|Home | About | Journals | Submit | Contact Us | Français|
This study investigates the implications of residential independence, enrollment in postsecondary education, employment, marital status, and parenthood for contact with, and closeness to, grandparents. Data come from 1,507 young adults interviewed in Wave 3 of the National Survey of Families and Households. Findings suggest that adult roles can be either negatively or positively associated with grandparent-grandchild ties, depending on specific configurations among such factors as the adult role in question, a particular dimension of intergenerational solidarity, lineage, and grandchild’s and grandparent’s gender. Young adults’ ties to parents can mediate the adverse consequences of residential independence for contact with grandparents.
The relationship between grandparents and grandchildren remains important even after the grandchild reaches adulthood (Giarrusso, Feng, Silverstein, & Bengtson, 2001). This relationship can influence grandparents’ and grandchildren’s well-being (Brussoni & Boon, 1998; Hartshorne & Manaster, 1982). Grandparents and adult grandchildren provide each other with emotional and instrumental support (Ashton, 1996; Harwood & Lin, 2000). Grandchildren can also become primary caregivers or at least, co-caregivers for their aging grandparents (Piercy, 1998; Dellman-Jenkins, Blankemeyer, & Pinkard, 2000). Grandparent-grandchild interactions and the availability of assistance for grandparents can be contingent, however, on life course events among grandchildren (Silverstein, Giarrusso, & Bengtson, 1998). Although grandchildren’s adult roles are a likely source of the dynamic nature of grandparent-grandchild ties over the life course, this topic has received relatively little attention from researchers, with two notable exceptions (Crosnoe & Elder, 2002; Mills, 1999).
The main goal of the present study is to examine whether and how grandchildren’s adult roles are linked to their perceptions of contact with, and closeness to, their grandparents. Drawing on data from Wave 3 of the National Survey of Families and Households (NSFH), this paper considers residential independence, enrollment in higher education, employment, marital status, and parenthood among a group of young adults aged 18 to 34 (N = 1,507). In an area of very limited research, this study advances our knowledge by focusing on young adults’ relationships with all living grandparents, by considering whether young adults’ ties to their parents can mediate the associations between grandchildren’s adult roles and their ties to grandparents, and by assessing whether these associations vary by grandchild’s gender.
This study is guided by Rosow’s (1985) role framework and the intergenerational similarity hypothesis (Bengtson & Black, 1973) paired with a model of family stress (McCubbin & Patterson, 1983). These theoretical perspectives are useful for specifying the mechanisms through which grandchildren’s adult roles can shape their ties to grandparents and offer two competing approaches to this issue. Rosow’s role framework suggests that tenuous or ambiguously defined roles, such as grandchild, become less salient to individuals when they acquire roles that are better regulated by social norms and expectations (e.g., roles of worker, spouse, or parent). Responsibilities related to adult roles can limit the amount of time and energy that grandchildren can invest in their relations with grandparents.
In contrast to Rosow’s role framework, the intergenerational similarity perspective and the family stress theory imply that grandchildren’s adult roles can strengthen their ties to grandparents. Bengtson and Black (1973) proposed the intergenerational similarity hypothesis to understand the implications of offspring’s adult roles for parent-child relations. This perspective states that experiences related to adult roles can help offspring grow in their understanding of, and appreciation for, their parents. Through the same process, grandchildren’s adult roles may create extra linkages with grandparents.
The family stress model (McCubbin & Patterson, 1983) suggests that different events in people’s lives can alter family ties by creating stress. However, the direction of change in these ties depends on how relevant stressors are perceived by family members. Thus, although grandchildren’s adult roles can cause stress in the family system, they may represent positive events and thereby, strengthen grandparent-grandchild relations. It should be noted that similarity in roles between generations and positive evaluations of these roles by family members may not always agree, because grandchildren are likely to experience some adult roles that were not encountered by their grandparents (e.g., enrollment in higher education or women’s full-time employment).
The two available studies in this area provided mixed support for the opposing theoretical contentions discussed above. Mills (1999) specifically tested Rosow’s role framework and found that grandchildren’s transitions into adult roles were not necessarily associated with declines in intergenerational solidarity with grandparents, while role losses did not always predict increased solidarity. Overall, Mills’s study suggests that the association between adult roles and grandparent-grandchild relations is complex, and varies by the role in question, by the specific aspect of intergenerational solidarity, and by the grandparent’s gender.
