The words you are searching are inside this book. To get more targeted content, please make full-text search by clicking here.

1246 Journal of Family Issues 1989), most studies indicate that on average divorce is associated with nega-tive short- and long-term effects for children, such as ...

Discover the best professional documents and content resources in AnyFlip Document Base.
Search
Published by , 2016-12-29 22:40:03

Journal of Family Issues

1246 Journal of Family Issues 1989), most studies indicate that on average divorce is associated with nega-tive short- and long-term effects for children, such as ...

Journal of Family Issues
http://jfi.sagepub.com/

Grandmother−−Grandchild Relationship Quality Predicts
Psychological Adjustment Among Youth From Divorced Families

Craig E. Henderson, Bert Hayslip, jr, Leah M. Sanders and Linda Louden
Journal of Family Issues 2009 30: 1245 originally published online 17 April 2009

DOI: 10.1177/0192513X09334913
The online version of this article can be found at:

http://jfi.sagepub.com/content/30/9/1245
Published by:

http://www.sagepublications.com

Additional services and information for Journal of Family Issues can be found at:
Email Alerts: http://jfi.sagepub.com/cgi/alerts

Subscriptions: http://jfi.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://jfi.sagepub.com/content/30/9/1245.refs.html

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Grandmother–Grandchild Journal of Family Issues
Relationship Quality Volume 30 Number 9
Predicts Psychological
Adjustment Among Youth September 2009 1245-1264
From Divorced Families © 2009 Sage Publications
10.1177/0192513X09334913
Craig E. Henderson
Sam Houston State University, Huntsville, TX http://jfi.sagepub.com
Bert Hayslip Jr. hosted at
University of North Texas, Denton
Leah M. Sanders http://online.sagepub.com
Sam Houston State University, Huntsville, TX
Linda Louden
Texas Woman’s University, Denton

This study investigates maternal grandmother–grandchild relationship quality
as a predictor of psychological adjustment among youth from divorced fami-
lies. Three hundred twenty-four adolescents aged between 17 and 20 report on
the quality of their relationships with their maternal grandmothers and their
relational competence, self-efficacy, and psychological symptoms. Structural
equation modeling analyses support a model in which participants’ relation-
ships with grandmothers predict their psychological adjustment. Family back-
ground (divorced vs. intact families) moderates the relationship between
relationship quality and adjustment; youth from divorced families indicate that
their relationships with their maternal grandmothers are more salient to their
adjustment than do youth from intact families. These findings suggest that the
bonds young people develop with their maternal grandmothers following their
parents’ divorce may positively affect their psychological functioning.

Keywords:  divorce; grandparents; well-being; structural equation modeling

Each year, the parents of 2% of children residing in the United States
divorce, affecting 1.5 million children (Department of Health and Human
Services, 2005). Although some studies suggest that not all children suffer
from negative long-term effects following parental divorce (Hetherington,

1245

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1246   Journal of Family Issues

1989), most studies indicate that on average divorce is associated with nega-
tive short- and long-term effects for children, such as difficulty relating to
peers and family members, as well as increases in psychological symptoma-
tology (Hetherington, Bridges, & Insabella, 1998; Wallerstein, 1987). Two
meta-analyses examining the effects of divorce on children have indicated
that children whose parents divorced had more difficult psychological adjust-
ment, lower scholastic achievement, lower self-esteem, and more strained
peer and parent–child relationships than their same-age peers from intact
families (Amato, 2001; Amato & Keith, 1991). Adolescents are likewise
vulnerable to negative consequences of parental divorce, even if the divorce
occurred several years in the past (Wallerstein, 1987).

A possible mechanism by which children exhibit postdivorce maladjust-
ment is through deteriorated parent–child relationships (Amato & Booth,
1996). Parents who have divorced report less affection and interaction with
their children than parents who have not divorced (Amato & Booth, 1996).
Recent studies applying attachment theory to broader family relationships
and ages beyond infancy suggest its relevance to negotiating family transi-
tions such as divorce (Benoit & Parker, 1994). From this theoretical perspec-
tive, attachment bonds with parents may be disrupted as children experience
multiple transitions, including separation of their biological parents, living
in a single-parent household, and forming new parental relationships with
stepparents (Lussier, Deater-Deckard, Dunn, & Davies, 2002). Such life
transitions underscore the importance of attachment relationships outside of
children’s divorcing parents (Kenny & Barton, 2003). Indeed, grandchildren
often become closer to their grandmothers following divorce (Gladstone,
1988), particularly those on their mothers’ side, a trend that may be related
to parental custody (Cherlin & Furstenberg, 1986).

Grandchildren from divorced families are likely to benefit from this
increased involvement with their grandchildren by gaining more emotional
support and having less strain in their lives (Johnson, 1998; Lussier et al.,
2002). However, the results from studies examining whether having strong
relationships with grandmothers helps reduce the negative effects of divorce
on children are equivocal. Some studies have shown that better relation-
ships with grandmothers following divorce is associated with grandchil-
dren having fewer internalizing and externalizing problems (Lussier et al.,
2002). Furthermore, Kennedy and Kennedy (1993) hold that grandmother

Authors’ Note: Please address correspondence to Craig E. Henderson, Department of
Psychology, Sam Houston State University, Campus Box 2447, Huntsville, TX 77341; e-mail:
[email protected].

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1247

involvement may compensate for parents’ physical and emotional unavail-
ability as the family adjusts to new configurations and economic hardship.
In other cases, even a symbolic, sentimental role may help ameliorate nega-
tive aspects of divorce (Johnson, 1998).

