term care facility with at least one activity of daily living (e.g., bathing, feeding, and dressing) or two or more instrumental activities of daily living (e.g., shopping, transportation, cooking) on a daily basis and had been doing so for at least 6 months (Stone, Cafferata, & Sangl, 1987). They could be living with or apart from their caregiver.

The sample was obtained using random digit dial procedures. The listed household database used was the Electronic Database of Listed Telephone Number Households that includes all phone numbers published in the electronic white pages. Approximately 63% of U.S. households are represented in this listed household database. The proportion of the 60,993 phone numbers purchased and entered into a computer-assisted telephone interviewing (CATI) system was determined based on calculations of working phone rates and expected study eligibility rates. CATI staff was able to successfully screen between 38.2% and 61.7% of all households, depending on the particular list purchased. Once an eligible woman was identified, between 36.3% and 46.3%, depending on the list, actually completed an interview. The 583 respondents to the survey had a mean age of 55.7 years (SD = 3.8); 76% were White, 55% were married, 57% had a bachelor’s or master’s degree, and 73% had a household income between $25,000 and $100,000, 10% more than $100,000, and 3% not reported. The majority of care recipients (73%) were female and had a mean age of 78.4 years (SD = 10.9). The most common relationship of the care recipient to the caregiver was mother (57%), followed by spouse or male partner (11%), and father (9%). Two thirds of the care recipients did not live with the caregiver at the time of the study. Caregivers indicated that they provided an average of 16.9 hours of caregiving per week.

Measures

Where possible, pre-existing scales with high validity were used to measure each variable. Confirmatory factor analyses (CFA) indicated strong loadings of each item on the latent variable being measured. CFA was used instead of coefficient alpha because coefficient alpha overestimates reliabilities when all assumptions are not met; also, CFA offers more conservative estimates, greater flexibility for complex models, and more transparent results (Green & Yang, 2008). Reliabilities reported for each scale are the estimates of the structural equation model implied reliabilities of scales with continuous indicators from the confirmatory factor analysis model.

Role involvement

We measured four aspects of role involvement: behavioral job involvement, psychological job involvement, behavioral caregiving involvement, and psychological caregiving involvement. To normalize the distribution, we used the log of the number of hours spent at work each week and the log of the number of hours spent in caregiving each week to assess behavioral involvement in work and caregiving, respectively (Hepburn & Barling, 1996). Psychological job involvement was assessed using Frone et al.’s (1992) adaptation of the five-item Job Centrality Scale (Kanungo, 1982), a widely used measure of individuals’ beliefs about the value of their present job and its centrality in their life. Respondents indicated on a 6-point scale from completely disagree to completely agree the extent to which they agreed with statements such as “the most important things that happen to me involve my present job.” A job involvement score was calculated by adding the scores for the six items and ranged from 6 to 30, with higher scores reflecting greater job involvement (reliability from CFA = .77). Caregiver involvement was assessed by modifying four of the job involvement items so that they referred to a set of respondents’ caregiving duties. Respondents indicated on a 6-point scale from completely disagree to completely agree the extent to which they agreed with statements such as “the most important things that happen to me involve my caregiving.” A psychological caregiving involvement score was calculated by adding the scores for the four items and ranged from 6 to 24, with higher scores reflecting greater caregiving involvement (reliability from CFA = .72). For this and other scales related to caregiving, the respondents were directed to focus on the caregiving of a specific elder whom they named earlier in the phone interview.

Work-caregiving conflict

We measured four components of work-caregiving conflict: external work interfering with caregiving (external WIC), internal work interfering with caregiving (internal WIC), external caregiving interfering with work (external CIW), and internal caregiving interfering with work (internal CIW). All scales demonstrated strong content validity and high internal reliability. External WIC was a seven-item scale developed using items from Netemeyer, Boles, and McMurrian’s (1996) work-family conflict scale and Carlson and Frone’s (2003) work interference with family scale. The scale measures the extent to which the time and real demands involved in work interfered with caregiving responsibilities and includes items such as “the demands of my work interfere with my caregiving responsibilities.” Respondents rated each item on a 6-point scale from completely agree to completely disagree, with higher scores indicating a greater feeling that work interfered with caregiving (reliability from CFA = .91).

