| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
Department of Social Work, California State University, Long Beach.
Address correspondence to Catherine Chase Goodman, DSW, Professor of Social Work, California State University, 1250 Bellflower Boulevard, Long Beach, CA 90840. E-mail: Goodman{at}csulb.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Methods. Interviews were conducted with 987 grandmothers recruited through schools and media, including 512 custodial grandmothers raising their grandchildren and 475 coparenting grandmothers helping their adult sons or daughters to raise their grandchildren.
Results. Contrary to expectation, triangulation involving a weak parent–grandmother relationship was not related to significantly lower grandmother well-being. In contrast, the emotionally isolated parent, particularly common in custodial families, was related to lower grandmother well-being, whether or not the parent lived in the household. The configuration in which the parent provided a link for both other generations was frequent in coparenting families, but it was not significantly different in grandmother well-being compared to other linking triads.
Discussion. Intergenerational triad types appear to be broadly relevant emotional configurations that transcend living situations but appear with different frequencies in custodial and coparenting families. Results affirm the utility of examining multirelationship patterns and describe the restructuring of intergenerational relationships characteristic of parental absence.
TRADITIONALLY, grandparents have stepped in to assist their children and grandchildren during times of crisis. During the past 30 years, this tradition has become the safety net for children whose parents are unable to provide parenting or shelter for them. The number of children raised in grandparent-headed families has more than doubled over the past 30 years, from 2.2 million in 1970 (3.2% of all children) to over 4.5 million in 2000 (6.3% of all children; see Bryson, 2001
; Casper & Bryson, 1998). Recently, research has begun to determine the needs of family members and the dynamics of this increasingly prevalent family structure (see Cox, 2000; Hayslip & Goldberg-Glen, 2000; Szinovacz, 1998). This study examines patterns of emotional bonds—alliances and coalitions—among the three generations in relationship to the well-being of the caregiving grandmother when circumstances lead to the parent's absence (custodial families) and when the three generations live together (coparenting families).
Grandparents in custodial families assume care when parents are unable to care for their children, typically in response to problems such as parental drug addiction, mental instability, or abuse or neglect of the grandchild. In contrast, grandparents in coparenting families are more likely to be assisting parents and grandchildren in response to financial need, parental divorce, or the parent's work or school responsibilities (Goodman & Silverstein, 2002). In both situations, the grandmother's role is transformed to include parenting of her grandchild, although the extent varies from complete care in custodial families to supplemental care in coparenting families. Comparisons of coparenting and custodial families have shown that coparenting grandmothers have less caregiver burden (Pruchno & McKenney, 2000), greater instrumental support (Musil & Ahmad, 2002), and higher well-being among Latina, but not African American or White, grandmothers (Goodman & Silverstein, 2002). Studies on national samples determined that the assumption of the custodial role by grandparents has been related to greater depression (Minkler, Fuller-Thomson, Miller, & Driver, 1997) and lower health status (Solomon & Marx, 2000) compared with the enactment of more traditional roles by grandparents. In particular, the grandchild's transition into the grandparent's household has been related to the grandmother's depression (Szinovacz, DeViney, & Atkinson, 1999). For coparenting grandparents, greater depression (Musil & Ahmad, 2002) and lower health status (Giarrusso, Feng, Wang, & Silverstein, 1996) have also been identified compared with grandparents in traditional roles.
The grandparent's well-being is shaped by the quality of relationships with parents and grandchildren. In custodial families, the relationship with the parent is often disrupted by crises, which have led to the assumption of care by the grandmother. Nevertheless, close relationships between custodial grandmother and parent were related to higher caregiver satisfaction (Pruchno & McKenney, 2002). Furthermore, conflicted grandmother–parent relationships were related to lower well-being in custodial and coparenting families (Goodman & Silverstein, 2002). In coparenting families, conflict and low cohesion have been related to maternal or grandmother depression (Caldwell, Antonucci, & Jackson, 1998; Kalil, Spencer, Spieker, & Gilchrist, 1998). The grandchild's behavior is also important for the grandmother's well-being (Hayslip, Shore, Henderson, & Lambert, 1998; Pruchno, & McKenney, 2002), but research strategies do not include a simultaneous examination of patterns of relationship—alliances and coalitions—among members of all three generations and the relationship of these patterns to grandmother well-being.