In contrast, other research has found that the grandmother–grandchild
relationship had little impact on the postdivorce adjustment of the grand-
children (Cherlin & Furstenburg, 1986; Hetherington, 1989). Hetherington
and colleagues have conducted some of the most comprehensive, longitu-
dinal work examining vulnerability and protective factors that contribute to
children’s long-term adjustment to divorce and remarriage (Hetherington,
1989). Hetherington reports that although grandmothers offer support to
their divorced and remarried children, as well as their grandchildren, “there
was little evidence that grandparents play a potent role in the social, emo-
tional, and cognitive development of their grandchildren unless they live in
the home” (p. 10).

However, some caveats are in order. First, Hetherington’s findings imply
that grandmothers can have greater impact on their grandchildren when
they live with them. Indeed, a cohort shift has occurred in the 15-plus years
since Hetherington conducted her research, with grandparents today being
more likely to assume full-time care of their grandchildren (Hayslip &
Goldberg-Glen, 2000) as well as perceive their role as more hands-on
(Hayslip, Henderson, & Shore, 2003). Cohort differences in the meaning
and activities of grandparenting may partially explain discrepancies in the
literature between studies suggesting that the quality of grandmother–
grandchild relationships may diminish adjustment problems (Lussier et al.,
2002) and those that do not (Cherlin & Furstenberg, 1986). Second, grand-
mothers’ active involvement in the lives of their grandchildren postdivorce
may be a mixed blessing: Although mothers report appreciating the support
and assistance, they also report conflict with their mothers when grand-
mothers take an active role in their grandchildren’s lives (Hetherington,
1989). Therefore, studies examining the impact of grandmother–grandchild
relationships on grandchildren’s psychological adjustment may be needed
at a point of development during which grandchildren are more autono-
mous and the postdivorce adjustment less acute. Few studies have com-
pared emerging adults from intact and divorced homes with respect to the
quality of their relationships with their grandparents. One exception to this
general trend is the work of Cooney and Smith (1996), who found that
young adults did not report closer relationships with their grandparents fol-
lowing a recent divorce. However, Cooney and Smith limited their partici-
pants to those whose parents divorced within 15 months. Furthermore, they

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1248   Journal of Family Issues

did not examine the impact that relationships with grandparents may have
on the participants’ psychological adjustment.

This study compares participant reports of their relationship quality with
their maternal grandmothers and psychological adjustment among youth in
late adolescence or emerging adulthood (ages 17-20) from divorced and
intact families. By studying older adolescents, we minimize the likelihood
of parent–grandparent conflict over child rearing. Furthermore, studying
older adolescents extends the literature on what we know about grandmoth-
ers’ roles in buffering the impact of divorce on children, which Hetherington
(1989) and others (Wallerstein & Lewis, 1998) indicate may persist even
into young adulthood. Only data on maternal grandmothers were selected
for analyses in this study based on previous research, which suggests the
following: (a) Grandparent contact following divorce is much higher for
maternal grandmothers than other grandparents (cf. Johnson, 1998); (b) the
vast majority of young people have reported that their maternal grandmoth-
ers are significantly more influential in their current lives than are other
grandparents (Kennedy, 1992); and (c) relationships with maternal grand-
parents are associated with psychological adjustment to a greater extent
than other grandparent configurations (Lussier et al., 2002). The particular
indices of psychological adjustment were selected because meta-analyses
have suggested that they are some of the most important in understanding
youths’ postdivorce adjustment (Amato, 2001; Amato & Keith, 1991).

In light of the above literature highlighting the role that grandparents may
play in the adjustment of youth to the divorce of their parents, we propose
the following hypotheses: (1) Among youth from divorced families, higher
levels of relationship quality with grandmothers will be associated with
lower levels of psychological symptoms and higher levels of relational com-
petence and self-efficacy, and (2) the relationship between relationship qual-
ity and psychological symptoms, relational competence, and self-efficacy
will be moderated by family type; that is, they will be stronger for youth
from divorced families relative to youth from intact families.

Method

Participants and Procedure

Three hundred twenty-four youth aged between 17 and 20 whose mater-
nal grandmothers were living at the time the study was conducted were
recruited from local high schools and from entry-level psychology classes

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1249

at a university located in north Texas. We intentionally targeted these two
populations because many studies have not explicitly focused on this devel-
opmental period (emerging adulthood) and studies with this particular age
group are mostly absent in the divorce literature. Although we attempted to
make the recruitment strategies as similar as possible (identical recruitment
scripts and study descriptions), it was necessary to make some modifica-
tions to recruit the separate populations. With respect to the high school
students, meetings were arranged with principals at local high schools to
secure their agreement to announce the study in classes and at Parent
Teacher Association meetings. In compliance with research ethical stand-
ards, written parent permission was obtained for students under age 18, and
these students also provided their informed assent to participate in the
research. Because we could not provide extra credit for high school stu-
dents (which we did as an incentive for the college students), a raffle for six
$50 gift certificates was offered as an incentive for participation.

With respect to the college students, researchers presented the study in
entry-level psychology classes; class extra credit commensurate with the
amount of time it took to complete the research procedures was offered as an
incentive for participation. These participants provided their informed con-
sent prior to participating in the research. Analyses comparing high school
and college students on demographic variables, proportion of students with
divorced parents, the primary predictor and criterion variables (discussed
below), and multiple group structural equation models (as described below)
indicated no differences between the high school and college students (results
available by authors on request). The study was approved and monitored by
the institutional review board located at the university at which the research
was conducted.

