Grandmothers, Caregiving, and Family Functioning

HHS Public Access Author manuscript Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Published in final edited form as: J Gerontol B Psychol Sci Soc Sci. 2006 March ; 61(2): S89–S98. doi:10.1093/geronb/61.2.s89. Grandmothers, Caregiving, and Family Functioning Carol M. Musil1, Camille B. Warner1, Jaclene A. Zauszniewski1, Alexandra B. Jeanblanc1, Kyle Kercher2 1School of Nursing, Case Western Reserve University, Cleveland, Ohio. 2Department of Gerontology, University of Nebraska at Omaha. Author Manuscript Abstract Objectives—We used McCubbin’s Resiliency Model of Family Stress, Adjustment and Adaptation (McCubbin, Thompson, & McCubbin, 2001) to examine how demographic factors, family stress, grandmother resourcefulness, support, and role reward affect perceptions of family functioning for grandmothers raising grandchildren, grandmothers living in multigenerational households, and grandmothers not caregiving for grandchildren. Author Manuscript Methods—A sample of 486 grandmothers completed a mailed questionnaire. We used structural equation modeling to (a) test the effects of demographic factors (i.e., grandmother’s age, race, marital status, and employment), family stressful life events and strain, grandmother’s resourcefulness, subjective and instrumental support, and role reward on perceptions of family functioning for each grandmother group; (b) evaluate differences in the measurement and structural models between the grandmother groups using multisample analysis; and (c) test the model on the full sample, coding for caregiver status. Results—The models did not differ significantly by grandmother group; therefore we assessed the composite model using a multisample analysis. We found general support for the resiliency model and equivalence of the models across grandmother groups. Less support, resourcefulness, and reward, and more intrafamily strain and stressful family life events contributed to perceptions of worse family functioning. Discussion—Findings demonstrate the importance of the quality of family functioning for grandmothers in all types of families. Author Manuscript HOW do grandmothers who raise grandchildren perceive their families as functioning? Do their perceptions differ from those of grandmothers who live apart from grandchildren or who live with grandchildren and the children’s parents in multigenerational homes? This article addresses these questions and examines family functioning (e.g., the way family members interact, solve problems, respond to one another, and communicate; Miller, Epstein, Bishop, & Keitner, 1985). Poor family functioning is associated with problems of individual family members and the family as a whole (Lange et al., 2005; Petrocelli, Calhoun, & Glaser, 2003; Tamplin, Goodyer, & Herbert, 1998). How the family is perceived as functioning is particularly relevant for the 1.5 million American grandmothers raising Address correspondence to Carol M. Musil, School of Nursing, Case Western Reserve University, Cleveland, OH 44106-4904. [email protected]. Musil et al. Page 2 Author Manuscript grandchildren and 2.2 million American grandmothers living in multigenerational homes (Simmons & Dye, 2003), because their families often face unique challenges (Goodman, 2003; Pruchno, 1999). However, a grandmother’s involvement with her children and grandchildren—whether direct (e.g., childcare) or indirect (e.g., emotional or financial support)—affects individual family members and their interactions. There has been little attention to nonresident grandmothers’ perceptions of the family even though these grandmothers see themselves as part of an extended family system (Fingerman, 1996). This article explores grandmothers’ perceptions of family functioning based on their caregiving to grandchildren (i.e., grandmothers raising grandchildren, grandmothers in multigenerational homes, and noncaregivers to grandchildren), and whether the same factors contribute to perceptions of family functioning in the three groups. Literature Review Author Manuscript Differences in perceptions of family functioning may reflect the type of caregiving to grandchildren, family stresses and strains, and resources and characteristics of the grandmother. Although family composition may affect family functioning, family stress also may differ by family composition (Garnefski & Diekstra, 1997; McFarlane, Bellisimo, & Norman, 1995). Author Manuscript Grandmothers raising grandchildren, either alone or with a partner, assume care of grandchildren when the parent(s) are unable or unwilling to do so, often related to substance abuse. Although many custodial grandmothers find their role rewarding (Pruchno, 1999), their complex situations place them at risk for more perceived problems in family functioning. Compared with other grandmothers, they often have more health problems and depression (Minkler & Fuller-Thomson, 1999; Musil & Ahmad, 2002) and are more likely to live in poverty (Simmons & Dye, 2003). Relationships with the children’s parent(s) and child-care issues are common stresses (Goodman, 2003). Furthermore, 13% of children in kinship care have emotional or behavioral problems compared with 7% of children living with parents (Billing, Ehrle, & Kortenkamp, 2002), which complicates family interactions. Some custodial grandparents rely on grandchildren for social and emotional support (Hayslip & Kaminski, 2005), disturbing the hierarchy of intergenerational exchanges. Thus, these grandmothers’ perceptions of their family may reflect objectively difficult situations. Author Manuscript In multigenerational homes, the middle generation often has always lived in or returns to the grandparents’ home. The shared living arrangements help the adult child, but may also be of mutual benefit (Choi, 2003). Family members must negotiate who will do what and how decisions will be made. If young children are in the home, the grandmother is often expected to help with child care (Jendrek, 1994). In other homes, the grandmother may inhibit a teen daughter’s role as mother, and the role overlap will create ongoing tension (Black & Nitz, 1996; Chase-Lansdale, Brooks-Gunn, & Zamsky, 1994). Life events, such as divorce or job loss, also may prompt the living arrangement and the need to establish new patterns of family interaction. However, continued coresidence of adult children and grandchildren may increase grandmothers’ depression (Szinovacz, DeViney, & Atkinson, 1999) and reduce grandparents’ marital quality (Ward & Spitze, 2004), which may lead to less reward and worse perceived family functioning in some multigenerational homes. J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 3 Author Manuscript Grandmothers who live in a separate residence from adult children and grandchildren see their grandmother role as an extension of their parent role and regard it as more significant than other nonfamily adult roles (Fingerman, 1996, 2004; Reitzes & Mutran, 2004b). They view their families as including adult children (and their spouses and grandchildren) and often feel responsible for grandchildren’s successes and failures, as they do for their own children (Fingerman, 1996, 1998; Reitzes & Mutran, 2004b). Nonresident grandparents often report good relationships with adult children. If the adult child is married, relationships with grandchildren are frequently linked to those with the adult child’s spouse (Fingerman, 2004; Uhlenberg & Hammill, 1998); these relationships may be a source of intrafamily conflict. This article, which pays attention to these grandmothers’ perceptions of family functioning, extends research that recognizes a grandmother’s ongoing connection to her family. Author Manuscript Theoretical Model In order to consider differences in family functioning across the three grandmother groups, we drew upon the Resiliency Model of Family Stress, Adjustment and Adaptation (McCubbin et al., 2001) to examine how family stress and grandmothers’ social support and resourcefulness affect their role reward and view of family functioning. The model (Figure 1), applicable to families undergoing transition and adjustments, suggests that family stresses and strains, if not mediated by resources, can lead to problems in family functioning. Positive situational appraisals (e.g., reward) are thought to improve perceptions of family functioning. Demographic factors with relationships to stress and support and potential relationships with reward and family functioning are included in the model. Author Manuscript Support, Resourcefulness, Reward, and Family Functioning Social support improves perceptions of family functioning in both caregiving and noncaregiving families (Ergh, Rapport, Coleman, & Hanks, 2002; Wilhelm, Brownhill, & Boyce, 2000). For grandmother caregivers, subjective support reduces the effects of stress on mental health (Kelley, Whitley, Sipe, & Yorker, 2000; Musil & Ahmad, 2002) and may improve perceptions of family functioning. Inadequate support is associated with less grandparent role satisfaction and less tolerance for problem behavior in grandchildren (Hayslip & Kaminski, 2005). Less support may detract from family functioning, both because of one’s perception of family responsiveness and because with less instrumental support, the mundane aspects of family life may be neglected or less carefully managed, especially in times of stress. Author Manuscript Learned resourcefulness is a tendency to use cognitive-behavioral strategies for managing responses to difficult situations. These include (a) the self-control skills of cognitive reframing, problem-solving strategies, and the delay of immediate gratification, and (b) a generalized belief in one’s ability to self-regulate internal responses (self-regulatory efficacy; Rosenbaum, 1980; Rosenbaum & Ben-Ari, 1985). Higher overall resourcefulness is related to better adaptive functioning, mental health, task performance, and social role functioning (Fingerman, Gallagher-Thompson, Lovett, & Rose, 1996; Zauszniewski & Chung, 2001; Zauszniewski, Chung, & Krafcik, 2001). Self-control strategies that use J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 4 Author Manuscript cognitive reframing and expectations of eventual gain have been linked with ability to delay reward (Sugiwaka & Okouchi, 2004), and we hypothesize that they contribute to immediate and long-term reward in the grandmother role and to family interaction patterns that promote open communication and reflective decision making. Self-regulatory efficacy is a confidence in one’s own ability to control emotional responses, and it differs from the domain and taskspecific self-efficacy constructs described by Bandura (1977) and others (Steffen, McKibbin, Zeiss, Gallagher-Thompson, & Bandura, 2002). Although higher domain-specific selfefficacy is related to less caregiver burden (Intrieri & Rapp, 1994) and to better relationships with grandchildren (King & Elder, 1998), researchers have not yet studied the links between self-regulatory efficacy and role reward or family functioning. Preoccupation with worry reflects low self-regulatory efficacy, which may detract from feeling satisfied in the grandmother role or from appraising family exchanges as positive. Author Manuscript Reward in the grandmother role may be important to the appraisal of family functioning. Many custodial grandparents report satisfaction from helping their grandchild (Pruchno, 1999), but may face challenging conditions that temper their reward (Kelley et al., 2000). Role reward is linked to role centrality and identity for nonresident grandparents (Reitzes & Mutran, 2004a) and to family intimacy and satisfaction in non-grandparent caregivers (Bulger, Wandersman, & Goldman, 1993; Carruth, Tate, Moffett, & Hill, 1997). It is expected that grandmothers who derive greater reward in their role will view their families more positively, regardless of structure. Author Manuscript The grandmother’s age, race, and marital and job status may affect her perceptions. Custodial grandmothers, who are often younger than other grandmothers but older than most parents, may have less energy to meet caregiving demands. Black grandmothers are more likely than White grandmothers to become custodial caregivers and live in multigenerational homes, but they are also less burdened than White grandmothers (Pruchno & McKenney, 2002) and may report more reward. Goodman and Silverstein (2002) found no racial differences in mental health by family structure; thus race may not affect perceived family functioning. Married and employed grandmothers often have more resources than those who are not (Szinovacz et al., 1999), which may contribute to greater reward and better views of family functioning. Author Manuscript We hypothesized that custodial and multigenerational grandmothers would report more stressful family life events and more strain than noncaregivers. We expected more stressful family life events and strain to reduce support, resourcefulness, and reward; we also expected more stressful family life events and strain, combined with less resourcefulness, support and reward, to contribute to perceptions of worse family functioning. We expected this model to operate similarly across all groups, which would facilitate conceptualizing these relationships and formulating future interventions. Methods A sample of 486 grandmothers from a midwestem state completed a mailed survey on themselves and their families. Women were eligible to participate if they had one or more grandchild(ren) aged 16 or younger. All women were assigned to one of three caregiving J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 5 Author Manuscript groups: custodial, multigenerational, or noncaregiver. Custodial grandmothers were raising coresident grandchildren with no parents present. Multigenerational grandmothers lived in the same home as a grandchild and one or both of the child’s parents. Noncaregiver grandmothers lived within 50 miles of their grandchildren but did not live with or regularly babysit them. Author