Internal WIC was a two-item scale developed from items in Carlson and Frone’s (2003) work interference with family scale. The scale measures the extent to which the individual thinks about her work responsibilities when interacting with the care recipient and includes items such as “when I am helping [name of care recipient], I think about work-related problems.” Respondents rated each item on a 6-point scale from completely agree to completely disagree, with higher scores indicating a greater feeling that they thought about their work while doing their caregiving (reliability from CFA = .84).

External CIW was a seven-item scale developed from items in Netemeyer et al. (1996) and Carlson and Frone (2003) scales and included items such as “the demands of my caregiving interfere with my work responsibilities.” Respondents rated each item on a 6-point scale from completely agree to completely disagree, with higher scores indicating a greater feeling that work interfered with caregiving (reliability from CFA = .92).

Internal CIW was a two-item scale developed from items in Carlson and Frone’s (2003) family interference with work scale. The scale measures the extent to which the individual thinks about her caregiving responsibilities when working and includes items such as “when I am at work I think about [name of care recipient].” Respondents rated each item on a 6-point scale from completely agree to completely disagree, with higher scores indicating a greater feeling that they thought about their work while doing their caregiving (reliability from CFA = .73).

Work costs

Unlike other measures utilized in this study, “work costs” was conceived as a second order factor that had two different but related indicators of work costs: short term and long term. We assessed short-term work costs or the extent to which an individual experienced interruptions on the job as a result of her caregiving responsibilities. Items were derived from questions about the impact of caregiving on work in small-scale studies and national surveys of caregivers (Arksey, 2002; Georgetown University Center on an Aging Society, 2005; Metropolitan Life Insurance Company, 1999). Respondents indicated for eight items whether or not (scaled as 1 or 0) they had experienced various job interruptions, such as “come late to work,” “miss work,” and “leave work for a doctor’s appointment for the care recipient” because of their responsibilities to the care recipient. The scores on these were summed to yield a total score for the scale, which ranged from zero to eight.

We assessed long-term work costs or the extent to which an individual changed jobs or careers as a result of her caregiving responsibilities. Items were derived from questions about the impact of caregiving on work in small-scale studies and national surveys of caregivers (Arksey, 2002; Georgetown University Center on an Aging Society, 2005; Metropolitan Life Insurance Company, 1999). Respondents indicated for seven items whether or not (scaled as 1 or 0) they had experienced various job changes, such as “change job within the organization” and “change your organization or employer” because of their responsibilities to the care recipient. Each was rated yes or no (1 or 0), and the scores were summed to yield a job change score, which ranged from zero to seven.

Control variables

We included five control variables in our analysis: (1) marital status, which was coded as 1 = married and 0 = unmarried; number of children under the age of 18; age; education, which was coded into seven categories—less than high school, high school graduate or GED, some college, 2-year college degree, 4-year college degree, master’s degree, doctoral degree; and household income, which was coded into four categories—less than $25,000, $25,000 to $50,000, $50,001 to $100,000, and more than $100,000.

Data Analysis Plan

We used Mplus (Muthén & Muthén, 2010) to create a series of structural equation models to test our hypotheses. Mplus provides a generalized latent variable modeling framework that can accommodate both continuous and categorical observed or latent variables. All analyses used a listwise deletion of missing values, which reduced the sample by 57 to 526. A generalized latent variable model is appropriate in this context as the outcome variable, work costs, has binary indicators, which do not follow a normal distribution. The models were estimated using Weighted Least Squares Mean Variance (WLSMV; Muthén, du Toit, & Spisic, 1997). WLSMV is a robust estimator appropriate for models with categorical indicators, which utilizes a diagonal weight matrix and applies a mean and variance adjustment to the chi-square model test statistic. When using the WLSMV estimator, special procedures are required to conduct model comparisons, as the difference in chi-square values are not themselves distributed as chi-square. In Mplus, this is done with “chi-square difference testing” that compares model derivatives. We first assessed only within-domain hypotheses—that is, job involvement was related to WIC and family involvement was related to CIW. We then tested a partial cross-domain model.