Participants were primarily Caucasian (72%) and female (75%). Nine
percent of the participants were African American, 7% were Hispanic, 6%
were Asian, and 4% reported being from other ethnic backgrounds. Overall,
the participants averaged 18.7 years of age (SD = 0.9), with the youth from
divorced families slightly older (but not significantly so, M = 18.8, SD = 1.0)
than youth from intact families (M = 18.7, SD = 0.9). Seventy-one percent of
youth reported that their parents were married, and 29% reported that their
parents were either divorced or separated. Of those whose parents had divorced,
on average the divorce occurred approximately 10 years ago (SD = 5.2; range
<1 to 20 years). Furthermore, of the youth from divorced families, the major-
ity reported they lived with a parent (63% mothers, 9% fathers), and a very
small proportion (2%) reported that they lived with a grandparent.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1250   Journal of Family Issues

Measures

Participants were asked to complete questionnaires beginning with a short
demographic section (requesting information on the participants’ demo-
graphic backgrounds, family structure, the number of years since divorce if
one had occurred, etc.) and a variety of measures targeting the participants’
relationship with their grandmothers and their psychological adjustment.

Inventory of Parent and Peer Attachment (IPPA). The IPPA (Armsden &
Greenberg, 1987) is a 53-item self-report inventory measuring adolescents’
and young adults’ relationship quality with parents and peers. For the current
study, we used the trust, communication, and alienation subscales for parents,
modified to be appropriate for grandmothers. The alienation subscale was
reverse scored to make the interpretation consistent with the other subscales.
Samples of modified items include “My grandmother [formerly parent]
respects my feelings” and “I feel my grandmother [formerly parent] is suc-
cessful as a grandparent [formerly parent].” Participants completed this meas-
ure with respect to each grandparent (maternal grandmother and grandfather
and paternal grandmother and grandfather) although we only used data for
maternal grandmothers for the reasons described previously. Furthermore,
participants reported many different combinations of living grandparents,
making consistent comparisons unwieldy from an analytic standpoint.

Previous validity studies conducted by Armsden and Greenberg (1987)
have shown that parent attachment scores correlated significantly with fam-
ily support as well as with a tendency to seek out parents in times of need.
Coefficient alphas were .90, .87, and .84 for the trust, communication, and
alienation subscales, respectively.

Contact with grandmothers. Extent of contact between grandchildren
and maternal grandmothers at the time the study was conducted was
assessed by two items, one indicating the current frequency of visits and the
other the current frequency of phone conversations with their maternal
grandmothers. Response options included daily, at least once per week, at
least once per month, and less than once per month. These items were
summed to provide an index of contact the participants had with their
grandmothers, with larger values indicating more contact. We also created
a similar variable representing parents’ contact with maternal grandmothers
by summing items indicating the parents’ frequency of visits and phone
contact with the participants’ maternal grandmothers. Coefficient alphas
were .70 for participants’ contact with their maternal grandmothers and .80
for their parents’ contact with them.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1251

Relationship Competence Scale (RCS). The RCS (Hansson, Jones, &
Carpenter, 1984) is a 100-item measure of competence in interpersonal
skills. Relational competence measures abilities that serve to maintain and
enhance relationships and is composed of the intimacy, trust, interpersonal
sensitivity, empathic concern, and perspective-taking subscales. Internal
consistency coefficients for the subscales range from .76 to .90. The inter-
personal sensitivity and trust subscales were used in the current study, and
the scales showed adequate reliability (αs = .81 and .76, respectively).
Responses range on a scale from 1 to 4, with 4 meaning strongly agree or
very much like me and 1 meaning strongly disagree or very much unlike me.
Sample items include “People are usually very dependable and trustwor-
thy” and “I really am thoughtful and considerate of others.”

Self-Efficacy Scale (SES). The Self-Efficacy Scale (Sherer et al., 1982) is
a 23-item scale that was used to measure an individual’s estimate of his or her
ability to cope with a situation. Woodruff and Cashman (1993) found that a
general self-efficacy subscale could be divided into three subscales (general
efficacy magnitude, strength, and competence). Concurrent validity estimates
are reported by Woodruff and Cashman (1993), who used Pearson correla-
tions to demonstrate moderate correlations between the SES and the Mastery
Scale (.54; Pearlin, Lieberman, Menaghan, & Mullan, 1981) and the
Rosenberg Self-Esteem Scale (.54; Rosenberg, 1965). The magnitude,
strength, and competence subscales of general self-efficacy were used in the
current study. Alphas for these scales were .82, .78, and .66, respectively.

Hopkins Symptom Checklist (HSCL). The HSCL (Derogatis, Lipman,
Rickels, Uhlenhuth, & Covi, 1974) is a 58-item self-report symptom inven-
tory using a 5-point anchored rating scale and covering five symptom
dimensions: somatization, obsessive–compulsive, interpersonal sensitivity,
anxiety, and depression. It is the precursor—and therefore very similar in
terms of item content and targeted assessment areas—to the Symptom
Checklist 90 (Derogatis, Lipman, & Covi, 1973) and the Brief Symptom
Inventory (Derogatis & Melisaratos, 1983), which have been used in numer-
ous studies. Because the participants were not sampled from a clinical
population, subscales were selected on the basis of their relevance to non-
clinical adolescent and young adult populations and included depression,
anxiety, and somatization. Derogatis et al. (1974) report alpha coefficients
from .84 to .87 for individual HSCL subscale scores. Several studies have
supported the criterion-related validity of the measure by demonstrating its
sensitivity to a variety of treatment effects (see Derogatis et al., 1974, for a
review). The interpersonal sensitivity subscale was not used because of its

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1252   Journal of Family Issues

overlap with the RCS, nor was the obsessive–compulsive subscale because
of the low probability of occurrence in the targeted population. Coefficient
alphas for the depression, anxiety, and somatization subscales were .87, .89,
and .85, respectively.

