Understanding the Correlates between Care-Recipient Age and Caregiver Burden, Work-family Conflict, Job Satisfaction, and Turnover Intentions

Abstract

Although the caregiver literature has explored a wide array of different variables, there is a gap in research on how demographics of the care recipient affect their caregiver. Using data from a diverse sample of 1,007 unpaid caregivers, individuals were separated into four groups based on care recipient age; childcare, adultcare, eldercare, and those with care recipients from multiple age categories. Then, following previous literature, childcare was split into four groups based on the age of the youngest child cared for (0–2 years; 3–5 years; 6–12 years; and 13–17 years). Group differences were found in work-family conflict (time) and five types of caregiver burden. Specifically, time-dependence burden differed most between groups with caregivers of children (0–2 years) having the highest levels and caregivers of children (13–17) having the lowest levels. There were no differences found between groups for job satisfaction, turnover intentions, or family-work conflict. Additional differences were found when considering caregivers’ gender. This research is an exploratory step in understanding how age of care recipient relates to different outcomes for caregivers. Implications are discussed.

Shifting the Burden: A Study of the Effects of Care Recipient Age on Caregivers

Although there is literature focusing on the experience of employed caregivers (e.g., Lilly, Laporte, & Coyte, 2007; Pinquart & Sorensen, 2006), there is a dearth of research examining the effects that caring for different types of recipients has on caregivers. Informal caregiving is specifically defined as providing unpaid support to family members or acquaintances with special needs (Roth, Fredman, & Haley, 2005). A recent survey reported that an estimated 43.5 million adults in the United States provide unpaid care to either an adult or a child (i.e., 39.8 million caring for an adult and approximately 34.2 million caring for an adult age 50 or older; National Alliance for Caregiving, 2015). These numbers are expected to increase as advances in medicine have increased life expectancy (DePasquale et al., 2015). Accordingly, employed individuals are not only having to care for children and/or spouses, but they are also providing care for their aging parents and other family members. Along with an expected rise in the number of individuals providing or needing caregiving, the landscape of who provides care has changed. Women’s labor force participation has expanded significantly since the 1970s (U.S. Department of Labor, 2017), which has led to an increase in dual-earning households. As such, the caregiving role which was formerly associated primarily with women has become more balanced with men taking a more active role (Fox & Brenner, 2012). These trends demand that researchers examine whether men and women experience different outcomes associated with caregiving responsibilities.

The two most common types of care recipient are children and the elderly (Kossek, Colquitt, & Noe, 2001). Several studies have focused on caregiving outcomes associated with these two types of care recipients (e.g., Emslie, Hunt, & Macintyre, 2004). In addition, there has been some focus on ‘sandwiched’ caregivers (adults caring for both children and aging parents; e.g., Hammer & Neal, 2008; Neal & Hammer, 2017). Researchers have found that caregivers juggling multiple caregiving roles (i.e., sandwiched caregivers) tend to experience higher levels of stress and report more absenteeism than those having only one caregiving role (e.g., Chapman, Ingersoll-Dayton, & Neal, 1994; Fredriksen-Goldsen & Scharlach, 2001). However, these studies focused on examining caregiving differences by comparing number of caregiving roles, rather than focusing on caregiving differences based on age of care-recipient.

In another study, the mental health and physical health of grandparent caregivers (individuals raising a grandchild under the age of 18), adult-child caregivers (individuals helping to care for a parent), and spouse caregivers were compared to non-caregivers (Strawbridge, Wallhagen, Shema, & Kaplan, 1997). Another study looked at childcare, eldercare, and ‘sandwiched caregivers’ in IT professionals and found that the sandwiched group exhibited a higher level of family-work conflict (FWC) compared to non-caregivers (DePasquale et al., 2015). Again, the focus of these studies was not to observe differences between care-recipient type, instead the focus was to observe differences between caregivers and non-caregivers. Further, there was no separation of adultcare and eldercare in either of these studies.

Together, these data suggest that caregiving experiences may differ depending on age of the care recipient. Because there is little research in this area, the primary aim of the present study is to examine how caregivers are affected by their caregiving responsibilities depending on care-recipient age. Specifically, we seek to explore whether age of the care recipient (i.e., childcare: 0–2 years; childcare 3–5 years; childcare 6–12 years; childcare 13–17 years; adultcare: 18–49 years; eldercare: 50 years or above; mixed-care: caring for more than one age category) is associated with differences in caregiver burden, work-family conflict, family-work conflict, job satisfaction, and turnover intentions. Delineating differences in effects of caring for individuals of different ages is important because it allows employers to consider offering more comprehensive work-life balance programs within the organization, rather than offering programs that are one-size-fits-all. Understanding differences in burden between caregiver types may enable employers to reduce costs associated with turnover and unproductivity that result from employed caregivers.

Caregiver Burden

Caregiver burden refers to the overall impact of the demands associated with caregiving such as physical, psychological, social, and financial demands (George & Gwyther, 1986). While previous studies have utilized a unidimensional measure of caregiver burden, emerging research has established that caregiver burden is a multi-faceted concept consisting of separate components that elaborate on subjective and objective burden: time-dependence, developmental, physical, social, and emotional burden (e.g., Novak & Guest, 1989; Rospenda, Minich, Milner, & Richman, 2010). Time-dependence burden refers to the perceived impact of caregiving on a caregiver’s time (i.e., number of hours spent caregiving). Developmental burden refers to a caregiver’s perception that caregiving responsibilities put them at a social disadvantage (i.e., missing out on life experiences that their peers experience). Physical burden refers to feelings of fatigue from caregiving. Social burden refers to role conflict (i.e., limiting time and energy to invest in relationships or jobs due to caregiving responsibilities). Lastly, emotional burden refers to a caregiver’s negative feelings about caregiving or the care recipient (Novak & Guest, 1989).

Nearly 1 in 4 caregivers spend 41 hours or more per week providing care to a family member (National Alliance for Caregiving, 2015), making caregiving a full-time job. Caregiving burden has been linked with mental and physical health problems (Clipp & George, 1990; Lieberman & Fisher, 1995; Skaff & Pearlin, 1992) and unhealthy behaviors, such as alcohol misuse (Rospenda et al., 2010). Indeed, numerous studies on informal caregivers have found several negative outcomes associated with caregiving, such as poorer health conditions, poorer quality of life, and increased physical and psychiatric morbidity across different care recipient samples, such as those with dementia, and stroke patients (Baumgarten et al., 1992; Clyburn, Stone, Hadjistavropoulos, & Tuokko, 2000; Connell, Janevis, & Gallant, 2001; Rose-Rego, Strauss, & Smyth, 1998; Schulz, O’Brien, Bookwala, & Fleissner, 1995). Moreover, findings show that as a care recipient’s physical and mental function decreases, perceived caregiver burden and depression increases (Grunfeld, et. al., 2004).

When considering that individuals of different ages require different degrees of care, it is likely that burden-related differences will be found between our care-recipient age groups. For example, following mortality salience literature (e.g., Martens, Greenberg, Schimel, & Landau, 2004), emotional burden may be higher for those caring for the elderly as a fear of losing that person in the near future may be more prevalent. The role reversal of caring for someone who once provided the caregiving may be emotionally taxing, and, it may be burdensome to watch loved ones lose both physical and mental capabilities that they once had (Albert & Brody, 1996; Scharlach, 1994). Physical burden may be less for those caring for adults as higher levels of self-care or independence may be prevalent, whereas time dependence burden may be higher for those caring for children who are unable to yet care for themselves in most capacities. However, given the lack of research in this area to date, and the exploratory nature of the current study, we do not make any specific predictions. Instead, we posit the following research question:

Research Question 1: Does caregiver burden (time-based, developmental-based, physical-based, social-based, emotional-based) differ by care-recipient type (childcare, adultcare, eldercare, and mixed-care)?