Crosnoe and Elder (2002) focused on grandchildren’s transition from high school to college and provided some support for the idea that grandchildren’s adult roles have positive implications for their ties with grandparents. Both grandchildren and grandparents reported better relationships when the grandchild enrolled in higher education. However, other adult roles (i.e., marriage, parenthood, and employment) were not predictive of the grandparent-grandchild relationship in their study. The latter findings may be explained by the limited age range of the study grandchildren. Few college-age individuals have married, become parents, and obtained a full-time job within two-to-three years after finishing high school.
In sum, theoretical frameworks and prior research suggest two competing hypotheses. According to Rosow’ role framework, grandchildren’s adult roles can be negatively associated with contact and closeness to grandparents. In contrast, the intergenerational similarity framework and the family stress model imply that grandchildren’s adult roles can be positively associated with contact and closeness to grandparents.
Inconsistent results of previous studies suggest that our understanding of these associations is far from complete. Specifically, prior research has three major limitations that the present study improves upon. First, this study takes into account lineage by analyzing young adults’ relationships with all available grandparents, which has not been done before. Literature indicates that because women are more invested in family ties than are men, grandchildren tend to have better relations with maternal grandparents than paternal ones (e.g., Hodgson, 1992). It is unclear, however, whether lineage matters for the associations between grandchildren’s adult roles and their relationships with grandparents.
Second, there is compelling evidence that parents’ intergenerational ties matter for adult grandchildren’s relationships with grandparents (e.g., Brown, 2003; Hodgson, 1992; Monserud, 2008). Yet, prior research did not consider whether parents’ relationships with the grandchildren and grandparent generations can account for the associations between grandchildren’s adult roles and their ties to grandparents. This study explores whether young adults’ relations with their parents can mediate these associations. Parents’ relationships with the grandparent generation were not included in the analyses because these measures were available only for a limited number of grandchildren in the sample.
Third, in contrast to prior research, this study explores whether grandchild’s gender moderates the association between grandchildren’s adult roles and their relations with grandparents. Literature on grandparent-grandchild relations indicates that grandchild’s gender makes a difference. Compared to grandsons, granddaughters have stronger ties to grandparents (e.g., Ashton, 1996; Kennedy, 1991). Furthermore, gender has implications for the process, timing, and consequences of transitions into adult roles due to socialization, cultural norms, and structural factors (Hogan & Astone, 1986; Mahaffy, 2003). Men and women can face different time demands and constraints related to adult roles. For instance, a study by Gauthier and Furstenberg (2002) demonstrates that transitions to partnership and parenthood are associated with major increases in time spent on routine housework for women and with only small increases for men. In contrast, after the transition to parenthood, time devoted to paid work decreases for women and increases for men.
This study uses data from Wave 3 of the NSFH conducted in 2001–2003 (Sweet & Bumpass, 2002). The NSFH data are a stratified, multistage area probability sample of the noninstitutional U.S. population age 19 and older, with an oversampling of some racial groups and family types (Sweet, Bumpass, & Call, 1988). At Wave 1 of the NSFH (1987–88), respondents were interviewed about one of the children under the age 18 residing in the household (i.e., the “focal child”; N = 3,808). At Wave 3 of the NSFH, interviews were conducted with those focal children who were age 18 and older at that time (N = 1,952). The final sample of this study are those focal children who had at least one living biological grandparent at Wave 3 (N = 1,507; ages 18–34). This study refers to focal children as young adults or grandchildren.
Contact with maternal/paternal grandparents represented young adults’ answers to one question capturing how often they saw, talked on the phone, or received a letter or e-mail from grandparents of a given lineage during the last year (1 = not at all, 6 = more than once a week). Closeness to each grandparent measured how close grandchildren felt to a specific grandparent on a scale from 0 = not at all close to 10 = extremely close. Because grandchildren reported on closeness with none to four grandparents, depending on the number who were still living, closeness to each grandparent was considered in turn.