Manuscript We used random-digit dialing (RDD) as the main sampling method, with supplemental snowball-on-random and convenience sampling (i.e., letters to members of a statewide grandparent–kincare coalition group) to recruit custodial and multigenerational grandmothers. Potential study participants were mailed a questionnaire packet, as well as a reminder postcard two weeks later, and a replacement packet four weeks later, if necessary. Participants received $15 after returning the questionnaire. Of the 665 women who agreed to participate, 486 returned completed questionnaires, for a response rate of 73%, consistent with other studies using mailed surveys (Dillman, 2000). The final sample included 183 custodial grandmothers (74 obtained through RDD, 9 through snowball on random, and 100 through convenience sampling), 136 multigenerational grandmothers (125 through RDD, 4 through snowball on random, and 7 through convenience sampling), and 167 noncaregiver grandmothers secured through RDD. Author Manuscript The grandmothers’ mean age was 57 years (SD = 10.1 years), and 54.7% of the women were married or living with a partner (Table 1). Custodial grandmothers were raising grandchildren because of the parents’ neglect, abuse, or abandonment of the children (35.3%); substance abuse (15.6%); work, school, or military duty (9.9%); incarceration (7.8%); need for financial or other help (7.6%); marital disruptions (7.4%); mental or physical illness or disability (6.7%); death (3.8%); or for other reasons (5.9%). Of grandmothers in multigenerational households, 60% babysat or supervised grandchildren. Noncaregiver grandmothers had varying involvement with grandchildren, including remaining in phone or e-mail contact, attending grandchildren’s sports and school events, and babysitting. Grandmothers had an average of 3.3 grandchildren (SD = 3.6) ranging in age from 1 month to 16 years. There were significant differences between the grandmother groups in age, race, and income. Multigenerational grandmothers were slightly younger than grandmothers in the custodial and noncaregiver groups, and noncaregiver grandmothers had a greater monthly household income than the custodial grandmothers. Grandmothers of color were more likely than White grandmothers to be custodial caregivers; grandmothers in multigenerational homes were most likely to be employed. Author Manuscript Measures We used a modified version of the Family Inventory of Life Events (McCubbin et al., 2001) to identify the “pile up” of family stresses and strains experienced by family members in the past year. Two factors—stressful family life events (9 dichotomous items) and intrafamily strain (10 dichotomous items)—were each represented by a summary score in the analysis. Items in the Stressful Family Life Events subscale included: “Family member ran away from home,” “A child died,” and “A family member went to jail.” The number of reported stressful family life events ranged from 0 to 8 (M = 1.7, SD = 1.5). Examples of items in the J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 6 Author Manuscript Intrafamily Strain subscale included: “Increase in conflict among children,” “Increased disagreement about a family member’s friend(s) or activities,” and “A family member appeared to depend on alcohol or drugs.” Strain scores ranged from 0 to 10 (M = 3.7, SD = 2.7). Author Manuscript We measured learned resourcefulness with a 25-item modified version of the Self-Control Schedule (Rosenbaum, 1980, 1990; Zauszniewski, 1997). Respondents indicated how well each item described them, with 0 = not at all like me and 5 = always like me. We used exploratory factor analysis (EFA) to examine the factor structure of the Self-Control Schedule with principal axis extraction and direct oblimin rotation. We identified three factors on which 14 items had substantial (> .4) and clean factor loadings: self-regulatory efficacy (6 items), self-control (6 items), and pain management (2 items). The first two factors reflect Rosenbaum’s self-control strategies (Rosenbaum, 1990; Zauszniewski, 1997) and self-regulatory efficacy (Gruber & Wildman, 1987; Sugiwaka & Okouchi, 2004). The third factor included two pain management items but was not used in analysis. Selfregulatory efficacy items (α = .78) included: “I often find it difficult to overcome my feelings of nervousness and tension without outside help,” “I can’t avoid thinking about mistakes I made in the past,” and “Although it makes me feel bad, I cannot avoid thinking about all kinds of possible catastrophes.” Negatively worded items were reverse scored. Scores ranged from 0 to 30 (M=20.6, SD=6.7); higher scores indicate greater self-regulatory efficacy. The self-control factor contained items (α = .86) that reflect self-control of mood and thought, such as “When I am feeling depressed, I try to think of pleasant events,” and “When in a low mood, I try to act cheerful so that my mood will change.” Self-control scores ranged from 0 to 30 (M = 20.1, SD=5.5); higher scores indicate the use of more selfcontrol skills. Self-regulatory efficacy and self-control were represented as latent factors. Author Manuscript We assessed subjective and instrumental support with two scales from the Duke Social Support Index (Hughes, Blazer & Hybels, 1990). The two social support dimensions were supported by EFA. The Instrumental Support subscale included 12 dichotomous items, such as “Do family and friends help out when you are sick?” and “Do they shop or run errands for you?” We summed items (α = .83) to form a composite score. Scores ranged from 0 to 12 (M = 8.6, SD = 3.0). Subjective support was measured by 7 questions about feelings of support and involvement with family and friends, with responses on a scale (0=hardly ever to 2=most of the time). The Subjective Support subscale included: “Do you feel you have a definite role in family and among friends?” and “Does it seem that your family and friends understand you?” Subjective support scores ranged from 0 to 14 (M = 11.7, SD = 2.7). We summed items to form a composite score (α = .85). For both subscales, higher scores indicate more support. Author Manuscript Role reward was measured by using a visual analog scale adapted from a measure of stress in the grandmother role (Musil & Ahmad, 2002). We asked participants to mark a 10-cm line in response to the question “How much reward do you receive from your role as a grandmother?” The line was anchored with endpoints of 0 = not at all rewarding and 10 = extremely rewarding. Role reward ranged from 0 to 10 (M = 7.8, SD = 2.6) in the sample. J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 7 Author Manuscript We measured perceived family functioning with the General Family Functioning and Communication subscales of the Family Assessment Device (FAD; Epstein, Baldwin, & Bishop, 1983). The FAD is consistent with assessments by outside observers (Miller et al., 1994) and has been used with extended family members (Edwards & Clarke, 2004). Respondents indicated how well statements described