Work-to-Family Conflict/Family-to-Work Conflict

In the United States, 70% of individuals report conflict between the responsibilities of work and non-work (Schieman, Glavin, & Milkie, 2009). Greenhaus and Beutell (pg. 77, 1985) defined work-family conflict as “a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect.” It is understood that this construct is bi-directional consisting of work-to-family conflict (WFC) and family-to-work conflict (FWC; e.g., Netermeyer, Boles, & McMurrian, 1996). WFC commonly predicts affective and behavioral outcomes that are family-related while FWC is commonly associated with work-related outcomes (Hammer, Neal, Newsom, Brockwood, & Colton, 2005).

In a survey conducted by the National Caregiving Alliance and AARP (2015), six out of ten employees reported that their caregiving responsibilities affected their work and created a need to adjust their work schedule, take a leave of absence, pass up a promotion, or retire earlier than planned. A large body of research has shown the negative impact of caregiving at the individual (i.e., diminished physical well-being) and organizational (i.e., increased absenteeism) level (National Alliance for Caregiving, 2009; Scharlach & Fredrisken, 1994). Few studies, however, have looked at the relationship between caregiver burden and WFC or FWC, particularly in large, diverse samples. Of the few studies conducted, most focused on specific sub-populations of caregivers and/or of employees. For example, one study on eldercare responsibilities found that FWC was significantly positively correlated with activities of daily living (e.g., personal hygiene and dressing) and managerial activities (e.g., assistance with finances) for women only, and involvement in eldercare acted as a predictor for FWC (Gignac, Kelloway, & Gottlieb, 1996). A second study on eldercare (care recipients ranging from 60 to 99 years old) found that caregivers were more likely to experience interference between their jobs and family responsibilities and more likely to miss work than non-caregivers (Scharlach & Boyd, 1989). Moreover, previous research has found that parental overload has been associated with higher WFC and FWC (Frone et al., 1997). Indeed, employees with children tend to miss more days of work, experience more conflict between work and family, and attend to family concerns more often during the workday than employees without children (Emlen, 1987; Scharlach & Boyd, 1989; Voydanoff, 1988).

Based on the Conservation of Resources theory (COR; Halbesleben, Neveu, Paustian-Underdahl, & Westmas, 2014; Hobfoll, 1989), high caregiving demands may deplete employee psychological resources, which can lead to poorer work performance (Zacher, Jimmieson, & Winter, 2012). The COR theory proposes that WFC/FWC leads to stress because important resources (e.g., time, energy) are lost when individuals have to juggle work and family roles (Hobfoll, 1989). Employed informal caregivers also have to deal with juggling work and caregiving roles, which can potentially threaten the loss of resources. Previous research in eldercare found that high caregiving demands lessened employees’ psychological resources to successfully invest in work-related responsibilities (Barling, MacEwen, Kelloway, & Higginbottom, 1994; Gottlieb, Kelloway, & Fraboni, 1994; Stephens, Townsend, Martire, & Druley, 2001), supporting the appropriateness of examining FWC from within a COR framework. It is also possible that COR theory can apply to WFC scenarios as stressful work environments can deplete psychological resources, which can result in poorer caregiving capabilities. Previous research (e.g., Grandey & Cropanzano, 1999), has looked at WFC/FWC from within the COR framework but we seek to further this previously developed theory. Each type of burden can create, or be considered, a resource deficit. By studying how different types of caregiving are associated with multiple categories of burden there is a deeper understanding of which resources might be useful to ameliorate burden.

Further, the stress-process model of caregiving suggests that stressors can be divided into primary and secondary stressors (Pearlin et al., 1990). Primary stressors stem directly from the care recipient’s needs and the magnitude of those demands (e.g., problematic behavior of care recipient). Primary stressors yield the mobilization of secondary stressors, as they are derived from the hardships and problems caused by the demands of the care recipient (e.g., WFC/FWC). In conjunction with the caregiving time-dependence burden theory, it’s possible that caring for children will yield more interrole conflict than caring for adults or elderly individuals: children demand more caregiving time as they are more likely to be unable to care for themselves at all. For example, children are less likely to be able to drive themselves to a doctor appointment. This is in line with calls for research in the work-life balance literature, as researchers have posited that the “question is not whether WFC is experienced throughout the lifespan, but rather when is this conflict greatest, and what factors explain the relationship between age and WFC” (Huffman, Culbertson, Henning, & Goh, 2013, p. 4). Thus, the present study aims to shed light on any differences that may exist between care-recipient type, in terms of caregiver WFC and FWC. In addition, exploring these differences will enable us to understand whether conflict is greatest when caring for children, adults, elderly individuals, or a combination of age groups.

Research Question 2: Does time-or strain-based work-to-family conflict (WFC) differ by care-recipient type (childcare, adultcare, eldercare, and mixed-care)?

Research Question 3: Does time- or strain-based family-to-work conflict (FWC) differ by care-recipient type (childcare, adultcare, eldercare, and mixed-care)?

Work Outcomes: Job Satisfaction and Turnover intentions

In recent literature, job satisfaction and turnover intentions outcomes have been heavily highlighted as organizations are becoming more interested in these constructs. However, neither construct has been thoroughly explored in relation to caregiving. From the stress-process model perspective, the degree to which primary stress is taxing can yield to secondary stress that may potentially manifest itself in detrimental work outcomes. For example, caring for someone with extreme demands (e.g., a young child) may increase role strain (e.g., FWC), which can potentially ultimately relate to lower job satisfaction and higher turnover intentions. Past research has indicated that role strain (i.e., conflict and ambiguity) intensifies the amount of burnout and job dissatisfaction (e.g., Myung-Yong Um & Harrison, 1998), as well as turnover intentions (e.g., Boyar, Maertz Jr., Pearson & Keough, 2003). From a COR perspective, the extent to which time is a limited resource is a critical focus among employed caregivers, as they may have difficulty allocating time to adequately fulfill both caregiving and job responsibilities. In addition, caregiving responsibilities can exhaust mental and physical resources needed for work. Employees may be more inclined to quit their jobs due to their inability to properly allocate such resources. It is also possible that these responsibilities may relate to a decrease in job satisfaction due to the toll work may have on an individual’s psychological well-being. Moreover, caregiving may also deplete financial resources. Given that a job is typically a family’s main source of income, individuals may experience loss aversion and project those emotions toward their place of employment, resulting in poorer job satisfaction and increased turnover intentions.

Locke (1969) defines job satisfaction as a pleasurable emotional state that results from one’s positive appraisal of one’s job or experiences at work. Balancing the employee role and the role of informal caregiver by taking the responsibility of caring for someone incapable of caring for themselves may impact job satisfaction when time-related resources become strained, as these individuals may negatively appraise their job experiences. Meta-analytic research (Tait, Padgett, & Baldwin, 1989) has found that there is a significant relationship between life satisfaction and job satisfaction, such that life satisfaction may positively influence employee work satisfaction. This is particularly important because satisfaction with both caregiving and work has also been associated with greater individual well-being (Martire, Stephens, & Atienza, 1997). Given that caregiving causes a great deal of personal stress and burden (Haley, Levine, Brown, & Bartolucci 1987; National Alliance for Caregiving, 2015), it would be expected that caregivers will have fewer available mental and physical resources, which could potentially result in lower levels of life and job satisfaction. This is supported by findings that caregiving has been associated with poorer quality of life (Grover & Dutt, 2011; Wong, Lam, Chan, & Chan, 2012) and poorer life satisfaction (Borg & Hallberg, 2006). It is also possible that depletion of time resources can cause strain that manifests in poorer life satisfaction. Therefore, based on previous findings and theoretical support, it is important to investigate whether caring for people of different types of age groups will yield differences in job satisfaction among employed caregivers.