Intergenerational Triads
The conceptual model for the study is based on two aspects of family structure. The model is developed from sociological literature on family solidarity (Silverstein & Bengtson, 1997; Silverstein, Lawton, & Bengtson, 1994) and from family systems literature on cohesion (Olson, 2000) and triangulation (Bowen, 1978; Guerin, Fogarty, Fay, & Kautto, 1996). Household structure is defined by the presence (coparenting) or absence (custodial) of the parent, reflecting the degree of proximity and opportunity for interaction with the middle generation.
Intergenerational triads, defined by affective bonds among grandmother, parent, and grandchild, involve the intricate dynamics of coalitions and alliances within a three-person social system. As represented by sociologist Georg Simmel (see Wolff, 1950), the triad is an unstable social system because it can subgroup as a dyad and an isolated individual. Alternatively, one member may mediate between the other two or generate conflict or competition in order to gain influence and power. In the model examined here, triads may be completely connected or cohesive, in which all members are close. At the other extreme, weak intergenerational relationships constitute a disconnected triad. Between these extremes are triads in which one generation has close bonds and links between two estranged or conflicted generations (linking triads) and triads in which one generation is emotionally isolated from both of the other two (isolated triads; see Figure 1). Thus, depending on which family member provides a link or is emotionally isolated, there are eight possible triad types: connected; child, parent, or grandmother linked; child, parent, or grandmother isolated; and disconnected.
|
; Silverstein et al., 1994). Research on the three generations has identified the middle generation as a link, occupying a "gatekeeper" function (Uhlenberg & Hammill, 1998, p. 277) between adjacent generations. Good relationships between parent and grandparent have been related to more frequent grandchild–grandparent contact and better quality relationships (King & Elder, 1995; Uhlenberg & Hammill, 1998), emphasizing the role of the parent as the emotionally influential mediator in the triad. In grandparent-headed families, the parent-linking role is more feasible in coparenting families than in custodial families.
Bowen, a family systems theorist, referred to emotional triangles as the building blocks of families (Bowen, 1978; Guerin et al., 1996). He used the concept of triangulation to refer to a dysfunctional pattern in which distant or conflicted parents manage their own tensions by scapegoating their child or creating an alliance between one parent and child. According to recent research, higher ratings of triangulation have been related to depression and poorer development among children (Bell, 2001; Wang & Crane, 2001). Furthermore, cross-generational coalitions, one form of triangulation, have been related to lower well-being and health status of the parent (Flemons & Tsai, 1992); triangulation in parenting has also been related to marital conflict and stress (Margolin, Gordis, & John, 2001). In survey research, affective aspects of relationship (agreement, closeness, or warmth) have been used to define triangulation (Bell, 2001), cross-generational coalitions (Flemons & Tsai, 1992), and interdependence in the parent–child triad (White, 1999), although conceptualization may contain other aspects, such as parental manipulations (Margolin et al., 2001). Triangulation has been alluded to in relationship to grandparent families, but it has seldom been researched (Goodman & Silverstein, 2001).
Hypotheses
This study attempts to capture the complex alliances and coalitions in grandparent-headed families and to examine the relationship between triad types and grandmother well-being. Although causality between well-being and relationship configurations may be circular and complex, the goal is to establish associations rather than imply causation. Beyond expectations that extreme triad types (connected and disconnected) will be related to higher and lower well-being respectively, three hypotheses address well-being in linking or isolated triads.
1. Child-linked triads (in which the grandmother–parent relationship is weak in spite of the child's close relationships) will be related to lower grandmother-well-being for both custodial and coparenting grandmothers. This triad represents triangulation, in which a weak relationship between parent and grandmother undermines parental authority and the child is caught in the middle with divided loyalties.
2. Parent-linked triads (in which only the parent has strong relationships with both other generations) will be more prevalent and favorable for grandmothers in coparenting than in custodial families. The strong position of the parent in the parent-linked triads emphasizes the parent's traditional role. For the custodial grandmother, a weak relationship with her grandchild would be detrimental to her caregiving function.
3. Triads in which the parent is emotionally isolated will be related to lower well-being in both custodial and coparenting families. Emotional isolation of the parent may be a consequence of the crisis in which the grandmother assumed care or a survival strategy to minimize the disruption caused by a troubled family member in custodial families. In coparenting families, paternal emotional disengagement would be a constant source of stress.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
The sample was recruited through the Los Angeles Unified School District (LAUSD; 63.4%) and through media announcements (36.6%). School recruitment consisted of sending notices home with 32% of the children in the LAUSD, including all students in 223 elementary, middle, and high schools. The notices announced the study and incentive ($15 and a $5 McDonald's gift certificate). Media recruitment consisted primarily of public service TV and radio announcements and newspaper advertisements. Face-to-face structured interviews, lasting roughly 1 hr, were conduced by a team of ethnically and linguistically matched, trained interviewers through the Survey Research Center, Institute for Social Science Research, University of California at Los Angeles, between November 1998 and January 2001.