Separate residence measured whether young adults were living separately from their parent(s) (0 = no, 1 = yes). Enrolled in school captured whether grandchildren were enrolled in any kind of postsecondary school (0 = no, 1 = yes). Three dummy variables were created to measure whether young adults had a full-time job (i.e., 30 hours or more per week), part-time job, or were not employed (reference category). Three dummy variables captured grandchildren’s marital status: married, cohabiting, or single (reference category). Parent measured whether young adults had children (0 = no, 1 = yes). Parent of an infant reflected whether young adults had a child under the age 1 (0 = no, 1 = yes).
Young adults ranked their relationship with each parent on a scale from 0 = really bad to 10 = absolutely perfect. Contact with each parent was created by averaging young adults’ reports on two questions regarding two types of contact with each parent over the last 3 months: face-to-face contact and communication by phone, letter, or e-mail (1 = not at all, 5 = more than once a week).
Granddaughter measured grandchild’s gender (0 = male, 1 = female). Grandchild’s age was measured in years (18–34). As no item regarding race or ethnicity was asked of grandchildren, race/ethnicity of the parent from Wave 1 interviews was used as a proxy measure of grandchild’s race. Race was coded 1 for White and 0 for non-White. It was not feasible to distinguish among the racial/ethnic backgrounds of non-Whites in the sample. The sample was over 86% White, with the remaining 14% split over a number of groups leaving no sizable sample in any one minority group, particularly because models for closeness to grandparents were conducted for each living grandparent.
Number of siblings measured how many brothers and sisters, including step- or half-siblings, grandchildren had (0 = no siblings, 4 = 4 or more siblings). Grandchild’s education and parental education captured the highest level of education completed by grandchildren (6 = 6th grade, 20 = doctorate) and by one of their parents (3 = 3rd grade, 20 = doctorate). Parental education was taken from parents’ interviews at Wave 1 because some parents did not participate in Waves 3 (15%) and 2 (7%). This variable captured parental education of only one parent because this information was available for both parents only if they were married to each other at the time of the interview.
Marital status of young adults’ biological parents was taken from parents’ interviews at Wave 3 and was measured by three dummy variables: parents married, parents not married (reference category), and missing marital status. Missing values for parents’ marital status were not imputed because about 15% of parents did not participate in Wave 3. The non-participation of parents can be an indicator of marital problems. It might have been more difficult to locate those parents who had divorced and moved somewhere else. Both maternal/paternal grandparents alive measured whether both grandparents of a given lineage were still living (0 = no, 1 = yes).
Except for parents’ marital status, missing values on all other independent, intervening, and control variables were handled using the Stata command ice for multiple imputation (Acock, 2005). Individual variables had between 0% to 2% missing values. Descriptive statistics of all variables are presented in Table 1.
Bivariate analyses via zero-order correlations (not shown) indicated that none of the correlations among the independent, intervening, and control variables included in the same regression model exceeded .60. That is, there was no multicollinearity present in the regression models. Models predicting contact with grandparents as a couple were estimated using Ordinal Logistic Regressions, whereas models for closeness to each living grandparent were conducted employing Ordinary Least Squares Regressions. The analysis used weights constructed by NSFH researchers to compensate for unequal probabilities of selection across demographic subgroups. The sample size in the regression models varied reflecting the number of young adults with each type of grandparent(s) still living. Three models were used in the regression analysis for each dependent variable. Model 1 contained measures of grandchildren’s adult roles entered as a block. Model 2 added measures of young adults’ ties to their parents. Model 3 added socio-demographic variables and retained significant interaction terms between adult roles and grandchild’s gender.
Table 2 presents results for contact with grandparents. Model 1 shows independent effects of adult roles on contact with grandparents. Separate residence was predictive of less contact with grandparents, regardless of lineage. Enrollment in postsecondary education had divergent effects on contact with maternal and paternal grandparents. It was related to less contact with maternal grandparents, but to more contact with paternal grandparents. Compared to non-employed grandchildren, those who worked full-time had less contact with maternal grandparents. Unlike single grandchildren, those who were cohabiting reported less contact with maternal grandparents. Parents of children under the age 1 perceived more contact with maternal grandparents.
Model 2 examines whether parent-child ties can mediate the associations between grandchildren’s adult roles and their ties to grandparents. Contact with mothers was positively related to contact with grandparents, regardless of lineage, whereas relationship quality and contact with fathers was positively associated with contact with paternal grandparents. Excluding separate residence and parent of an infant, those adult roles that mattered in Model 1 remained predictive of contact with grandparents after the inclusion of parent-child relations.
