Close
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Psychological well -being in non -traditional families
(USC Thesis Other)
Psychological well -being in non -traditional families
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Copyright 2002
PSYCHOLOGICAL WELL-BEING IN
NON-TRADITIONAL FAMILIES
by
Tara Rose
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)
December 2002
Tara Rose
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3093813
UMI
UMI Microform 3093813
Copyright 2003 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 90089-1695
This dissertation, written by
T a ra Rose
under the direction o f h e r dissertation committee, and
approved by all its members, has been presented to and
accepted by the Director of Graduate and Professional
Programs, in partial fulfillment of the requirements for the
degree of
DOCTOR OF PHILOSOPHY
Director
Date Decem ber 1 8 , 2002
Dissertation Committee
Chair
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ii
Table of Contents
List of Tables..................................................................................................................... v
List of Figures......................................................................................................................viii
Abstract...................................................................................................... ............. ........ix
Introduction.......................................................................... 1
Theory and Research on Family Structure and Well-Being............................2
Research with Heterosexual Parented Households............................................5
Research with Lesbian, Bisexual and Gay Parented Households ..................6
Family Background Variables that Influence Well-Being ..............................9
Parental Divorce and Adult Children's Well-Being ..........................................10
Family Relationship Quality and Well-Being ................................. 10
Family Structure and Family Process.................................................................12
Methodological Considerations.......................................................................................14
Well-being as a Multidimensional Construct ................................................. 16
The Present Study............................................................................................................. 17
Methods..............................................................................................................................21
Participants ...........................................................................................................21
Lesbigay mothers and their children ..........................................................21
Heterosexual mothers and their children.................... 27
Materials ............................................................. 29
Demographic and Family Background Measures ....................... 29
Age ........................................................................................... 29
Racial identity..........................................................................................29
Partner status ...........................................................................................29
Income ......................................................................................................30
Education................................................................................................. 30
Dyad (mother and child) cohabitation status........................................ 30
Mother's sexual orientation.....................................................................31
Child's gender............................................................. 31
Child's parental status ..................................................................31
Lesbigay mother's age at coming out .................................................... 31
Family Relationship Quality.......................................................................... 32
Affectual solidarity........................................................ 32
Conflict...................... 32
Emotional Support................................ 33
Psychological Well-Being ............................................................................. 33
Depressive symptoms............................................................................. 34
Negative and positive affect .................................. 34
Self-esteem .............................................................................................. 34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Data Analysis and Statistics .............................................................................. 35
Results ..............................................................................................................................40
Descriptive Findings........................................................................................... 41
Characteristics of study participants..................................................... 41
Preliminary Analyses........................................................................................ ..46
Distributional assumptions ....................................................................46
Verifying convergent validity of measures ......................................... 47
Inter-judge agreement on family relationship quality measures ....... 52
Family Structural Comparisons .........................................................................53
Comparison of lesbigay and heterosexual mothers and their children 56
Comparison of single and married/partnered mothers and their
children ................................................................................................... 56
Assessing for differences in demographic variables for possible
covariates ................................................................................................65
Family Relationship Quality and Mother's Sexual Orientation.................. . 69
Dyad position and family type ............................................................. 74
Summary of hypothesis I ....................................................................... 75
Family Relationship Quality and Mother's Partner Status ..............................75
Summary of hypothesis II .....................................................................81
Mother and Child Well-being........................................................... ................ 81
Individual well-being and family relationship quality .... .................. 83
Individual well-being and family relationship quality without
demographics.......................................................................................... 99
Summary of hypothesis III ................................................................... 102
Individual well-being and family structure ......................................... 103
Individual Family Member's Well-being and Family Structure......................104
Individual well-being and family structure without demographics .... 108
Summary of hypothesis I V ....................................................................113
Age and Gender Differences ............................................................................. 114
Age and Gender with Maternal Sexual Orientation ........................................ 116
Discussion.........................................................................................................................116
Family Structure Theory.....................................................................................122
Theoretical Limitations ............................................................ .........................129
Strengths and Limitations of the Study ............................................................ 132
Implications and Further Research....................................................................136
References ........................................................................................................................140
Appendices ................................................................................................................. 149
Appendix A: Demographics ................................................................................ 149
Appendix B: Lesbigay Mother’s Age at Coming O u t..................................... 151
Appendix C: Family Relationship Q uality........................................................152
Appendix D: Well-Being Measures................................................................... 159
Appendix E: Correlation Matrices of All Main Variables for Full Sample ... 162
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Appendix F: Correlation Matrices of All Main Variables for Lesbigay Mothers
and Their Children............................................................................................... 165
Appendix G: Correlation Matrices of All Main Variables for Heterosexual
Mothers and Their Children ............................................................................... 168
Appendix H: Correlation Matrices for Single Mothers and Their Children ... 171
Appendix I: Correlation Matrices for Married/Partnered Mothers and Their
Children ................ 174
Appendix K: Follow-up Models for Hypothesis I .............................................177
Appendix L: Follow-up Models for Hypothesis I I ............................................183
Appendix M: Analyses for the Third Specific Aim ..........................................188
Appendix N: Summary of the Hierarchical Regression Models for Mother
Dyadic Family Relationship Quality on Age, Gender and Interactions ..........195
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
V
List of Tables
Table 1: Age o f Lesbigay Mother in the Coming Out Process.........................................23
Table 2: Specific Aims and Hypotheses..............................................................................37
Table 3: Means and Standard Deviations o f Demographics Major Study
Variables fo r Full Sample, Lesbigay Parented Families and Heterosexual
Parented Families.................................................................................................. 43
Table 4: Counts and Percentages o f Demographics Major Study Variables
for Full Sample, Lesbigay Parented Families, and Heterosexual
Parented Families................................... 44
Table 5: Means and Standard Deviations o f Demographics Major Study Variables for
Single Parented and Married/Partnered Parented Families....................................46
Table 6: Counts and Percentages o f Demographics Major Study Variables for Children
from Single Mothers and Children o f Married/Partnered Mothers........................47
Table 7: Transformations o f Study Variables................................................................... 48
Table 8: Correlation Matrix o f Transformed Dyadic Relationship Quality for the Full
Sample........................................................................................................................... 50
Table 9: Correlation Matrix o f Transformed Mother’ s Perceived Relationship Quality
fo r Full Sample.............................................................................................................50
Table 10: Correlation Matrix o f Transformed Child’ s Perceived Relationship Quality
fo r Full Sample............................................................................................................51
Table 11: Correlation Matrix o f Transformed Mother’ s Well-Being Variables o f Full
Sample..................................................................................... 52
Table 12: Correlation Matrix o f Transformed Child’ s Well-Being Variables o f Full
Sample.............................................................. 53
Table 13: Means and Standard Deviations o f Major Study Variables for Lesbigay
Headed (NTF) and Heterosexual Headed (LSOG) Single Parented Families and
Married/Partnered Parented Households ......................................................56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VI
Table 14: Means and Standard Deviations o f Major Study Variables for Full Sample,
Lesbigay Parented Families and Heterosexual Parented Families........................ 60
Table 15: T-tests for Groups on Family Relationship Quality for Mother’ s Sexual
Orientation.................................................................................................................... 61
Table 16: T-tests fo r Groups on Well-Being o f Lesbigay Mothers and Their Children
and Heterosexual Mothers and Their Children ....................................................62
Table 17: Means and Standard Deviations o f Transformed Major Study Variables for
Full Sample, Single Parented Families and Married/Partnered Parented
Households.................................................... 64
Table 18: T-tests fo r Ratings o f Groups on Family Relationship Quality fo r Mother’ s
Relationship Status....................................................................................................... 65
Table 19: T-tests fo r Groups on Well-Being o f Single Mothers and Their Children and
Married/Partnered Mothers and Their Children..................................................... ,66
Table 20: T-tests and Chi-Squares for Demographic Variables and the Mother’ s Sexual
Orientation..................................................................................... 68
Table 21: Summary o f Hierarchical Regression Models o f Transformed Dyadic
Affectual Solidarity on Demographic, Mother’ s Sexual Orientation with Beta
Weights........................................... 72
Table 22: Summary o f Hierarchical Regression Models o f Transformed Dyadic
Conflict on Demographic, Mother’ s Sexual Orientation with Beta Weights 73
Table 23: Summary o f Hierarchical Regression Models o f Transformed Dyadic
Emotional Support on Demographic, Mother’ s Sexual Orientation with Beta
Weights.................................................................................. 74
Table 24: Analysis o f Variance for Transformed Affectual Solidarity........................... 75
T able 25: Summary o f Hierarchical Regression Model at Step 2 o f Transformed Dyadic
Affectual Solidarity on Demographic, Mother's Relationship Status with Beta
Weights...........................................................................................................................78
Table 26: Summary o f Hierarchical Regression Model at Step 2 o f Transformed Dyadic
Conflict on Demographic, Mother’ s Relationship Status with Beta Weights 79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Vll
Table 27: Summary o f Hierarchical Regression Model at Step 2 o f Transformed
Dyadic Emotional Support on Demographic, Mother’ s Relationship Status with
Beta Weights. ............................................................................................................ 80
Table 28: Summary o f Hierarchical Regression Models at Step 2 o f Transformed
Mother’ s Conflict on Demographic, Mother’ s Relationship Status with Beta
Weights.......................................................................................................................... 82
Table 29: Summary o f Hierarchical Regression Models o f Transformed Mother’ s
Depressive Symptoms on Demographic, Family Relationship Quality and Family
Type with Beta Weights................................................................................................86
Table 30: Summary o f Hierarchical Regression Models o f Transformed Mother’ s
Negative Affect on Demographic, Family Relationship Quality and Family Type
with Beta Weights............................................................................................ 88
Table 31: Summary o f Hierarchical Regression Models o f Transformed Mother’ s
Positive Affect on Demographic, Family Relationship Quality and Family Type
with Beta Weights...................... 90
Table 32: Summary o f Hierarchical Regression Models o f Transformed Mother’ s Self-
Esteem on Demographic, Family Relationship Quality and Family Type with Beta
Weights.......................................................................................................................... 92
Table 33: Summary o f Hierarchical Regression Models o f Transformed Child’ s
Depressive Symptoms on Demographic, Family Relationship Quality and Family
Type with Beta Weights................................................................................................ 94
Table 34: Summary o f Hierarchical Regression Models o f Transformed Child’ s
Negative Affect on Demographic, Family Relationship Quality and Family Type
with Beta Weights........................................................................................................ 96
Table 35: Summary o f Hierarchical Regression Models o f Transformed Child’ s
Positive Affect on Demographic, Family Relationship Quality and Family Type
with Beta Weights..................................................................... 98
Table 36: Summary o f Hierarchical Regression Models o f Transformed Child’ s Self-
Esteem on Demographic, Family Relationship Quality and Family Type with Beta
Weights............... 100
Table 37: Significant Variables in Hierarchical Regression Analyses for Transformed
Data fo r Specific Aim II............................................................................................ 118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
viii
List of Figures
Figure 1: Map of location of Study Lesbigay Mothers and Their Children in the
United States...................................................................................................... 25
Figure 2: Alternative Theoretical Explanation for the Relationships Between
Well-Being and Family Structure............................................................,...131
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ix
Abstract
This study examined the relationships among individual well-being, family
relationship quality and family structure (mother’s sexual orientation and mother’s
relationship status) in dyads from non-traditional families, in which each dyad was
composed of a mother and an adolescent or adult child. The study included 112
mother/child dyads with an additional 29 families who had at least one family member
participating. All mothers had legally divorced or ended a significant relationship with
a man in which a child was conceived. In regard to family structure, 75 families had
mothers who subsequently were in relationships exclusively or primarily with other
women, while 66 families had heterosexual mothers. Families were further categorized
by mother’s relationship status: mothers who were single (never remarried) versus
mothers who had married or partnered again. Results showed some support for the
hypothesis that family relationship quality is related to mother and child’s individual
well-being: dyadic affectual solidarity for mothers and children and dyadic conflict for
mothers. In regard to the family structure hypothesis, mother’s sexual orientation was
not related to quality of the dyad’s relationship, mother’s well-being or child’s well
being. Mother’s relationship status was significantly related to mother’s well-being.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Introduction
Researchers and social commentators have widely debated the role of family in
determining individual well-being. This discussion is particularly relevant when
families have become so diverse that there is no single portrait of the average American
family. Two major hypotheses have been advanced concerning how family may
influence individual well-being. The first hypothesis, family structure hypothesis,
focuses on how a family member’s well-being is related to the structure of the family,
operationalized as the number of parents in the household and/or the parents’ sexual
orientation. The other major hypothesis, based on family process theory, suggests
individual’s well-being is predicted by the quality of his or her relationships with other
family members. This introduction will review each of these literatures while also
addressing the question of how these literatures should be interpreted with respect to
non-traditional families - particularly families led by lesbigay (lesbian, bisexual and
gay individuals) mothers.
Little is known about how family structure and relationship quality affect
parents’ and children’s well-being in non-traditional families. Yet, on the subject of
children with lesbian and gay parents, the American Academy of Pediatrics recently
stated “children’s optimal development seems to be influenced more by the nature of
the relationships and interactions within the family unit than by the particular structural
form it takes” (Perrin & the Committee on Psychosocial Aspects of Child and Family
Health, 2002, p. 341). Previous literature examining families with heterosexual parents
found that family structure alone (i.e., whether the parents are first married, divorced,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
remarried or single) is a poor predictor of well-being. Parental sexual orientation is
also a poor predictor of well-being. However, after controlling for demographics,
family relationship quality in heterosexual parented families does predict individual
well-being. For example, heterosexual divorced single mothers and their children
report the same overall level of relationship quality as heterosexual remarried mothers
and their children. In other words, partnership status is not associated with relationship
quality in these studies (e.g., Demo & Acock, 1994), but relationship quality is relevant
to well-being.
To date, there are few studies that examine both lesbian mothers’ and their adult
children’s relationships, and even fewer studies comparing family types (i.e., family
relationship quality and well-being of lesbian mothers and their children compared to
divorced heterosexual mothers and their children.) The exception is research findings
that children in lesbian-parent households feel better about their relationships with their
stepparents than children in heterosexual-parent households. Overall, there is limited
understanding of how these non-traditional families’ well-being is affected by family
structure and family relationship quality.
Theory and Research on Family Structure and Well-Being
A number of researchers have discussed predictors of child’s well-being,
focusing on structural variables such as the number of parents in the home (Ross &
Mirowsky, 1999; Weinraub & Gringlas, 1995) and the sexual orientation of parents
(Patterson, 1992, 1995). Several competing models predict influences on family
members’ individual well-being. Some theorists focus on family structure as the most
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
important predictor of parents’ and children’s well-being. For children,
psychoanalytic theory historically focused on the Oedipal and Electra dramas, in
which parents of both genders must be present to facilitate the successful
resolution of children’s conflicts (Bronfenbrenner, 1960). Feminist psychoanalytic
writers emphasize the different influences of male and female parents on child
development, suggesting a child be exposed to both (Chodorow, 1980). Social learning
approaches assert that children are provided with different lessons, examples and
rewards from their male and female parents (Huston, 1983). Structural functionalism
theory (within the field of sociology and family studies) asserts that parents’ conformity
to traditional differentiated gender roles ensures family stability and societal order,
whereas deviance from these roles leads to instability and divorce (Kingsbury &
Scanzoni, 1993). The belief that a child universally needs a family structured along
Anglo-Saxon post-industrial traditional lines (the intact nuclear family), has been
articulated as recently as 1997 (Wardle, 1997) and 2000 (Lemer & Nagai, 2000).
Wardle, for example, writes:
children generally develop best, and develop completely, when
raised by both a mother and a father and experience regular
family interaction with both genders’ parenting skills during
their years of childhood. It is now undeniable that, just as a
mother’s influence is crucial to the secure, healthy, and full
development of a child, paternal presence in the life of a child
is essential to the child emotionally and physically
(p. 860).
Thus, it is clear that these reviews - despite the lack of supportive, careful empirical
studies - have widespread effects on the real and perceived situation confronting
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
lesbigay parents. Cameron and Cameron (1998) reviewed appellate custody dispute
cases and found that the accusations made by one parent against the other suggests
homosexual parents are disproportionately of poor character and associated with harm
to their children. Cameron, Cameron, & Landess (1996) also have compiled a catalog
of the theorized ill effects of lesbigay parenting, and their work, as documented by
Stacey and Biblarz (2001), has been frequently cited by court cases and child welfare
policymakers in their efforts to restrict lesbigay parenting and adoption rights. Wardle,
who cites Cameron et al. (1996) prominently in his 1997 review, also drafted Utah state
regulations to prevent adoption and foster care placement with lesbigay parents.
Cameron et al. (1996) were frequently cited during an Arkansas Child Welfare Agency
Review Board policy development process, which limited foster care placements to
heterosexual households.