Data Analysis

The objectives of this study were to compare youth from divorced and
intact families on the extent to which their relationships with their grand-
parents predicted three indices of psychological adjustment: self-efficacy,
relational competence, and psychological symptoms. Both hypotheses were
tested using structural equation modeling (SEM).

The primary statistical advantage of SEM is that complex models can be
tested for their goodness of fit to the observed data. Models explicitly speci-
fying the number of latent variables (LVs), the manifest (measured) variable
to LV relationships, and the LV to LV relationships are developed a priori,
after which the model is tested in terms of model fit criteria (Kline, 1998). To
assess model fit, in addition to the model χ2 we examined two fit indices, the
root mean square error of approximation (RMSEA) and the comparative fit
index (CFI), with RMSEA values of .08 and below and CFI values of .90 and
above being indicative of good model fit (Browne & Cudeck, 1993). It
should be noted that in moderately sized samples (N > 200), the chi-square
test tends to be inflated and may reach significance even with adequately fit-
ting models (Kline, 1998). Therefore, when discrepancies between the chi-
square test and other fit indices arose, we relied more heavily on the RMSEA
and CFI fit indices.

First, we used confirmatory factor analyses (CFAs) to determine
whether the subscales corresponding to each LV adequately loaded on the
latent construct. Four LV measurement models were created from the sub-
scales of the IPPA, RCS, SES, and HSCL to yield LVs corresponding to
grandmother–grandchild relationship quality, relational competence, self-
efficacy, and psychological symptoms, respectively. The robust maximum
likelihood estimator was used to fit the model to the sample covariance
matrix using the software Mplus (version 4.20; Muthén & Muthén, 1998-
2005). Modification indices (Lagrange Multiplier and Wald tests) were
examined to determine if model fit could be improved by either constrain-
ing or freeing parameters; however, no modifications were made without
having a theoretical justification for doing so (Silvia & MacCallum, 1988).
All SEMs controlled for the participant’s gender, mother’s educational
level (a proxy for socioeconomic status), amount of contact youth had with

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1253

their grandparents, and for youth from divorced families, the number of
years since the divorce by including them in the models as exogenous
manifest variables.

Next, we tested a sequence of SEMs. The first (Testing Hypothesis 1)
examined the relationship between grandmother–grandchild relationship
quality and adjustment specifically among the youth from divorced fami-
lies. Second, we wanted to examine if the same configural pattern fit for the
youth from intact families; therefore, we tested a second SEM model exam-
ining the same relationship configuration among the youth from intact
families. Finally (Testing Hypothesis 2), we examined whether family type
(divorced vs. intact) moderated the relationship between grandmother rela-
tionship quality and adjustment, conducting a multiple group analysis in
which the relationship between grandmother–grandchild relationship qual-
ity and adjustment was compared in the two family types.

We then placed stronger invariance constraints on the models to focus
more closely on the path coefficients between grandmother relationship
quality and the adjustment variables. We constrained the measurement
model parameters and all variances and covariances to equality between
the two models, with the exception of the relationships between relation-
ship quality and adjustment. This parameter was allowed to be free to
permit a test of both within-group and between-group hypotheses regard-
ing relationship quality–adjustment associations. This was followed by a
series of analyses in which we systematically freed one relationship
quality–adjustment path coefficient at a time. Chi-square difference tests
were used to compare the nested models, with a significant test indicating
that the relationship between grandmother relationship quality and adjust-
ment reliably differed for the grandchildren from intact versus divorced
families. Finally, we examined the extent to which results were influ-
enced by the amount of contact participants’ parents had with participants’
maternal grandmothers.

Results

Table 1 shows descriptive statistics for the subscales of the IPPA as well
as all adjustment variables by family type. As shown in Table 1, youth from
both family types reported similarly close relationships with their grand-
mothers and similar values on the adjustment variables; however, partici-
pants from divorced families reported higher levels of psychological
symptoms. Table 2 shows the zero-order correlations among the relationship

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1254   Journal of Family Issues

Table 1
Descriptive Statistics for Primary Variables and Results

of Statistical Tests of Mean Comparisons for Youth
From Intact and Divorced Families

Intact Divorced F
Variable Families, M (SD) Families, M (SD)
0.38
Grandmother–grandchild 1.19
0.14
  relationship quality
0.16
Trust 39.41 (7.98) 40.01 (7.56) 0.02
26.73 (7.09)
Communication 25.82 (6.42) 34.82 (7.50) 0.02
0.86
Alienation 35.18 (7.71) 27.00 (5.41) 0.14
31.34 (5.62)
Relational competence 6.24*
18.43 (3.85) 6.24*
Trust 27.25 (4.76) 28.48 (4.98) 6.56*
14.84 (2.91) NA
Sensitivity 31.44 (5.01) 0.07
39.34 (7.88)
Self-efficacy 31.02 (7.51)
20.92 (4.25)
Magnitude 18.50 (3.51) 10.69 (5.24)

Strength 29.03 (4.70) 6.64 (1.62)

Competence 14.98 (2.76)

Psychological symptoms

Somatization 36.91 (7.81)

Depression 28.66 (7.93)

Anxiety 19.55 (4.44)

Years since divorce NA

Contact with maternal 6.70 (1.48)

  grandmother

Note: NA = not applicable.
*p < .05.

quality and adjustment variables for the two family types. Trends in Table 2
suggest that the relationships between relationship quality and adjustment
were stronger for youth from divorced families relative to those from intact
families.