In terms of turnover intentions, caregivers regularly struggle to balance the demands of work and caregiving, resulting in the depletion of cognitive and financial resources. This can result in caregivers changing jobs to a less demanding job, deciding to work closer to home, or finding a job that yields a greater financial return. In more strenuous situations, caregiving has caused many employees to quit their jobs indefinitely so that they are able to fulfill their caregiving responsibilities more appropriately. One study reported that 5–14% of caregivers quit their jobs so that they can work full-time as a caregiver (Stone, Cafferata, & Sangl, 1987). Other studies have reported that 14% to 20% of employed caregivers reported intention to quit their jobs or actually quitting to be able to provide caregiving services (Brody, Kleban, Johnsen, & Hoffman, 1987; Gibeau & Anastas 1989; Muttschler 1994; Petty & Friss 1987; Scharlach & Boyd 1989; Stone & Short 1990). It is possible that caregivers will differ in turnover intentions, as some younger/older recipients might require more mental and/or physical resources from the caregiver. Further, Neal, Chapman, Ingersoll-Dayton, and Emlen (1993) found that caregivers engaging in multiple caregiving roles reported more absenteeism. As such, it is expected that those providing mixed-care (i.e., more than one type of caregiving) will exhibit more turnover intentions.

Thus, job satisfaction and turnover intentions may be impacted by the loss of time-related resources. Specifically, as previously noted, different care-recipient types may require more of a caregiver’s time than others. However, to our knowledge, previous studies have not examined differences based on care recipient age but rather have focused on specific samples. As such, the present study seeks to fill this void in the literature by further exploring if differences exist in job satisfaction and turnover intentions among different care-recipient types.

Research Question 4: Does job satisfaction differ by care-recipient type (childcare, adultcare, eldercare, and mixed-care)?

Research Question 5: Does turnover intentions differ by care-recipient type (childcare, adultcare, eldercare, and mixed-care)?

Gender of Caregiver

The last few decades have seen an increased balance in housework (Bianchi, Milkie, Sayer, & Robinson, 2000) and caregiving roles (Fox & Brenner, 2012). However, despite these achievements, gender differences are still found. Women still tend to be more affected by caregiving roles than men. American mothers still spend nearly twice as much time as fathers per day in household activities and caring for household members (U.S. Bureau of Labor Statistics, 2014a), while men work one-half to one hour longer than women per day (U.S. Bureau of Labor Statistics, 2014b). Previous research has also demonstrated that women are more likely to adjust their work schedules (e.g., leave early, taking time off without pay) to fulfill caregiving responsibilities (Mutschler, 1994). This may be explained in part by wage disparity between men and women. As women currently earn less in comparable roles, by adjusting their schedule, it is possible that they are trying to mitigate financial losses as much as possible. Another possible explanation for these gender differences can be attributed to gender role theory (e.g., Rajadhyaksha, Korabik, & Aycan, 2015), which suggests that the primary domain for men is work, whereas for women it is family. As such men face more severe consequences at work for taking time off to care for the family as it is not perceived to be “men’s” work.

Despite evidence of similarity of WFC perceptions, it has been found that men and women have slight yet significant differences in multiple comparisons (Shockley, Shen, DeNuzio, Arvan, & Knudsen, 2017). Employed mothers report having greater FWC than employed fathers, supporting previous notions that the demands of childcare fall more heavily on women (Byron, 2005; Watai, Nishikido, & Murashima, 2008; Neilson & Stanfors, 2014). Moreover, previous research studies have found that men tend to report higher WFC than women (e.g., Kinnunen & Mauno, 1998; Watai, Nishikido, & Murashima, 2008). Finally, Hammer & Neal (2008) conducted a study on ‘sandwiched caregivers’ and found gender differences, such that women tend to experience higher degree of family-to-work conflict, as well as report higher levels of absenteeism than men. These findings provide support for gender role theories, suggesting it is possible that women will experience higher FWC, whereas men will experience higher WFC.

Furthermore, women tend to hold jobs that are considered “inferior” to those held by men. Nonetheless, women consistently report higher job satisfaction than men (Clark, 1997; Hodson, 1989). Previous data has also shown that men and women value job flexibility differently, such that women value jobs with higher job flexibility (Bender, Donohue, & Heywood, 2005). As women are more likely to seek jobs that offer job flexibility (Bender, Donohue, & Heywood, 2005), women caregivers may have higher degrees of job satisfaction than male caregivers. Having the flexibility to attend to caregiving responsibilities may help alleviate some of the strain associated with caregiving responsibilities.

Lastly, in terms of turnover intentions, some research has suggested that women are less likely to report intentions to quit than men (Moynihan & Landuyt, 2008). Previous studies with women samples have found that job dissatisfaction is an important determinant of turnover intentions (e.g., Kiyak, Namazi, & Kahana, 1997). Since women are more likely to have higher job satisfaction, it would make sense that they would be less inclined to leave their jobs. Further, as previously mentioned, it is also possible that women tend to seek jobs that offer job flexibility (Bender, Donohue, & Heywood, 2005), which can potentially discourage any intentions to quit because of caregiving responsibilities. Due to societal pressures related to gender roles, it is important to examine whether gender differences exist among all outcomes associated with care-recipient type, to allow for greater understanding with less confounding information.

Research Question 6: Are there gender differences in burden (time-based, developmental-based, physical-based, social-based, and emotional-based), WFC, FWC, job satisfaction, and turnover intentions between care-recipient type (childcare, adultcare, eldercare, and mixed-care) groups?

Method

Participants

Data were collected by the Survey Research Laboratory at the University of Illinois-Chicago. Participants were employed adults with unpaid caregiving responsibilities identified through purchasing random phone numbers from randomly selected block groups in the greater Chicago metropolitan area (2006–2008). All contacted participants were screened for eligibility (employed 20+ hours/week in past 12 months and having unpaid caregiver responsibilities). There were initially 2,114 caregivers who agreed to be mailed a survey, of which 1,007 returned the questionnaire (response rate of 47.2%). Participants received a $30 American Express gift card incentive. The average age was 43.0 years (SD = 9.9) while the ethnicity breakdown was 42.7% White, 37.0% African American, 16.0% Latino/a, and 4.3% American Indian, Alaskan Native, Asian, Native Hawaiian, or other race/ethnicity (0.7% missing). The majority of individual’s provided caregiving in their own home (79%) but some individuals were cared for in the care recipient’s home (17%) or in a care facility (4%). Further, the majority of individual’s worked during the day (62%) while some individual’s worked during the afternoon (8%) or at night (8%) while 21% either didn’t reply or worked an alternative type of shift (e.g., split or rotating shift).

There are inconsistencies in the literature regarding the cut-off ages of care recipient groups. We chose a cut-off score of 17 years for childcare, as individuals in this age group typically still live at home and receive parental care. Further, at age 18, individuals are legally treated as an adult. In terms of eldercare, we chose a cut off score of 50 because this is the age at which individuals can start receiving elderly-relevant benefits in the United States (AARP, 2015). This definition of eldercare is also supported by previous research (Anastas, Gilbeau, & Larson, 1990). Consequently, our adultcare bracket consists of individuals between these two groups. Finally, reviewers suggested for the childcare group to be more thoroughly assessed by delving into whether differences exist depending on the age of the child. One important point is that the demands associated with taking care a toddler will vary drastically from the demands associated with caring for a teenager. For example, a teenager is more likely to be able to feed themselves, whereas a child will depend on caretakers to be fed. As such, we decided to break the childcare category into four distinct age brackets focusing specifically on age of youngest child: 0–2 years; 3–5 years; 6–12 years; and 13–17 years. Existing caregiving literature (e.g., Baltes and Young, 2007; Erickson, Martinengo, & Hill, 2010; Allen & Finkelstein, 2014) supports the use of these age categories. Further, these age brackets are in alignment with Erikson’s stages of development. Seven comparison groups were created: caring for only children with the youngest child in each age grouping (0–2 years; n = 148), (3–5 years; n = 96), (6–12 years; n = 133), (13–17 years; n = 68), only caring for adults aged 18–49 years (adultcare; n = 54), only caring for elder-adults 50 years and over (eldercare; n = 115), and caring for multiple age ranges (mixed-care; n = 305).