The literature from this research can be characterized as the “family structure
hypothesis”, which suggests that children reared in households without their two
biological parents (a man and a woman) exhibit lower levels of well-being than children
raised in intact, first-married family units with heterosexual parents. For parents, this
framework suggests that married parents should have higher levels of well-being than
single parents. The family structure hypothesis has further application for lesbigay
parented households, predicting children in lesbigay parented households will have
lower levels of well-being than in heterosexual parented households, and lesbigay
mothers will experience lower levels of well-being than heterosexual mothers (Cameron
& Cameron, 1997; for reviews see Patterson, 1994a; Tasker & Golombok, 1998).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Despite the theories, empirical evidence is scarce. There are two bodies of
literature: the first compares divorced and single heterosexual parents to first-married
heterosexual mothers, and the second compares lesbian mothers to heterosexual
mothers. In both cases, authors have concluded that although some evidence suggests
family structure may predict well-being when not controlling for other critical variables,
the research in support of this family structure hypothesis has been seriously flawed
(Blechman, 1982; Demo & Acock, 1994; Tasker & Golombok, 1998; Cameron &
Cameron, 1997; Cameron, Cameron, & Landess, 1996).
Research with Heterosexual Parented Households
Research within heterosexual parented families primarily focuses on whether
mothers are divorced and the effects on family members. The greatest difference among
children from different family structures is that children in families with mothers who
are married for the first time do significantly better on well-being measures than
children in divorced families (Hetherington, Cox, & Cox, 1982; Hess & Camera, 1979).
The effect of parental relationship status was examined in one study that compared
children of heterosexual first married mothers, children of continuously single mothers,
children of divorced mothers, and children of remarried mothers (Demo & Acock,
1994). For mothers, controlling for variables other than structure consistently found
heterosexual first married mothers report significantly higher levels of well-being,
followed closely by stepmothers, compared to divorced or continuously single mothers
(Demo & Acock, 1994; Acock & Hurlbert, 1993). Another study that addresses family
structure, compared five different family structures (adoptive, two-parent biological,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
single-mother, stepfather, and stepmother households) and found few structural
differences, suggesting the importance of family processes rather than family structure
(Lansford, Ceballo, Abbey, & Stewart, 2001). In conclusion, heterosexual partnered
women and their children fared better than heterosexual non-partnered women and their
children. However, researchers generally conclude that being a single divorced mother
is a stressor, often including financial hardships (Amato & Booth, 1997). The life
stressors, rather than the family structure, may have caused family type effect in well
being differences. Focusing only on divorced mothers, the present study will analyze
family structure in a sample of mothers who have experienced the stressor of divorce.
Socioeconomic aspects of the stressor are accounted for as covariates in the analyses
conducted in the present study, allowing us to examine the effects of socioeconomic
strain and the effects of current partnership status (single versus partnered) separately
from the effects of divorce per se.
Examining families with heterosexual parents, and children and adolescents still
living at home, there have been significant associations among relationship process
variables, as well as with child’s well-being. Among heterosexual parents and their
children, parental relationship quality has been found to have significant associations
with child outcomes (Cowan, Cowan, & Kerig, 1993; Emery, 1982; Emery & O’Leary,
1984; for review see Erel & Burman, 1995).
Research with Lesbian. Bisexual and Gay Parented Households
A recent review cites 21 studies examining families with a lesbigay parent, and
seven of these examine well-being (Stacey & Biblarz, 2001). These studies
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
conspicuously fail to provide support for the family structure hypothesis, at least with
respect to parental sexual orientation. Research on well-being has generally relied on
samples of lesbian mothers and their children, not gay fathers and their children.
Researchers have used comparison samples composed of divorced heterosexual women
and their children, hence, controlling for divorce, but not necessarily for current marital
status (Golombok et al., 1983). Researchers examining children of lesbian parents
compared to children of heterosexual mothers report no significant differences on
measures of mental health and social competence based on their mothers' sexual
orientation (Patterson, 1994b; Chan, Raboy, & Patterson, 1998; Golombok et al., 1983;
Kirkpatrick, Smith, & Roy, 1981). Differences in perception failed to emerge in
lesbians’ adult children and divorced heterosexual mothers’ children in regard to a self
esteem inventory (Huggins, 1989). A study of adolescent children of lesbian mothers
without a comparison sample found that those teenagers that perceived more stigma had
lower self-esteem (Gershon, Tschann, & Jemerin, 1999). There also were no
differences for adult children in reports of life satisfaction based on family type
comparing traditional, single parent, and homosexual households (Zweig, 2000).
Studies also report that lesbian mothers are at least as psychologically healthy as
matched heterosexual mothers and larger standardized samples on emotional
adjustment, self-acceptance, self-concept, self-confidence, self-esteem, and depressive
symptoms (Dundas & Kaufman, 2000; Golombok et al., 1983; Green, Mandel, Hotvedt,
Gray, & Smith, 1986; Mucklow & Phelan, 1979; Rand, Graham, & Rawlings, 1982).
Studies of lesbian women’s well-being have included lesbian mothers in the sample but
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
have not conducted separate analyses examining this subgroup. Researchers have
found some small differences that were attributed to the added stressors lesbians and
minorities face, but that overall lesbians were similar to heterosexual women on
measures of well-being (Rothblum & Factor, 2001; Flughes, Pollinger Haas, Razzano,
Cassidy, & Matthews, 2000; Hughes, Pollinger Haas, & Avery, 1997; Cochran & Mays,
2000; Thompson, McCandless, & Strickland, 1971). In summary, comparisons of
lesbian mothers and their children with heterosexual mothers and their children suggest
that family structure differences based on parental sexual orientation predict neither the
children’s nor parents’ well-being (Patterson, 1992).
Stacey and Biblarz (2001) have criticized this characterization of no differences
between the two family types. They suggest that differences in favor of lesbigay
parents have been downplayed and “heterosexism has hampered intellectual progress in
the field” (p. 159). Writers may simply be expressing appropriate interpretive caution.
To elaborate, these previous studies have several limitations, including the fact
that all samples were comprised of white, relatively affluent, middle-class convenience
samples. Additionally, lesbian and heterosexual mothers, while matched on divorce,
were not matched on current partner/non-partner status. Perhaps the researchers’ most
severe criticism focuses on the studies’ small sample size. Out of the studies cited by
Stacey and Biblarz (2001) that focus on well-being, a total of 52 lesbian mothers have
been compared to 52 heterosexual mothers. Participants in children studies total only
75 children with lesbian mothers and 75 children with heterosexual mothers. The small
samples in each study leave room to question the power of the analyses and possible
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
results, especially where no differences were found. This study will address those
limitations by using a larger sample and a matched comparison group of heterosexual
mothers and their children that include mothers’ partnered/non-partnered status.
Family Background Variables that Influence Well-Being
In contrast to the psychological theories concerning the influence of family
process and family structure on well-being, a sociological perspective suggests that
well-being is largely a function of access to power and resources, and position in the
larger social structure. In this tradition, researchers have typically controlled for a
number of background or demographic variables that are considered to influence
parents’ and children’s well-being through their influence on variables such as access to
social and economic resources, and stigmatized status (or lack thereof). These variables
include: parents’ age and divorce history, race, and socioeconomic status. These
variables were chosen based on previous studies that found them to be important when
comparing different family types (Voydanoff, 1990; Weitzman, 1985; Weiss, 1984).
These variables may explain variation in well-being within each family type. For
example, among divorced women, loss of income is an important predictor of post
divorce adjustment (Burkhauser & Duncan, 1989). Additionally, many studies report
that competence, self-esteem and well-being increase with age through early and middle
adulthood (Flerzog, Rodgers, & Woodworth, 1982). These background variables have
been controlled for in two ways: matching families on particular variables and using
statistical control. This study used statistical control for most background variables. In
previous studies, divorced status has been controlled by statistical regression or
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
matching. In the present study, we restrict our sample to households in which the
biological parents were divorced (in most cases) or separated from an opposite sex
relationship so that the child was raised by his or her biological mother and any partners
the mother may have had.
Parental Divorce and Adult Children’s Well-Being
Researchers have identified some long-term effects of divorce on adult children
by comparing children with divorced families to those with intact families. Adult
children of divorce are more likely to have problematic relationships with their parents
and to have greater difficulties in their own marital relationships (Hetherington, 1999;
Zill, Morrison, & Coiro, 1993). Amato’s (2001) meta-analysis (combining studies of
older and younger children) found they scored significantly lower on psychological
adjustment, self-concept and social relations. However, other reviews of the literature
conclude that by adulthood the long-term outcome of divorce for the majority of
children is resiliency rather than dysfunction and that controlling for socioeconomic
status accounts for differences in well-being (Bryner 2001; Kelly, 2000).
Family Relationship Quality and Well-Being
While many family researchers have focused on the association between family
structure and well-being, others have examined the association between family
relationship quality and well-being. This body of work, which could be characterized
as family process theory, suggests that the quality of relationships has more influence
on well-being than the structure of the family. Thus, according to systems theorists, a
strong association exists between family relationships and the psychological well-being
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
of those family members (Carter & McGoldrick, 1989). Theorists speculate that
family affectual solidarity and emotional support enhance well-being, while conflict
reduces well-being for both parents and children (Silverstein & Bengtson, 1997; Emery,
1988; Fustenberg, Morgan, & Allison, 1987).
Family relationship quality can be defined as the nature and the extent of social
relationships among members of a family, and this construct can be assessed using a
variety of subjective and objective indicators, such as self-reported instrumental help
and support given and received by each family member (objective) and perceived
affectual solidarity or conflict among different dyads within the family (subjective).
Typically, a family under consideration is divided into a subset of possible dyads, and a
subset of these dyads is assessed from both constituents’ points of view (i.e., mother
and son are both asked to characterize objective and subjective aspects of their
relationship). In most studies, only some respondents within the family are available to
give reports, although the ideal study would examine all possible dyads taken from the
perspective of each family member, as in the Longitudinal Study of Generations
(Bengtson & Mangen, 1988).
Research from this tradition demonstrates that children are happier when they
get along well with others in the family, and that they are less anxious and fearful if
they have enjoyable relationships with their mothers (Demo & Acock, 1994). Children
and adolescents experiencing lower levels of conflict with their parents also have a
stronger sense of well-being (Fustenberg et al., 1987; Shek, 1998). Conflict is a key
factor associated with reduced psychological well-being of children (Emery, 1982;
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
Howes & Markman, 1989) and negative parental relationships in adolescence is
related to lower levels of well-being for young adults (Aquilino & Supple, 2001).
Adolescents and young adults who reported stronger affectual solidarity to parents
reported greater self-esteem at baseline and at 17-year and 20-year follow-ups (Roberts
& Bengtson, 1996). Moreover, conflict with children has a negative influence on
mothers’ well-being (Demo & Acock, 1994). Further, emotional support reduces
mothers’ depression, anxiety and other psychological problems (Ross, Mirowsky, &
Goldsteen, 1991). Mothers who report close and supportive relationships with their
spouses evidence well-being, suggesting a benefit to heterosexual women in “good”
relationships (Grove, Hughes, & Style, 1983). Additionally, relationships with aging
parents are an important factor in influencing adult children’s well-being (e.g., Mancini
& Blieszner, 1989). In summary, family system theory suggests that family relationship
quality predicts family members’ well-being, a position empirically supported in the
literature examining heterosexual mothers and their children across the life span.
Family Structure and Family Process
Taken together, family structure theory and family process theory naturally
suggest another question: how does composition relate to process? This is a particularly
interesting question, at least with respect to parental sexual orientation. A comparison
of lesbigay and heterosexual parented families is a natural laboratory in which to
examine this question. Few studies have examined family member’s relationships and
feelings about each other in lesbian compared to heterosexual mother headed
households. In general, studies including parent-child relationship measures for both
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
children and mothers have shown no differences between lesbian and heterosexual
parented households. Adults raised in lesbian households generally report positive
relationships with their mother and her partner (Tasker & Golombok, 1998). They also
scored higher on an attachment-related assessment than did children raised by
heterosexual mothers (Golombok, Tasker, & Murray, 1997). Researchers found that
adult children of lesbian mothers were able to forge closer relationships with their
mothers’ new female partner than those children from heterosexual households with
their mothers’ new male partner (Tasker & Golombok, 1994). Some investigators have
suggested that children from lesbian households could more easily accommodate
change in the family structure, as their mothers’ partner could be regarded as an
addition to the parental structure, rather than as direct competition to the fathers (Tasker
& Golombok, 1994; Tasker & Golombok, 1998).
Several researchers have studied family relationships in young children across
different family structures. Brewaeys, Ponjaert, Van Hall and Golombok (1997a) found
that when asked about their perceptions of parent-child relationships, all children
reported positive feelings about their parents, and there were no differences in child’s
report as a function of family type (comparing children conceived via donor
insemination with lesbian parents to children conceived by partnered heterosexual
parents). The original Golombok et al. study (1983) and Kirkpatrick et al. study (1981)
collected data on children’s family relationships and reported no differences across
family types. These researchers have also examined the mothers’ perspective of the
parent-child relationship and found no differences between lesbian and heterosexual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
mothers (Brewaeys, Ponjaert, Van Hall, & Golombok, 1997b; Golombok et al., 1983;
Kirkpatrick et al., 1981; Shavelson, Biaggio, Cross, & Lehman, 1980). In sum, there is
minimal research comparing family relationships and well-being in lesbian mothers and
their children to heterosexual mothers and their children. This study will provide more
information about family relationship quality and its influence on well-being for older
children.
Methodological Considerations
The research on lesbian headed households is limited. While the population of
lesbian parents and their children offers an opportunity to test psychological theories
about family structure, it is a difficult population to sample given the long history of
legal discrimination and stigmatization (Parks, 1998; Patterson & Redding, 1996). By
necessity, previous researchers have relied on snowball samples and convenience
sampling, and this study has taken a similar approach. As noted in Stacey and Biblarz
(2001), there are intrinsic sampling difficulties involved in studying lesbigay families,
given the stigmatization and civil liberties issues that these families face in many
jurisdictions. These issues are most pronounced for the older cohort of parents who
initially were involved in heterosexual marriages and then subsequently divorced.
Stacey and Biblarz (2001) note that, until recently, it was relatively common for
lesbigay individuals to marry heterosexually before coming out. The younger cohort of
lesbigay individuals who are relatively less stigmatized in this way have not yet had
time to raise children to adulthood.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
Partly as a result of the intrinsic difficulties in sampling, we chose to examine
a group identified through a national social organization for children of lesbigays. In
this study, the families represent the population of families in which at least one family
member had some connection to a social support organization for either gay, lesbian,
bisexual and transgender parents (Family Pride Coalition) or for children who have gay,
lesbian, bisexual or transgender parents (COLAGE-Children of Lesbians and Gays
Everywhere). This population may be the only one in which it is practical and possible
to sample within the current cultural context. Unlike traditional snowball samples used
in prior studies, our sample includes families from across the United States, as well as
rural, urban and suburban settings. Prior studies have been constrained in their
geographic representativeness by a snowball sampling approach starting within a
defined region (i.e., usually a large, multicultural urban coastal city such as Los
Angeles, San Francisco or New York).
We chose families with teenage and adult children partly because there is so
little written about them in the literature, and also because we were interested in issues
related to the adult adjustment of people who have lesbigay parents, rather than in
tapping the short-term effects of adjusting to stigmatization faced by lesbigays and their
family members. At the same time, we sampled lesbigay mother headed households,
excluding gay fathers because they are even more difficult to locate than lesbian
mothers (Stacey & Biblarz, 2001). Gay fathers combat not only heterosexual bias but
also gender bias in custody cases (i.e., the assumption that women are better parents
than men). As a result, the current cohort of gay fathers (with teenagers and young
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
adult children), who came out after they had children may not necessarily have had
custody of their children, and their level of contact may not be parallel to the same
cohort of lesbigay mothers.
This sample includes two subsets, each of which is considered nontraditional, in
the sense that it is composed of parents who divorced or ended a relationship with a
man as if they had been married. However, as Wallerstein (2001) notes, divorce is
increasingly more commonplace in this culture, making this children of divorce sample
representative of a large segment of the population. All of our subjects have a common
experience: the end of the pairing that produced the study child. This allows us to
examine the point raised by Wardle, who suggests that healthy child development
depends on parenting by a married heterosexual couple (Wardle, 1997).
Well-Being as a Multidimensional Construct
Since the previous literature has concerned itself so much with the mental health
of children - and, at times, of their parents - we examine well-being as the main
outcome. Well-being is a multi-dimensional construct in the literature, not simply the
absence of depression or negative affect (Lucas, Diener, & Suh, 1996). In this study,
we examine both the positive and negative components of well-being: on one side
depressive symptoms and negative affect, and on the other side, positive affect and self
esteem. We wanted to capture the complexity of well-being in relation to multiple
aspects of family.