Confirmatory Factor Analyses

Confirmatory factor analyses supported the validity of each of the four
measurement models (χ2 ps > .05, CFIs > .99, RMSEAs < .08). For
grandmother–grandchild relationship quality, each of the manifest varia-
bles was strongly associated with the latent variable (trust, β = .95; com-
munication, β = .76; alienation, β = .46). The same was the case for each of

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Table 2
Bivariate Correlations Among Relationship Quality and Adjustment Variables by Family Type

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010 1 2 3 4 5 6 7 8 9 10 11 12 13

  1. Trust − .751** .412** .349** .395** .330** .287** .203* –.262** –.174 –.166 .212* –.098

  2. Communication .721** − .321** .200* .322** .135 .208* –.033 –.079 –.001 –.070 .359** .026

  3. Alienation .506** .508** − .265** .486** .340** .356** .340** –.401** –.388** –.383** .280** –.174

  4. Sensitivity .211** .187** .434** − .516** .203* .151 .204* –.119 –.122 –.187* .261* –.099

  5. Trust .272** .295** .433** .467** − .187 .249** .290** –.406** –.419** –.398** .157 –.184

  6. Magnitude .093 –.019 .354** .245** .133* − .546** .520** –.217* –.248** –.297** .029 –.104

  7. Strength .115 .106 .299** .268** .095 .626** − .459** –.254** –.329** –.261** .180 –.086

  8. Competence .123 .035 .419** .190** .176** .656** .507** − –.370** –.333** –.418** .024 –.280**

  9. Somatization –.118 –.089 –.296** –.119* –.068 –.211** –.192** –.222** − .729** .795** –.030 .293**

10. Depression .187** –.180** –.441** –.182** –.248** –.266** –.265** –.323** .716** − .777** –.101 .149

11. Anxiety –.092 –.074 –.315** –.104 –.125* –.275** –.198** –.310** .791** .754** − –.096 .161

12. Contact with .319** .401** .240** .031 .088 .093 .154* .056 –.151* –.184** –.120 − –.127

maternal

grandmother

13. Years since NA NA NA NA NA NA NA NA NA NA NA NA −

divorce

Note: Correlations among variables reported by youth from intact families appear below the diagonal. Correlations among variables reported by youth from
divorced families appear above the diagonal. The sample covariance matrix is available from the authors by request. NA = not applicable.
*p < .05. **p < .01.

1255

1256   Journal of Family Issues

the adjustment variables: relational competence (sensitivity, β = .90;
trust, β = .53), self-efficacy (magnitude, β = .90; strength, β = .73; compe-
tence, β = .72), and psychological symptoms (somatization, β = .79;
depression, β = .81; anxiety, β = .92). The full measurement model, in
which the data were fit to all four CFAs simultaneously, also demonstrated
adequate model fit, χ2(38) = 162.65, p < .001, CFI = .90, RMSEA = .09,
although the RMSEA was slightly higher than desired. The modification
indices suggested that fit could be improved by allowing the error variances
for two of the grandmother relationship quality variables (trust and com-
munication) to correlate. Given the similarity of the two constructs, we
allowed the errors to correlate, resulting in improved model fit, χ2(38) =
99.83, p < .001, CFI = .95, RMSEA = .07.

Structural Equation Models

To test Hypothesis 1, an SEM was tested in the youth from divorced
families. The data demonstrated good fit to the model, χ2(56) = 73.64, p = .06,
CFI = .96, RMSEA = .07. The path coefficients between the grandmother–
grandchild relationship quality LV and the LVs representing participant
adjustment were strong (relational competence, β = .53; self-efficacy,
β = .79; psychological symptoms, β = –.68; see Figure 1), supporting our
first hypothesis. Relationship quality with maternal grandmothers accounted
for 33% of the relational competence LV, 57% of the self-efficacy LV, and
43% of the psychological symptoms LV.

Next, we tested the configural invariance of the model by testing the
model with youth from intact families. Again, the error variances between the
two grandmother relationship quality variables were allowed to correlate.
The data also fit this model adequately, χ2(48) = 118.13, p < .001, CFI = .92,
RMSEA = .08, indicating that similar overall structural relationships between
relationship quality with maternal grandmothers and adjustment existed for
the youth from intact families as well. However, the path coefficients
between the grandmother–grandchild relationship quality LV and the LVs
representing participant adjustment were, with the exception of relational
competence, weaker than those found with the youth from divorced fami-
lies (relational competence, β = .77; self-efficacy, β = .55; psychological
symptoms, β = –.47; see Figure 2). For the youth from intact families, rela-
tionship quality with maternal grandmothers accounted for 59% of the
relational competence LV, 31% of the self-efficacy LV, and 22% of the
psychological symptoms LV.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1257

Figure 1
Path Model Showing Relationship Between Attachment to Maternal

Grandparents and Relational Competence, Self-Efficacy, and
Psychological Symptoms for Youth From Divorced Families

D1

Psychological .84 Depression e4
Symptoms .84 e5
e6
Anxiety
e7
.89 e8

Som. e9
e10
e1 e2 e3 D2 e11

–.68

Trust Comm Alien

.55 .37 .64

Relationship .53 Relational .71 Sensitivity
Quality Competence .84 Trust

.79 D3

Self .86 Magnitude
Efficacy .71

Strength

.74

Competence

Note: Control variables are not shown in the figure, but they were included in the model as exog-
enous variables. Comm = Communication; Alien = Alienation; Som = Somatization; e indicates
error variance (e.g., e1), and D indicates disturbance (e.g., D1).