Measures
Measures not already in Spanish were translated by a bilingual psychologist and reviewed by bilingual human service professionals from different nationalities to account for regional differences (Mexican, Spanish, and Salvadoran). Demographic and social descriptors were grandmother's age, ethnicity, education, employment status, income, country of birth, and key reasons for assuming care that differentiate household structures. An additional descriptor, physical health, was measured by use of the SF-36 (Ware, Kosinski, & Keller, 1994). One target grandchild, either the child identified through school recruitment or a random grandchild, was described in detail by the grandmother, including gender, age, time living with the grandmother, and whether the child had behavioral or emotional problems. Coparenting and custodial families were distinguished by whether or not a parent currently lived in the household. Respondents also made open-ended comments about their relationships.
Intergenerational closeness and triad measurement
Five items from the Bengtson Scale of Intergenerational Solidarity were used to measure closeness (Mangen, Bengtson, & Landry, 1988). Ratings were on a 6-point scale from 1 (not at all) to 6 (extremely) and included questions on emotional closeness, getting along together, affection, communication, and understanding. Therefore, this measure of emotional closeness also includes aspects of compatibility, communication, and understanding. The grandmother rated her relationships with the target grandchild and the parent (her adult son or daughter), as well as the relationship between this parent and the target grandchild. The coefficient alpha was.84 for grandmother–grandchild,.92 for grandmother–parent, and.94 for grandchild–parent relationships.
Grandmother well-being
Two indices represented the grandmother's well-being. Depression was measured by use of the Center for Epidemiologic Studies Depression (CES-D) scale, which is a widely used 20-item measure extensively used in the general population, as well as in clinical samples (Radloff, 1977). Respondents used 4-point scales from 0 (rarely, none of the time) to 3 (most or all of the time) to rate how they felt during the past week. The coefficient alpha for this sample was.89. Life satisfaction was measured by use of the Satisfaction With Life Scale (Diener, Emmons, Larsen, & Griffin, 1985), a 5-item measure scored on a 7-point Likert scale, from 1 (strongly agree) to 7 (strongly disagree). The coefficient alpha for this sample was.84.
Analysis
An SPSS cluster analysis for k means was used to develop triad clusters for the eight theoretical triad types. Data were classified according to predetermined, researcher-set cluster centers using z scores for relationship closeness between the grandmother and parent, grandmother and grandchild, and parent and child. Classification was based on a process of successive iterations using an algorithm, with cluster centers computed with simple Euclidean distance. Although the procedure may be used in an exploratory way to identify clusters within data, it was used here to confirm theoretical triad types and partial iteration was allowed (two iterations; convergence set at.10). This number of iterations allowed for a theoretically consistent set of clusters, which were distinct and coherent (see Figure 2).
|
). |
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
|
Simple triad main effects within custodial and coparenting households
Simple main effects for triads within household structures were calculated, which was an approach indicated by the significant interaction. All pairwise comparisons were conducted, and the Bonferroni method was used to adjust alpha for the number of multivariate pairwise tests (
=.05/56 =.001). For ease of presentation, triad simple main effects comparisons are displayed to show individual linking and isolated triads against extreme categories (connected and disconnected), which served as index categories (see Table 4). Other significant pairwise comparisons appear in the narrative. It is notable that these index categories performed differently in custodial and coparenting families. Custodial connected families scored particularly high, yielding more significant distinctions from other triad types, whereas coparenting disconnected families scored particularly low, resulting in a greater number of significant comparisons. Effect size (
2), although somewhat low, was greater for triad types than for any of the covariates. As a follow-up, all pairwise comparisons were conducted for each dependent variable separately in order to display the direction of differences. Alpha values for these univariate tests were adjusted for two dependent variables and two levels of household structure ( =.05/2 =.025/2 =.01), and Tukey's post hoc adjustment was used to control for multiple tests (see Table 5). Although univariate tests help to clarify the direction of results, the two dependent variables were correlated (r = –.362): Univariate tests treat them as if they were separate measures, not taking this overlap into consideration.
|
|
Hypothesis 2: Parent connection
No significant differences in well-being emerged between custodial and coparenting grandmothers within parent-linked families. Although redundant statistically, comparisons were conducted between custodial and coparenting groups within each triad type. These showed no significant differences after alpha was adjusted for the number of tests ( =.05/8 =.006). Therefore, the hypothesis was not confirmed that there would be definite differences between custodial and coparenting grandmothers when the parent played the primary linking role. However, as predicted, few custodial families were classified as being parent linked. This small group of custodial grandmothers had lower life satisfaction than those in connected triads. In coparenting families, grandmothers in parent-linked triads had less depression and higher life satisfaction compared with those in disconnected triads (see Table 5).