their family, with 1 = very accurately and 4 = does not describe at all; higher scores indicate worse perceived family functioning. We substantiated the factor structure with EFA. The 12 General Family Functioning subscale items (α = .86) included: “In times of crisis we can turn to each other for support,” “We are able to make decisions about how to solve problems,” and “Individuals are accepted for what they are.” Mean scores are used and ranged from 1.0 to 3.8, with a mean item score of 1.9 (SD = .60). The 8 Communication subscale items included: “We often don’t say what we mean,” and “When someone is upset we know why.” The item “We don’t talk to each other when we are angry” reduced subscale reliability and was eliminated, yielding α = .71. Mean scores ranged from 1.0 to 4.0, with a mean of 2.0 (SD=.56). We used mean scores of the subscales to represent the latent factor: overall family functioning. Author Manuscript The demographic variables in the analysis included age in years, race, marital status, and employment. Race was coded as a dichotomous variable, with 0 = people of color and 1=White race, based on the grandmothers’ self-identification. Marital status was coded 0 = not married and 1 = married or with a partner. Employment was assessed by the question “Are you employed in a paying job?” with 0=no; 1=yes, part-time; and 2=yes, full-time. Analysis Procedures Author Manuscript We compared the three groups’ mean scores on all study variables by using analysis of variance with Scheffe’s post hoc test. We used AMOS 5.0 (Assessment Systems Corporation, St. Paul, MN) to test the fit of saturated, sequential models for each caregiving subgroup in which the following were hypothesized to affect perceived family functioning: (a) grandmother’s age, race, marital status, and employment, (b) stressful family life events and intrafamily strain, (c) grandmother resourcefulness and support, and (d) role reward. The subgroup models were combined into a composite model incorporating each subgroup’s significant paths and covariances (p < .05), and a multisample analysis evaluated the equivalence of the model across the subgroups. We assessed model fit with the Comparative Fit Index (CFI), the Tucker–Lewis coefficient (TLI), the root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR; Hu & Bentler, 1999). CFI and TLI values of ≥ .90 (preferably .95 or greater) and RMSEA and SRMR values ≤ .08 indicate good fit (Hu & Bentler). Author Manuscript Results Custodial grandmothers reported less subjective support and worse perceptions of family functioning than did noncaregiver grandmothers (Table 2). Custodial grandmothers also reported less role reward than did the grandmothers in the other two groups. Grandmothers in multigenerational homes reported the most instrumental support. Noncaregivers to grandchildren reported less intrafamily strain than did primary caregivers. J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 8 Grandmother Subgroup Models Author Manuscript For custodial caregivers, the initial saturated structural model yielded fit indices of χ2(173, n = 183) = 242.79, p < .001, CFI = .94, TLI = .91, RMSEA = .05, SRMR = .05, suggesting a need to improve fit. We added covariances between error terms identified in the modification indices when theoretically and logically appropriate. We then eliminated insignificant paths, achieving a final model with χ2(209, n = 183) = 240.97, p < .05, CFI = .97, TLI = .97, RMSEA = .03, SRMR = .07, with all paths and covariances significant (p < .05). More life events—and less self-control, self-regulatory efficacy support, and role reward—worsened perceptions of family functioning (Table 3). Author Manuscript For the multigenerational group, the saturated model yielded χ2(174, n = 135) = 253.12, p < .001, CFI = .91, TLI = .86, RMSEA = .06, SRMR = .06, indicating a need to improve fit. After we added error covariances and eliminated insignificant paths, the final model achieved good fit: χ2(217, n = 136) = 255.47, p < .05, CFI = .95, TLI = .95, RMSEA = .04, SRMR = .08. More intrafamily strain and less subjective support led to perceptions of worse family functioning. For the noncaregiver group, the initial saturated model showed a need to improve fit: χ2(173, n =167) = 246.54, p < .001, CFI = .92, TLI = .89, RMSEA = .05, SRMR = .06. We added two error covariances, eliminated insignificant covariances and paths, and achieved χ2(213, n = 167) = 269.09, p < .01, CFI = .94, TLI = .93, RMSEA = .04, SRMR = .08. Less subjective support, more intrafamily strain, and less self-regulatory efficacy led to worse perceptions of family functioning. Less role reward also contributed to perceived problems in family functioning. Author Manuscript Multi-Sample Comparison To test for differences in the measurement and structural models across the grandmother caregiving subgroups, we used multisample AMOS analysis to compare increasingly constrained models. We began with an unconstrained model in which all parameters were allowed to be freely estimated for each group. Next, the measurement (factor) weights were constrained to be equal and were found not to differ across the groups. The structural weights (paths) between variables in the model were constrained to be equal and were found not to differ significantly across groups. Finally, the structural covariances were constrained to be equal and also were found not to differ between the groups. Because the results indicated that the subgroup models did not differ significantly, we evaluated a single model comprising the entire sample. Author Manuscript Final Model We combined the subgroups into a single sample, with caregiver status dummy coded (custodial vs else and multigenerational vs else). The initial saturated structural model showed adequate fit: χ2(195, N = 486) = 378.29, p < .001, CFI = .94, TLI = .91, RMSEA = . 04, SRMR = .04. After eliminating insignificant paths, the final model (Figure 2) achieved good fit: χ2(239, N = 486) = 361.38, p < .001, CFI = .96, TLI = .95, RMSEA = .03, SRMR = .05. Less self-control, self-regulatory efficacy, subjective and instrumental support, and role reward—but more intrafamily strain and stressful family life events—contributed to J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 9 Author Manuscript perceptions of worse family functioning. Custodial caregiving led to less instrumental support and more intrafamily strain and self-regulatory efficacy. Multigenerational caregiving was related to more instrumental support and intrafamily strain. Discussion Author Manuscript This study examined (a) the perception of family functioning as appraised by grandmothers whose caregiving responsibilities to grandchildren differed, and (b) how family stress, support, resourcefulness, and role reward affect these perceptions. The consistency of the model across groups strengthens its applicability, which is important because our longitudinal data suggest that grandmothers living with grandchildren may experience changes in household composition at