There are, however, some limitations on the variables that were chosen for this
study. The measures given to the individuals in the lesbian headed households were the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
same measures in the longitudinal study of families. This panel study was already in
place, for nearly 30 years, and since we wanted to compare to this sample, we were
limited in our questions. In this case, the trade-off for the opportunity to compare the
family types meant that we were not able to use other measures of family relationship
quality or well-being.
The Present Study
This study examined the relationship among family structure, family
relationships and psychological well-being of individual family members. As reviewed,
family structure has long been used as a predictor variable for psychological well-being,
although research has found little differences among family types. Therefore, the
question becomes: how important is family structure? Structure tells us very little about
how family members feel about themselves, how they provide support and how much
conflict they have among each other. These family relationship factors may vary across
family type and account for differences in parents’ and children’s well-being.
Researchers concede that this does not mean that family structure is unimportant for
well-being, but rather that there are more proximate factors (such as family relationship
quality) that are influenced by family structure and which then, in turn, influence family
well-being. Researchers have asserted that for a study to understand the most important
forces influencing well-being, it must move beyond previous studies designed around
two or more family types (Bronfenbrenner & Crouter, 1983). Researchers must
“accept the challenge of identifying more proximate familial experiences... of children’s
lives” (Demo & Acock, 1994, p. 223-224).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
The present study, like Demo and Acock’s (1994), examined family structure
compared to family relationships. It also went further by examining two variations in
family structure: one variation involved the sexual orientation of the mother, and the
other variation involved whether the mother had a partner. Furthermore, this study
went beyond many previous research projects by examining family functioning from
both the mother’s and adult child’s perspective, as well as by taking multiple measures
of well-being directly from each member of these dyads. This study explored measures
of family relationship quality, comparing them with lesbian parented and heterosexual
parented families and between families in which the mother had a partner versus
families in which she did not. Next, affectual solidarity, conflict, and emotional support
were used as predictors of well-being. Analyses were conducted to determine the
relative importance of predictors of family relationship quality and well-being using a
variance accounted for strategy. Finally, exploratory analyses determined whether
predictors account for different variances between the two family types and if there
were age and gender differences in family relationship and well-being variables.
The present study had three major aims. The first aim was to examine mothers’
and children’s perceived relationship quality across lesbigay mother and heterosexual
mother households, and across single and dual parent households. In hypothesis I it was
predicted that after controlling for demographics, mothers’ and their children’s
perceptions of the quality of their relationships with each other would show no
significant differences, regardless of mothers’ sexual orientation. In other words,
family relationship quality would not differ as a function of maternal sexual orientation,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
after controlling for demographics. In hypothesis II, the family relationship quality
was again examined, comparing families in which the mother was married or partnered
(remarried or repartnered) and those in which the mother was single. It was predicted
that mothers’ and their children’s perceptions of the quality of their relationships with
each other would show no significant differences, regardless of the number of parents in
the household (mothers’ partner status), after controlling for demographics. Stated
differently, after controlling for demographics, family relationship quality- will not differ
according to the number of parents in the household.
The second aim of the study explored the two competing theories on the
influences of individual well-being based on alternative family structures. The second
aim was to determine the contribution of mother-child relationship quality and family
structure (mother’s sexual orientation and partner status) to the prediction of both
children’s and mother’s well-being. Based on the family process theory, hypothesis III
predicted that family relationship quality would significantly predict mother’s and
children’s well-being across all family types. Again, based on family process theory
and in contradiction to the family structure theory, hypothesis IV predicted that family
type (lesbian versus heterosexual parented families, and single versus dual parent
families) would not make an independent contribution to the prediction of well-being
after controlling for family relationship quality (affectual solidarity, emotional support
and conflict) and family backgroimd variables (demographics).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Finally, the third specific aim was exploratory in nature, determining whether
there were age and gender differences for children’s well-being and perception of their
relationships with their mothers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Methods
Participants
One hundred and forty-one families (with at least one family member reporting)
were included in this study, drawing from a pool of mother and teenage/adult child
dyads. Mothers were previously divorced or had ended a heterosexual relationship in
which a child was conceived. Two types of families were examined: 77 families (48
complete dyads and 29 other families with one person reporting) with lesbigay mothers
and their teenage/adult children recruited from two national lesbigay support
organizations and 66 families (66 dyads) with divorced heterosexual mothers and their
teenage/adult children from a subset of the Longitudinal Study of Generations (LSOG).
From these two samples we further subdivided the groups into married/partnered
mothers and their children, and single mothers and their children. The percentage of
single mothers was surprising small, 17.7 % (n = 12) for the lesbigay mothers and
33.3 % (n = 22) for the heterosexual mothers. This small number of single mothers
limited the statistical comparisons that could have been made examining interaction
effects (mother’s sexual orientation by mother’s partner status) and these analyses were
not run because of power issues.
Lesbigay mothers and their children. Forty-eight complete lesbigay mother/child dyads
were drawn from the Non-traditional Family Study (NTF), a study of gay, lesbian and
bisexual parents and their children. Sixty-eight mothers and 57 teenage/adult children
(43 daughters, 75.4% and 14 sons, 24.6%) were recruited through Gay and Lesbian
Parents Coalition International (GLPCI), now called Family Pride Coalition, and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
Children of Lesbians and Gays Everywhere (COLAGE), for a total of 77 families
with at least one family member reporting.
In order to gain an understanding of the social context of the lesbigay mothers’
lives, questions were asked about their sexual identity. The majority of the women in
the NTF sample defined their sexual orientation as lesbian (n = 56, 87%). Two women
defined themselves as gay and seven women defined themselves as bisexual (six of
whom were either with a woman or had recently been with a woman, and one who was
not with anyone at the time of the study).
Women were asked the ages they had come out in four different contexts, in
order to understand the individual’s position in her coming out process (the degree with
which one acknowledges one’s own sexual orientation, as well as telling others), often
an important aspect of examining gay and lesbian samples (Herek, 1991). In this
sample, the range of ages at which the women recognized themselves as non
heterosexual and disclosed this information to others ranged widely from early
childhood (4 years old for same sex attraction) to midlife (48 years old for first same
sex relationship). In what is sometimes seen as the earliest stages of coming out, the
women noticed they were attracted to a member of their own sex when they were
between 4 years old and 45 years old, with a mean of 17 years old (S.D. = 10.7). The
age that they began to define themselves as something other than heterosexual (that they
came out to themselves) ranged from 11 to 47 years old, with a mean of 30 years old
(S.D = 8.9), a median of 13 years later after noticing their same sex attraction.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
Table 1
Age o f Lesbigay M other in the Coming Out Process.
Process N M ‘
S.D. M edian Range
Attraction to sam e sex 65 17.46 10.64 14.00 4 -45
Come out to self 64 29.73 8.89 30.00 11-47
Same sex relationship 65 28.26 10.08 29.00 11-48
Age told another person 65 31.20 8.53 31.00 14-47
The age at which they “told another person” was soon after their first relationship and
coming out to themselves, with a mean of 31.2 years (S.D. = 8.5 years), with a range
from 14 to 47 years old. The age at which participants first disclosed their lesbian,
bisexual or gay sexual orientation to another person ranged from 14 to 47 years old,
with a median of 31 years. Ages for various facets of the coming out process are shown
in Table 1. Participants reported that they first disclosed their orientation to their study
children at ages ranging from child’s birth (never keeping it a secret) to 25 years, with a
standard deviation of child’s age of 5.7 years and a median of 10.6 years. All mothers
had disclosed their orientation at least two years before participating in the study.
There were 57 teenage/adult children of lesbigay mothers who ranged in age
from 17 (one individual) to 38 with a mean of 25.79 (S.D. = 5.00). The majority of
them were single (71.93%), while 28.07% were either married or partnered. They were
a predominantly Caucasian sample, with only two Hispanic/Latino, and four mixed
ethnic background (one Caucasian/African American, two Caucasian/Hispanic/Latino,
and one Caucasian/Middle Eastern).
This sample cannot be used to study the effect of lesbigay parents on child’s
sexual orientation. COL AGE contains a subgroup devoted to providing a forum for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
sexual minority children of sexual minority parents, and adult children from this
COLAGE subgroup (Second Generation) cannot be distinguished from children from
COLAGE who do not belong to that subgroup. Effectively, homosexual children were
oversampled, but the extent of the oversampling is unknown. Across the entire sample
of children, three identified as lesbian or gay (5.5%), and four identified as bisexual
(7.3%), and four identified as “not sure” (7.3%). Thus, the sample of adult children was
predominantly composed of respondents who identified as heterosexual.
All but one of the families was from the United States, with one dyad
responding from Canada (see Figure 1). Participants met each of three criteria to be
included in the study. First, the adult children were 17 years old or older. Second,
mothers were divorced or permanently separated from a heterosexual relationship that
had been comparable to a marriage. Third, each lesbigay mother must have come out to
herself and disclosed her sexual orientation to her adult child at least two years prior to
the study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
Figure 1.
Map o f Location o f Study Lesbigay Mothers and Their Children in the United States.
CO LU M B IA
® C - A N A;D A Q U E B E C rfS j-s,:
' ^ : : C a lg a ry : ONTARIO • • ■ ■ * ' » r
^ S e a t t l e ' - - - M . B gS xA _ SA SK . ; ^ W in n ip eg \ Q u e b e c / ^ ' T j . s , ""
a WASH Helena ' ’ ■ ~ -' --------------------------------- '' M ontreal
: , 0H elerkl . DAKOTA ; - «
S a lem . a . .... i. -a--...x w Ottawa' s~ 5 V R ;’ - • < ;
■ : MONTANA BismoreK M1WI 1 - ■ jna"V 1 ■
OREGON ■ " : ; " v .® ; P a ! jl T o ro n to ., - r t p c t a n
:®AHO , m o m n p . §©u.TH OAKOtA ' MICWGAli P ^ S a n f . m
. . « « j . t i o - . r * T . . a r rt , ■ < ? L r :'s u ' T“
. > P1 \ J ^ a it Lake City NEBRASKA lcgyA Ch,« ^ | o l < j L s v * Philadelphia
< 3 i
tcia c M , > r t " • ^ ~ S alt L ak e City? ' - ' t i e H B M n * / ' , l0 W 4 .'; ' - t i o : • Q
a C £ ; a.. ............. X:; !U iN Q is ;.® ::r,S ' , ^ v tts n m a ttm
' % i — i'*"' ■ {... T o p e k a _ In d ian ap o lis 4
% U * ■ * - ; C O L O R A D O ; r /•-5st.'L o ^C -V -'^^E N iyC K Y .r-' ■ ...... ‘ ......... a
, , \ % ®LasVfegas T " -" L " oo-..K A » S. -.M IS S O U R I; ua-mvme
lo s A n g e B ^ - ARIZONA,-:- . _ » S a n t a F e Memphis . Columbia v '
S a n D le g o * . p h o e n i x 3 OKLAHOMA P • Atlarit e Co,™ bl« C
~®Mexicall * NEW MEXICO ■ * L U f " ARKANSAS ; S-0,
- ;• \ R p ™ Dallasaf ALAr'SEORGI*
Legend
Push pins = Mothers
Flags = Teenage/adult children
Squares = Mulitple participants in one
city
1 ; \ _ Jacksomife
■ A u sU n / ’ T allan atsee
6 - "■»* ^ V , Orleans
San Antonio
V fX A S •' ’ _ © ^ N a ssa u :
C U « ' ' '
Potential participants were recruited over a two-year period. They were
recruited at two national annual conferences (one held in Los Angeles, California, and
one in Provincetown, Massachusetts), and members of each organization were sent
letters asking if they were interested in participating. While 1000 letters were sent to
parents and 900 letters were sent to the children, it is impossible to know how many
were actually mothers and children of lesbigay mothers and how many met criteria to be
in the study since no statistics were kept on the organizations’ membership.
Additionally, the letters had to be sent out through the organizations’ offices, rather than
by the investigator because their mailing lists were not made directly available and it is
not clear to whom the letters were sent. For example, many of the members of GLPCI
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
are lesbigay parents with younger children, some are gay fathers and many of the
children of COLAGE are under 17 years of age. Thus, it is impossible to calculate any
response rate based on recruitment letters.
After each of the potential recruits stated an interest in filing out a questionnaire
and believed that their mother or adult child also would be interested in the study, the
other half of the dyad was then sent a letter asking if he or she was interested in the
study. Out of a possible 151 dyads represented in the preliminary inquiries, a minimum
of 100 dyads (a 66% response rate) were expected to participate. However, 66 possible
participants (one half of the dyad) were either unable to be located or were not
appropriate for the study (for example, mothers had children who were too young to
complete a survey, families may not have met the inclusion criteria, or one member of
the dyad was out of the country) when follow-up inquires were made.
Questionnaires were sent out only after both halves of the dyad agreed to
participate in the study. Eighty-five mother/child dyads agreed to participate and were
sent surveys (170 surveys), for a total response rate of 74.0%, 80.0% for mothers and
67.9% for children (one of the children who agreed to the study was on active military
duty and out of the country during the data collection time period).
Data collection procedures for NTF families followed procedures for LSOG
families. All respondents received a letter requesting permission to contact their
respective mother or child. With permission, the parent or child was contacted and
asked if they were interested in participating. Interested respondents received mailed
questionnaire packets similar to the LSOG questionnaires. As a token of appreciation,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
NTF participants were offered $10 or a one-year membership to COLAGE.
Participants returned the completed questionnaire using self-addressed stamped
envelopes supplied by the investigator.
Heterosexual mothers and their children. Sixty-six complete heterosexual mother -
child dyads were drawn from the USC Longitudinal Study of Generations (LSOG). The
dyads were selected based on mother and child’s ages and mother’s history of divorce.
All mothers reported being divorced and under 65 years of age and all children were 17
years or older. While there were many more dyads (n=T06) in the LSOG sample with
divorced mothers, the decision to drop dyads with mothers over 65 years of age was
based on the information that none of the NTF mothers were over 65 years of age, and
social science studies often see individuals over 65 as a separate cohort from midlife
individuals, particularly with regard to the central variables in this study (family
relationships and well-being). The data from these dyads was collected in 1997 from a
sample of 1724 individuals, 299 extended families. The 66 completed dyads of
divorced mothers and teenage/adult children included 46 daughters (69.7%) and 20 sons
(30.3%). Mother’s age averaged 49.52 (S.D. = 7.37) and child’s age averaged 27.56
(S.D. = 8.99). Based on a question asked in the 1991 survey data, none of the LSOG
divorced mothers self-identified as lesbian, gay or bisexual.
The LSOG sample was derived from an ongoing longitudinal project that has
followed intergenerational families (four generations) for almost 30 years. The original
1971 sample was drawn from a population of Los Angeles-area health maintenance
organization members. Within a random sample of male subscribers (1 out of every 6),
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
families were eligible for inclusion if they had a grandfather age 55 or older with a
dependent (usually a wife), a living child and at least one grandchild between the ages
of 16 and 26. In each family, participants include grandparents (first generation), their
children (second generation), their grandchildren (third generation) and when available,
their great grandchildren (fourth generation) who were added in 1991.
The overall response rate for the Time 1 survey was 64% (n = 2044 individuals
from 407 families). Time-2 (1985), Time-3 (1988) and Time-4 (1991) waves had
response rates of 81%, 75%, and 74%, respectively. In each new wave, the sample
included those family members eligible at the previous wave; new first- and second-
generation spouses who married into the family; and second-, third- and fourth-
generation children. A few exceptions to eligibility at each wave were made. A
number of subjects at each wave were no longer eligible because they had died, become
incapacitated, divorced out of the family or could not be located. It is primarily a
European-American (Caucasian) sample (84%).
Participants in the LSOG study were mailed a questionnaire in 1994 that took
approximately 45 to 60 minutes to complete. The questionnaire included topics about
family relationships, family background and demographics, and psychological well
being. LSOG participants completed the questionnaire and returned it using a self-
addressed stamped envelope supplied by the investigators. They were offered $10 for
their participation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
Materials
LSOG Participants completed a questionnaire that included demographic
measures and scales assessing psychological well-being and family relationships.
Participants from the NTF study completed a similar survey with questions reworded to
reflect their unique family structure and experiences.
Demographic and Family Background Measures
Demographic variables, which are often used for sample matching and comparisons
between samples, were collected. These variables were major demographic variables
that have previously been associated with family relationships and psychological well
being (Acock & Demo, 1994; White, 1992), and used as a basis to match heterosexual
mothers with lesbian mothers and their children in previous studies (Green et al., 1986;
Golombok et al., 1983). (See Appendix A.)
Age. Participants were asked to give their date of birth. Their ages were calculated as
the date on which they completed their questionnaire minus their date of birth. Mothers
were also asked their study children’s ages in case their children did not participate in
the study.