Measures

Burden.

Caregiver burden was measured using a modified 19-item version of the Caregiver Burden Inventory (Novak & Guest, 1989) which measures burden in 5 areas: time-dependence (e.g., “I don’t have a minute’s break from my caregiving chores when I’m not at work”), developmental (e.g., “I feel that I am missing out on life because of caregiving”), physical (e.g., “My health has suffered because of caregiving”), social (e.g., “I’ve had problems with my spouse/partner because of caregiving”), and emotional (e.g., “I resent my care recipient(s)”). Each scale was measured by 4-items except for physical (3-items) and scored on a scale from 1 (not descriptive at all) to 4 (very descriptive).

WFC/FWC.

The 22-item Work-Family Conflict Scale (Kelloway, Gottlieb, & Barham, 1999) was used to assess conflict between work and family roles in one’s current job. The scale contains subscales for time-based WFC (5-items; e.g. “job demands keep me from spending the amount of time I would like with my family”), strain-based WFC (6-items; e.g., “I think about work when I am home”), time-based FWC (5-items; e.g., “My family demands interrupt my workday”), and strain-based FWC (6-items; “When I am at work, I am distracted by family demands”). All items were scored on a scale from 1 (never) to 4 (almost always).

Satisfaction.

Job Satisfaction was measured using 5-items (e.g., “I feel fairly well satisfied with my present job”) from the Brayfield-Rothe (1951) measure of job satisfaction as modified by Judge, Locke, Durham, and Kluger (1998) using a scale ranging from 1 (strongly disagree) to 10 (strongly agree).

Turnover Intent.

Intent to change jobs within the next 12 months was measured by 3-items from Konovsky and Cropanzano (1991) (e.g., “How often do you think about quitting your job”) scored on a scale from 1 (hardly ever) to 7 (very often).

Results

Bivariate correlations and Cronbach’s alpha reliability coefficients for study variables can be found in Table 1. One-way ANOVA with Tukey post-hocs were used to examine differences in caregiver outcomes by care recipient age group. Group differences were found for time-dependence (F(6, 877) = 48.93, p < .001), developmental (F(6,873) = 8.90, p < .001), physical (F(6,875) = 6.49, p < .001), social (F(6,864) = 4.08, p < .001), and emotional (F(3,868) = 6.23, p < .001) burden (see Table 2 for details). For research question 2, differences between groups were not found for WFC-strain but were found for WFC-time (F(6,883) = 2.65, p < .05). For research question 3, marginal differences were found for FWC-time (F(6,865) = 1.86, p = .085) but no differences were found for FWC-strain. No differences between groups were found for job satisfaction (research question 4) or turnover intent (research question 5).

Table 1.

Zero-order correlations.

	1	2	3	4	5	6	7	8	9	10	11
1. TI	(.83)
2. JS	−.64**	(.88)
3. TDB	.04	−.00	(.85)
4. DB	.10**	−.08*	.32**	(.85)
5. PB	.04	−.04	.39**	.55**	(.86)
6. SB	.09**	−.10**	.24**	59**	.57**	(.73)
7. EB	.05	−.01	.02	.51**	.27**	.46**	(.77)
8. WFCT	.19**	−.16**	.15**	.23**	.25**	.27**	.07*	(.84)
9. WFCS	.25**	−.21**	.14**	.33**	.28**	.32**	.14**	.66**	(.89)
10. FWCT	.08*	−.05	.24**	.42**	.38**	.37**	.15**	.50**	.54**	(.89)
11. FWCS	.11**	−.14**	.14**	.41**	.40**	.43**	.22**	.49**	.60**	.69**	(.82)

Note:TI = turnover intentions, JS = job satisfaction, TDB = time-dependence burden, DB = developmental burden, PB = physical burden, SB = social burden, EB = emotional burden, WFCT = work-family conflict time, WFCS = work-family conflict strain, FWCT = family-work conflict time, FWCS = family-work conflict strain. Cronbach’s coefficient alpha reliability for each variable is in the parentheses along the diagonal.

^*= significant at .05,

^**= significant at the .01, n = 953.

Table 2.

Differences between caregiving type groups.

	0–2 years(n = 148)	3–5 years(n = 96)	6–12 years(n = 133)	13–17 years(n = 68)	Adultcare(n = 54)	Eldercare(n = 115)	Mixed-care(n = 305)
Age of caregiver	35.28 (6.78)	38.01 (7.39)	39.71 (7.93)	48.56 (7.08)	46.91 (10.86)	49.36 (10.77)	42.39 (9.55)
Household Income
10,000 or less	4.8	2.1	10.8	4.8	9.6	4.5	7.8
10,001 – 20,000	6.2	8.5	9.2	6.3	5.8	4.5	8.2
20,001 – 30,000	8.3	12.8	8.5	7.9	17.3	9.8	10.5
30,001 – 50,000	10.3	10.6	19.2	23.8	23.1	24.1	22.8
50,001 – 70,000	11.7	12.8	15.4	11.1	11.5	18.8	16.7
70,001 – 90,000	16.6	14.9	8.5	14.3	13.5	12.5	11.2
90,001 or above	42.1	38.3	28.5	31.7	19.2	25.9	22.8
Marital status
Married	82.4	75.8	46.2	45.5	37.0	38.1	60.7
Committed	8.8	9.5	13.6	10.6	24.1	16.8	12.3
Widow/Divorce/Separated	5.4	7.4	19.7	34.8	27.8	23.9	16.0
Never Married/ Single	3.4	7.4	20.5	9.1	11.1	21.2	11.0
Ethnicity
White	60.0	57.3	39.0	44.1	44.4	48.7	34.1
Black	20.0	20.8	38.2	42.6	44.4	36.5	42.0
Latino/a	16.0	18.8	21.3	10.3	9.3	7.8	17.4
Asian, AI/AN/NH	4.0	3.1	1.5	2.9	1.9	7.0	6.6
Caregiver Burden
Tme-dependence	3.41 (.76)234567	2.80 (.84)134567	2.33 (.83)124	1.68 (.50)12367	2.11 (.89)	2.17 (.83)124	2.38 (.88)124
Developmental	1.56 (.50)346	1.52 (.50)6	1.34 (.50)16	1.28 (.50)16	1.43 (.57)6	1.81 (.82)123457	1.49 (.58)6
Physical	1.58 (.58)3457	1.47 (.53)4	1.29 (.58)1	1.15 (.31)1267	1.25 (.38)1	1.41 (.65)4	1.40 (.56)14
Social	1.41 (.51)	1.44 (.45)	1.31 (.50)6	1.25 (.34)67	1.28 (.39)	1.54 (.62)34	1.47 (.58)4
Emotional	1.07 (.28)67	1.13 (.29)6	1.10 (.29)6	1.07 (.20)6	1.26 (.40)	1.31 (.50)1234	1.21 (.47)1
Work-family Conflict Time	2.37 (.76)	2.55 (.72)5	2.29 (.78)	2.45 (.78)	2.04 (.66)1	2.30 (.84)	2.38 (.82)
Work-family Conflict Strain	2.37 (.62)	2.44 (.65)	2.24 (.63)	2.39 (.70)	2.22 (.68)	2.38 (.73)	2.35 (.69)
Family-work Conflict Time	2.02 (.71)	2.09 (.68)	1.93 (.73)	1.83 (.69)	1.74 (.65)	2.01 (.79)	1.96 (.71)
Family-work Conflict Strain	1.93 (.60)	2.01 (.62)	1.90 (.69)	1.92 (.67)	1.74 (.62)	2.07 (.76)	1.98 (.68)
Job Satisfaction	3.66 (.90)	3.85 (.72)	3.72 (.88)	3.60 (.93)	3.86 (.88)	3.75 (.89)	3.61 (.88)
Turnover Intentions	3.15 (1.97)	3.18 (1.92)	3.22 (1.99)	3.18 (2.04)	2.76 (1.82)	3.01 (1.90)	3.29 (1.86)