a greater rate than do other grandmothers. A model of family functioning that can incorporate variations in grandmothers’ caregiving to grandchildren will advance the work in this area. Author Manuscript There were group differences in the grandmothers’ perceptions of family functioning. In ANOVA comparisons of the Family Functioning subscales and additional structural equation modeling analyses of the latent construct (not shown in this article), grandmothers raising grandchildren perceived more difficulties in overall family functioning than noncustodial grandmothers, mostly due to differences in general patterns of family interactions. Custodial grandmothers had a mean score of 2.0 on the General Family Functioning subscale; Miller and colleagues (1985) identified this cut-off score as “unhealthy” (p. 353). In the present sample, 53% of custodial grandmothers reported scores in the unhealthy range, compared with 44% of multigenerational and 31% of noncaregiver grandmothers. This report may reflect a history of persistent problems in family interaction among the families with custodial grandmothers; yet the percentage of multigenerational and noncaregiver grandmothers expressing concerns about their family dynamics is considerable. The FAD has shown good correlation with outside clinical raters (Miller et al., 1994), between spouses (Wilhelm et al., 2000), and between adolescents and their parents (Tamplin et al., 1998). Although the present study data represent only one party’s view, they are likely a reasonably accurate description of the family situation. Identifying the concordance in viewpoints of grandmothers and their families, as well as differences in family concerns according to family structure, could serve as the basis for interventions—ranging from multigenerational family interventions to educational programs delivered through senior centers—that are tailored to meet the needs of various family groups. Author Manuscript Consistent with our hypotheses about family stress, custodial grandmothers reported more stressful family life events than did noncaregiver grandmothers, and custodial and multigenerational caregivers both reported more daily intrafamily strain than did noncaregivers. Others have noted the prevalence of stressful events in the lives of custodial grandmothers (Pruchno, 1999), but the degree of intrafamily strain has received less attention. Specific family strains—such as arguments between parents and children and concerns about family members’ alcohol and drug use or emotional problems—were most prevalent in the custodial caregiver group, but were noted by at least 30% of the multigenerational and 25% of noncaregiver grandmothers. Because these concerns cut J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 10 Author Manuscript across all groups, there is a need for continued community-health efforts focused on intergenerational communication and strategies for dealing with impaired family members. Author Manuscript There may have been differences by caregiving group in the frame of reference that grandmothers used for reporting on their family. We did not ask participants to identify specific family members as other researchers have done (Goodman, 2003; Pruchno & McKenney, 2002), but we instead asked them to respond according to how they, as grandmothers, see their family. Thus, primary caregivers may have weighed more heavily on grandchildren living with them than they did on those living outside the home. Yet, nonresident adult children and grandchildren retain a prominent place in the grandmother’s view of family and are a source of strain and reward (Fingerman, 1998; Reitzes & Mutran, 2004a, 2004b). Although we suspect that grandmothers considered coresident family members and those outside the home when making their evaluations, future research might address this with greater precision. The current wave of data collection assesses closeness of the grandmother with the parent(s) of each grandchild and may provide some insight into this question. Author Manuscript Across the groups, two factors in particular affect perceptions of family functioning: intrafamily strain and subjective support. Strain contributed to concerns about family functioning both directly and indirectly through reduced subjective support and reward. Previous research has shown that family strains take their toll on mental health (Kelley et al., 2000; Musil & Ahmad, 2002), and the present findings indicate that strain also affects support and perceptions of family interactions. In terms of subjective support, grandmothers who did not feel they had a definite role in the family or who felt misunderstood had less reward and perceived more problems with family functioning. Goodman’s (2003) study of triads in custodial and multigenerational families had similar conclusions: that family structure may be less important than cohesive intergenerational relationships to the grandmother and to family well-being. Longitudinal studies of strain, support, and perceptions of family functioning might help clarify these relationships, especially in multigenerational homes where a myriad of factors may affect the household composition. Nevertheless, the finding that custodial grandmothers have less subjective support than other grandmothers is consistent with findings of other studies (Musil & Ahmad, 2002) and suggests that a support group may continue to be a useful mechanism to assist grandmothers who are raising grandchildren. Many custodial and multigenerational grandmothers commented that they would also benefit from affordable respite care so that they could attend to their own needs (e.g., health care) or have time alone; although respite services are less widely available, such services may be critical for grandmothers who are responsible for the daily care of grandchildren. Author Manuscript In contrast to our expectations, greater self-regulatory efficacy, or confidence in one’s ability to control emotional responses, contributed to decreased reward for women who lived with their grandchildren. Others have noted that becoming a grandmother either earlier than anticipated or with more responsibility than expected may lead to personal dissonance about one’s role (Burton, 1996; Landry-Meyer & Newman, 2004), which could affect reward. Conversely, persons with lower self-regulatory efficacy have a tendency to worry and dwell on the past and feel less in command of their anxiety, so living with grandchildren allows J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 11 Author Manuscript them to take an active role in the children’s daily life. Many grandmothers raising grandchildren regard this as a “second chance” that may lead to greater reward. This relationship deserves further exploration and replication. Author Manuscript Both dimensions of learned resourcefulness—self-regulatory efficacy and self-control— affected perceived family functioning in the composite model, although there were some differences by caregiving group. Higher self-regulatory