Racial identity. Participants’ race was identified through response to the open-ended
question, “What is your racial or ethnic background?”
Partner status. Partner status or relationship status was identified through response to
the question, “What is your CURRENT marital status or living arrangement?”
Participants who responded 1 = “single”, “dating”, “separated”, or “divorced” were
placed in one group, while those who responded 2 = “married” or “with partner as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
though married”, were placed in a second group. Mother and child were each asked
this question about their own relationship status.
Income. Participants were asked, “What was YOUR TOTAL HOUSEHOLD INCOME
for the last year?” Participants chose one of 16 possible answers, starting with “Less
than $10,000” and moving in $10,000 increments to “$150,000 or more.” Responses
were divided into three categories: 1 = “low”, 2 = “medium” and 3 = “high.” Those in
the low-income range reported $0 to $29,999 a year. Those in the moderate-income
range reported more than $30,000 and less than $69,999 a year. Those in the high range
reported more than $70,000 in annual household income. Mother and child were each
asked this question. Children’s income was based on their separate household income,
unless they lived at home or were in college and reported no full-time job, in which
case, total household income was left blank.
Education. Educational attainment was determined by asking, “What is the highest
level of education you have attained?” Responses were placed on an eight-point scale,
ranging from 1 = “8th grade or less” to 8 = “Post-graduate degree (Ph.D., M.D., J.D.,
etc.).” This question was asked of both mother and adult child.
Dyad (mother and child) cohabitation status. The dyad’s current status was determined
by asking, “Who lives with you?” This question was asked of both mother and child. If
either mother or child (or both) responded that they lived with the other, the dyad was
classified as living together, while if mother and/or child did not list the other as people
they lived with, the dyad was classified as living separately.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Mother’s sexual orientation. One question assessed mother’s sexual orientation,
“How do you define your sexual orientation?” Respondents indicated “heterosexual”,
“gay or lesbian”, “bisexual”, “no label” or wrote in an open-ended response. Those
who identified as “gay”, “lesbian” or “bisexual” and had chosen primarily women with
which to have intimate relationships were labeled 1 = lesbigay mothers, and those who
identified as 2 = “heterosexual” formed a second group. As described above, this
question was asked of both samples and served as the basis on which to divide the two
samples. All LSOG mothers responded “heterosexual”, while all NTF mothers
responded “gay”, “lesbian” or “bisexual”.
Child’s gender. Children were asked their gender and could respond, 1 = “male” or 2 =
“female.” NTF children were also given an unused third option, “other” with space to
specify. Mothers were also asked the gender of their study children with 1 = ’’ male” or
2 = “female”, in case their children did not participate in the study.
Child’s parental status. Child’s parental status was determined by asking, “Do you have
any children?” Teenage/adult children who responded “yes” were placed in one group
(those with children), while those who responded “no” were placed in a second group
(those without children).
Lesbigay mother’s age at coming out. Four questions were asked only of lesbigay
mothers with regard to coming out in order to determine the age at which certain
coming out markers took place for the woman: “When did you first feel an attraction to
a member of the same sex?” “What age did you first come out to yourself?”, “When did
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
you have your first experience/relationship with a person of the same sex?” and
“When did you first tell another person about your GLB orientation.” (See Appendix
B.)
Family Relationship Quality
Dyadic family relationship quality was composed of three constructs: affectual
solidarity, conflict and emotional support. Mother and child’s perceptions of their
relationships with each other were examined separately and as composite scores
summing both mother’s and child’s scores. The method of summed scores is based on
the Bengtson’s research group’s examination of intergenerational relationships (i.e.,
Bengtson & Roberts, 1991). Separate scores for child and mother were also examined
where appropriate. (See Appendix C.)
Affectual solidarity. A five-item solidarity scale for mother and child was used
(Bengtson & Roberts, 1991). Each mother or adult child responded to five items that
assess his or her perceptions of mother-child affectual solidarity. Response categories
range from 1 = “not at all close/well” to 6 = “extremely close/well” on a six-point scale.
The responses were summed, producing a rating of perceived affectual solidarity;
higher scores indicated stronger feelings of affectual solidarity. The potential range for
this dyadic measure is 6 to 36; higher scores indicated higher levels of solidarity.
Individual scores range from 3 to 18.
Conflict. A three-item measure developed for the LSOG study to determine level of
conflict between mother and adult child, was given to both mother and child. This
measure asks mothers and children to rate the amount of conflict, criticism and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
arguments in their relationship with each other on a six-point scale ranging from 1 =
“not at all” to 6 = “a great deal.” It was adapted from several sources including
Wortman and Dunkel-Shetter (1987); House, Kahn, McLeod and Wiliams (1985), Kahn
and Antonucci (1981) and Moos (1974). The responses from all conflict questions were
summed to produce a rating of perceived conflict. The potential range for this dyadic
measure is 6 to 36; higher scores indicated higher levels of conflict. Individual scores
range from 3 to 18.
Emotional Support. Emotional support offered to and received from mother and adult
child was assessed using six items from an emotional support measure developed for
the LSOG sample. The source of the items was Pearlin (1983). Each mother and adult
child was asked to rate whether they received or provided “emotional support”,
“information and advice”, and “discuss important life decisions.” The 1 = “yes” or 0 =
“no” responses to these items were summed to represent the amount of emotional
support between the dyad. The potential range for this dyadic measure was 0 to 6;
higher scores indicated higher levels of support. Individual scores range from 0 to 3.
Psychological Well-Being
Psychological well-being was composed of four constructs: depressive
symptoms, negative affect, positive affect and self-esteem. Subjective well-being is
often measured with a variety of indicators and there is no one standard indicator
(Headey & Wearing, 1991). Depression (depressive symptoms), self-esteem, positive
moods (positive affect) and negative moods (negative affect) have all been used as
measures of psychological well-being in studies examining the association between
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
family relationships and individual well-being (Lewinsohn, Redner, & Seeley, 1991;
Amato & Booth, 1997). Both mother and child were given these measures separately.
(See Appendix D.)
Depressive symptoms. The Center for Epidemiological Studies - Depression Scale
(CES-D; Radloff, 1977) is a self-report, 20-item measure of depressive symptoms.
Respondents were asked to rate the frequency with which they had experienced
depressive symptoms in the past week on a four-point scale, ranging from 0 = “rarely or
none of the time” to 3 = “most or all of the time.” Four positively worded items were
reversed scored. The responses were summed to produce a rating of depressive
symptoms. The total score may range from 0 to 60, with higher scores indicated higher
levels of depressive symptoms.
Negative and positive affect. Positive affect and negative affect were measured by the
Bradburn (1969) Affect Balance Scale. There were 10 items total, five for positive
affect and five for negative affect. Respondents are asked, “during the past few weeks,
did you feel...” and then presented with different affectual experiences, such as proud,
pleased or excited for positive affect and bored, lonely, or upset for negative affect. The
scales were based on 1 = “yes” or 0 = “no” response to the questions. The positive
affect scale has five items and can range from 0 to 5; and the negative affect has five
items and can range from 0 to 5. Higher scores indicate increased levels of positive and
negative affect, respectively.
Self-esteem. The Rosenberg (1965) Self-esteem measure consists of 10 items on a four-
point scale, with responses ranging from 1 = “strongly agree” to 4 = “strongly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
disagree.” This scale was inadvertently modified to nine items for the NTF sample.
While the LSOG sample received the original measure, nine items were used for
comparison purposes. Positively worded items were reverse scored, so that higher
scores indicated higher levels of self-esteem. The potential range of the scale was from
10 to 40.
Data Analysis and Statistics
Participants who failed to answer more than 20% of questions on any scale were
omitted from analyses for that variable, while participants missing 20% or fewer items
were included with the average item score for the individual replacing the missing
items.
Analyses to determine demographic covariates used in the first two specific
aims were conducted in two steps by examining t-tests and correlation matrices. First,
the two groups of interest (lesbigay mothers and their children versus heterosexual
mothers and their children) were compared to determine differences in demographic
composition. Those demographics that distinguished the two groups were then
examined in a correlation matrix with the well-being variables. In order for a
demographic to be included as a covariate, it needed to be statistically significantly
different between the two groups and correlate with at least one well-being variable.
The decision to use these demographics was made because we were most interested in
differences among the two family types on measures of well-being.
Originally, we had hoped to conduct a sample matching procedure matching
heterosexual mothers with lesbian mothers and their children as has been done in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
previous studies (Green et al., 1986; Golombok et al., 1983), with demographic
variables that have previously been associated with family relationships and
psychological well-being (Acock & Demo, 1994). However, our samples proved too
small to complete the procedure. We were still able to examine the following variables
for differences between the two family types: (1) child’s gender (male, female); (2)
mother’s and child’s ages (years); (3) mother’s current partner status (married or with
partner as though married, single or dating); (4) mother’s and child’s cohabitation status
(lived together, lived separately); (5) mother’s and child’s socio-economic status (high,
medium, low; based on reported household income); (6) mother’s and child’s
educational attainment; (7) child’s current partner status (married or with partner as
though married, single or dating); and (8) child’s parental status (had a child or children,
had no child or children).
The first specific aim was to examine mother and child’s relationship quality
with each other across lesbigay headed and heterosexual headed households
(determined by mother’s sexual orientation), and across single and dual parent
households (determined by mother’s current partner status). (See Table 2.) Family
relationship quality was assessed by three measures (dyadic affectual solidarity, dyadic
conflict, and dyadic emotional support) summing reports from both sides of the mother-
child dyad using Bengtson et al.’s approach to dyadic family relationship quality. Six
hierarchical multiple regression analyses were used to address the hypotheses I and II.
Results with Bonferroni correction (2 independent variables (maternal sexual
orientation, maternal relationship status) x 3 relationship quality measures (dyadic
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
affectual solidarity, dyadic conflict, dyadic emotional support)) are presented.
However, because the hypotheses focus on the absence of group differences it would be
unfairly advantageous to use an overly conservative Bonferroni adjustment. For this
reason unadjusted results are presented along with Bonferroni adjusted results.
Table 2
Specific Aims and Hypotheses.
•
Specific Aim Hypothesis Independent
Variable(s)
Dependent
Variable(s)
First I Mother’s sexual orientation
(demographics covariate)
Perceived
relationship quality
II Mother’s partner status
(demographics covariate)
Perceived
relationship quality
Second III Perceived relationship quality
(demographics covariate)
Mother and child’s
well-being
IV Mother’s sexual orientation
and mother’s partner status
(demographics and
relationship quality covariate)
Mother and child’s
well-being
Third Exploratory Age, gender (with mother’s
sexual orientation for
interaction)
Child’s well-being
and perceived
relationship quality
Hypothesis I stated that family relationship quality would not differ across
lesbigay headed compared to heterosexual headed households after controlling for
demographics. This hypothesis was tested using three hierarchical regression analyses,
with demographic covariates entered on the first step, and family type (mother’s sexual
orientation, lesbigay versus heterosexual headed households) were added to
demographic variables and entered on the second step. Hypothesis II stated that family
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
relationship quality would not differ across single versus dual parent households.
This hypothesis was tested using three hierarchical regression analyses, with
demographic covariates entered on the first step and family type (mother’s relationship
status, single versus dual parent (married/partnered)) added to demographic covariates
entered on the second step.
Follow-up analyses of hypothesis I and II included re-running analyses to
separately examine mother’s and child’s perspective on family relationship quality.
Additionally, hypothesis I included a chi-square examining lesbigay mothers and their
children’s difference in perspective on their dyadic relationship compared to
heterosexual mothers and their children’s difference in perspective on their dyadic
relationship.
The second specific aim was to examine the contribution of family relationship
quality and family structure (both maternal sexual orientation and maternal relationship
status) to the prediction of both mother’s and child’s well-being. Family relationship
quality was assessed by three measures (dyadic affectual solidarity, dyadic conflict, and
dyadic emotional support) summing reports from both sides of the mother-child dyad
using Bengtson et al.’s approach to dyadic family relationship quality. Well-being was
assessed by four measures (depressive symptoms, negative affect, positive affect, self
esteem); thus, for each participant, there were three measures of dyadic family
relationship quality and four measures of well-being. Eight hierarchical multiple
regression analyses were used to explore hypotheses III and IV. Both hypotheses were
examined as separate steps in the same analyses; hypothesis III focused on the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
contribution of family relationship quality to well-being and hypothesis IV centered
on the contribution of family type to well-being after controlling for demographics and
family relationship quality. In step one of the hierarchical regression, demographics
were entered into the model. In step two, demographics were entered as well as family
relationship quality (dyadic affectual solidarity, dyadic conflict and dyadic emotional
support). In step three, demographics, family relationship quality and family structure
(mother’s sexual orientation and mother’s relationship status) were entered in the
models.
Follow-up analyses of hypothesis III and IV included re-running analyses
without controlling for demographics. For these eight hierarchical regressions, family
relationship quality was entered in step one, and family relationship quality, as well as
family structure, was entered in step two.
The goal of the third specific aim was to determine whether there were age and
gender differences for children’s well-being and perception of their relationships with
their mothers. Specifically, we were interested in whether age and gender interact with
maternal sexual orientation to predict child well-being and perceived relationship
quality. Exploratory analyses were conducted using children’s age and gender as
predictors of individual well-being and relationship quality, measured by child’s
affectual solidarity, child’s emotional support and child’s conflict. This aim was
examined using a series of hierarchical regression models. In each model, child’s age,
child’s gender and maternal sexual orientation were entered on step one. In addition to
these variables, interaction terms representing the product of child age and maternal
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
sexual orientation, and child gender and maternal sexual orientation were entered in
step two. The interaction terms used variables that were centered as suggested by
Pedhazur (1997). The contribution of the interaction terms was assessed using the F-
test of change in variance accounted for between step one and step two. This was done
in order to determine whether better well-being outcomes were observed for older
versus younger children or boys versus girls raised by mothers of one sexual orientation
or another. This is analogous to a treatment aptitude interaction in randomized clinical
trials. If we had been able to randomly assign children to conditions, this term would be
telling us whether lesbigay mothers make girls healthier than boys, for example,
“happier” girls.
Follow-up analyses for the third specific aim included running two additional
series of hierarchical regression models, one for mother’s perception of family
relationship quality (mother’s affectual solidarity, conflict and emotional support) and
dyad’s perceptions (dyadic affectual solidarity, conflict and emotional support). The
only difference in the regression models from the child’s analyses, was that for mother’s
family relationship quality, child’s age and gender were supplemented with mother’s
responses when data were missing from child.
Results
Three types of findings are presented: descriptive findings characterizing the
sample, preliminary analyses of the data, and results of tests of hypotheses. T-tests,
Pearson product-moment correlations, and hierarchical multiple regressions were used
to test the hypotheses. For hypothesis III and for the exploratory analyses
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
corresponding to specific aim three, experimentwise type I error was controlled using
a Bonferroni adjustment procedure (Rosenthal & Rosnow, 1991). An alpha level of .05
was used for all statistical tests unless otherwise noted. Results are presented with and
without a Bonferroni adjustment procedure. For this adjustment, tests were grouped
into families based on logical considerations of the variables involved. Since
Bonferroni adjustment is a conservative procedure, where there was ambiguity about
how to allocate tests to families for the purposes of adjustment, small families of tests
were chosen rather than larger ones. In effect, this offsets the conservatism of the
Bonferroni procedure. Furthermore, hypotheses I, II, and IV were written based on the
expectation of no differences (support for null hypothesis). It would be unfairly
advantageous to these hypotheses to use an overly conservative Bonferroni adjustment.
For this reason unadjusted results are presented along with Bonferroni adjusted results.
Descriptive Findings
Characteristics of study participants. The total number of mothers was 134; the total
number of children was 123. Tables 3 and 4 present demographic information for the
full sample, lesbigay mothers and their children, and heterosexual mothers and their
children. In these tables, demographic information about children were obtained from
mothers when children did not respond for a total of 143 children. Table 5 and 6
present demographic information for single mothers and their children and
married/partnered mothers and their children (not distinguishing between lesbigay
headed and heterosexual headed households). The mothers in this study were an
average age of 51.1 years (S.D. = 6.4). They were mostly white and well educated,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
95% Caucasian, an average of some college education, and an average income in the
$30,000-69,000 household range. The teenage/adult children in this study were
between the ages of 17 and 45 with an average of 26.77 years (ST). = 7.49). They, like
their mothers, were also mostly white and well educated, with 97% Caucasian and an
average of some college education, with many still completing their education at the
time of the study.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 3
Means and Standard Deviations o f Demographics Major Study Variables fo r Full Sample, Lesbigay Parented Families and Ffeterosexual Parented Families.
Variable
Full
N
NTF LSOG Full
M
NTF LSOG Full
S.D.