Those adult roles that were significant in Model 2 still predicted contact with grandparents after the introduction of control variables in Model 3. Also, being a parent became positively related to contact with paternal grandparents. The associations between adult roles and contact with maternal grandparents did not vary by grandchild’s gender. On the other hand, tests of interaction terms between adult roles and grandchild’s gender indicated that cohabitation was related to more contact with paternal grandparents only for grandsons.
Table 3 presents results for closeness to maternal grandparents. Model 1 shows that enrollment in postsecondary education was associated with less close ties to both maternal grandparents. Also, young adults with infants reported less close relationships with their maternal grandfathers. With the inclusion of parent-child ties in Model 2 and control variables in Model 3, these associations remained predictive of closeness to maternal grandparents. Model 2 indicates that relationship quality with mothers was positively related to closeness to maternal grandmothers. Frequent contact with mothers was predictive of closer ties to both maternal grandparents, whereas frequent contact with fathers was related to less close ties to maternal grandfathers. Tests of interaction terms indicated that enrollment in school and cohabitation were predictive of less close ties to maternal grandfathers only for grandsons.
Table 4 presents results for closeness to paternal grandparents. In Model 1, separate residence was associated with less close relationships with paternal grandmothers. Enrollment in higher education was positively related to closeness with both paternal grandparents. Compared to non-employed grandchildren, those who worked part-time had less close relationships with paternal grandmothers. Except for the association between enrollment in postsecondary education and closeness to paternal grandmothers, these associations remained significant after the inclusion of parent-child ties in Model 2, although relationship quality with fathers was positively related to closeness with both paternal grandparents. Tests of interaction terms indicated some differences by grandchild’s gender (Model 3). Enrollment in higher education was positively related to closeness with paternal grandmothers only for granddaughters. Marriage was associated with less close ties to paternal grandmothers only for grandsons. Cohabitation was predictive of less closeness with paternal grandfathers only for granddaughters. Finally, parenthood had positive implications for closeness with paternal grandfathers only for grandsons.
On the basis of two competing theoretical approaches, this study examined how important role markers of adulthood including residential independence, enrollment in postsecondary education, employment, marital status, and parenthood are related to grandchildren’s perceptions of contact with, and closeness to, grandparents. Consistent with previous research in this area (Crosnoe & Elder, 2002; Mills, 1999), mixed support was found for whether adult roles can weaken or strengthen grandchildren’s ties to grandparents. This study, however, advances our understanding of this issue in three major ways. Unlike prior research, this study considered grandparent’s lineage, took into account young adults’ ties to parents, and investigated the moderating effect of grandchild’s gender. This study indicates that lineage and grandchild’s gender can make a difference in the linkages between adult roles and the grandparent-grandchild ties. Also, young adults’ ties to parents can account for the negative associations between separate residence and contact with grandparents, regardless of lineage.
Findings demonstrate that consistent with Rosow’s role framework, residential independence, enrollment in postsecondary education, full-time and part-time employment, marriage, cohabitation, and parenting an infant can lead to weaker relations between grandchildren and certain grandparents. On the other hand, in support of the intergenerational similarity hypothesis and the family stress model, enrollment in higher education, cohabitation, parenthood, and parenting an infant were predictive of enhanced ties between young adults and certain grandparents. The exact mechanisms underlying the implications of part-time employment and cohabitation remain uncertain in this study. According to Rosow’s role framework, the role of grandchild can become less important to individuals when they are cohabiting or working part-time. At the same time, the family stress model suggests that cohabitation and part-time employment may adversely affect ties to grandparents because young adults may be aware that these events are viewed negatively by grandparents.
This study indicates that parenthood in general as well as parenting an infant can create additional opportunities for young adults’ interactions with grandparents. These findings are consistent with Hodgson’s (1992) research in which adult grandchildren reported that they had become closer to their grandparents because they wanted their own children to get to know their great-grandparents. Another plausible reason for the positive association between parenthood and grandchildren’s contact with paternal grandparents as well as grandsons’ closeness to paternal grandfathers may relate to the paternal grandparents’ aspiration that the family name would carry on. However, this study suggests that time and energy constraints related to parenting an infant can diminish closeness to maternal grandfathers.