efficacy improved perceptions of family functioning in the composite model and was significant for the noncaregiver and custodial subgroups. Having greater confidence in their ability to handle emotional responses may be especially relevant for women who live apart from their families and must trust that the families can manage on their own. For example, some grandmothers report concerns about young parents being able to handle the responsibilities of raising children, holding a job, and handling problems with spouses or boyfriends. Although all grandmothers need to balance connection with, and distance from, their families, those with lower self-regulatory efficacy may be less able to tolerate a more peripheral role ascribed to them by their children who live outside their home. Further research in this area may indicate that opportunities for grandmothers to learn specific resourcefulness skills might benefit family relationships and the grandmothers’ own mental health (Zauszniewski, 1997). Author Manuscript Overall, the analytic strategy supported a single model, but this approach obscured some caregiving group differences in the effects of reward and resourcefulness on perceived family functioning. For example, self-control showed a weak effect on family functioning in the multisample analysis but had stronger effects for custodial grandmothers. The use of self-control skills may enhance family functioning through better problem-solving skills and positive reframing, which may lead to more effective communication and interactions. In addition, although we did not find differences in perceptions between White women and women of color, racial and ethnic differences in household structure and grandmother role expectations (Landry-Meyer & Newman, 2004; Simmons & Dye, 2003) indicate a need for more work in this area. Author Manuscript This study extended other research by examining perceptions of family functioning according to grandmother caregiver status. A major strength was the multisample comparison that showed some differences in grandmothers’ perceptions of family functioning based on family structure, but also showed that subjective social support has consistent effects on perceptions of family functioning and role reward. Grandmothers play a vital role in the lives of their grandchildren and other family members, and their familial relationships warrant further attention. The fact that women are living longer and are likely to spend three or more decades as grandmothers points to the importance of their continuing influence in families. Acknowledgments This research was supported by Grant 5RO1-NR05067, “Intergenerational Caregiving to Youth At-Risk” from the National Institute of Nursing Research. We gratefully acknowledge Theresa S. Standing for her insightful comments on earlier versions of this article. We also acknowledge the editor and three anonymous reviewers of the manuscript for their constructive feedback. J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 12 Author Manuscript References Author Manuscript Author Manuscript Author Manuscript Bandura A (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Billing A, Ehrle J, & Kortenkamp K (2002). Children cared for by relatives: What do we know about their well-being? New Federalism National Survey of America’s Families, No. B-46. Black MM, & Nitz K (1996). Grandmother co-residence, parenting, and child development among low income, urban teen mothers. Journal of Adolescent Health, 18, 218–226. [PubMed: 8777198] Bulger MW, Wandersman A, & Goldman CR (1993). Burdens and gratifications of caregiving: Appraisal of parental care of adults with schizophrenia. American Journal of Orthopsychiatry, 63, 255–265. [PubMed: 8484431] Burton LM (1996). Age norms, the timing of family role transitions, and intergenerational caregiving among aging African American women. The Gerontologist, 36, 199–208. [PubMed: 8920089] Carruth AK, Tate US, Moffett BS, & Hill K (1997). Reciprocity, emotional well-being, and family functioning as determinants of family satisfaction in caregivers of elderly parents. Nursing Research, 46 (2), 93–100. [PubMed: 9105332] Chase-Lansdale PL, Brooks-Gunn J, & Zamsky ES (1994). Young African-American multigenerational families in poverty: Quality of mothering and grandmothering. Child Development, 65 (Special Issue II), 373–393. [PubMed: 8013228] Choi N (2003). Coresidence between unmarried aging parents and their adult children: Who moved in with whom and why? Research on Aging, 25, 384–404. Dillman D (2000). Mail and internet surveys: The tailored design (2nd ed.). New York: Wiley. Edwards B, & Clarke V (2004). The psychological impact of cancer diagnosis on families: The influence of family functioning and patients’ illness characteristics on depression and anxiety. Psycho-Oncology, 13, 562–576. [PubMed: 15295777] Epstein NB, Baldwin LM, & Bishop DS (1983). The McMaster family assessment device. Journal of Marital and Family Therapy, 9, 171–180. Ergh T, Rapport L, Coleman R, & Hanks R (2002). Predictors of caregiver and family functioning following traumatic brain injury: Social support moderates caregiver distress. Journal of Head Trauma Rehabilitation, 17 (2), 155–174. [PubMed: 11909512] Fingerman KL (1996). Sources of tension in the aging mother and adult daughter relationship. Psychology and Aging, 11, 591–606. [PubMed: 9000292] Fingerman KL (1998). The good, the bad, and the worrisome: Emotional complexities in grandparents’ experiences with individual grandchildren. Family Relations, 47, 403–14. Fingerman KL (2004). The role of offspring and in-laws in grandparents’ ties to their grandchildren. Journal of Family Issues, 25, 1026–1049. Fingerman KL, Gallagher-Thompson D, Lovett S, & Rose J (1996). Internal resourcefulness, task demands, coping, and dysphoric affect among caregivers of the frail elderly. International Journal of Aging & Human Development, 42, 229–249. [PubMed: 8805085] Garnefski N, & Diekstra RF (1997). Adolescents from one parent, stepparent and intact families: Emotional problems and suicide attempts. Journal of Adolescence, 20 (2), 201–208. [PubMed: 9104655] Goodman CC (2003). Multigenerational triads in grandparent-headed families. Journal of Gerontology: Social Sciences, 58B, S281–S289. Goodman CC, & Silverstein M (2002). Grandmothers raising grandchildren: Family structure and well-being in culturally diverse families. The Gerontologist, 42, 676–689. [PubMed: 12351803] Gruber VA, & Wildman BG (1987). The impact of dysmenorrhea on daily activities. Behaviour Research and Therapy, 25 (2), 123–138. [PubMed: 3593165] Hayslip B, & Kaminski PL (2005). Grandparents raising their grandchildren: A review of the literature and suggestions for practice. The Gerontologist, 45, 262–269. [PubMed: 15799992] Hu LT, & Bentler P (1999). Cut off criteria for fit indexes in covariance structure analysis: Conventional criteria versus alternatives. Structural Equation Modeling 6 (1), 1–55. Hughes DC, Blazer D, & Hybels C (1990). Duke Social Support Index (DSSI): A