NTF LSOG Range
M other’s Demographics
Age 134 68 66 51.14 52.72 49.52 6.41 4.87 7.37 37 to 64
Length o f current relationship 115 60 55 10.70 9.52 11.99 8.09 9.52 9.68 .5 to 44
Education 133 67 66 6.04 7.57 4.48 1.97 7.57 ,1.29 2 to 9
Income 124 66 58 2.22 2.44 1.97 0.73 2.44 0.75 1 to 4
Child’s Demographics
Age 142 76 66 26.61 25.79 27.56 7.16 5.00 8.99 17 to 45
Length o f current relationship 48 17 31 6.59 4.53 7.73 6.87 2.55 8.17 15 to 24
Age at mother’s divorce 133 71 62 8.80 7.33 10.48 7.18 5.92 8.12 0 to 36
Education 130 65 65 5.48 6.40 4.57 1.74 1.66 1.27 2 to 9
Income 111 49 69 1.69 1.80 1.61 0.75 0.82 0.69 1 to 3
44
Table 4
Counts and Percentages o f Demographics Major Study Variables fo r Full Sample, Lesbigay Parented Families, and
Heterosexual Parented Families.
.Variable Full NTF LSOG
n(% ) n(% )
fl(% )
N = 134 n for M = 68 n for M = 66
n for C = 57 n for C = 66
M other
Race
Caucasian 127 (94.78) 62 (91.18) 65 (98.48)
African-American 1 (.75) 1 (1.47) 0 (.00)
Hispanic/Latino 2(1.50) 2 (2.94) 0 (,0‘ 0)
Multi-racial 4(3.00) 3 (4.41) 1 (1.52)
Partner Status
Single 34 (25.56) 12(17.65) 22 (33.85)
Married/P artnered 99 (74.44) 56 (82.35) 43 (66.15)
Missing 1 (.75) - 1 (1.52)
Child
Gender
Male 34 (27.64) 14 (24.56) 20 (30.30)
Female 89 (72.36) 43 (75.44) 46 (69.70)
Racial identity
Caucasian 116(94.31) 51 (89.47) 65 (98.48)
African-American 0 (.00) 0 (.00) 0 (.00)
Hispanic/Latino 2(1.63 2(3.51) 0 (.00)
Multi-racial 5 (4.07) 4 (7.02) 1 (1.52)
Partner status
Single 75 (60.98) 41 (71.93) 34 (51.52)
Marri ed/Partnered 48 (39.02) 16(28.07) 32 (48.48)
Parental Status
No child(ren) 82(67.21) 47 (82.46) 35 (53.85)
Has child(ren) 40 (32.79) 10(17.54) 30 (46.15)
Missing 1 (.75) - 1 (L52)
Dyad
Mother/child cohabitation status
No 106 (76^81) 42 (73.68) 48 (76.19)
Yes 32 (23.19) 15 (26.32) 15 (23.81)
Note. NTF = Lesbigay parented sample LSOG = Heterosexual parented sample M = Mothers C = Children
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 5
Means and Standard Deviations o f Demographics Major Study Variables fo r Single Parented and Married/Partnered Parented Families.
Variable
Single
N
Married/
Partnered
Single
M
Married/
Partnered
Single
S.D.
Married/
Partnered
Range
M other’s Demographics
Age 34 99 52.65 50.69 7.49 5.95 37- 64
Education 34 98 5.76 6.16 1.79 2.01 2 - 9
Income 31 92 1.68 2.40 .70 .65 1 -4
Child’s Demographics
Age 34 98 28.53 25.98 8.47 6.68 17-45
Length o f current relationship 16 27 9.25 5.41 8.44 6.01 .50 - 24
Age at mother’s divorce 30 93 10.82 8.30 7.64 6.92 0-36
Education 31 89 5.32 5.42 1.56 1.74 2 - 9
Income 30 72 1.70 1.69 .79 .74 1 -3
C /1
46
Table 6
Counts and Percentages o f Demographics Major
Children o f Married/Partnered Mothers (n = 99).
Study Variables fo r Children from Single Mothers (n = 29) and
Variable Single Married/Partnered
n (%) n(%)
Child
Gender
Male 9(31.03) 23 (27.38)
Female 20 (68.97) 61 (72.62)
Partner status
Single 15 (51.72) 54 (64.29)
Married/Partnered 14(48.28) 30(35.71)
Parental status
No child(ren) 11 (39.29) 63 (75.00)
Has child(ren) 17 (60.71) 21 (25.00)
Missing 1 (3.45)
Dyad
Mother/child cohabitation status
No 22 (75.86) 60 (74.07)
Yes 7(24.14) 21 (25.93)
Preliminary Analyses
Distributional assumptions. Prior to substantive analyses, score distributions of the
study variables were examined. Statistical analyses used in the study rely on normally
distributed variables. The distribution of the family relationship quality and the well
being measures were skewed for all but one of the measures (child’s negative affect)
and transformations were completed on each measure based on the recommendations of
Tabachnick and Fidel (1994). See Table 7. For each measure, both mother’s and
child’s perspective had the same transformation. For family relationship quality
variables, reflective inverse transformations were used to reduce skewness effects for
mother and child’s affectual solidarity measures. Similarly reflective inverse
transformations were used to reduce skewness effects for mother and child’s emotional
support. Inverse transformations were used for mother and child’s conflict measures.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
For individual well-being, log transformations were used to reduce skewness
effects for depressive symptoms. Square root transformations were used to reduce
skewness for mother’s negative affect, while reflective inverse transformations were
used for mother’s and child’s positive affect. Reflective square root transformations
were used to reduce skewness effects for mother and child’s self-esteem. Note that all
analyses were run with both transformed and untransformed data. Transformations did
not alter the fundamental findings of the study. In order to meet distributional
assumptions of the tests, however, the results of transformed data are presented.
Table 7
Transformations o f Study Variables
M other’s Variable Transform ation C hild’s Variable Transform ation
Family Relationship Fam ily Relationship
Affectual solidarity Reflective inverse A ffectual solidarity Reflective inverse
Conflict Inverse Conflict Inverse
Emotional Support Reflective inverse Em otional Support Reflective inverse
W ell-being W ell-being
Depressive sym ptom s Log D epressive symptoms Log
N egative affect Square root Negative affect N o transfonnation
Positive affect Reflective inverse Positive affect Reflective inverse
Self-esteem Reflective square root Self-esteem Reflective square root
Verifying convergent validity of measures. Prior to addressing the specific aims, we
used correlation matrices to examine the family relationship quality and well-being
measures in the full sample in order to determine that they were appropriate to consider
as constructs. Within each mother-child dyad unit, there were three measures of dyadic
family relationship quality, four measures of mother’s well-being and four measures of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
child’s well-being. Also examined were mother’s and child’s individual reports of
family relationship quality. Appendix E provides correlation matrices of transformed
family relationship quality and well-being variables. Note that family relationship
quality related to more child well-being variables (eight out of a possible 12
correlations) as compared to maternal well-being variables (two out of a possible 12
correlations were significant). A similar pattern exists for both lesbigay headed
households and heterosexual headed households. Appendix F provides the same
matrices for lesbigay mothers and their children while Appendix G provides the same
matrices for heterosexual mothers and their children. Appendix H and I provide the
matrices for single mothers and their children and married/partnered mothers and their
children, respectively.
For family relationship quality, Table 8 presents the correlation matrix of
transformed dyadic family relationship quality variables for the full sample. Dyadic
affectual solidarity and conflict were negatively correlated at the p = .05 level (r = -.46,
P < .0001) and dyadic affectual solidarity and dyadic emotional support were
significantly correlated at the p = .05 level (r = .45, p < .0001). Dyadic conflict and
emotional support were nonsignificantly related (r = -.09, ms,).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
Table 8
Correlation Matrix o f Transformed Dyadic Relationship Quality for the Full Sample
_ _ 2 ' 3
1. Dyadic affectual solidarity — _o.46**** 0.45****
2. Dyadic conflict — -0.09
3. Dyadic em otional support
_ _ _ _ _ _ _ _ __ _ _ _
Table 9 presents the correlation matrix of transformed family relationship
quality variables for the full sample of mothers. Mother’s report of affectual solidarity
and conflict were negatively correlated at the p = .05 level (r = -.35, p = < .0001) and
mother’s report of affectual solidarity and emotional support were significantly
correlated at the p = .05 level (r = .31, p = .0005). Mother’s report of conflict and
emotional support were nonsignificantly related (r = .02, n.s.).
Table 9 ~
Correlation Matrix o f Transformed Mother’ s Perceived Relationship Quality for Full
Sample
_ . . . _ . _ _ _ . . .
1. Affectual solidarity with child — . 3 ]***
2. Conflict with child — r -.02
3. Emotional support with child
_____ _
****g <0001
Table 10 presents the correlation matrix of transformed child’s report of family
relationship quality. Child’s report of affectual solidarity and conflict with mother were
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
negatively correlated at the 2 = -05 (r = -.35, p < .0001) and child’s report of affectual
solidarity and emotional support were positively correlated at the p < .05 level (r = .31,
P = .0005). Child’s report of conflict and emotional support were nonsignificantly
related (r = -.08, ms.). Thus, the correlation matrices of mother’s and child’s family
relationship quality provide some evidence of convergent validity.
Table 10
Correlation Matrix o f Transformed Child’ s Perceived Relationship Quality for Full
Sample
1 2 3
1. Affectual solidarity with m other -0.35****
Q 2 1 ***
2. Conflict with m other - -0.08
3. Emotional support with m other -
***g <.001
* * * * 2 <.0001
For individual well-being measures, Table 11 shows that all four transformed
mother’s well-being measures (depressive symptoms, negative affect, positive affect
and self-esteem) correlated significantly with each other at least at the p = .001 level.
There were positive correlations between negatively focused well-being measures, such
as mother’s depressive symptoms and negative affect (r = .57, p < .0001) and positive
correlations between positively focused well-being measures, such as mother’s positive
affect and self-esteem (r = .51, p <.0001). There were significant negative correlations
between negatively focused measures and positively focused measures. All correlations
were in the expected direction. While it was not expected that positive affect and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
negative affect, from the Bradburn Affect Balance (BAB) scale, would correlate with
each other, in fact they did.
Table 1 1
Correlation Matrix o f Transformed Mother’ s Well-Being Variables o f Full Sample
1 2 3 4
1. M other depressive sym ptom s — .57**** - 36**** - 56****
2. Mother negative affect
_3 j *** - 4 4 ****
3. M other positive affect --
4. M other self-esteem -
* * * £ < .001
* * * * E < 0001
Table 12, the correlation matrix of the child’s well-being measures, shows a
similar pattern to that of mothers in terms of direction of significant correlations. The
only difference was that unlike the mother’s scores on the BAB scale, child’s negative
and positive affect did not significantly correlate with each .other. There were positive
correlations for depressive symptoms and negative affect (r = .59, p < .0001) and
positive affect and self-esteem (r = .35, p < .0001), while again, opposing well-being
measures were negatively con-elated, depressive symptoms with positive affect and self-
esteem (r = -.38, p < .0001 and r = .38, p < .0001, respectively). Child’s negative affect
and positive affect were not significantly related (r = -.1 5 ,p = .ll) which as discussed
in the mother’s section, was as expected. Overall, these correlation matrices suggest
convergent validity for mother’s and child’s well-being. The measures behave the same
in different groups suggesting consistency for the construct.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
Table 12
Correlation Matrix o f Transformed Child’ s Well-Being Variables o f Full Sample
M easure 1 2 3 4
C hild’s depressive sym ptom s -- 5 9**** _ g g * * * * .5 4 ****
C hild’s negative affect — -.15 . 4 4 * * * *
C hild’s positive affect — .35****
C hild’s self-esteem
**** g < .0001
There appears to be convergent validity for both mothers’ and children’s family
relationship quality and well-being measures. In summary, the correlations among
family relationship quality variables and among well-being measures show that the
variables within each group are related to each other in the predicted fashion and
suggest that they are appropriate to consider as constructs of family relationship quality
variables and individual well-being variables.
Inter-judge agreement on family relationship quality measures. Mother’s and child’s
reports on measures of family relationship quality (affectual solidarity, conflict, and
emotional support), were significantly correlated with each other at the p < .0001 level
(rs = .37, .44, .37, respectively) (see Appendix E). In other words, mother and child’s
perception of the relationship with one another is similar in perspective and provides
modest evidence of inter-judge reliability for the same relationship described by two
different participants.
It is worth noting that, from the correlation matrices of family relationship
quality and family member’s well-being in Appendix E, we can see that there are eight
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
instances where both mother’s and child’s perceived relationship quality are
correlated with a well-being measure. Specifically, the instances of these cases are for
all of the children’s well-being measures and for mother’s negative affect and mother’s
self-esteem score. This was of concern, since mother’s and child’s perspectives of the
relationship correlated with each other and if all the separate family relationship quality
variables were entered into the equation, multicollinearity problems would have
resulted. The results confirmed the logic of using the Bengtson’s research group’s
dyadic family relationship quality variables, or composite scores for both members of
the dyad.
Family Structural Comparisons
The present study focused on two types of nontraditional families: families with
differences in mother’s sexual orientation (lesbigay headed households versus
heterosexual headed households) and mother’s partner or relationship status (single
versus married/partnered) in two samples, for a total of four family types: lesbigay
headed, single parent households; lesbigay headed, partnered parent households;
heterosexual headed, single parent households; and heterosexual headed,
married/partnered parent households. Data screening revealed unequal cell sizes among
the four family types. Due to some small cell sizes, we did not have sufficient statistical
power to examine the possible interaction effects between mother’s sexual orientation
and mother’s relationship status.
We did not expect any interactions, and examination of individual cell means
revealed few instances of possible mean differences among the four family types (See
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
Table 13). We ran t-tests (comparison of medians with transformed data) on what
appeared to be a significant interaction, understanding that we could potentially inflate
the risk of Type I error in the results. For family relationship quality, dyadic emotional
support appeared to be equivalent for families with single lesbigays and their children
and partnered lesbigay mothers and their children (t(46) = -.03, p = .97). In contrast, in
heterosexual headed families, single heterosexual mothers and their children appeared
to reported less dyadic emotional support than partnered heterosexual mothers and their
children but the t-test was nonsignificant (t(63) = -1.03, p = .31). For dyadic conflict, it
appears that conflict levels between lesbigay mothers and children are equivalent,
whether the mother is single or partnered (t(45) = -.31, p = .76); whereas single
heterosexual mothers and their children appeared to report more conflict than married
heterosexual mothers and their children, but the t-test was nonsignificant (t(62) = .74, p
= .45). Additionally, it appeared that lesbigay mother perception of affectual solidarity
with child was equivalent whether they were single or partnered (t(66) = .44, p = .66);
as compared to heterosexual mother perception, where single mothers appeared to
report less affectual solidarity with their children compared to a married mothers who
may be reporting higher affectual solidarity with children, again, however the t-test was
nonsignificant (t(63) = -1.39, p = .17).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 13
Means and Standard Deviations o f Major Study Variables fo r Lesbigay H eaded (NTF) and Heterosexual Headed (LSOG) Single Parented Families and
M arried/Partnered Parented Households.
Variable N ■ M SD
Single Partnered Single Partnered Single Partnered
NTF LSOG NTF LSOG NTF LSOG NTF LSOG NTF LSOG NTF LSOG
Dyadic Family Relationship Quality
Affectual solidarity 7 2 2 41 42 48.00 47.05 50.98 49.71 6.76 9.77 6.55 8.97
Conflict 7 2 2 40 42 1 2 .0 0 13.41 12.40 12.33 4.76 5.57 3.77 5.32
Emotional support 7 2 2 41 43 9.86 7.41 6.55 8.97 1 .6 8 3.55 2.6 8 3.77
Mother’s Family Relationship Quality
Affectual solidarity w/child 1 2 2 2 56 43 25.58 24.77 25.73 26.58 3.03 4.24 3.14 3.70
Conflict w/child 1 2 2 2 56 43 5.67 6.18 6.05 5.72 2.02 2.74 2.42 2.75
Emotional support w/child 1 2 2 2 56 43 5.25 3.50 4.86 4.40 0.97 2.15 1.61 1.98
Child’s Family Relationship Quality
Affectual solidarity w/mother 7 2 2 41 42 23.43 22.27 25.15 23.14 4.93 6.56 4.51 6.08
Conflict w/mother 7 2 2 40 42 6.00 7.23 6.48 6.60 2.89 3.48 2.30 3.35
Emotional support w/mother 7 2 2 41 43 4.57 3.91 4.54 3.77 1.51 2.09 1.80 2.37
Mother’s Well-being
Depressive symptoms 1 2 2 2 55 43 13.42 18.11 8.61 9.47 9.92 13.19 7.60 9.57
Negative affect 1 2 2 2 56 43 1.67 2.64 1.58 1.70 1.56 1.47 1.37 1.49
Positive affect 1 2 2 2 56 43 3.92 3.05 4.33 3.79 1 .0 0 1.43 0.92 1.28
Self-esteem 1 2 2 2 56 42 34.17 30.85 34.82 33.92 3.85 5.09 4.94 3.69
Child’s Well-being
Depressive symptoms 7 2 2 41 43 12.14 15.15 1 2 .1 0 13.05 5.90 1 0 .2 0 9.38 11.31
Negative affect 7 2 2 41 43 2.71 2.59 2.39 2.56 1.80 1.30 1.59 1.67
Positive affect 7 2 2 41 43 4.07 3.85 3.90 4.05 1.30 1.37 1 .1 1 1.07
Self-esteem 7 2 2 40 42 34.54 33.74 32.91 33.98 4.53 4.90 5.16 4.62
Note. Due to missing data n ’s are provided for each variable.