Adequacy of model fit was assessed by examining the comparative fit index (CFI), the Tucker-Lewis Index (TLI), and the root-mean-square error of approximation (RMSEA). Desirable models yield CFI and TLI indices with values greater than .90, and RMSEA less than .08. The chi-square/df ratio was also used. It is suggested that a chi-square not larger than 2 to 5 times the degrees of freedom is acceptable (Bollen & Long, 1993). Although there is limited research showing that these thresholds hold when the estimator is WLSMV, as opposed to maximum likelihood, which assumes indicators are distributed multivariate normal, Yu (2002) suggests that they do hold, at least under some circumstances.

Results

Table 1 displays the means, standard deviations, and zero-order correlations among all variables for the sample. Scale reliabilities from the CFA are shown on the diagonal of the correlation matrix.

Table 1.

Means, Standard Deviations, Correlations, and Reliability Coefficients (N = 583)

	M	SD	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.
1. Age	55.61	3.73	—
2. Education	—	—	−.045	—
3. Income	—	—	−.127**	.460**	—
4. Log work hours	44.40	8.34	−.088	.111**	.082	—
5. Psych job involvement	21.11	4.97	−.007	.098*	.050	.182**	.77
6. Log caregiving hours	2.32	1.05	.072	−.044	−.094*	.034	.016	—
7. Psych caregiving involvement	18.77	3.90	.030	−.224**	−.202**	−.086*	.110**	.226**	.72
8. External WIC	18.68	9.16	−.080	.175**	.110**	.080	.065	.052	−.061	.91
9. Internal WIC	5.37	−.075	−.075	.246**	.121**	.113**	.143**	.149**	−.064	.545**	.83
10. External CIW	15.89	7.88	−.071	.188**	.113**	.086*	.077	.204**	−.024	.584**	.603**	.92
11. Internal CIW	8.25	2.42	−.099*	.024	.021	−.052	−.054	.207**	.158**	.363**	.310**	.324**	.73
12. Job interruptions	1.00	2.47	−.011	.057	−.027	.039	−.050	.176**	.112**	.125**	.129**	.239**	.059
13. Job changes	.37	4.60	−.115**	.227**	.192**	−.023	−.024	.135**	−.081	.383**	.443**	.537**	.352**	.179**

Note. Reliabilities are shown on the diagonal of the correlation matrix.

p < .05, ** p < .01.

Hypothesis Testing

As our tests of the study hypotheses, we first modeled the within-domain bivariate relationships between involvement and work-family conflict, as well as the hypothesized relationships with work costs, using job interruptions and job changes as manifest variables (CFI =.912, TLI = .903, RMSEA = .026, chi-square/df ratio 1390.19/1021 = 1.36). The final model, as shown in Figure 2, included the hypothesized set of cross-domain relationships between involvement and work-caregiving interference, as well as the hypothesized relationships with work costs (CFI = .913, TLI = .905, RMSEA = .026, chi-square/df ratio 1380.87/1017 = 1.36). A chi-square test for difference testing showed an improvement from the original model (chi-square = 13.12, df = 4; p < .01). These results suggest confirmation of Hypotheses 1b-d, 2a, 3c, and 4b-d. Control variables had limited significance in the final model, with only the relationships between education and external WIC (β = .16, p < .01), education and internal WIC (β = .26, p < .001), education and external CIW (β = .20, p < .001), age and internal CIW (β = −.11, p < .05), and income and job interruptions (β = .11, p < .05) showing significance.

Figure 2.