working paper (revised). Unpublished manuscript. J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 13 Author Manuscript Author Manuscript Author Manuscript Author Manuscript Intrieri RC, & Rapp SR (1994). Self-control skillfulness and caregiver burden among help-seeking elders. Journal of Gerontology: Social Sciences, 49, S19–S23. Jendrek MP (1994). Grandparents who parent their grandchildren: Circumstances and decisions. The Gerontologist, 34, 206–216. [PubMed: 8005493] Kelley SJ, Whitley D, Sipe TA, & Yorker BC (2000). Psychological distress in grandmother kinship care providers: The role of resources, social support, and physical health. Child Abuse & Neglect, 24,311–321. [PubMed: 10739075] King V, & Elder GH (1998). Perceived self-efficacy and grandparenting. Journal of Gerontology: Social Sciences, 53B, S249–S257. Landry-Meyer L, & Newman BM (2004). An exploration of the grandparent caregiver role. Journal of Family Issues, 25, 1005–1025. Lange G, Sheerin D, Carr A, Dooley B, Barton V, Marshall D, et al. (2005). Family factors associated with attention deficit hyperactivity disorder and emotional disorders in children. Journal of Family Therapy, 27 (1), 76–96. McCubbin H, Thompson A, & McCubbin M (2001). Family measures: Stress, coping and resiliency inventories for research and practice [CD-ROM]. Honolulu, HI: Kamehameha Schools. McFarlane AH, Bellissimo A, & Norman GR (1995). Family structure, family functioning and adolescent well-being: The transcendent influence of parental style. Journal of Child Psychology & Psychiatry, 36, 847–864. [PubMed: 7559849] Miller IW, Epstein NB, Bishop DS, & Keitner GI (1985). The McMaster Family Assessment Device: Reliability and validity. Journal of Marital and Family Therapy, 11, 345–356. Miller IW, Kabacoff RI, Epstein NB, Bishop DS, Keitner GI, Baldwin LM, et al. (1994). The development of a clinical rating sale for the McMaster Model of Family Functioning. Family Process, 33 (1), 53–69. [PubMed: 8039568] Minkler M, & Fuller-Thomson E (1999). The health of grandparents raising grandchildren: Results of a national study. American Journal of Public Health, 89, 1384–1389. [PubMed: 10474557] Musil C, & Ahmad M (2002). Health of grandmothers: A comparison by caregiver status. Journal of Aging and Health, 14 (1), 96–121. [PubMed: 11892763] Petrocelli JV, Calhoun GB, & Glaser BA (2003). The role of general family functioning in the quality of the mother-daughter relationship of female African American juvenile offenders. Journal of Black Psychology, 29, 378–392. Pruchno R (1999). Raising grandchildren: The experiences of black and white grandmothers. The Gerontologist, 39, 209–221. [PubMed: 10224717] Pruchno RA, & McKenney D (2002). Psychological well-being of black and white grandmothers raising grandchildren: Examination of a two-factor model. Journal of Gerontology: Social Sciences, 57B, 444^52. Reitzes DC, & Mutran EJ (2004a). Grandparenthood: Factors influencing frequency of grandparentgrandchildren contact and grandparent role satisfaction. Journal of Gerontology: Social Sciences, 59B, S9–S16. Reitzes DC, & Mutran EJ (2004b). Grandparent identity, intergenerational family identity, and wellbeing. Journal of Gerontology: Social Sciences, 59B, S213–S219. Rosenbaum M (1980). A schedule for assessing self-control behaviors: Preliminary findings. Behavior Therapy, 11, 109–121. Rosenbaum M (1990). The role of learned resourcefulness in self-control of health behavior In Rosenbaum M (Ed.), Learned resourcefulness: On coping skills, self-control, and adaptive behavior (pp. 3–30). New York: Springer. Rosenbaum M, & Ben-Ari K (1985). Learned helplessness and learned resourcefulness: Effects of noncontingent success and failure on individuals differing in self-control skills. Journal of Personality and Social Psychology, 48 (1), 198–215. [PubMed: 3981388] Simmons T, & Dye JL (2003). Grandparents living with grandchildren: 2000. Washington, DC: U.S. Census Bureau. Steffen AM, McKibbin C, Zeiss AM, Gallagher-Thompson D, & Bandura A (2002). The revised scale for caregiving self-efficacy: Reliability and validity studies. Journal of Gerontology: Social Sciences, 57B, 74–86. J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 14 Author Manuscript Author Manuscript Sugiwaka H, & Okouchi H (2004). Reformative self-control and discounting of reward value by delay or effort. Japanese Psychological Research, 46 (1), 1–9. Szinovacz ME, DeViney S, & Atkinson MP (1999). Effects of surrogate parenting on grandparents’ well-being. Journal of Gerontology: Social Sciences, 54B, S376–S388. Tamplin A, Goodyer IM, & Herbert J (1998). Family functioning and parent general health in families of adolescents with major depressive disorder. Journal of Affective Disorders, 48 (1), 1–13. [PubMed: 9495597] Uhlenberg P, & Hammill BG (1998). Frequency of grandparent contact with grandchild sets: Six factors that make a difference. The Gerontologist, 38, 276–285. [PubMed: 9640847] Ward RA, & Spitze GD (2004). Marital implications of parent-adult child coresidence: A longitudinal view. Journal of Gerontology: Social Sciences, 59B, S2–S8. Wilhelm K, Brownhill S, & Boyce P (2000). Marital and family functioning: Different measures and viewpoints. Social Psychiatry & Psychiatric Epidemiology, 35, 358–365. [PubMed: 11037305] Zauszniewski J (1997). Evaluation of learned resourcefulness for elders. Journal of Nursing Measurement, 5 (1), 71–86. [PubMed: 9505470] Zauszniewski J, & Chung C (2001). Resourcefulness and health practices of diabetic women. Research in Nursing and Health, 24 (2), 113–121. [PubMed: 11353459] Zauszniewski J, Chung C, & Krafcik K (2001). Social cognitive factors predicting the health of elders. Western Journal of Nursing Research, 23, 490–503. [PubMed: 11482053] Author Manuscript Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 15 Author Manuscript Author Manuscript Figure 1. Theoretical model. Author Manuscript Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 16 Author Manuscript Author Manuscript Figure 2. Final structural equation model with standardized regression weights. Author Manuscript Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Author Manuscript Author Manuscript 57 (10.1) Age 167 72 154 69 124 27 26 Some HS HS graduate Business or trade school Some college College graduate Post graduate college 1 97 52 44 99 19 6 $2,001-$3,000 $3,001-$4,000 $4,001-$5,000 > $5,000 Missing J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. 