CO
56
For well-being variables, it appeared that lesbigay mothers’ self-esteem scores
were equivalent whether or not they were single or partnered, while heterosexual
mothers were more likely to have higher self-esteem if they were married than if they
were single. Additionally, it appeared that depressive symptoms of children of lesbigay
mothers were equivalent whether or not their mothers were in relationships, while
children of married heterosexual mothers were more likely to have lower depressive
affect than children of single heterosexual mothers. Again, we were able to test main
effects for mother’s sexual orientation and mother’s relationship status, but not the
interaction between mother’s sexual orientation and mother’s relationship status
because of the small samples.
Comparison of lesbigay and heterosexual mothers and their children. Lesbigay mothers
and their children were compared to heterosexual mothers and their children for
differences in demographics and major study variables. (See Tables 3 and 4.)
Comparing the two family types on demographics, lesbigay mothers were, on average,
older (M = 52.7 years (S.D. - 4.87) versus M = 49.5 years (S.D. = 7.37)) than the
heterosexual mothers, t (112) = 2.96, p = .004 using the Satterthwaite t-test for unequal
variances. Lesbigay mothers were more likely to be married/partnered ( 82.35%) rather
than single (17.65%) compared to heterosexual mothers (6.15%, 33.85%, respectively; t
(131) = 2.16, p = .03). This finding suggests that that with a larger sample, we could
have tested for interaction effects. Lesbigay mothers also had significantly higher
levels of education and higher levels of income than heterosexual mothers in this
sample.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
Children of lesbigay mothers were on average, younger at the time of parental
divorce (or end of their mother’s significant heterosexual relationship) (M = 7.40 years,
S.D. = 6.48) compared to children of heterosexual mothers (M = 11.89 years, S.D. =
13.76), t (91.1) = -2.31, p = .02 using the Satterthwaite t-test for unequal variances.
Overall, a significantly larger proportion of children of lesbigay mothers were single at
the time of the study (71.93%) compared to children of heterosexual mothers (51.52%)
although there were no significant differences on child’s age at the time of the study.
They were also less likely to have children (82.46%) than children of heterosexual
mothers (53.85%). Finally, children with lesbigay mothers had attained a higher
educational level than children of comparable age with heterosexual mothers. There
were no significant differences between the two family types on child’s age and gender,
and mother/child cohabitation status.
With regard to the major study variables, Table 14 presents the means and
standard deviations for mother and child’s family relationship quality and well-being in
lesbigay and heterosexual headed households. Group differences in participant’s self-
report scores were analyzed using t-tests. Results are presented based on pooled
variance estimations unless otherwise noted. When variance differed significantly
across the two groups, Satterthwaite t-tests for unequal variances were used.
On measures of family relationship quality, three t tests were used to compare
lesbigay and heterosexual headed families, and six follow-up t-tests were used to
separately examine mother’s and child’s perspective (three scales rated by each member
of the mother-child dyad yields six dependent variables). See Table 15. None of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
dyadic family relationship quality variables were significantly different between
lesbigay headed households and heterosexual headed households. Examining mother
and child’s perceptions separately revealed two of the tests indicated significant
differences in family relationship quality across the two family types before Bonferroni
adjustment, but none after Bonferroni adjustment. Before Bonferroni correction,
lesbigay mothers (M = 25.71, S.D. = 3.10) reported lower (t (119) = -2.13, p = .035)
levels of affectual solidarity than heterosexual mothers (M = 25.92, S.D. = 3.94).
Lesbigay mothers (M = 4.93, S.D. = 1.52) reported higher (t (132) = 2.33, p = .02)
levels of emotional support than heterosexual mothers (M = 4.12, S.D. = 2.06) using
Satterthwaite t-tests for unequal variances.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 14
M eans a n d Standard D eviations o f Transform ed M ajor Study Variables fo r F ull Sample, Lesbigay Parented Families and Fleterosexual P arented
Families.
V ariable n M S.D.
Possible
Full N TF LSOG Full N TF LSOG Full N TF LSOG Range
Dyadic Fam ily Relationship Quality
Affectual solidarity 113 48 65 49.47 50.54 48.68 8.24 6.59 9.24 1 0 to 60
Conflict 112 47 65 12.59 12.34 12.77 4.79 3.87 5.37 1 0 to 60
Em otional support 114 48 66 8.61 9.52 7.94 3.32 2.55 3.67 O to 12
M other’s Fam ily Relationship Quality
A ffectual solidarity w/child 134 68 66 25.81 25.71 25.92 3.53 3.10 3.94 1 to 30
C onflict w/child 134 68 66 5.95 5.99 5.91 2.53 2.35 2.72 1 to 30
Em otional support w/child 134 68 66 4.53 4.93 4.12 1.85 1.52 2.06 0 to 6
C hild’s Fam ily Relationship Quality
A ffectual solidarity w /m other 122 57 66 23.76 24.89 22.77 5.54 4.49 6.19 1 to 30
Conflict w /m other r n 56 65 6.63 6.38 6.85 3.02 2.56 3.36 , 1 to 30
Em otional support w /m other 123 57 66 4.T5 4.54 3.82 2.05 1.73 2.25 0 to 6
M other’s W ell-being
D epressive sym ptom s 133 67 66 10.87 9.47 12.30 10.05 8.19 11.52 0 to 60
N egative affect 133 67 66 1.79 1.60 1.99 1.47 1.39 1.52 0 to 40
Positive affect 133 67 66 3.91 4.25 3.56 1.22 0.94 1.37 1 to 5
Self-esteem 133 68 65 33.78 34.71 32.80 4.74 4.75 4.57 1 to 5
C hild’s W ell-being
D epressive symptoms 122 56 66 13.23 12.19 14.11 10.13 8.67 11.20 0 to 60
N egative affect 123 57 66 2.51 2.40 2.61 1.54 1.52 1.56 0 to 40
Positive affect 123 57 66 3.95 3.94 3.95 1.14 1.12 1.17 1 to 5
Self-esteem 1 2 1 56 65 33.62 33.32 33.87 4.72 4.82 4.65 1 to 5
Note. Due to missing data n’s are provided for each variable ^
vo
60
Table 15
T-tests for Groups on Transformed Family Relationship Quality for Mother’ s Sexual
Orientation
Scale and Group df
t
p-value
Affectual solidarity
Dyadic
111 .63 .53
Conflict 110 -.26 .80
Emotional support 112 1.09 .28
Affectual solidarity w/child
Mothers
119 -2.13U .04
Conflict w/child 132 .57 .57
Emotional support w/child 132 2.3 3u .02
Affectual solidarity w/mother
Children
120 1.76 .08
Conflict w/mother 119 -,05u .96
Emotional support w/child 121 1.07 .29
a T-tests em ployed pooled variance estim ation unless otherw ise noted. T- tests m arked with a u used
Satterthw aite’s m ethod for com pensating for unequal variances
u Used Satterthwaite t-tests for unequal variances
On measures of maternal well-being, two of the four t tests comparing the two
groups of mothers were significant after Bonferroni correction (see Table 16). Lesbigay
mothers (M = 4.25, S.D. = .94) reported higher positive affect (t (131) = 2.90, p = .004)
than did heterosexual mothers (M = 3.56, S.D. = 1.37). Similarly, lesbigay mothers (M
= 34.71, S.D. = 4.75) reported higher self-esteem for (t (131) = 2.61, p = .01) than did
heterosexual mothers (M = 32.80, S.D. = 4.57). No between group differences in
reported well-being of adult children were observed based on mothers’ sexual
orientation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
Table 16
T-tests for Groups on Well-Being o f Lesbigay Mothers and Their Children and
Heterosexual Mothers and Their Children
Measure and Group d f
f
p-value
Mothers
Depressive symptoms 120 -.45“ .65
Negative affect 131 -1.43 .15
Positive affect 131 2.90 .004
Self-esteem 131 2.61 .01
Children
Depressive symptoms 120 -.37 •71
Negative affect 121 -.73 .47
Positive affect 121 -.45 .65
Self-esteem 119 -.61 .54
a T-tests em ployed pooled variance estim ation unless otherw ise noted. T- tests m arked with a u used
Satterthw aite’s m ethod for com pensating for unequal variances
Used Satterthwaite t-tests for unequal variances
Comparison of single and married/partnered mothers and their children. Single mothers
and their children were compared to married or partnered mothers and their children for
mean differences of major study variables. Table 17 presents the means and standard
deviations for mother and child’s family relationship quality and well-being in single
and married/partnered mothers and their children (within the NTF and LSOG families).
Group differences in participant’s self-report scores were analyzed using t-tests.
Results are presented based on pooled variance estimations.
On measures of family relationship quality, three t tests were used to compare
single mother and married/partnered mother headed families, and six follow-up t-tests
were used to separately examine mother’s and child’s perspective (three scales rated by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
each member of the mother-child dyad yields six dependent variables). See Table 18.
None of the dyadic or individual family relationship quality variables were significantly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 17
M eans and Standard D eviations o f Transform ed M ajor Stud)! Variables fo r Fidl Sample, Single Parented Families and M arried/Partnered Parented
Households.
Variable N
M
SD
Single Partnered Single Partnered Single Partnered Range
Dyadic Fam ily Relationship Quality
Affectual solidarity 29 83 47.28 50.34 9.03 5.34 23-60
Conflict 29 82 13.07 12.37 7.85 4.60 6-30
Em otional support 29 84 8.00 8.88 3.35 3.33 0-12
M other’s Fam ily Relationship Quality
A ffectual solidarity w/child 34 99 25.06 26.10 3.83 3.40 17-30
Conflict w /child 34 99 6.00 5.91 2.49 2.56 3-16
Em otional support w /child 34 99 4.12 4.66 2.00 1.79 0-6
C hild’s Fam ily Relationship Quality
Affectual solidarity w /m other 29 83 22.55 24.13 6.14 5.42 5-30
Conflict w /m other 29 82 6.93 6.54 3.34 2.87 3-18
Em otional support w /m other 29 84 4.07 4.14 1.96 2.13 0-6
M other’s W ell-being
D epressive sym ptom s 34 i 98 16.46 8.98 12.19 -8.49 0-50
N egative affect 34 9$ ’ 2.29 1.64 1.55 1.40 0-5
Positive affect 34 98 3.35 4.09 1.35 1.12 0-5
Self-esteem 34 98 31.70 34.43 4.99 4.45 15.56-40
C hild’s W ell-being
Depressive sym ptom s 29 84 14.43 12.59 9.34 10.36 0-54
N egative affect 29 84 2.62 2.48 1.40 1.62 0-5
Positive affect 29 84 3.88 3.98 1.33 1.09 1-5
Self-esteem 29 82 33.93 33.46 4.75 4.89 21.11-40
Note. Due to missing data n’s are provided for each variable.
Os
UJ
64
different between single mother headed households and married/partnered mother
headed households.
Table 18
T-tests for Ratings o f Groups on
Status
Family Relationship Quality for Mother’ s Relationship
Scale and Group df f p-value
Affectual solidarity
Dyadic
110 -1.43 .16
Conflict 109 .57 .57
Emotional support 111 -1.12 .27
Affectual solidarity w/child
Mothers
131 -.55 .58
Conflict w/child 131 .34 .74
Emotional support w/child 131 -1.63 ,11
Affectual solidarity w/mother
Children
110 -.54 .59
Conflict w/mother 109 -.01 .99
Emotional support w/child 111 -.56 .58
a T-tests em ployed pooled variance estim ation unless otherw ise noted. T- tests m arked with a u used
Satterthw aite’s m ethod for com pensating for unequal variances
u Used Satterthwaite t-tests for unequal variances
On measures of maternal well-being, three of the four t tests comparing the two
groups of mothers were significant after Bonferroni correction (see Table 19). Single
mothers (M = 12.30, S.D. = 11.52) reported more depressive symptoms (t (130) = 3.28,
P = .001) than did married/partnered mothers (M = 9.47, S.D. = 8.19). At the same
time, single mothers (M = 3.56, S.D. = 1.38) reported lower positive affect
(t (130) = -3.09, p = .003) than did married/partnered mothers (M = 4.25, S.D. = .94).
Single mothers (M = 32.80, S.D. = 4.57) also reported lower self-esteem (t (130) = -
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
2.91, 2 = -004) than did married/partnered mothers (M = 34.71. S.D. = 4.75). On
measures of children’s well-being, no differences between groups were observed based
on mother’s relationship status.
Table 19
T-testsfor Groups on Well-Being o f Single Mothers and Their Children and
Married/Partnered Mothers and Their Children
Measure and Group d f
t
p-value
Mothers
Depressive symptoms 130 3.28 .001
Negative affect 130 1.91 .06
Positive Affect 130 -3.09 .003
Self-Esteem 130 -2.91 .004
Children
Depressive Symptoms 111 1.29 .20
Negative Affect 111 .43 .67
Positive Affect 111 .49 .62
Self-Esteem 109 .42 .68
Assessing for differences in demographic variables for possible covariates.
In this study, demographic covariates were chosen based on variables that might
confound any possible relationship between maternal sexual orientation and well-being
of mothers or children. In order to maintain conceptual consistency, we sought a single
set of demographic covariates to use in all analyses. For this reason, we screened
demographic variables to find those that were related simultaneously to maternal sexual
orientation and to any of the four well-being variables measured by report of the child
or the mother. As it turns out, the set of covariates resulting from this procedure
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
overlaps considerably with the set one would get by seeking all demographics that
are related to maternal sexual orientation. This method also overlaps considerably with
the set of demographic variables simultaneously related to maternal sexual orientation
and any of the family relationship quality variables (i.e., related to maternal sexual
orientation and related to at least one family relationship quality variable). Thus, this
set of covariates is appropriate in analyses in which family relationship quality is the
outcome variable.
Procedurally, in order to determine which demographic covariate variables
would go into later analyses, demographic variables were examined in two groups. The
first group of variables was examined across mother’s sexual orientation using t-tests
and chi-squares. Lesbigay mothers and their children were compared to heterosexual
mothers and their children for differences on demographic variables (see section on
comparison of lesbigay and heterosexual mothers and their children above) including:
child’s age, child’s gender, child’s partner status, child’s parental status, child’s age at
divorce, child’s education, child’s income age, mother’s age, mother’s education,
mother’s income, and mother/child cohabitation status. T-tests were conducted to
compare the two samples on measurement variables, and chi-square analyses of
independence were used across lesbigay headed and heterosexual headed households.
(See Table 20.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
Table 20 9
T-tests and Chi-Square analyses for Demographic Variables and the Mother’ s Sexual
Orientation (SO)
Candidate n y? t
£
Related Related
Variable for to to
Maternal Maternal Well
SO SO being
Mothers
Mother’s age 134 — -.25 .003 Yes No
Mother’s education 133 — -.79 <0001 Yes Yes
Mother’s income 124 — -.33 .0002 Yes Yes
Children
Child’s age 142 — .12 .14 No No
Child’s gender 121 .50 — .48 No No
Child’s age at
mother’s divorce
1
— .22 .01 Yes Yes
Child ’ s partner status 141 12.23 — .0005 Yes Yes
Child’s parental status 118 11.28 — .0008 Yes Yes
Child’s education 130 — -.52 <.0001 Yes Yes
Child’s income 111 — -.12 .20 No No
Dyads
Mother/child
cohabitation 138 .03 " .88 No Yes
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The group of demographic characteristics were then were examined with
respect to well-being variables. Seven demographic variables were significantly related
to well-being (see Table 20), with six of them overlapping with the demographic
variables significantly related to mother’s sexual orientation. Those variables were
child’s partner status, child’s parental status, child’s age at mother’s divorce, child’s
education, mother’s education, and mother’s income. Because mother’s education was
highly correlated with mother’s income and we had concerns of multicollinearity with
both of those variables in the models, mother’s education was dropped. Additionally,
child’s education was highly correlated with both mother’s income and mother’s
education and child’s age. This created a problem and given the choice of eliminating a
covariate, child’s education was the best option because we wanted to keep the
covariate (mother’s income) earlier in the causal path. Furthermore, for theoretical
simplicity, we wanted a set of covariates that could be used in all analyses. Because
mother’s income affects both the mother and the child, it was chosen over child’s
education. The final set of variables was child’s partner status, child’s parental status,
child’s age at mother’s divorce and mother’s income. When the analyses were run with
SAS Version 8.12, tests of multicollinearity for each hierarchical regression were
examined in the final step of each model. None of the analyses met the criteria for
multicollinearity as suggested by Tabachnick and Fidell (1996) based on the work of
Belsely, Kuh, and Welsch (1980).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dividing a continuous variable into three levels such as we did with mother’s
income tends to reduce the variability of that variable. In this case, we felt that
mother’s income showed so much variability, that it would compete with other
demographic predictors, such as child’s partner status and child’s parental status which
were true dichotomies, that we decided to trichotomize income in order to provide more
of an opportunity for the remaining demographic predictors to exert an effect.