Note:Total n = 822 – 915. Ethnicity, Marital status, and Household income are percentages while the others are the mean. Standard deviations are presented in parentheses. AI is American Indian, AN is Alaskan Native, and NH is Native Hawaiian or other Pacific Islander. Group means with difference subscripts differ significantly at the p < .05. Subscripts: 1 = childcare 0–2 years, 2 = childcare 3–5 years, 3 = childcare 6–12 years, 4 = childcare 13–17 years, 5 = adultcare, 6 = eldercare, and 7 = mixed-care, respectively.

To examine Research Question 6, analyses were run for gender separately. Sample size is small for the adultcare group and interpretations are limited. For men, time-dependence (F(6,400) = 22.82, p < .001), developmental (F(6,398) = 3.80, p = .001), physical (F(6,400) = 3.57, p < .005), and emotional (F(6,389) = 2.25, p < .05) burden were significantly different between care recipient age groups, but social burden was not. WFC-time was marginally different (F(6,402) = 2.04, p = .059) while job satisfaction, turnover intent, WFC-strain, and FWC (time and strain) showed no differences between groups (see Table 3).

Table 3.

Differences between caregiving type groups split by gender.

Men	Childcare 0–2 years (n = 80)	Childcare 3–5 years (n = 58)	Childcare 6–12 years (n = 52)	Childcare 13–17 years (n = 29)	Adultcare 18–49 years (n = 26)	Eldercare 50+ years (n = 43)	Mixed-care (n = 135)
	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)
Caregiver Burden
Time-dependence	3.14 (.88)34567	2.75 (.78)34567	2.09 (.85)12	1.60 (.51)127	2.09 (.88)12	1.93 (.67)12	2.18 (.84)124
Developmental	1.50 (.49)	1.49 (.53)	1.24 (.44)6	1.27 (.35)6	1.50 (.64)	1.64 (.75)347	1.36 (.42)6
Physical	1.52 (.61)34	1.41 (.47)	1.17 (.52)1	1.15 (.35)1	1.32 (.43)	1.27 (.40)	1.33 (.51)
Social	1.34 (.49)	1.39 (.41)	1.27 (.50)	1.25 (.32)	1.29 (.44)	1.52 (.61)	1.39 (.46)
Emotional	1.06 (.22)	1.08 (.19)	1.11 (.38)	1.04 (.09)	1.22 (.34)	1.23 (.48)	1.17 (.42)
Work-family Conflict
Time	2.51 (.73)	2.52 (.72)	2.21 (.78)	2.62(.74)	2.13 (.62)	2.32 (.87)	2.47 (.75)
Strain	2.44 (.58)	2.43 (.69)	2.14 (.67)	2.48 (.75)	2.43 (.64)	2.36 (.71)	2.41 (.64)
Family-work Conflict
Time	1.94 (.66)	2.01 (.70)	1.75 (.65)	1.88 (.65)	1.77 (.65)	2.00 (.76)	1.90 (.62)
Strain	1.89 (.57)	1.97 (.64)	1.77 (.67)	1.94 (.63)	1.88 (.63)	2.00(.66)	1.97 (.64)
Job Satisfaction	3.36 (1.00)	3.88 (.73)	3.88 (.89)	3.45 (.89)	3.66 (1.01)	3.76 (.99)	3.69 (.88)
Turnover Intent	3.05 (1.88)	3.01 (1.84)	2.96 (1.98)	3.52 (2.14)	3.38 (1.95)	2.73 (1.82)	2.96 (1.76)
Women	(n = 69)	(n = 37)	(n = 82)	(n = 39)	(n = 28)	(n = 72)	(n = 166)
	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)	Mean(SD)
Caregiver Burden
Time-dependence	3.73 (.42)234567	2.91 (.94)1456	2.53 (.76)14	1.75 (.51)12367	2.13 (.90)12	2.31 (.89)124	2.53 (89)14
Developmental	1.64 (.52)	1.55 (.48)	1.41 (.53)6	1.31 (.60)6	1.37 (.50)6	1.92 (.86)3457	1.59 (.68)6
Physical	1.66 (.55)45	1.56 (.63)4	1.37 (.60)	1.14 (.29)12	1.19 (.32)1	1.50 (.76)	1.46 (.59)
Social	1.49 (.52)	1.53 (.50)	1.34 (.50)	1.26 (.37)	1.27 (.36)	1.55 (.63)	1.53 (.66)
Emotional	1.09 (.35)5	1.19 (.40)	1.09 (.23)5	1.09 (.25)5	1.30 (.45)	1.35 (.51)134	1.24 (.52)
Work-family Conflict
Time	2.22 (.77)	2.56 (.72)	2.36 (.77)	2.33 (.80)	1.98 (.70)	2.29 (.83)	2.31 (.88)
Strain	2.31 (.67)	2.47 (.60)	2.33 (.58)	2.32 (.67)	2.07 (.67)	2.39 (.76)	2.32 (.73)
Family-work Conflict
Time	2.10 (.77)	2.23 (.61)	2.06 (.75)	1.80 (.72)	1.71 (.66)	2.01 (.82)	2.01 (.77)
Strain	1.97 (.64)	2.08 (.61)	2.00 (.69)	1.90 (.72)	1.64 (.61)	2.12 (.82)	1.98 (.71)
Job Satisfaction	3.68 (.76)	3.77 (.72)	3.60 (.84)	3.70 (.97)	4.08 (.69)	3.74 (.85)	3.56 (.88)
Turnover Intentions	3.33 (2.08)	3.47 (2.06)	3.38 (1.95)	2.94 (1.99)	2.22 (1.55)7	3.17 (1.94)	3.61 (1.92)5

Note: Men are above the dotted line and women are below the dotted line. * For time-dependence burden, adultcare is marginally higher (p = .053). ** For social burden, the differences are marginal (p = .055). *** For turnover intent, the difference is marginal for childcare. Standard deviations are presented in parentheses. Subscripts indicate that the value is significantly different from the indicated subscript value. Subscript 1, 2, 3, 4, 5, 6, 7 are childcare 0–2 years, childcare 3–5 years, childcare 6–12 years, childcare 13–17 years, adultcare, eldercare, and mixed-care, respectively.

For women, time-dependence (F(6,462) = 32.80, p < .001), developmental (F(6,459) = 5.92, p < .001), physical (F(6,459) = 4.22, p < .001), social (F(6,455) = 2.63, p < .05), and emotional (F(6,463) = 3.80, p = .001) burden were significantly different for care recipient age groups. There were no differences between groups found for WFC or FWC. There was a significant difference between groups for turnover intent (F(6,421) = 2.14, p < .05) but no difference between groups for job satisfaction.