Final Model With Parameter Estimates

Tests of the indirect effects confirmed some of the mediating hypotheses. External CIW mediated the relationship between behavioral work involvement and work costs (β = .05, p < .05) and between behavioral caregiving involvement and work costs (β = .09, p < .001). Internal CIW mediated the relationship between psychological job involvement and work costs (β = −.04, p < .05) and between psychological caregiving involvement and work costs (β = .07, p < .01).

Discussion

This study makes several important contributions to the literature. First, it expands the analysis of an individual’s involvement with work and caregiving to include both behavioral and psychological dimensions. Second, it demonstrates a more nuanced understanding of family involvement by specifically focusing on involvement with elder caregiving. Third, it considers the cross-domain influences of involvement on work-caregiving conflict. Finally, it examines a more complex variable of work costs and suggests a mediating relationship between selected aspects of involvement and such costs.

Our work builds on the work of Carlson and Frone (2003) and Barling et al. (1994) by focusing on the elder caregiving domain and examining the links of four types of involvement and four types of work-caregiving conflict to work costs. It adds to the limited literature that studies the impact of family involvement on role conflict (e.g., Aryee et al., 2005; Graves et al., 2007) and on job-related outcomes (e.g., Adams et al., 1996; ten Brummelhuis et al., 2008), which do not focus on work costs but on other job-related outcomes. Findings from our analysis suggest that the relationships among various types of work and family involvement and work-family conflict are complex. Cross-domain relationships exist, specifically from the work domain to the caregiving domain. In particular, behavioral job involvement is associated with external CIW, although the magnitude of the standardized beta is small. Psychological job involvement is associated with internal CIW. Psychological job involvement is the only type of involvement with both within-domain and across-domain consequences for work-elder caregiving conflict. Contrary to our hypothesis, it was negatively associated with CIW. This result suggests that increasing psychological job involvement might offer a way to counteract the demands of caregiving by actually reducing work-elder caregiving conflict. Identification with the job may overshadow the time and psychological demands of caregiving.

Our research indicates that internal and external work-caregiving conflict has different antecedents and consequences. Using conservation of resources and identity theories, we predicted these differences—that is, that behavioral involvement is linked to external conflict, whereas psychological involvement is linked to internal conflict, although the psychological links are stronger than the behavioral ones. Both external and internal CIW seem to play a significant mediating role in the relationship between involvement and work costs. These results suggest that work roles and family roles may differ in their salience to individuals, and perhaps if family is a more important or salient role, then the extent to which work interferes with family may be more apparent or distressing compared to the extent to which family interferes with work.

Although some research has investigated daily disruptions among elder caregivers (e.g., Barling et al., 1994; Barnett, 2005; Wakabayashi & Donato, 2005) and even the effects of part-time employment for elder caregivers (Stone & Short, 1990), our study assesses both job disruptions and job changes in the same study. Furthermore, it examines the relationship of various types of work and family involvement on the job-related outcome of work costs, a focus relatively unique to our study. Our research suggests that behavioral caregiving involvement has direct links to work costs, but so do three types of work-caregiving conflict. Such links have not been previously studied or shown to relate to work costs. The lack of strong direct relationships between both behavioral and psychological job involvement and work costs and between psychological caregiving involvement and work costs suggests that the amount of time devoted to caring for an elder overrides all other types of involvement in importance for affecting work costs.

Finally, our work considers the extent to which work-family conflict plays a mediating role between types of work and caregiving involvement and the work costs of job interruptions and job changes. Whereas previous research has examined the mediating role of work-family and even work-caregiving conflict (Greenhaus, Collins, Singh, & Parasuraman, 1997; Guerts, Kompier, Roxburgh, & Houtman, 2003; Michel et al., 2009; Thomas & Ganster, 1995), this study is the first to show the role of such conflict as a mediator between a more comprehensive set of involvement variables and work costs, which include job disruptions and job changes.

Practically, the results suggest that human resources practices that focus on increasing behavioral job involvement increase CIW. Organizations must find ways to increase psychological job involvement if they want to reduce work-family conflict, but such reduction will only occur in internal CIW. Companies must note that as employees increase their behavioral caregiving involvement—that is, devote more time to caregiving—work costs increase. No other type of involvement has a direct influence on work costs. Finding ways to support elder caregiving, such as through providing benefits that help defer the time required for such caregiving, is one way companies can reduce the impact on the work place.