Family Relationship Quality and Mother’s Sexual Orientation
Hypothesis I suggested that family relationship quality would not differ across lesbigay
headed households and heterosexual headed households, after controlling for
demographics. Three hierarchical regression analyses (dyadic affectual solidarity,
dyadic conflict and dyadic emotional support) were examined, with demographic
predictors entered into the first step (child’s partner relationship status, child’s parental
status, child’s age at mother’s divorce and mother’s income) and mother’s sexual
orientation with demographic variables entered on the second step. The F-tests for R-
square change and individual variable t-test analyses revealed that sexual orientation did
not make a significant contribution to family relationship quality, supporting hypothesis
I. Follow-up analyses of mother’s and child’s individual perspective on the relationship
with each other (mother’s perspective of affectual solidarity, conflict and emotional
support, and child’s perspective of affectual solidarity, conflict and emotional support)
also supported the hypothesis that family relationship quality did not differ across
lesbigay headed and heterosexual headed households. ). It would be unfairly
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
advantageous to for this hypothesis to use an overly conservative Bonferroni
adjustment for this reason unadjusted results are presented along with Bonferroni
adjusted results.
In the hierarchical regression analysis of dyadic affectual solidarity as the
dependent variable, demographic variables assessing child’s relationship status, child’s
parental status, and child’s age at divorce and mother’s income were entered for step
one (for complete regression table, see Table 21). The overall model significance test
for step one was nonsignificant (F (4,95) = 1.31, p = .27). R-square for this step was
.0521, indicating that only 5.2% of the variance in dyadic affectual solidarity was
accounted for by the demographic predictors entered. The overall model significance
for step two was also nonsignificant (F (5,94) = 1.35, p = .25). The R-square for step
two was .0671, indicating that 6.7% of the variance in dyadic affectual solidarity was
accounted for by demographic variables and mother’s sexual orientation. The F-test for
t 2
the change in variance accounted for (R ) between step 1 and step 2 showed that
maternal sexual orientation failed to enhance the prediction of dyadic affectual
solidarity, after controlling for relevant demographic predictors (F = 4, 94) = 1.51, g =
.22). So although individual t-tests of groups on mother’s sexual orientation (see Table
15) for affectual solidarity showed that lesbigay mothers had lower affectual solidarity
than heterosexual mothers, once demographics were entered in the hierarchical
regression model, there were no significant differences. Although the demographics
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
were not themselves significantly different between the two groups, they accounted
for differences observed on the t-tests.
Table 21
Summary o f Hierarchical Regression Models o f Transformed Dyadic Affectual
Solidarity on Demographic, Mother’ s Sexual Orientation with Beta Weights.
Variable B SE B B
Step 1
Child’s partner status 0.16 0.09 .21
Child’s parental status -0.05 0.09 -.07
Child’s age at divorce 0.00 0.00 .01
Mother’s income 0.06 0.05 .13
Step 2
Child’s partner status 0.17 0.09 .23
Child’s parental status -0.03 0.09 -.04
Child’s age at divorce 0.00 0.00 .03
Mother’s income 0.05 0.05 .10
Mother’s sexual -0.10 0.08 -.13
orientation
Note R2 = .0521 Step 1; A R2 = .015 (ps > .05).
In the hierarchical regression analysis of dyadic conflict as the dependent
variable, the overall model significance test for step one was nonsignificant (F (4, 94) =
1.52, p = .20). (See Table 22.) R-square for this step was .0608, indicating that only
6.1% of the variance in dyadic affectual solidarity was accounted for by the
demographic predictors entered. The overall model significance for step two was
nonsignificant (F (5, 93) = 1.53, p = .19). The F-test for the change in variance
accounted for (R ) between step 1 and step 2 showed that maternal sexual orientation
failed to enhance the prediction of dyadic conflict, after controlling for relevant
demographic predictors (F - 4, 93) = 1.55, p = .22).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
Table 22
Summary o f Hierarchical Regression Models o f Transformed Dyadic Conflict on
Demographic, Mother’ s Sexual Orientation with Beta Weights.
Variable B SEB fi
Step 1
Child’s partner status -0.24 0.15 -.19
Child’s parental status 0.04 0.16 .03
Child’s age at divorce -0.01 0.01 -.14
Mother’s income 0.04 0.09 .04
Step 2
Child’s partner status -0.26 0.15 -.20
Child’s parental status 0.00 0.17 .00
Child’s age at divorce -0.01 0.01 -.16
Mother’s income 0.06 0.09 .07
Mother’s sexual 0.17 0.14 .14
orientation
Note. R2 = .1085 S te p l;A R 2 = ' .005 (ps > .05).
In the hierarchical regression analysis of dyadic emotional support as the
dependent variable, the overall model significance test for step one was significant (F
(4, 96) = 2.92, p = .025). (See Table 23.) R-square for this step was .1085, indicating
that 10.9% of the variance in child’s emotional support was accounted for by the
demographic predictors entered. A t-test on the individual predictor parameter
estimates showed that mother’s income was also significantly different from zero (t (1)
= 3.36, p = .001) before Bonferroni correction. The overall model significance for step
two was significant (F (5, 95) = 2.42, p = .041). The R-square for step was .1130,
indicating that 11.3% of the variance in child’s emotional support was accounted for by
demographic variables and mother’s sexual orientation, but less than 1% added with
mother’s sexual orientation. The F-test for the change in variance accounted for (R )
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
between step 1 and step 2 showed that maternal sexual orientation failed to
significantly enhance the prediction of dyadic emotional support, after controlling for
relevant demographic predictors (F (4, 95) = .48, p = .49).
Table 23
Summary o f Hierarchical Regression Models o f Transformed Dyadic Emotional
Support on Demographic, Mother’ s Sexual Orientation with Beta Weights.
Variable B SE B B
Step 1
Child’s partner status -0.01 0.08 -.02
Child’s parental status 0.00 0.09 .00
Child’s age at divorce 0.00 0.00 .05
Mother’s income 0.16 0.05 .33
Step 2
Child’s partner status -0.01 0.08 -.01
Child’s parental status 0.01 0.09 .02
Child’s age at divorce 0.00 0.00 .06
Mother’s income 0.15 0.05 .32
Mother’s sexual -0.05 0.07 -.07
orientation
Note. R2 = ,1085for Step 1; A R2 A 0045 (ps > .05).
** P < .001._________________________________
Follow-up analyses examining only the mother’s perspective involved three
additional hierarchical regressions (mother’s perception of affectual solidarity, conflict
and emotional support) and analyses examining only the child’s perspective involved
three more hierarchical regressions (child’s perception of affectual solidarity, conflict
and emotional support). In step one of the mother’s analyses, demographic variables
(child’s relationship status, child’s partner status, child’s age at divorce, and mother’s
income) were entered, and in step two mother’s sexual orientation as well as
demographic variables were entered into the model. In step one of the child’s analyses,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
demographic variables (child’s relationship status, child’s partner status and child’s
age at divorce) were entered, and in step two mother’s sexual orientation was added to
the model. All results were consistent with the dyadic relationship quality analyses.
For complete regression tables, see Appendix K.
Dyad position and family type. Mean scores before adding covariates revealed that,
while mother’s reports show heterosexual headed households to be characterized by
greater affectual solidarity than lesbigay headed households, reports made by children
show a non-significant tendency in the opposite direction. This suggested the
possibility that dyad position (mother versus adult child) and family type (heterosexual
versus lesbigay headed) interact to predict affectual solidarity. We ran a post-hoc test
of this hypothesis using a 2x2 ANOVA with one between subjects effect (family type)
and one within subjects effect (dyad position), and,an interaction term. These results
are summarized in Table 24. As this table shows, the effect of family type was
Table 24
Analysis o f Variance fo r Transformed Affectual Solidarity
Source d f F T |
P
Family Types (MSO) 1
Between Subjects
0.08 0.78
Dyad Position
Parent v. Child (PC) 1
Within Subjects
6.14 .015
MSO x PC 1 12.00 0.0008
MSO x PC 111 (0.053)
within-group error
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
non-significant (F (1, 111) = 0.08, g = .78). The effect of dyad position was significant
(F (1, 111) = 6.14, p = .015), with mothers reporting greater affectual solidarity than
their children. The interaction effect (family type by dyad position) was also significant
(F (1, 111)= 12.00, p = .0008), indicating that the discrepancy between mother’s and
child’s reports, within the same dyads, differs across family type. Visual inspection of
means shows that, in lesbigay headed households, mothers give nearly the same
quantitative estimates of relationship affectual solidarity as their children do, while for
heterosexual headed households, the discrepancy between a mother’s report and that of
her child is wider.
Summary for hypothesis I. In summary, no differences were found in family
relationship quality for lesbigay versus heterosexual headed families after controlling
statistically for demographic variables associated with family relationship quality.
Indeed, even without demographic controls there were no significant differences after
correcting for multiple statistical tests. This finding supported the hypothesis that there
are no significant differences between perceived family relationship quality across
families with lesbigay or heterosexual mothers. Additional post hoc analyses indicated
that the discrepancy between mother’s and child’s reports of affectual solidarity, within
the same dyads, was greater for heterosexual than lesbigay headed households.
Family Relationship Quality and Mother’s Partner Status
Family relationship quality would not differ depending on the mother’s partner
status (single versus married/partnered), according to hypothesis II. Three hierarchical
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
regression analyses (dyadic affectual solidarity, dyadic conflict, dyadic emotional
support) were examined, with demographic predictors entered into the first step (child’s
relationship status, child’s parental status, child’s age of divorce and mother’s income)
and mother’s partner status entered on the second step with the demographics variables.
The first step was essentially the same as that presented in the analyses testing
hypothesis I, with hierarchical regressions of family relationship status with
demographic variables entered on the first step where none of the regressions were
significant. As for hypothesis I, it would be unfairly advantageous to for this hypothesis
to use an overly conservative Bonferroni adjustment for this reason unadjusted results
are presented along with Bonferroni adjusted results.
Of the three regressions run, none of the regressions were significant at step two
with regard to R-squared change, supporting the hypothesis that mother’s partner status
did not make a significant contribution to family relationship quality. Follow-up
analyses of mother’s and child’s individual perspectives on their relationship with each
other were examined with six hierarchical regression analyses (mother’s perspective of
affectual solidarity, conflict and emotional support, and child’s perspective of affectual
solidarity, conflict and emotional support) were conducted. The results supported the
hypothesis that family relationship quality did not differ across families with single
mothers and married/partnered mothers.
In the hierarchical regression analysis of dyadic affectual solidarity as the
dependent variable, the overall model significance test for step one was nonsignificant
(F (4, 94) = 1.39, 2 = .24). R-square for this step was .0560, indicating that only 5.6%
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
of the variance in dyadic affectual solidarity was accounted for by the demographic
predictors entered. The overall model significance for step two was also nonsignificant
(F (5, 93) = 1.19, p = .32). (See Table 25.) The F-test for the change in variance
accounted for (R ) between step 1 and step 2 showed that maternal partner status did not
enhance the prediction of dyadic affectual solidarity, after controlling for relevant
demographic predictors (F (4, 93) = .41, p - .53).
Table 25
Summary o f Hierarchical Regression Model at Step 2 o f Transformed Dyadic Affectual
solidarity on Demographic, Mother’ s Relationship Status with Beta Weights
Variable B SEB B
Step 2
Child’s partner status 0.16 0.09 .22
Child’s parental status -0.02 0.10 -,03
Child’s age at divorce 0.00 0.01 .00
Mother’s income 0.05 0.05 .09
Mother’s relationship 0.06 0.10 .07
status
Note. A R2 = .0205 (p > .05).
In the hierarchical regression analysis of dyadic conflict as the dependent
variable, the overall model significance test for step one was nonsignificant (F (4,93) =
1.57, p = .19). R-square for this step was .0634, indicating that 6.3% of the variance in
dyadic conflict was accounted for by the demographic predictors entered. The overall
model significance for step two, was nonsignificant (F (5,92) = 2.00, p = .086). (See
Table 26.) The R-square for this step was .0981, indicating that 9.8% of the variance in
dyadic conflict was accounted for by demographic variables and mother’s partner
status. The F-test for the change in variance accounted for (R2) between step 1 and step
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
2 showed that maternal partner status approached significance for enhancing the
prediction of dyadic conflict, after controlling for relevant demographic predictors (F
(4, 92) = 3.54, p = .06).
Table 26
Summary o f Hierarchical Regression Model at Step 2 o f Transformed Dyadic Conflict
on Demographic, Mother’ s Relationship Status with Beta Weights._________________
Variable B SE B fi
Step 2
Child’s partner status -0.23 0.15 .04
Child’s parental status -0.07 0.17 .00
Child’s age at divorce -0.01 0.01 .02
Mother’s income 0.11 0.09 .00
Mother’s relationship -0.32 0.17 .03
status
Note. A R2 ~ .0438 (p > .05).
In the hierarchical regression analysis of dyadic emotional support as the
dependent variable, the overall model significance test for step one was significant (F
(4, 95) - 2.87, p = .027). R-square for this step was .1079, indicating that 10.8% of the
variance in dyadic emotional support was accounted for by the demographic predictors
entered. A t-test on the individual predictor parameter estimates showed that mother’s
income was significantly different from zero (t (1) - 3.34, p - .0012). The overall
model significance for step two approached significance (F (5, 94) = 2.30, p = .051).
(See Table 27.) The R-square for step was .1092, indicating that 10.9% of the variance
in child’s emotional support was accounted for by demographic variables and mother’s
partner status. The F-test for the change in variance accounted for (R2) between step 1
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
and step 2 showed that maternal relationship status failed to enhance the prediction of
dyadic emotional support, after controlling for relevant demographic predictors (F (4,
94) = .14, p = .71), income was the demographic variable accounting for the variance.
Table 27
Summary o f Hierarchical Regression Model at Step 2 o f Transformed Dyadic
Emotional Support on Demographic, Mother’ s Relationship Status with Beta Weights.
Variable B SEB B
Step 2
Child’s partner status -0.01 0.08 -.01
Child’s parental status -0.01 0.09 -.01
Child’s age at divorce 0.00 0.00 .04
Mother’s income 0.16 0.05 .35 **
Mother’s relationship -0.03 0.09 -.04
status
Note. A R2=.0048 (2 >.05). **P<.01.
In follow-up analyses we separately examined mother’s and child’s perception
of the relationship. Six regressions were run, three for the mother’s perception
(mother’s affectual solidarity, conflict and emotional support) and three for the child’s
perception (child’s affectual solidarity, conflict and emotional support). (See Appendix
L.) No significant findings emerged for mother’s affectual solidarity or mother’s
emotional support. However, there were some results worth reflecting on for mother’s
conflict that are described below. No significant findings emerged in the hierarchical
regression analysis of child’s affectual solidarity, conflict and emotional support, and no
results were of note.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
In the hierarchical regression analysis of mother’s conflict as the dependent
variable, demographic variables assessing child’s partner status, child’s parental status,
child’s age at divorce, and mother’s income were entered for step one. The overall
model significance test for step one was nonsignificant (F (4, 95) = 1.50, p = .21).
R-square for this step was .0593, indicating that only 5.9% of the variance in mother’s
conflict was accounted for by the demographic predictors entered. The overall model
significance for step two approached significance (F (5, 94) = 2.16, p = .065). (See
Table 28.) The R-square for step two was .1031, indicating that 10.3% of the variance
in mother’s conflict was accounted for by demographic variables and mother’s partner
status. In the second model, a t-test on the individual predictor parameter estimate
showed that mother’s partner status was significantly different from zero (t (1) = -2.14,
P = .035), suggesting that mothers perceived less conflict with their children when they
were married/partnered as compared to when they were single. The F-test for the
change in variance accounted for (R“) between step 1 and step 2 showed that maternal
partner status enhanced the prediction of conflict, as rated by mother, after controlling
for relevant demographic predictors (F (4, 94) = 4.59, p = .035). These findings were
not significant after Bonferroni correction. Analyses without Bonferroni correction
reveal that it is the mother’s perception of conflict that accounts for the significant
differences in dyadic conflict findings.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
Table 28
Summary o f Hierarchical Regression Models at Step 2 o f Transformed Mother’ s
Conflict on Demographic, Mother’ s Relationship Status with Beta Weights
Variable B SEB 1 3
Step 2
Child’s partner status -0.02 0.02 -.10
Child’s parental status -0.03 0.02 -.15
Child’s age at divorce 0.00 0.00 -.11
Mother’s income 0.01 0.01 .13
Mother’s relationship status -0.05 0.02 -.24 *
Note. A R 2 = = .0438 (p = .035). *p< .05.