Supplemental Analysis

T-tests revealed several differences between how men and women experience caregiving burden, WFC, FWC, and turnover intent. Men experienced lower levels of: (1) time-dependence burden for childcare: 0–2 years (p < .001), eldercare (p < .05), and mixed-care (p < .001); (2) developmental burden for mixed-care (p < .001); (3) physical burden for childcare: 6–12 years (p = .05) and mixed-care (p < .05); (4) social burden for mixed-care (p < .05); (5) FWC-time for childcare: 6–12 years (p < .05); (6) turnover intentions for mixed-care (p < .005). Men experienced higher levels of: (1) WFC-strain for adultcare (p < .05); (2) turnover intentions for adultcare (p < .05).

To aide in interpretation, Cohen’s D effect sizes were computed for all research questions but only variables with significant effect sizes (.20 or greater; Cohen, 1988) are presented (see Tables 4 & 5). Each group (childcare: 0–2, childcare: 3–5, childcare: 6–12, childcare: 13–17, adultcare, eldercare, and mixed-care) was compared for each variable of interest. This resulted in 693 comparisons. Of these comparisons, 392 (approximately 57%) were significant (Cohen’s D value of .20 or higher) with a total of 265 small effect sizes (.20-.49), 92 medium effect sizes (.50-.79) and 35 large effect size (.80 or greater).

Table 4.

Cohen’s D of caregiving type groups.

All	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
TD	.76	1.36	2.69	1.57	1.56	1.25	.56	1.62	.80	.75	4.9	.95	.26			.60	.72	.98		.31	.25
Dev		.44	.56	.24	.37		.36	.48		.43				.69	.28	.28	.78	.39	.54		.45
Phy	.20	.50	.92	.67	.28	.32	.32	.74	.48			.30				.29	.51	.55	.30	.31
Soc		.20	.37	.29	.23		.27	.48	.38					.41	.30		.58	.46	.50	.38
Emo	.21			.55	.59	.36		.24	.37	.44	.20		.46	.51	.28	.60	.63	.39			.21
WFCT	.24			.46			.35		.74	.32	.22	.21	.35			.57			.34	.46
WFCS		.21		.23			.31		.33			.23		.21		.25			.23
FWCT			.27	.41			.23	.38	.53				.27				.24		.37	.32
FWCS				.31	.20				.44				.24	.23		.28	.21		.48	.37
JS	.23			.22				.30			.30					.29				.28
TI				.21					.22				.24							.29

Note:TD is time-dependence burden, Dev is developmental burden, Phy is physical burden, Soc is social burden, Emo is emotional burden, WFCT is work-family conflict time, WFCS is work-family conflict strain, FWCT is family-work conflict time, FWCS is family-work conflict strain, JS is job satisfaction, and TI is turnover intentions. Columns 1, 2, 3, 4, 5, 6 are childcare 0–2 years compared to childcare 3–5 years, childcare 6–12 years, childcare 13–17 year, adultcare, eldercare, and mixed-care, respectively. Columns 7, 8, 9, 10, 11 are childcare 3–5 years compared to childcare 6–12 years, childcare 13–17 years, adultcare, eldercare, and mixed-care, respectively. Columns 12, 13, 14, 15 are childcare 6–12 years compared with childcare 13–17 years, adultcare, eldercare, and mixed-care, respectively. Columns 16, 17, 18 are childcare 13–17 compared to adultcare, eldercare, and mixed-care, respectively. Columns19, 20 are adultcare compared to eldercare and mixed-care, respectively. Column 21 is eldercare compared to mixed-care. Cohen’s D effect sizes above .20-.49 are considered small, .50-.79 is medium (bold/italicized), and .80 or above is large (bold/underlined).

Table 5.

Cohen’s D of caregiving type groups split by gender.

Men	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
TD	.47	1.21	2.14	1.19	1.55	1.12	.81	1.75	.79	1.13	.70	.70		.21		.68	.68	.83	.20		.33
Dev		.56	.54		.22	.31	.51	.49		.23	.27		.47	.65	.28	.45	.63	.23	.20	.26	.46
Phy	.20	.62	.74	.38	.48	.34	.48	.63	.20	.32			.31	.22	.31	.43	.32	.41
Soc			.22		.33		.26	.38	.24	.25				.45	.25		.55	.35	.43	.22	.24
Emo				.56	.46	.33		.27	.51	.41	.28	.25	.31	.28		.72	.55	.43
WFCT		.40		.56	.24		.41		.58	.25		.54			.34	.72	.37	.20	.25	.49
WFCS	.25			.36			.24	.24	.63		.22		.41			.37			.45	.36
FWCT			.40	.54			.25	.64	.82	.30	.32	.35	.50				.27	.28	.40	.42
FWCS				.53	.20			.27	.72				.55			.39	.29		.66	.51
JS				.55			.22		.44		.26		.62			.45			.44	.66	.21
TI				.61				.26	.69			.22	.66			.40		.34	.54	.80	.23
Women	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
TD	.1.13	.95	4.24	2.28	2.04	1.72	.44	1.53	.85	.66	.42	1.21	.48	.27		.52	.77	1.08	.20	.45	.25
Dev		.44	.59	.53	.39		.28	.44	.37	.53				.71	.30		.82	.44	.78	.37	.43
Phy		.50	1.18	1.04	.24	.35	.31	.86	.74			.49	.37				.63	.69	.53	.57
Soc		.29	.51	.49			.38	.61	.60					.37	.32		.56	.50	.55	.49
Emo	.27			.52	.59	.34	.31	.30	.26	.35			.59	.66	.37	.58	.65	.37			.21
WFCT	.46			.33			.27	.30	.82	.35	.31		.52			.47			.40	.42
WFCS	.25			.36			.24	.24	.63		.22		.41			.37			.45	.36
FWCT			.40	.54			.25	.64	.82	.30	.32	.35	.50				.27	.28	.40	.42
FWCS				.53	.20			.27	.72				.55			.39	.29	.53	.66	.51
JS				.55			.22		.44		.26		.62			.45			.44	.66	.21
TI				.61				.26	.69			.22	.66			.40		.34	.54	.80	.23

Note: TD is time-dependence burden, Dev is developmental burden, Phy is physical burden, Soc is social burden, Emo is emotional burden, WFCT is work-family conflict time, WFCS is work-family conflict strain, FWCT is family-work conflict time, FWCS is family-work conflict strain, JS is job satisfaction, and TI is turnover intentions. Columns 1, 2, 3, 4, 5, 6 are childcare 0–2 years compared to childcare 3–5 years, childcare 6–12 years, childcare 13–17 year, adultcare, eldercare, and mixed-care, respectively. Columns 7, 8, 9, 10, 11 are childcare 3–5 years compared to childcare 6–12 years, childcare 13–17 years, adultcare, eldercare, and mixed-care, respectively. Columns 12, 13, 14, 15 are childcare 6–12 years compared with childcare 13–17 years, adultcare, eldercare, and mixed-care, respectively. Columns 16, 17, 18 are childcare 13–17 compared to adultcare, eldercare, and mixed-care, respectively. Columns19, 20 are adultcare compared to eldercare and mixed-care, respectively. Column 21 is eldercare compared to mixed-care. Cohen’s D effect sizes above .20-.49 are considered small, .50-.79 is medium (bold/italicized), and .80 or above is large (bold/underlined).

In an attempt to be thorough, socio-demographic variables were tested to see if differences occurred for the variables of interest. Only when a socio-demographic variable was significantly related to a variable of interest was it controlled for in an ANCOVA. Age, hours worked, marital status, household income, and ethnicity were significantly related to various variables and were controlled for accordingly. Although a direct measurement of caregiver hours was not assessed, a question did ask about the amount of caregiving provided by the participant compared to other members of the household. Response options ranged from (1) all/almost all caregiving is done by someone else to (5) all/almost all caregiving is done by me. Flextime was also considered. This was assessed by asking “how frequently do you use flextime” and was rated as (0) flextime is not available, (1) flextime is available but have never used, (2) flextime is available and used sometimes, and (3) flextime is available and used frequently. Next, we considered if caring for an individual with a disability or illness had differing effects on our variables of interest. When there was a significant relationship between flextime, caregiver comparative hours, or caring for an individual with a disability/illness and any of the variables of interest, it was controlled for. Outputs of supplemental analyses are available upon request.