Limitations and Future Research

This study has several limitations in its sample, measurement, and data analysis, which suggest directions for future research. Although the random selection of the sample is a strength of the research, this study did not compare different occupational, racial, ethnic, educational, income, or age differences. We controlled for marital status, number of children under the age of 18, education, income, and age, but future research should examine the impact of these demographic variables on the relationships studied here.

The study is cross-sectional and so the cause-effect relationships are difficult to determine. Although we suggest that job and caregiving involvement precede work-caregiving conflict which in turn precedes work costs, the order could be reversed, with conflict causing involvement, or work costs causing involvement and conflict. The temporal sequence of job changes is particularly complex because the measurement of past changes was used, rather than intended changes. A longitudinal analysis can help tease apart the causal relationships. In addition, the significant relationships vary in strength. Future research can confirm which links among variables are strongest and most resilient.

Although the reliability of all measures is satisfactory, some, particularly psychological caregiving involvement and internal CIW, could be strengthened. Although we believe that the internal and external WIC and CIW scales showed content validity, as well as strong associations as reflected in exploratory and confirmatory factor analyses, additional tests of the validity of items that compose these scales could occur. The number of items in the internal scales could also be increased. We grouped job interruptions and job changes into the single latent variable of work costs, which may have masked interesting effects on specific interruptions and changes. Analyzing each type of job interruption or job change as a separate dependent variable might also occur.

Future research needs to continue to examine the relationships with different samples, better measures, and more sophisticated analyses. In spite of these limitations, this research suggests that both work and caregiving involvement are significant for work-caregiving conflict and for work costs. The interaction of job involvement and caregiving involvement is an important area of future study within the elder care context.

Conclusion

Increasing numbers of employed caregivers make the interaction of work and elder caregiving a significant concern for both employees and their employers. A better understanding of involvement, conflict, and costs can help create a more productive and humane workplace.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This study was funded by a grant from the National Institute on Aging (Work-Family Conflicts of Older Women, R01 AG020695). We thank Rachel Pruchno, the Principal Investigator of the grant, and Steven Lacey, who assisted with the data analysis for the article.

Bios

Judith R. Gordon is a professor of management in the Carroll School of Management at Boston College and Chairperson of the Management and Organization Department. Her research and publication interests focus on career development of women, the special issues faced by midlife and older working women, work-life balance, and organizational change.

Elizabeth D. Rouse is a doctoral student in organization studies at Boston College. Her research focuses on creators and creativity. She considers both the process of creativity and how new creations emerge, but she also considers the role of creators’ identities in “letting go” to move on to new ventures.

References

Adams

G. A.

King

L. A.

King

D. W.

(1996). Relationships of job and family involvement, family social support, and work-family conflict with job and life satisfaction. Journal of Applied Psychology, 81(4), 411-420.

Allen

T. D.

Herst

D. E. L.

Bruck

C. S.

Sutton

(2000). Consequences associated with work-to-family conflict: A review and agenda for future research. Journal of Occupational Health Psychology, 2(278), 308.

Arksey

(2002). Combining informal care and work: Supporting careers in the workplace. Health and Social Care in the Community, 10(3), 151-161.

Aryee

Srinivas

E. S.

Tan

H. H.

(2005). Rhythms of life: Antecedents and outcomes of work-family balance in employed parents. Journal of Applied Psychology, 90(1), 132-146.

Azarnoff

R. S.

Scharlach

A. E.

(1988). Can employees carry the eldercare burden? Personnel Journal, 9, 60-65.

Barling

(1990). Employment, stress and family functioning. New York: Wiley.

Barling

MacEwen

K. E.

Kelloway

E. K.

Higginbottom

S. F.

(1994). Predictors and outcomes of elder-care-based interrole conflict. Psychology and Aging, 9(3), 391-397.