We next tested a moderator model in which the association between rela-
tionship quality with maternal grandmothers and adjustment was moderated
by family type. The χ2 difference test for nested models revealed that the
model in which the associations between relationship quality with maternal
grandmothers and relational competence, self-efficacy, and psychological
symptoms was free to vary across family type fit significantly better than the
model in which these relationships were constrained to be the same across
youth from divorced and intact families, Δχ2(3) = 48.90, p < .001. These

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1258   Journal of Family Issues

Figure 2
Path Model Showing Relationship Between Attachment to Maternal

Grandparents and Relational Competence, Self-Efficacy, and
Psychological Symptoms for Youth From Intact Families

D1

.83 Depression e4
e5
Psychological .88 e6
Symptoms
Anxiety

.79

Som.

e1 e2 e3 D2

–.47

Trust Comm Alien

.69 .59 .75

Relationship .77 Relational .70 Sensitivity e7
Quality Competence .65 Trust e8

.55 D3

Self .87 Magnitude e9
Efficacy .77 e10
e11
Strength

.72

Competence

Note: Control variables are not shown in the figure, but they were included in the model as exog-
enous variables. Comm = Communication; Alien = Alienation; Som = Somatization; e indicates
error variance (e.g., e1), and D indicates disturbance (e.g., D1).

results supported our second hypothesis (please note that the absolute level
of this relationship is still negative for those in intact families, but that it
is stronger for those in divorced families). Tests in which one association
between relationship quality with maternal grandmothers and adjustment
variable was freed at a time revealed that the moderation was specific to
psychological symptoms, Δχ2(1) = 48.80, p < .001, as the relationships
were not significant for either relational competence, Δχ2(1) = 0.10, ns, or
self-efficacy, Δχ2(1) = 0.70, ns.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1259

Accounting for the Participants’ Parents’
Relationships with Maternal Grandmothers

As mentioned previously, all SEMs accounted for participants’ contact
with their maternal grandmothers, and in the case of youth from divorced
families, the number of years since the divorce occurred. However, because
parents of at least younger children serve as gatekeepers in the amount of
contact children have with their grandparents (Matthews & Sprey, 1984), we
examined the extent to which the relationship between participants’ relation-
ship quality with their maternal grandmothers and their adjustment was medi-
ated by the amount of contact participants’ parents had with their maternal
grandmothers. Because the results above indicated that differences between
youth from intact and divorced families with respect to maternal grandmother
relationship quality–adjustment relationships were specific to psychological
symptoms, we confined our mediation analyses to the relationship quality
and psychological symptom LVs. We examined mediation using the product
of coefficients (for the indirect paths) method (MacKinnon, Lockwood, &
Hoffman, 2002). As noted by MacKinnon et al. (2002), the joint significance
of the two paths comprising the indirect effect—in our case the paths between
(a) relationship quality and parents’ contact with maternal grandmothers
and (b) parents’ contact with maternal grandmothers and psychological
­symptoms—provides the “best balance of Type I error and statistical power”
(p. 83). As implemented in Mplus, the significance of the mediated effect is
tested by a pseudo z statistic in which the product of coefficients is divided
by its bootstrapped standard error (bootstrapping is used to compensate for
the standard error’s nonnormality; Shrout & Bolger, 2002), and the critical
value for determining statistical significance is ±1.96.

Results indicated that parents’ contact with maternal grandmothers did not
mediate the relationship between participants’ relationship quality with their
maternal grandmothers and psychological symptoms (pseudo z = 1.02, ns).
Although relationship quality was significantly associated with parents’ contact
with maternal grandmothers (pseudo z = –2.67, p < .01), the relationship between
parents’ contact with maternal grandmothers and participants’ report of psycho-
logical symptoms was not statistically significant (pseudo z = –1.10, ns).

Discussion

Among youth from divorced families, grandmother–grandchild relation-
ship quality was strongly related to adjustment, and moderator analyses

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1260   Journal of Family Issues

indicated that the relationships between relationship quality and adjustment
outcomes were stronger for the youth from divorced families than for youth
from intact families. Taken together, these results suggest that strong rela-
tionships with maternal grandmothers may help grandchildren adjust to the
divorce of their parents, independently of the contact the middle genera-
tion has with the grandmothers. As such, the results are consistent with
other research suggesting that youth from divorced homes are under
many circumstances resilient to the negative consequences that some youth
experience, especially when the divorcing parents are able to minimize
their conflict (Amato, Spencer Loomis, & Booth, 1995). For instance,
Hetherington’s (1989) landmark study highlights numerous individual,
familial, and extrafamilial factors that contribute to better psychosocial
adjustment among youth from divorced families. It is noteworthy that in
the current study among youth from divorced families, the participants’
relationships with their grandmothers strongly predicted their psychologi-
cal adjustment even when the divorce occurred many years ago. However,
conclusions remain tentative until they are confirmed by longitudinal stud-
ies conducted with larger samples.

The findings in the current study are similar to those of Lussier et al.
(2002), who found that closeness to maternal grandparents was associated
with fewer internalizing problems among children from single-parent
families. However, Lussier et al. did not compare the strength of closeness–
adjustment relationships between children from intact and single-parent
families. The current study also builds on Lussier et al. by replicating this
relationship in a larger and older sample, using a stronger measure of rela-
tionship quality with grandparents (the IPPA has more breadth of item
content and better psychometric properties), and using state-of-the-art SEM
analytic procedures. The findings are also consistent with Cooney and
Smith (1996) in that youth from divorced homes did not report that their
relationships with their maternal grandmothers were stronger than youth
from intact homes (a finding that may in part be driven by the homogeneity
of our sample). Nevertheless the current findings extend this research by
suggesting that the relationships that youth from divorced homes had with
their maternal grandmothers were more strongly related to their psycho-
logical adjustment.