Time-dependence, developmental, physical, and social burden, WFC-strain, FWC-strain, job satisfaction, and turnover intentions showed little to no difference when controlling for relevant socio-demographic variables. When controlling for age, marital status, and whether an individual is caring for someone with a disability/illness or not, emotional burden became non-significant. When controlling for hours worked, comparative caregiver hours, age, quadratic age, income, and ethnicity, WFC-time became non-significant. When controlling for comparative caregiver hours, use of flextime, income, ethnicity, and whether an individual is caring for someone with a disability/illness or not, FWC-time went from marginally significant (p = .085) to significant (F(11, 742) = 2.30, p < .05; see Table 5).

For men, time-dependence burden, developmental burden, physical burden, social burden, WFC-strain, FWC-time, FWC-strain, job satisfaction, and turnover intentions showed little to no difference. When controlling for marital status and whether an individual cared for someone with a disability/illness or not, emotional burden became non-significant. When controlling for hours worked, income, and ethnicity WFC-time went from marginally significant to non-significant.

For women, time-dependence burden, developmental burden, WFC-time, WFC-strain, FWC-time, FWC-strain, and job satisfaction showed little to no difference. When controlling for comparative caregiver hours, use of flextime, age, income, ethnicity, and whether an individual is caring for someone with a disability/illness or not, physical burden became non-significant. When controlling for use of flextime, ethnicity, and whether an individual is caring for someone with a disability/illness or not, social burden became non-significant. When controlling for whether an individual is caring for someone with a disability/illness or not, emotional burden became non-significant. When controlling for use of flextime, age, marital status, income, and ethnicity, turnover intentions became non-significant.

Discussion

The present paper focused on exploring whether differences in caregiver burden, WFC, FWC, job satisfaction, and turnover intentions exist between caregivers of different care recipient groups (e.g., childcare, adultcare, eldercare, & mixed-care). Although there is a wealth of knowledge on how caregiving affects an individual’s professional and personal life (e.g., Baumgarten et al., 1992; Connell et al. 2001), there is still much to learn on how individuals are affected by the roles in which they partake. To our knowledge, no previous study had investigated differences in the aforementioned outcomes between these care recipient age groups. Therefore, by using a diverse sample of caregivers, this paper was able to begin examining potential differences in caregiver outcomes by care-recipient types, rather than simply in comparison to non-caregivers.

Significant differences were found for all caregiver burden categories measured and for WFC-time and marginal differences for FWC-time, however, no differences were found for turnover intentions or job satisfaction. It is possible that turnover intentions and job satisfaction are just too distally located, or in terms of the stress-process model, it is likely that secondary stressors are less related to outcomes than primary stressors. This may also explain why smaller differences were found for WFC and FWC. It is likely that as distance from the original stress of caregiving increases, weaker differences between types of caregiving are found (DePasquale et al., 2015). Effects may be even more difficult to detect due to the sample size of the adultcare group (n = 54).

When assessing the differences found, it is clear that that those providing childcare for ages 0–2 years and 3–5 years experiences the highest levels of time-dependence burden while childcare 13–17 years had the lowest levels. Similarly, for physical burden, childcare 0–2 years and 3–5 years had the highest levels while childcare 13–17 years had the lowest levels. These findings are in support of COR theory, as children are likely to be the most care-dependent group. Children aged 13–17 years are also more likely to want independence from their caregivers as they develop their identities. For developmental burden, or rather the perception that caregiving responsibilities put an individual at a social disadvantage, those providing eldercare had the highest levels. This may be in line with mortality salience literature as caring for and being around those who are older may remind individual’s of the limited time left to spend with them. As such, individuals may feel more obligated to spend more time with these older family members and/or acquaintances and have limited time to spend with their friends. This is partly supported by the data that we have on age of caregiver. As the average age of those providing eldercare is almost fifty years old, it is possible that these individuals would be experiencing the empty nest phase of life, but instead are spending the time that could have been available for leisure activities (e.g., socializing with friends) on taking care of elderly parents or relatives. This may cause these individuals to feel that they are missing out when it comes to their social life. Further it is quite possible that elderly individuals may have needs that are routine and time sensitive, such as taking medication. It is also unlikely that caregivers will be able to bring older care recipients out with them to different events, as one may be able to do with a child. This finding is not as surprising as prior research among a sample of elderly care recipients suggests that developmental burden relates more specifically to the elderly (Razani et. al., 2007).

For social burden, those performing childcare for ages 13–17 had the lowest levels followed by those performing adultcare. Only those performing childcare for ages 13–17 displayed significantly lower levels of social burden than those performing mixed-care and eldercare. Further, childcare for 6–12 years was significantly lower than eldercare which had the highest level of social burden. As social burden refers to work-life balance conflict or role interference it is understandable why these groups vary the way they do. In the United States, childcare may be the most socially normative and common type of caregiving. As such, individuals caring for children are likely to have greater support from colleagues, spouses, and work policies that would decrease burden for this particular population. Although eldercare has been well-studied, it may still be stigmatized in social settings. As such, individuals may be perceived as still living at home or may have routine commitments that limit their ability to work. If a large number of workers performing eldercare are included in the mixed-care group, this might explain the higher burden found for this group of caregivers as well, or simply the fact these individuals have more people in their care.

Similarly, emotional burden was significantly higher for eldercare than for all groups of childcare. It is possible that individuals have higher levels of emotional burden when providing eldercare due to the possibility of death in the near/immediate future which would support mortality salience literature (e.g., Martens, Greenbarg, Schhimel, & Landau, 2004). It may also be possible that childcare creates the least amount of emotional burden, as it is considered “normal and expected” for adults to have to care for children. Further, children are expected to be less troublesome as they grow and develop both mentally and physically. One research participant who was caring for both a young child and her aging mother called the researchers to discuss the study, providing some context for how emotional burden can differ by age of the care recipient. She described how caring for her young child was time consuming, but not as emotionally demanding as caring for her aging mother, because she knows that her child will only become more self-sufficient, while her aging mother will only become less self-sufficient.

It should be noted that significant differences were found when splitting caregiving type by caregiver gender, but as sample sizes were smaller (n = 26 – 166), only some were found to be different when looking at post-hoc results. For example, social burden was not significantly different for men or women even though some differences were found when not splitting by gender.

WFC-time differences were found when comparing childcare ages 3–5 years and those performing adultcare. It is possible that although adultcare recipients require caregiving, they are more self-sufficient than these children, or that caring for an adult who might otherwise be taken care of by a third party strengthens the family unit and/or role.

Results differed when focusing on each gender separately. Women were found to have significant differences between care-recipient type groups for turnover satisfaction but only when comparing adultcare and mixed-care. Some research has suggested that women are less likely to report intentions to quit than men (Moynihan & Landuyt, 2008), other research yielded different results (i.e., Selvarani & Chandra, 2014). It could be that determinants of turnover intentions related to caregiving may be different for men and women and be situation-specific.