Hypothesis 3: Isolated parents
Isolated parent triads were related to lower well-being for custodial and coparenting grandmothers, confirming this hypothesis. Grandmothers in isolated parent triads had life satisfaction and depression levels comparable with those in disconnected triads and were more depressed than grandmothers in connected triads in both custodial and coparenting families (see Table 5). For coparenting families, the isolated parent triad stood out in multivariate comparisons as the only isolated triad to have lower grandmother well-being compared with the positive index (connected) and the only isolated triad to show no differentiation from the negative index (disconnected; see Table 4). As further evidence of their negative standing, isolated parent triads were also significantly differentiated from one of the linking triads, that is, grandmother linked, F = (2,957) = 5.46, p =.000, and 2 =.011 in coparenting families.
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
The study attempted to identify differences in well-being of the caregiving grandmother in eight intergenerational triads—from all close relationships to all weakened or disconnected relationships. Contrary to the hypothesis, triangulation, involving a weakened relationship between parent and grandmother, was a relatively favorable configuration for grandparent-headed families in both family structures and was not associated with significantly lower levels of well-being than connected or other linking triads. As hypothesized, triads in which the parent was isolated emotionally proved to be a negative configuration for both family structures, although this situation was infrequent in coparenting families. In contrast, parents seldom performed a linking function in custodial families. Overall, grandmothers in parent-linked triads had well-being comparable with those in other linking triads. The patterns of grandmother well-being between custodial and coparenting families were not clearly differentiated and household differences were more powerfully shown in the distribution of families to triad types. Therefore, the triad types appear to be broadly relevant emotional structures that transcend living situations but appear with different frequencies in custodial and coparenting families.
This study adds to the growing literature showing minimal or qualified grandmother well-being differences in comparisons of custodial and coparenting grandparent families: Differences have emerged in only a few areas (Musil & Ahmad, 2002; Pruchno & McKenney, 2000) or have depended on ethnic background (Goodman & Silverstein, 2002). Findings give credibility to the position that emotional bonds with the parent are more salient and powerful than physical proximity and opportunity for interaction, which distinguish custodial and coparenting families.
Triangulation, characterized as the child-linked triad, was comparable with other linking triads, in contrast to pilot research (Goodman & Silverstein, 2001). Previous research on parent-headed families showed triangulation related to lower parental well-being (Flemons & Tsai, 1992). Most literature on triangulation has been based on parent-headed families (Flemons & Tsai, 1992; Margolin et al., 2001) and child outcomes (Bell, 2001; Wang & Crane, 2001). The intergenerational parenting unit of grandmother and parent, which involves an age and relationship hierarchy, may have altered triangular processes. For example, the effects of triangulation may be attenuated as a result of parental absence in custodial families and the desire for appropriate parental authority in coparenting families. Comments by grandmothers illustrated ambiguous parental authority and the child's divided loyalties: A custodial grandmother stated, "My granddaughter shows off when her mother comes around. I don't care for that too much; it upsets me." A coparenting grandmother said, "Trying not to get between her and Mom and what she tells her [is the hard part]."
The relationship pattern of the parent-linked triad captures the parent's strong, emotionally close relationships with other generations. Although literature emphasizes parental influence in grandparent–grandchild relationships (King & Elder, 1995; Uhlenberg & Hammill, 1998), the parent link demonstrates a weakened grandmother–grandchild bond. In custodial families, parent-linked triads were infrequent and may reflect the grandchild's anger at a substitute parent or a family moving toward parent–child reunification. In the coparenting family, this triad may represent society's expectation of close parent–child bonds. A custodial grandmother said, "I feel he still doesn't trust me.... ‘You're not my mother, you don't mean nothing' attitude." A coparenting grandmother stated, "Being the grandmother, he doesn't open up as much as I would like. He goes with his mother and father. It's natural."