219 Not married 151 78 Employed Employed part-time Employment status 267 Married Marital status Refused 4 124 $1,000-$2,000 16 31 45 55 20 9 11 20 26 45 9 5 6 25 14 32 15 3 34 66 % < $1,000 Income (monthly) 14 ≤ 8 years Education Non-White N 319 31–87 Range White Race M (SD) Category Total (n = 486) 57 (9.4) M (SD) 31–87 Range 28 49 86 97 1 7 31 13 14 36 58 23 8 6 45 27 57 31 9 77 106 N Custodial (n = 183) Author Manuscript Comparison of Demographic Characteristics, by Caregiver Group 15 27 47 53 0.5 4 17 7 7.5 20 32 12 4 3 25 15 31 17 5 42 58 % 55 (11.4) M (SD) 31–86 Range 18 53 66 70 0 8 25 18 15 28 33 9 7 7 36 19 46 19 2 46 90 N 13 39 52 48 0 6 18 13 11 21 24 7 5 5 26 14 34 14 2 34 66 % Multigenerational (n = 136) 59 (9.3) M (SD) 38–83 Range 32 49 67 100 5 4 43 13 23 33 33 13 11 14 43 23 51 22 3 44 123 N Noncaregiver (n = 167) 19 29 40 60 3 2 25 8 14 20 20 8 7 8 26 14 30 14 2 26 74 % 7.3 2.6 19.0* 14.7 9.6** a 8.0 *** Test Statistic (χ2) Author Manuscript Table 1. Musil et al. Page 17 Author Manuscript p ≤ .001. Figure presented is F statistic. a *** p ≤ .01 ** p ≤ .05 * Note: HS = high school. Unemployed M (SD) Range 257 N 53 % Author Manuscript Category M (SD) Range 106 N 58 % M (SD) Range 65 N 48 % Multigenerational (n = 136) Author Manuscript Custodial (n = 183) M (SD) Range 86 N Noncaregiver (n = 167) 52 % Test Statistic (χ2) Author Manuscript Total (n = 486) Musil et al. Page 18 J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Musil et al. Page 19 Table 2. Author Manuscript Means and Standard Deviations for Study Variables, by Caregiver Group and F Statistic Variable Custodial (N = 183) Role reward 7.1 (3.1) Multigenerational (N = 136) a,c c Noncaregiver (N = 167) a 8.2 (2.0) 8.3 (2.3) 11.3*** Self-regulatory efficacy 20.8 (6.9) 19.7 (6.5) 21.2 (6.5) 2.1 Self-control 20.0 (5.7) 19.9 (5.4) 20.3 (5.4) 0.3 Instrumental support Subjective support 7.7 (3.9) a,c a 11.1 (3.1) a Intrafamily strain 4.4 (2.8) Stressful family life events General family functioning Author Manuscript Family communication b,c 9.7 (2.2) 11.8 (2.4) b 3.9 (2.9) a 1.7 (1.5) a 1.9 (0.6) 2.1 (0.6) 1.9 (0.6) 1.8 (1.6) 2.0 (0.6) a,b 19.2*** a 7.9*** a,b 18.4*** a 3.7* 1.8 (0.5) a 9.0*** 2.0 (0.5) 2.0 8.6 (2.8) 12.2 (2.6) 2.7 (2.3) 1.4 (1.3) Notes: For F statistic at p ≤ .05, the post hoc comparison reflects differences at p ≤ .05 unless otherwise noted. * p ≤ .05 ** p < .01 *** a F Statistic p < .001. Difference between custodial and noncaregivers. b Difference between multigenerational and noncaregivers. c Difference between custodial and multigenerational caregivers. Author Manuscript Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. Author Manuscript .71 .70 1.57 1.36 .90 1.76 SC3 SC4 SC5 .59 .94 .68 .73 .94 SE2 SE3 SE4 SE5 SE6 J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. — .03 .08 Age → Stressful life events Age → SE Age → Subjective support — −1.67 Race → Instrumental support Race → SC Race → Subjective support 1.13 −.31 Race → Intrafamily strain — Race → Stressful life events Relationship between variables 1.0 SE1 Self-regulatory efficacy SC6 .75 1.56 SC2 — — — −1.35 — — — −.57 .90 .90 .76 .82 .82 1.0 1.41 1.31 1.52 1.0 1.0 .96 1.0 M SC1 Self-control 1.0 Family communication C General family function Perceived family function Measurement model Variable Author Manuscript .06 .02 −.04 — 1.55 — — −.70 1.22 .72 .68 .96 .76 1.0 1.37 1.21 1.23 1.54 1.10 1.0 .73 1.0 NC Regression Weights .24 .19 — −.24 — −.24 .20 — .71 .55 .55 .73 .52 .72 .74 .51 .62 .80 .81 .49 .76 .96 C — — — −.29 — — — −.19 .64 .64 .50 .60 .65 .67 .35 .41 .71 .65 .91 .59 .79 .81 M .21 .20 −.27 — .26 — — −.24 .79 .50 .48 .57 .59 .64 .61 .62 .58 .74 .66 .54 .71 .94 NC Standardized Regression Weights .02 .01 — .44 — .10 .36 — .11 .11 .10 .11 .10 — .33 .17 .20 .25 .24 — .07 — C Author Manuscript Individual Group Model Results — — — .38 — — — .21 .20 .18 .18 .22 .24 — .15 .15 .16 .15 .16 — .10 — M SE .02 .01 .01 — .35 — — .20 .16 .13 .13 .14 .12 — .25 .21 .22 .25 .19 — .09 — NC 3.34 2.39 — −3.75 — −2.30 3.17 — 8.27 6.59 6.58 8.53 6.22 — 5.36 5.41 6.75 6.40 6.42 — 10.27 — C — — — −3.59 — — — −2.72 3.56 4.10 7.66 5.93 6.34 — 5.88 5.97 4.89 5.62 5.27 — 8.91 — M 2.96 2.21 −3.69 — 4.42 — — −3.46 7.64 5.44 5.25 6.92 6.26 — 5.44 5.75 5.50 6.27 5.77 — 8.58 — NC Critical Ratio Author Manuscript Table 3. Musil et al. Page 20 −.04 Instrumental support → Perceived family function Role reward → Perceived family function Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11. — 1.07 4.21 2.28 Subjective support ←→ SE Subjective support ←→ Instrumental support Stressful life events ←→ Intrafamily strain .08 Race ←→ Marital status Age ←→ Marital status −2.47 Employment ←→ Age Exogenous covariances — −.04 Instrumental support → Role reward .33 −.05 Subjective support → Perceived family function −.09 Subjective support → Role reward −.51 SE → Perceived family function Intrafamily strain → Role reward SE → Role reward −.34 Intrafamily strain → Subjective support — −.31 Intrafamily strain → Instrumental support −.12 −.21 Intrafamily strain → SE SC → Perceived family function −.16 Stressful life events → Perceived family function Intrafamily strain → Perceived family function — .09 Stressful life events → Role reward — 1.19 — Employment → SE Marital status → Instrumental support .56 Employment → Subjective support Marital status → SC — −.13 Employment → SC — Age → Role reward .05 Age → Intrafamily strain C Age → Instrumental support Author Manuscript Variable Author Manuscript 2.41 .95 .70 −1.41 .06 −3.64 — — — −.11 .30 — −.49 — .04 −.30 −.41 −.13 −.17 — .31 — — .22 — — — — — M .95 3.09 .69 −.82 .06 −3.47 −.03 — .19 −.09 .16 −.09 — — .04 — −.32 — −.16 — — — −.27 .45 .49 — .04 −.05 — NC .51 .47 .33 — .31 −.31 −.21 −.21 — −.25 .33 −.19 −.21 −.13 — −.31 −.28 −.18 −.35 .23 — .18 — — .16 −.18 — — .15 C .59 .21 .34 −.25 .27 −.35 — — — −.52 .36 — −.27 — .25 −.42 −.47 −.16 −.45 — .22 — — .18 — — — — — M .35 .47 .31 −.18 .28 −.43 −.14 — .23 −.45 .18 −.19 — — .17 — −.28 — −.34 — — — −.20 .37 .16 — .16 −.20 — NC Standardized Regression Weights .37 .71 .27 — .02 .63 .01 .01 — .02 .07 .04 .22 .06 — .08 .08 .09 .04 .02 — .43 — — .22 .06 — — .03 C .41 .38 .22 .46 .02 .93 — — — .02 .08 — .20 — .02 .08 .07 .06 .04 — .13 — — .10 — — — — — M SE .23 .55 .19 .31 .02 .67 .01 — .07 .01 .07 .04 — — .01 — .07 — .04 — — — .12 .11 .21 — .02 .02 — NC 6.11 5.90 4.01 — 3.95 −3.94 −3.30 −3.11 — −3.16 4.41 −2.53 −2.35 −2.09 — −4.32 −4.05 −2.52 −4.36 3.89 — 2.77 — — 2.61 −2.30 — — 2.13 C 5.92 2.50 3.21 −3.04 3.09 −3.91 — — — −5.56 3.87 — −2.50 — 2.71 −3.96 −6.26 −2.00 −4.60 — 2.39 — — 2.19 — — — — — M 4.20 5.62 3.54 −2.64 3.56 −5.17 −2.25 — 2.96 −6.29 2.22 −2.40 — — 2.60 — −4.38 — −3.90 — — — −2.27 3.93 2.34 — 2.15 −2.68 — NC Critical Ratio Author Manuscript Regression Weights Musil et al. Page 21 — SC2 ←→ SC6 SC5 ←→ SC6 .46 — .27 — .37 .36 — −.46 M — −.27 — — — — .36 — NC — −.58 .20 −.71 — — .32 — C .31 — .32 — .31 .26 — −.36 M Notes: C = custodial; M = multigenerational; NC = noncaregiver; SC = self-control; SE = self-regulatory efficacy. .26 −.43 SC1 ←→ SC4 — −.56 — SE3 ←→ SE6 SC3 ←→ SC6 .50 SE2 ←→ SE4 SC5 ←→ SC3 — C SE1 ←→ SE2 Covariance of residuals associated with latent constructs Author Manuscript Variable Author Manuscript — −.28 — — — — .25 — NC Standardized Regression Weights — .13 .11 .13 — — .13 — C .13 — .10 — .11 .15 — .15 M SE — .09 — — — — .13 — NC — −3.25 2.48 −4.25 — — 3.68 — C 3.57 — 2.68 — 3.29 2.36 — −3.14 M — −2.91 — — — — 2.75 — NC Critical Ratio Author Manuscript Regression Weights Musil et al. Page 22 Author Manuscript J Gerontol B Psychol Sci Soc Sci. Author manuscript; available in PMC 2019 August 11.