When comparing men and women, women experienced higher levels of time-dependence, developmental, physical, and social burden for some categories. Women were also found to have a higher level of FWC-time for the childcare 6–12 years and higher turnover intentions for the mixed-care group. Men experienced a higher level of turnover intentions and WFC-strain for adultcare. These findings align with our previous notions regarding differences between the factors women and men take into consideration when choosing a job. It is possible that women experience higher levels of job satisfaction and lower turnover intentions because they tend to seek jobs that offer more flexibility (Bender, Donohue, & Heywood, 2005). These findings provide further support for the gender-role theory suggesting that women experience higher FWC, while men experience higher WFC (Byron, 2005; Kinnunen & Mauno, 1998; Neilson & Stanfors, 2014).

When assessing what socio-demographic variables should be controlled for, interesting information was found. Those who frequently used flextime (M = 1.55, SD = .63) had higher physical burden than those who didn’t have the policy (M = 1.38, SD = .55, p < .01) and those who used it sometimes (M = 1.36, SD = .56, p < .05). For social burden, those who frequently used flextime (M = 1.53, SD = .57) had higher levels of burden than those who didn’t have the policy (M = 1.32, SD = .59, p < .05). For WFC-strain, those who frequently used flextime (M = 2.48, SD = .61) had higher WFC-strain than those who sometimes used flextime (M = 2.29, SD = .66, p < .05). It should be noted that those who never used flextime (M = 2.19, SD = .69) had a lower level of WFC-strain but this difference was not significant. This is most likely due to the small sample size of those who have the policy but never used it (n = 36). Also, those who don’t use available flextime are likely caregiving for more independent care recipients. For FWC-time, those who frequently used flextime (M = 2.29, SD = .74) had higher FWC-time than those who didn’t have the policy (M = 1.84, SD = .68, p < .001), those who had never used the policy (M = 1.87, SD = .67, p < .01), and those who sometimes used the policy (M = 2.00, SD = .66, p = .001). Those who had sometimes used the policy also had higher FWC-time than those who didn’t have the policy (p < .05). For FWC-strain, those who frequently used flextime (M = 2.16, SD = .64) had significantly higher FWC-strain than those who didn’t have the policy (M = 1.88, SD = .68, p < .001), those who had never used the policy (M = 1.81, SD = .65, p < .05), and those who had sometimes used the policy (M = 1.96, p < .05). For job satisfaction, those who didn’t have this policy (M = 3.54, SD = .90) had lower job satisfaction than those who sometimes used flextime (M = 3.80, SD = .81, p = .005) and those who frequently used flextime (M = 3.86, SD = .78, p < .001). For turnover intentions, those who didn’t have the policy (M = 3.37, SD = 2.00) had higher turnover intentions than those who frequently used flextime (M = 2.87, SD = 1.85, p < .05). These findings are consistent with attraction, selection, and attrition theory (Schneider, 1987), as those with higher levels of burden or conflict chose jobs in which they could use flextime frequently.

The majority of analyses were unaffected by controlling for socio-demographic variables. When controlling for significant socio-demographic variables, differences in emotional burden and WFC-time between care recipient type went from significant to non-significant and similarly FWC-time went from marginally significant to non-significant. For women, physical, social, and emotional burden, as well as turnover intentions, went from significant differences amongst care recipient type to non-significant. However, for men, although emotional burden went from significant between care recipient type to non-significant, WFC-time went from non-significant to significant. This ultimately shows the importance of ruling out confounding variables in analyses.

The last point of interest to discuss is the differences found between significance testing and effect size testing. When using effect sizes, differences between care recipient type were found for every variable although for some, specifically WFC-strain, FWC-strain, job satisfaction, and turnover intentions, were all small or lower or impractically different. The largest differences were found for time-dependence burden where effect sizes were large or medium for more than half of the group comparisons. Physical burden also displayed a large effect size and five medium effect sizes. Developmental, social, and emotional burden, as well as WFC-time showed primarily small to medium effect sizes. When splitting by gender, every variable had at least one comparison with a medium effect size with almost all categories containing multiple medium or large effect sizes. Similarly, time-dependence burden had the highest effect sizes for each gender but men had the largest differences. Further, men also had the largest differences for physical burden.

Limitations and Future Directions

The present study does warrant a few caveats given some of its potential limitations. Although this study used a diverse and large overall sample, when splitting caregivers into specific groups, group sizes were sometimes rather small (i.e., n = 26 and 28 in the adultcare group for men and women, respectively). Future studies should aim to compare care-recipient type groups with larger and similar group sizes. Second, this study utilizes a cross-sectional design that may allow for common-method bias. Future research in this area should employ longitudinal design to assess causality between these constructs and individual and organizational factors or differences that occur over time. Finally, the WFC/FWC measure used only assessed two types of conflict (time and stress) where other scales (Carlson, Kacmar, & Williams, 2000) include a third type (i.e., behavior).

Implications for Research and Practice

Nonetheless, the present study does have important research implications. These preliminary findings provide a foundation for future researchers to further examine relationships that were previously unexplored, and discover any underlying mechanisms that may explain differences found between care-recipient type groups. Further, these findings also have important organizational implications. They suggest that differences exist in WFC, FWC, job satisfaction, and turnover intentions among employed individuals providing different types of informal caregiving. Considering that role strain can potentially lower work performance, it is therefore important for organizations to offer and implement programs (i.e., flexible work arrangements) that help employees ease the burden, as well as the psychological and physical demands associated with caregiving. As such, organizations may benefit from offering employed caregivers different plan options based on care-recipient type. For example, several organizations already offer childcare services in the workplace, however, based on the present results that show that eldercare is particularly burdensome, it may also be crucial for employers to offer eldercare services in the workplace. It is possible that clearer policies and/or procedures for how management should interact and be notified of worker needs related to caregiving will help ameliorate some of the burden associated with caregiving. For example, if a company offers flexible hours to those with children, it may be beneficial to also offer this to caregivers who share similar burdens, but care for adults or the elderly. As no policy/benefit package is perfect, it is advisable to note that individuals are affected differently when caring for different people. For example, it may be best to provide those with eldercare responsibilities some sort of psychiatric help or emotional support system as these individuals consistently reported the highest degree of emotional burden. For those who are providing childcare, caregivers reported that time-dependence burden is consistently the highest. If possible, these individuals may benefit from flexible shifts. This may help decrease turnover while increasing job satisfaction, especially for women. Finally, it is possible that training of human resources and management may help decrease the conflict that these individuals experience between work and family. This training could include information on how to provide family supportive supervisor behaviors which has been show to relate to WFC, work-family positive spillover, job satisfaction, and turnover intentions (Hammer, Kossek, Yragui, Bodner, & Hanson, 2009). It should be advised that this training should be well thought out and tested as previous training on family-supportive supervision was beneficial to those with higher levels of FWC (i.e., improved work and health outcomes), but detrimental to those with lower levels of FWC (Hammer, Kossek, Anger, Bodner, & Zimmerman, 2011).

Conclusion

Overall, the present study serves as an initial investigation into the nature of the associations between different caregiver types and caregiver burden, WFC, FWC, job satisfaction, and turnover intentions. While the mechanisms behind some of these differences (or lack of differences) are not yet clear, it is clear there are differences between care-recipient type groups, in terms of burden. Although future research is needed to further explore the underlying nature of these relationships, the present study sheds light on one of many demographic variables associated with differences in proximal and distal outcomes. Collectively, these results provide a foundation for future research studies.

Table 6.

Analysis of covariance results.

				All participants
FWC-time	Sum of squares	df	Mean square	F	p
Intercept	14.18	1	14.18	33.09	.000
Caregiver hours	7.92	1	7.92	18.47	.000
Flextime use	13.28	1	13.28	31.00	.000
Income	14.48	1	14.48	33.80	.000
Ethnicity	2.21	1	2.21	5.15	.023
Disability	6.04	1	6.04	14.11	.000
Caregiver category	5.91	6	.98	2.30	.033
Error	317.90	742	.43
Total	3352.36	754