The findings are also consistent with other recent research highlighting
the importance of studying divorce from within an intergenerational frame-
work (Amato & Cheadle, 2005; Johnson, 1998). Attachment theory has been
posited as a framework for understanding the importance of intergenera-
tional relationships to adolescent and young adult development (Simpson,

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1261

Rholes, & Nilligan, 1992). Although the measure of relationship quality we
used in this study is not an explicit measure of attachment style, it has been
used in previous studies examining youths’ attachment to important family
members (Lopez & Gover, 1993). Applied to the current results, attachment
theory suggests that those with strong relationships with their maternal
grandmothers may have perceived them as a secure relationship with which
they could connect during the stress of the divorce. As a result, the findings
from this study suggest that these youth may have benefited from connect-
ing to their grandmothers specifically in terms of minimizing psychological
symptoms. However, because the research design is cross-sectional, the
direction of effects is not entirely clear (i.e., better psychological adjustment
may lead to closer grandmother–grandchild relationships). The results sug-
gest hypotheses to be tested in future prospective studies carried out using
longitudinal designs and using specific measures assessing attachment style
as well as covering more youth problems such as depression diagnoses.

In addition to the limitations listed above, namely the fairly small sample,
the cross-sectional research design, and our inability to definitively draw
conclusions regarding the direction of the effects, the study is limited in other
aspects as well. First, the sample was primarily Caucasian and female. Thus,
our findings may not generalize to youth from other sociodemographic back-
grounds. Furthermore, it is not clear whether the findings would generalize
to grandparent kinship patterns other than maternal grandmothers, or to
younger children and adolescents. In fact, with respect to younger children
and adolescents, the relationships between maternal grandmother relation-
ship quality and psychological adjustment may be stronger and also likely
more influenced by the relationship the middle generation (the grandchil-
dren’s parents) may have with the maternal grandmothers. Third, although
we controlled for participant gender, mother’s education level, and contact
with grandparents in our structural equation models, previous studies have
suggested that proximity to grandparents and socioeconomic status (with
which mother’s education level is highly correlated) are important variables
in understanding grandparent–grandchild relationships (Uhlenberg &
Hammill, 1998). Unfortunately, we did not collect data in this study so that
we could control for these variables as well. That said, it is likely that prox-
imity would be highly correlated with contact and mother’s education level
would be correlated with family socioeconomic status. Fourth, it is possible
that recruiting from two populations may have introduced selection bias into
the study results. For example, high school students under age 18 were
required to obtain parent permission as well as provide informed assent,
whereas college students for the most part only had to provide informed

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1262   Journal of Family Issues

consent. It is possible that we may have obtained different results if we had
used a single recruitment strategy. Fifth, all data were collected from the
grandchild’s perspective and therefore, the findings of the study may have
arisen in part due to common method variance. Therefore, we were unable to
control for some potentially important variables such as the quality of the
relationships between the participants’ mothers and their maternal grand-
mothers (albeit our findings suggest that the contact mothers had with mater-
nal grandmothers did not affect the results). Finally, in many cases, the data
were collected several years after the divorce had occurred. It is possible that
other findings may have emerged if the research were conducted with
younger children in the immediate aftermath of divorce.

Despite these limitations, these findings suggest interesting hypotheses to
be tested in future studies. It is noteworthy that we found adolescents’ rela-
tionships to their grandmothers to affect their adjustment several years later.
Indeed, one would assume that grandmothers would exert less influence in
the lives of 17- to 20-year-olds (Cooney & Smith, 1996; Kennedy, 1990). Our
findings suggest that these youth may have had a continuing bond with their
maternal grandmothers that transcended the divorce. Furthermore, many of
the participants were college students and as such may be more resilient to
hardships than the general population. Perhaps the influence of maternal
grandmothers may contribute in some way to the resilience of young people
from divorced homes and perhaps may diminish the impact of divorce on
their lives (Wallerstein & Lewis, 1998). Prospective studies incorporating
larger samples and longitudinally following children in the immediate after-
math of divorce, and the way in which their relationships with their grand-
mothers unfold over time, are needed to examine these possibilities.

References

Amato, P. R. (2001). Children of divorce in the 1990s: An update of the Amato and Keith
(1991) meta-analysis. Journal of Family Psychology, 15, 355-370.

Amato, P. R., & Booth, A. (1996). A prospective study of divorce and parent-child relation-
ships. Journal of Marriage and the Family, 58, 356-365.

Amato, P. R., & Cheadle, J. (2005). The long reach of divorce: Divorce and child well-being
across three generations. Journal of Marriage and Family, 67, 191-206.

Amato, P. R., & Keith, B. (1991). Parental divorce and the well-being of children: A meta-
analysis. Psychological Bulletin, 110, 26-46.

Amato, P. R., Spencer Loomis, L., & Booth, A. (1995). Parental divorce, marital conflict, and
offspring well-being during early adulthood. Social Forces, 73, 895-915.

Armsden, G. C., & Greenberg, M. T. (1987). The Inventory of Parent and Peer Attachment:
Individual differences and their relationship to psychological well-being in adolescence.
Journal of Youth and Adolescence, 16, 427-454.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

Henderson et al. / Grandmother–Grandchild Relationship Quality   1263

Benoit, D., & Parker, K. C. H. (1994). Stability and transmission of attachment across three
generations. Child Development, 65, 1444-1456.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen
& J. S. Long (Eds.), Testing structural models (pp. 136-162). Newbury Park, CA: Sage.

Cherlin, A., & Furstenberg, F. F. (1986). The influence of grandparents on grandchildren. In
A. Cherlin & F. F. Furstenberg (Eds.), The new American grandparent: A place in the fam-
ily, a life apart (pp. 165-184). New York: Basic Books.