The findings consistently demonstrate that young adults’ ties to parents appear to mediate only adverse implications of separate residence for contact with grandparents, regardless of lineage. Other adult roles seem to matter for relations with grandparents despite the importance of parent-child ties. Although it can be more difficult for parents to encourage and provide opportunities for grandparent-grandchild interactions when offspring do not reside in the parents’ household, stronger ties to parents can still help young adults maintain frequent contact with grandparents. Increased geographic distance could also account for the adverse effects of separate residence. This measure, however, is not available in the NSFH. Note that contact with grandparents in this study captured not only face-to-face contact but also communication by phone, letter, or e-mail. The latter types of interactions should be less dependent on proximity.
This study provides consistent evidence that lineage matters for the associations between enrollment in postsecondary education and contact and closeness with grandparents. Enrollment was negatively associated with ties to maternal grandparents and positively related to relations with paternal grandparents. These findings are inconsistent with Crosnoe and Elder (2002) who found only positive implications of enrollment in college. Crosnoe and Elder, however, did not consider lineage. One possible explanation for these lineage differences may be a general matrilineal bias in the grandparent-grandchild relationship. Similar to younger counterparts, adult grandchildren report a preference toward maternal grandparents (Mills, Wakeman, & Fea, 2001). At the same time, students may have fewer opportunities for maintaining strong relations with grandparents, regardless of lineage. College students also report being considerably less satisfied with the amount of contact that they have with grandparents while in college compared to the amount they had during high school or home full time (Hartshorne & Manaster, 1982). Grandchildren’s reports may reflect their greater “guilt” about adverse effects of being a student on their ties with maternal grandparents. Alternatively, grandchildren can perceive that positive evaluations of the student role by grandparents provide a certain boost for their relations with paternal grandparents. Future studies would benefit from taking into account geographic proximity and grandparents’ financial support to grandchildren. College students’ feelings of closeness to grandparents can be contingent on whether grandparents are available to help, encourage, and support them (Kennedy, 1991).
The tests of interaction terms provided some support for the idea that grandchild’s gender can moderate the associations between adult roles and grandchildren’s ties to grandparents. This study suggests that the implications of enrollment in postsecondary education, marriage, cohabitation, and parenthood can vary for grandsons and granddaughters. Additional research is needed for investigating specific mechanisms of these gender differences.
The present study has limitations that need to be considered. First, this study did not test why adult roles can have either negative or positive consequences for the grandparent-grandchild relationship because relevant measures are not available in the NSFH. Second, the paper did not take into account grandparents’ attributes because this information was available only for certain grandparents, depending on the marital status of young adults’ parents. Third, because of data limitations, I examined only the perspective of grandchildren. Grandparents’ perceptions may be less sensitive to the effects of grandchildren’s adult roles because grandparents tend to report higher relationship quality than do grandchildren (Bengtson & Kuypers, 1971).
Despite its limitations, this study sheds light on times in the life course when certain events experienced by grandchildren may make them less available to grandparents if the latter need help. Yet, intergenerational support for older family members is potentially more salient today than in the past because population aging is likely to entail shortages of national resources allocated to health care and other types of services for the elderly (Putney & Bengtson, 2003). It is important, therefore, to make additional resources (e.g., affordable caregiving services) available to families when circumstances prevent potential family caretakers from providing functional assistance to the elderly family members. Deteriorating relations with grandchildren may also affect grandparents’ overall well-being (Kivnick, 1985; Forsyth, 1994). However, possible adverse implications of adult roles do not necessarily mean that grandchildren do not want to maintain strong ties to grandparents. When problems arise, it can be important for family practitioners and mental health professionals to encourage grandparents to initiate interactions with their adult grandchildren on their own even if grandchildren appear to be “too busy” for their grandparents.
The preparation of this article was supported by a grant from NIH/NIA to Glen H. Elder, Jr. for his Demography of Aging and the Life Course training grant (5 T32 AG00155-14). An earlier version of this paper was presented at the 2009 Annual Meeting of the Population Association of America, Detroit, MI. I thank Glen H. Elder Jr., Peter Uhlenberg, and Monica Kirkpatrick Johnson for their comments on previous versions of this manuscript.