Isolated parent triads were negative for grandmothers in both household structures. Although isolated parents are infrequent in coparenting families, some grandmothers were contending with marginal, preoccupied, or immature parents who were poorly integrated into the emotional life of child and grandmother. Previous studies have identified disengaged or dependent mothers in three-generational families (Apfel & Seitz, 1991; Oyserman, Radin, & Saltz, 1994). Grief over loss of parent is frequently referred to in literature describing custodial families (Hirshorn, 1998). A coparenting grandmother said, "She can't be responsible for herself how can she be responsible for them? All she wants do is sit on couch and read ... no work, nor school, no job." A custodial grandmother stated, "Answering questions [is hard] ... ‘why doesn't my mom want me?'" Another custodial grandmother commented about her daughter, "I miss the person that she was."
Limitations and Implications for Research
This large sample provided considerable power to address eight triad configurations. However, in order to investigate intergenerational triads, the complex and rich family system was simplified, eliminating the influence of grandfathers, siblings, and others in the household. Grandfathers, especially, have parenting authority and impact the well-being of the grandmother by supporting or undermining her caregiving role. Because the assessment of emotionally close relationships would be different for infants and toddlers, only grandmothers caring for school-aged children were included. Furthermore, only the viewpoint of the caregiving grandmother was used instead of viewpoints from many family members. Her tendency to emphasize closeness to subsequent generations, referred to as intergenerational stake, may have influenced her ratings. In particular, the overall high rating by grandmothers of closeness in their relationship with their grandchildren may also indicate a substantial social desirability bias, although variability was sufficient to construct triads in spite of this skew. A fuller understanding of these multiperson family structures should include focus on the well-being of other family members—particularly the grandchild and parent.
Although effect size in this study is somewhat low, it is comparable with effect size of demographic and social variables, which typically together account for 10–15% of the variance in studies of well-being (Argyle, 1999). The focus on triads has precluded the addition of other influential factors to the model, such as experience of stress or support (Musil & Ahmad, 2002) or the grandchild's behavior problems (Hayslip et al., 1998; Pruchno & McKenney, 2002). Future research may further characterize important family types such as child-linked, parent-linked, or isolated parent triads and determine the importance of these structures relative to other social factors.
Implications for Practitioners
The triad model helps identify families in trouble in which low grandmother well-being might compromise caregiving, particularly those with isolated or disconnected family members. These families typically struggle with grief, overlapping parenting roles, and conflict between the generations (Hirshorn, 1998). Family therapists may assist in developing family rules and alliances that promote clear parenting authority (Qualls, 2000). They may support the entire family by encouraging emotional inclusion of the parent when possible; helping parents deal with substance abuse and developmental issues; and assisting grandmother and grandchild through the anger, guilt, and sadness of grief when relationships with the parent cannot be sustained. Service professionals need to assist grandparent-headed families through grief groups, family therapy, service advocacy, and legislative protection. As we continue to see the growth of diverse and adaptive family forms in response to social change, society should support their optimal functioning. The positive bonds between grandmother and grandchild create a legacy for future generations so that tomorrow's grandparents may also provide a safety net and sustain their families through intergenerational interdependence.
|
Acknowledgments |
|---|
I wish to thank Merril Silverstein, Andrus Gerontology Center, University of Southern California, for methodological and conceptual contributions to this article as well as support and collaboration throughout the grant.
|
Footnotes |
|---|
Received for publication November 6, 2002. Accepted for publication March 10, 2003.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. Kopera-Frye Needs and Issues of Latino and Native American Nonparental Relative Caregivers: Strengths and Challenges Within a Cultural Context Family and Consumer Sciences Research Journal, March 1, 2009; 37(3): 394 - 410. [Abstract] [PDF]
|
|
|
|
 |
|
 T. S. Standing, C. M. Musil, and C. B. Warner Grandmothers' Transitions in Caregiving to Grandchildren West J Nurs Res, August 1, 2007; 29(5): 613 - 631. [Abstract] [PDF]
|
|
|
|
 |
|
 C. C. Goodman Family Dynamics in Three-Generation Grandfamilies Journal of Family Issues, March 1, 2007; 28(3): 355 - 379. [Abstract] [PDF]
|
|
|
|
 |
|
 C. M. Musil, C. B. Warner, J. A. Zauszniewski, A. B. Jeanblanc, and K. Kercher Grandmothers, caregiving, and family functioning. J. Gerontol. B. Psychol. Sci. Soc. Sci., March 1, 2006; 61(2): S89 - S98. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