Cooney, T. M., & Smith, L. A. (1996). Young adults’ relations with grandparents following
recent parental divorce. Journal of Gerontology: Social Sciences, 51B, S91-S95.

Department of Health and Human Services. (2005). Births, marriages, divorces, and deaths:
Provisional data for 2004. Retrieved June 20, 2005, from http://www.cdc.gov/nchs/data/
nvsr/nvsr53/nvsr53_21.pdf

Derogatis, L. R., Lipman, R. S., Rickels, K., Uhlenhuth, E. H., & Covi, L. (1974). The Hopkins
Symptom Checklist: A self-report symptom inventory. Behavioral Science, 19, 1-15.

Derogatis, L. R., Lipman, R. S., & Covi, L. (1973). SCL-90: An outpatient psychiatric rating
scale—Preliminary report. Psychopharmacology Bulletin, 9, 13-27.

Derogatis, L. R., & Melisaratos, N. (1983). The Brief Symptom Inventory: An introductory
report. Psychological Medicine, 13, 595-605.

Gladstone, J. W. (1988). Perceived changes in grandmother-grandchild relations following a
child’s separation or divorce. The Gerontologist, 28, 66-72.

Hansson, R. O., Jones, W. H., & Carpenter, B. N. (1984). Relational competence and social
support. Review of Personality and Social Psychology, 5, 265-284.

Hayslip, B., Jr., & Goldberg-Glen, R. (Eds.). (2000). Grandparents raising grandchildren.
New York: Springer.

Hayslip, B., Henderson, C. E., & Shore, R. J. (2003). The structure of grandparental role
meaning. Journal of Adult Development, 10, 1-12.

Hetherington, E. M. (1989). Coping with family transitions: Winners, losers, and survivors.
Child Development, 60, 1-14.

Hetherington, E. M., Bridges, M., & Insabella, G. M. (1998). What matters? What does not?
Five perspectives on the association between marital transitions and children’s adjustment.
American Psychologist, 53, 167-184.

Johnson, C. L. (1998). Effects of adult children’s divorce on grandparenthood. In M. E. Szinovacz
(Ed.), Handbook on grandparenthood (pp. 184-199). Westport, CT: Greenwood.

Kennedy, G. E. (1990). College students’ expectations of grandparent and grandchild role
behaviors. The Gerontologist, 30, 43-48.

Kennedy, G. E. (1992). Quality in grandparent/grandchild relationships. International Journal
of Aging and Human Development, 35, 83-98.

Kennedy, G. E., & Kennedy, C. E. (1993). Grandparents: A special resource for children in
stepfamilies. Journal of Divorce & Remarriage, 19, 45-68.

Kenny, M. E., & Barton, C. E. (2003). Attachment theory and research: Contributions for
understanding late adolescent and young adult development. In J. Demick & C. Andreoletti
(Eds.), Handbook of adult development (pp. 371-390). New York: Kluwer.

Kline, R. B. (1998). Principles and practice of structural equation modeling. New York:
Guilford.

Lopez, F. G., & Gover, M. R. (1993). Self-report measures of parent-adolescent attachment
and separation-dividuation: A selective review. Journal of Counseling & Development,
71, 560-569.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010

1264   Journal of Family Issues

Lussier, G., Deater-Deckard, K., Dunn, J., & Davies, L. (2002). Support across two genera-
tions: Children’s closeness to grandparents following parental divorce and remarriage.
Journal of Family Psychology, 16, 363-376.

MacKinnon, D. P., Lockwood, C. M., & Hoffman, J. M. (2002). A comparison of methods to
test mediation and other intervening variable effects. Psychological Methods, 7, 83-104.

Matthews, S. H., & Sprey, J. (1984). The impact of divorce on grandparenthood: An explora-
tory study. The Gerontologist, 24, 41-47.

Muthén, L., & Muthén, B. (1998-2005). Mplus user’s guide [Computer software]. Los
Angeles: Muthén & Muthén.

Pearlin, L., Lieberman, M., Menaghan, E., & Mullan, J. (1981). The stress process. Journal of
Health and Social Behavior, 22, 337-356.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton
University Press.

Sherer, M., Maddux, J. E., Mercandante, B., Prentice-Dunn, S., Jacobs, B., & Rogers, R. W.
(1982). The Self-Efficacy Scale: Construction and validation. Psychological Reports,
51, 663-671.

Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies:
New procedures and recommendations. Psychological Methods, 7, 422-445.

Silvia, E. S. M., & MacCallum, R. C. (1988). Some factors affecting the success of specifica-
tion searches in covariance structure modeling. Multivariate Behavioral Research, 23,
297-326.

Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking and support giving
within couples in an anxiety-provoking situation: The role of attachment styles. Journal of
Personality and Social Psychology, 62, 434-446.

Uhlenberg, P., & Hammill, B. (1998). Frequency of grandparent contact with grandchild sets:
Six factors that make a difference. The Gerontologist, 38, 276-285.

Wallerstein, J. S. (1987). Children of divorce: Report of a ten-year follow-up of early latency-
age children. American Journal of Orthopsychiatry, 57, 199-211.

Wallerstein, J. S., & Lewis, J. (1998). The long-term impact of divorce on children: A first
report from a 25-year study. Family & Conciliation Courts Review, 36, 368-383.

Woodruff, S. L., & Cashman, J. F. (1993). Task, domain, and general efficacy: A reexamina-
tion of the self-efficacy scale. Psychological Reports, 72, 423-432.

Downloaded from jfi.sagepub.com at UNIV OF MIAMI on December 13, 2010


Click to View FlipBook Version