Barnett

R. C.

(2005). Ageism and sexism in the workplace. Generations, 29(3), 25-30.

Bollen

K. A.

Long

J. S.

(1993). Testing structural equation models. Newbury Park, CA: Sage.

10.

Carlson

D. S.

Frone

M. R.

(2003). Relation of behavioral and psychological involvement to a new four-factor conceptualization of work-family interference. Journal of Business and Psychology, 17(4), 515-535.

11.

Cooper

G. L.

Dewe

O’Driscoll

M. P.

(2001). Organizational stress: A review and critique of theory, research and applications. Thousand Oaks, CA: Sage.

12.

Ford

M. T.

Heinen

B. A.

Langkamer

K. L.

(2007). Work and family satisfaction and conflict: A meta-analysis of cross-domain relations. Journal of Applied Psychology, 92(1), 57-80.

13.

Frone

M. R.

Russell

Cooper

M. L.

(1992). Antecedents and outcomes of work-family conflict: Testing a model of the work-family interface. Journal of Applied Psychology, 77(1), 65-78.

14.

Frone

M. R.

Yardley

J. K.

Markel

K. S.

(1997). Developing and testing an integrative model of the work-family interface. Journal of Vocational Behavior, 50, 145-167.

15.

Georgetown University Center on an Aging Society. (2005). Caregiving and paid work: Are there trade-offs (Data Profile No. 4). Retrieved from http://ihcrp.georgetown.edu/agingsociety/pdfs/CAREGIVERS4.pdf

16.

Graves

L. M.

Ohlott

P. J.

Ruderman

M. N.

(2007). Commitment to family roles: Effects on managers’ attitudes and performance. Journal of Applied Psychology, 92(1), 44-56.

17.

Green

S. B.

Yang

(2008). Commentary on coefficient alpha: A cautionary tale. Psychometrika, 74(1), 121-135.

18.

Greenhaus

J. H.

Beutell

N. J.

(1985). Sources of conflict between work and family roles. Academy of Management Review, 10(1), 76-88.

19.

Greenhaus

J. H.

Collins

K. M.

Singh

Parasuraman

(1997). Work and family influences on departure from public accounting. Journal of Vocational Behavior, 50, 249-270.

20.

Greenhaus

J. H.

Parasuraman

(1994). Work-family conflict, social support and well-being. In Davidson

M. J.

Burke

R. J.

(Eds.), Women in management: Current research issues (pp. 213-229). London: Paul Chapman Publishing Ltd.

21.

Guerts

S. A. E.

Kompier

M. A. J.

Roxburgh

Houtman

I. L. D.

(2003). Does work-home interference mediate the relationship between work load and well-being? Journal of Vocational Behavior, 63(3), 532-559.

22.

Hammer

L. B.

Bauer

T. N.

Grandey

A. A.

(2003). Work-family conflict and work-related withdrawal behaviors. Journal of Business and Psychology, 17(3), 419-436.

23.

Hammer

L. B.

Neal

M. B.

Newsom

J. G.

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.

24.

Hepburn

C. G.

Barling

(1996). Eldercare responsibilities, interrole conflict, and employee absence: A daily study. Journal of Occupational Health Psychology, 1(3), 311-318.

25.

Hobfoll

S. E.

(1989). Conservation of resources: A new attempt at conceptualizing stress. American Psychologist, 44, 513-524.

26.

Hobfoll

S. E.

(2002). Social and psychological resources and adaptation. Review of General Psychology, 6, 307-324.

27.

Huang

Hammer

L. B.

Neal

M. B.

Perrin

N. A.

(2004). The relationship between work-to-family conflict and family-to-work conflict: A longitudinal study. Journal of Family and Economic Issues, 25, 79-100.

28.

Ingersoll-Dayton

Neal

M. B.

Hammer

L. B.

(2001). Aging parents helping adult children: The experience of the sandwiched generation. Family Relations, 30, 262-271.

29.

Kahn

R. L.

Byosiere

(1992). Stress in organizations. In Dunnette

M. D.

Hough

L. M.

(Eds.), Handbook of industrial and organizational psychology (2nd ed., pp. 571-648). Chicago: Rand McNally.

30.

Kanungo

R. N.

(1982). Measurement of job and work involvement. Journal of Applied Psychology, 67(3), 341-349.

31.

MacEwen

K. E.

Barling

(1991). Effects of maternal employment experiences on children’s behavior via mood, cognitive difficulties and parenting behavior. Journal of Marriage and the Family, 53, 635-644.

32.

McCarthy

(1997). 1-800-DIAL-DAYCARE. Business & Health, 15(4), 63(2).

33.

Metropolitan Life Insurance Company. (1999). The met-life juggling act study: Findings from a national study by the National Alliance for Caregiving and the National Center on Women and Aging at Brandeis University. Retrieved from http://www.geckosystems.com/downloads/juggling.pdf

34.

Michel

Mitchelson

J. K.

Kotrba

L. M.

LeBreton

J. M.

Baltes

(2009). A comparative test of work-family conflict models and critical examination of work-family linkages. Journal of Vocational Behavior, 74(2), 199-218.

35.

Muthén

L. K.

Muthén

B. O.

(2010). Mplus statistical analysis with latent variables user’s guide (6th ed.). Los Angeles: Muthén & Muthén.

36.

Muthén

du Toit

S. H. C.

Spisic

(1997). Robust inference using weighted least squares and quadratic estimating equations in latent variable modeling with categorical and continuous outcomes. Unpublished technical report.

37.

Neal

M. B.

Hammer

L. B.

(2007). Working couples caring for children and aging parents: Effects on work and well-being. Mahwah, NJ: Lawrence Erlbaum.

38.

Netemeyer

R. G.

Boles

J. S.

McMurrian

(1996). Development and validation of work-family conflict and family-work conflict scales. Journal of Applied Psychology, 81, 400-409.

39.

Reitzes

D. C.

Mutran

E. J.

Fernandez

M. E.

(1994). Middle-aged working men and women: Similar and different paths to self-esteem. Research on Aging, 16, 355-374.

40.

Stephens

M. A. P.

Franks

M. M.

(1999). Parent care in the context of women’s multiple roles. Current Directions in Psychological Science, 8, 149-152.

41.

Stephens

M. A. P.

Townsend

A. L.

Martire

L. M.

Druley

J. A.

(2001). Balancing parent care with other roles: Interrole conflict of adult daughter caregivers. Journal of Gerontology: Psychological Sciences, 56B(1), P24-P34.

42.

Stone

Cafferata

G. L.

Sangl

(1987). Caregivers of the frail elderly: A national profile. The Gerontologist, 27, 616-626.

43.

Stone

R. I.

Short

P. F.

(1990). The competing demands of employment and informal caregiving to disabled elders. Medical Care, 28(6), 513-526.

44.

ten Brummelhuis

L. L.

van der Lippe

Kluwer

E. S.

Flap

(2008). Positive and negative effects of family involvement on work-related burnout. Journal of Vocational Behavior, 73(3), 387-396.

45.

Thomas

L. T.

Ganster

D. C.

(1995). Impact of family-supportive work variables on work-family conflict and strain: A control perspective. Journal of Applied Psychology, 80(1), 6-15.

46.

Wakabayashi

Donato

K. M.

(2005). The consequences of caregiving: Effects on women’s employment and earnings. Population Research and Policy Review, 24, 467-488.

47.

Wegge

Schmidt

Parkes

van Dick

(2007). ‘Taking a sickie’: Job satisfaction and job involvement as interactive predictors of absenteeism in a public organization. Journal of Occupational and Organizational Psychology, 80(1), 77-89.

48.

C. Y.

(2002). Evaluating cutoff criteria of model fit indices for latent variable models with binary and continuous outcomes. Unpublished doctoral dissertation, University of California, Los Angeles.