Summary for hypothesis II. In summary, no differences were found in family
relationship quality for single versus married/partnered mother headed families after
controlling statistically for demographic variables. Indeed, even without demographic
controls there were no significant differences after correcting for multiple statistical
tests. This finding supported the hypothesis that there are no significant differences for
perceived family relationship quality across families with single or married/partnered
mothers. Additional post hoc analyses of mother and child’s individual perceptions
supported the findings on dyadic relationship quality.
Mother and Child Well-being
The second specific aim was to examine mother’s and child’s well-being across
lesbigay headed and heterosexual headed households (determined by mother’s sexual
orientation), and across single and dual parent households (determined by mother’s
current partner status). Individual well-being was assessed by four measures reported
by the targeted individual (mother or child) for a total of eight measures per dyad. The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
measures assessed depressive symptoms, negative affect, positive affect and self
esteem. Eight hierarchical multiple regression analyses were used to address the two
hypotheses subsumed under the second specific aim. Bonferroni correction procedures
were used for hypothesis III, but they were not used for hypothesis IV since it would be
unfairly advantageous to these hypotheses to use an overly conservative Bonferroni
adjustment. For this reason, unadjusted results are presented along with Bonferroni
adjusted results (2 independent dyad variables (mother, child) x 4 well-being measures).
Hypothesis III stated that family relationship quality would significantly predict
mother’s and child’s well-being across family types. Hypothesis IV stated that family
type (lesbigay headed compared with heterosexual headed households, and single
mother headed households compared with married/partnered mother headed
households) would not make an independent contribution to the prediction of well
being after controlling for family relationship quality and demographics. These two
hypotheses were tested using the eight hierarchical regression analyses, with
demographic covariates entered on the first step, demographics and family relationship
quality (dyadic affectual solidarity, dyadic conflict and dyadic emotional support) on
the second step; and demographics, family relationship quality and family type
(mother’s sexual orientation, lesbigay versus heterosexual headed household, mother
single to mother married/partnered households, on the third step. Hypothesis III was
examined by reviewing the second step in relation to the first step and hypothesis IV
was examined by reviewing the third step.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
Demographics as covariates were chosen for the first step because they were
related to mother’s sexual orientation and related to at least one well-being variable.
For analyses involving a maternal mental health variable as the dependent variable, the
demographic covariates were child’s partner status, child’s parental status, child’s age at
divorce and mother’s income. For analyses involving a child mental health variable,
demographic covariates were child’s partner status, child’s parental status, child’s age at
divorce. Child’s analyses did not use mother’s income (a variable generated by mothers
responses), so that all child respondents could be included in analyses, including those
whose mothers did not respond.
Individual well-being and family relationship quality. Hypothesis III stated that family
relationship quality would predict mother’s and child’s well-being, after controlling for
demographics. The unique contribution of family relationship quality variables, after
controlling for demographics, was tested using eight hierarchical multiple regression
models with two steps. On the first step, demographic variables were entered (child’s
relationship status, child’s parental status, and child’s age at divorce, as well as
mother’s income for mother’s analyses). On the second step, demographic variables
were entered along with dyadic family relationship quality variables (dyadic affectual
solidarity, dyadic conflict, dyadic emotional support) for each mother or child well
being measure (depressive symptoms, negative affect, positive affect and self-esteem).
In addressing hypothesis III, three of the eight regressions were significant at
step two, with regard to R-squared change, providing some support for the prediction
that family relationship quality was significantly related to mother’s and child’s well
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
being. Follow-up analyses were conducted by re-running the analyses without
controlling for demographics. These analyses were run to ensure that findings were not
a result of chosen demographics, and that if analyses were run with only family process
variables and family structure variables, the results would not reveal a different set of
findings. In fact, in the analyses without controlling for demographics, a similar, but
less strong pattern of findings emerged.
In the hierarchical regression controlling for demographic variables for mother’s
report of depressive symptoms, demographics (child’s relationship status, child’s
parental status, child’s age at divorce and mother’s income) were entered on step one.
On step two, dyadic family relationship quality variables were entered. (See Table 29.)
In this analysis, at step one, the overall model significance test was significant (F (4, 92)
- 4.97, p = .0011). R-square for this step was .1776, indicating that 17.8% of the
variability in maternal depressive symptoms was accounted for by demographic
predictors entered. In the first model, a t-test on the individual predictor parameter
estimate showed that three demographic variables were statistically different from zero,
child’s parental status, child’s age at divorce and mother’s income (t (1) = 2.29, p =
.025, t (1) = -2.96, p = .004, t (1) = -2.85, p = .005, respectively). At step two, the
model was significant (F (6, 89) = 5.14, p <.001), and the R-square for this step was
.2880. This indicates that the combination of demographic and family relationship
quality variables accounted for 28.8% of the variability in maternal depressive
symptoms, a change of 11%. In the second model including both demographics and
family relationship quality variables, t-tests on individual predictor parameter estimates
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
showed that two family relationship quality variables, dyadic affectual solidarity and
dyadic conflict were significantly different from zero (t (1) = 2.64, p = .010; t (1) =
3.58, g = .0006, respectively). Significance testing for the change in variance accounted
for indicated that this increment was significant (F-change (4, 89) = 4.60, g = .005).
Table 29
Summary o f Hierarchical Regression Models o f Transformed Mother’ s Depressive
Symptoms on Demographic, Family Relationship Quality and Family Type with Beta
Weights _______________ ___________ ___________________
With Demographics Without Demographics
Variable B SEB 1 3 B SEB 1 3
Step 1
Child’s partner status -0.01 0.10 -.01 - - -
Child’s parental status 0.23 0.10 .27
*
- - -
Child’s age at divorce -0.02 0.01 -.30
**
- - -
Mother’s income -0.16 0.06 -.28
**
Step 2 - - -
Child’s partner status -0.01 0.09 -.01 - - -
Child’s parental status 0.24 0.10 .28
*
- - -
Child’s age at divorce -0.01 0.01 -.25
*
- - -
Mother’s income -0.17 0.05 -.30
**
- - -
Dyad, affectual solidarity 0.38 0.15 .32
**
0.18 0.15 .14
Dyad, conflict 0.26 0.07 .38 0.20 0.07 .30 **
Dyad, emotional support -0.10 0.13 -.08 -0.13 0.13 -.10
Step 3
Child’s partner status 0.01 0.09 .01 - - -
Child’s parental status 0.18 0.10 .21 - - -
Child’s age at divorce -0.01 0.01 -.26
**
- - -
Mother’s income -0.11 0.06 -.19 - - -
Dyad, affectual solidarity 0.38 0.14 .31
**
0.21 0.14 .17
Dyad, conflict 0.23 0.07 .34
-t*
0.20 0.07 .30 **
Dyad, emotional support -0.12 0.13 -.10 -0.12 0.13 -.09
M ’s sexual orientation -0.05 0.08 -.06 -0.03 0.08 -.03
M’s relationship status
....... --------- -------- ^
-0.28 0.10 -.28
r , r.2
**
-0.29 0.09 -.29 **
; r ,2 = ..................
Note. W ith dem ographics: R 2 = .1776 for Step 1 (g = .001); A R 2 = .l 104 for Step 2 (2 — .005); A R2 ‘
.0595 for Step 3 (g = .022). W ithout dem ographics: R2 = .0737 for Step 2 (g = .043); A R2 = .0783 for
Step 3 (g = .01). *g< .05, **g<.01 ***g< .001.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In the hierarchical regression controlling for demographic variables for
mother’s report of negative affect, demographics (child’s relationship status, child’s
parental status, child’s age at divorce and mother’s income) were entered on step one.
On step two, dyadic family relationship quality was entered. (See Table 30.) The
overall model significance for step one approached significance (F (3, 93) = 2.37, p =
.058). R-square for this step was .0925, indicating that 9.3% of the variability in
maternal negative affect was accounted for by demographic predictors entered. In the
first model, a t-test on the individual predictor parameter estimate showed that two
demographic variables were statistically different from zero, child’s parental status and
child’s age at divorce and (t (1) = 2.18, p = .032, t (1) = -2.64, p = .010, respectively).
The overall model significance for step two was significant (F (7, 90) = 2.42, p = .026),
and the R-square for this step was .1580. This indicates that the combination of
demographic and family relationship quality variables accounted for 15.8% of the
variability in maternal negative affect, a change of 6.6%. T-tests on individual predictor
parameter estimates showed that one family relationship quality variable, dyadic
conflict was significantly different from zero (t (1) = 2.21, p = .029). Significance
testing for the change in variance accounted for indicated that this increment
approached significance (F-change (4, 90) = 2.35, p = .08).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
Table 30
Summary o f Hierarchical Regression Models o f Transformed Mother’ s Negative Affect
on Demographic, Family Relationship Quality and Family Type with Beta Weights.
With Demographics Without Demographics
Variable B SEB fi B SEB 1 3
Step 1
Child’s partner status -0.10 0.17 -.07 - - -
Child’s parental status 0.39 0.18 .27
*
- - -
Child’s age at divorce -0.03 0.01 -.28
**
- - -
Mother’s income -0.03 0.10 -.04 - - -
Step 2
Child’s partner status -0.02 0.17 -.02 - - -
Child’s parental status 0.39 0.18 .26
*
- - -
Child’s age at divorce -0.02 0.01 -.24
*
- - -
Mother’s income -0.04 0.10 -.04 - - -
Dyad, affectual solidarity -0.01 0.26 -.01 -0.04 0.23 -.02
Dyad, conflict 0.29 0.13 .26
*
0.25 0.12 .23 *
Dyad, emotional support -0.05 0.23 -.03 -0.01 0.21 -.01
Step 3
Child’s partner status -0.01 0.17 .00 - - -
Child’s parental status 0.27 0.18 .18 - - -
Child’s age at divorce -0.02 0.01 -.26
*
- - -
Mother’s income 0.07 0.11 .07 - - -
Dyad, affectual solidarity -0.01 0.26 -.01 0.00 0.23 .00
Dyad, conflict 0.23 0.13 .21 0.24 0.11 .22 *
Dyad, emotional support -0.08 0.23 -.04 0.02 0.21 .01
M’s sexual orientation 0.04 0.15 .03 0.07 0.14 .05
M’s relationship status
x r _ . _ i w t j . J _____________t.* __ t »2 _ .
-0.42 0.18 -.26
.■; f r i
*
-0.33 0.15 -.20 *
...........................
Note. W ith dem ographics: R 2 = .0925 for Step 1 (g = .058); A R? .0659 for Step 2 (g = .078); A R2
.0486 for Step 3 (g = .073). W ithout dem ographics: R2 = .0563 for St e’ p 2 (g = .10); A R2 = .0471 for Step
3 (g = .068). *g< .05.
In the analysis of mother’s report of positive affect as the dependent variable,
demographic variables (child’s relationship status, child’s parental status, child’s age at
divorce and mother’s income) were entered on step one. On step two, family
relationship quality variables were entered along with the demographic variables. (See
Table 31.) In this analysis, at step one, the overall model significance test was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
nonsignificant (F (4, 93) = 1.84, g = .13). R-square for this step was .0733, indicating
that only 7.3% of the variability in maternal positive affect was accounted for by
demographic predictors entered. The overall model significance for step two was also
nonsignificant (F (7, 90) = 1.74, £=.11), and the R-square for this step was .1198. This
indicates that the combination of demographic and family relationship quality variables
accounted for 11.9% of the variability in maternal positive affect. Significance testing
for the change in variance accounted for indicated that this increment was
nonsignificant (F-change (4, 90) = 1.55, p = .21).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
Table 31
Summary o f Hierarchical Regression Models o f Transformed Mother’ s Positive Affect
on Demographic, Family Relationship Quality and Family Type with Beta Weights.
With Demographics Without Demographics
Variable B SEB B B SEB B
Step 1
Child’s partner status 0.01 0.07 .02 - - -
Child’s parental status -0.15 0.08 -.23 - - -
Child’s age at divorce 0.00 0.00 .06 - - -
Mother’s income 0.06 0.04 .14 - - -
Step 2
Child’s partner status 0.00 0.07 -.01 - - -
Child’s parental status -0.15 0.08 -.23 - - -
Child’s age at divorce 0.00 0.00 .03 - - -
Mother’s income 0.04 0.04 .10 - - -
Dyad, affectual solidarity -0.02 0.12 -.03 0.07 0.10 .08
Dyad, conflict -0.08 0.06 -.17 0.00 0.05 -.01
Dyad, emotional support 0.14 0.10 .16 0.13 0.09 .15
Step 3
Child’s partner status 0.00 0.07 .01 - - -
Child’s parental status -0.09 0.08 -.15 - - -
Child’s age at divorce 0.00 0.00 .07 - - -
Mother’s income 0.00 0.05 -.01 - - -
Dyad, affectual solidarity -0.03 0.11 -.03 0.05 0.10 .06
Dyad, conflict -0.06 0.06 -.12 0.00 0.05 .00
Dyad, emotional support 0.14 0.10 .16 0.10 0.09 .12
M’s sexual orientation -0.12 0.06 -.19 -0.12 0.06 -.20 *
M’s relationship status
' - k 'T , .X 2 'L
0.12
< * > r * _n
0.08 .17 0.15 0.06 .21 *
TW T=......
Note. W ith dem ographics: R2 = 0.0733 for Step 1 (2 > .05); A R 2 .0456 for Step 2 (g > .05); A R2 "
.0516 for Step 3 (g > .05). W ithout dem ographics: R2 = .0376 for Step 2 (g > .05); A R2 = .1020 for Step 3
(g = . 003). *g< .05.
In the analysis of mother’s report of self-esteem as the dependent variable,
demographic variables (child’s relationship status, child’s parental status, child’s age at
divorce and mother’s income) were entered on step one. On step two, family
relationship quality was entered along with the demographic variables. (See Table 32.)
In this analysis, at step one, the overall model significance test was significant (F (4, 92)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
= 5.37,2 = .0006). R-square for this step was .1539, indicating that 15.4% of the
variability in maternal self-esteem was accounted for by demographic predictors
entered. T-tests on individual predictor parameter estimates showed that two
parameters, child’s parental status and mother’s income, were significantly different
from zero (t (1) = -3.10, p = .0025, t (1) = 2.11, p = .037, respectively). At step two,
the model continued to be significant (F (7, 89) = 3.51, p = .002), and the R-square for
this step was .2166. This indicates that the combination of demographic and family
relationship quality variables accounted for 21.7% of the variability in maternal self
esteem. This also indicates that 6.3% of the variability in maternal self-esteem was
uniquely accounted for by family relationship quality. Significance testing for the
change in variance accounted for indicated that this increment was nonsignificant (F-
change (4, 89) = 1.04, p = .38) and that it was not family relationship quality accounting
for the variance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
Table 32
Summary o f Hierarchical Regression Models o f Transformed Mother’ s Self-Esteem on
Demographic, Family Relationship Quality and Family Type with Beta Weights
With Demographics Without Demographics
Variable B SEB B B SEB B
Step 1
Child’s partner status -0.05 0.21 -.02 - - -
Child’s parental status -0.71 0.23
_ 37 **
- - -
Child’s age at divorce 0.02 0.01 .15 - - -
Mother’s income 0.26 0.12 .20 * - - -
Step 2
Child’s partner status -0.05 0.22 -.03 - - -
Child’s parental status -0.73 0.23
_ 38 **
- - -
Child’s age at divorce 0.01 0.01 .11 - - -
Mother’s income 0.21 0.13 .17 - - -
Dyad, affectual solidarity -0.16 0.34 -.06 0.05 0.32 .02
Dyad, conflict -0.22 0.17 -.15 -0.05 0.16 -.03
Dyad, emotional support 0.36 0.30 .13 0.46- 0.29 .17
Step 3
Child’s partner status -0.04 0.22 -.02 - - -
Child’s parental status -0.62 0.24 -.32 * - - -
Child’s age at divorce 0.02 0.01 .14 - - -
Mother’s income 0.12 0.14 .09 - - -
Dyad, affectual solidarity -0.16 0.34 -.06 0.00 0.30 .00
Dyad, conflict -0.17 0.17 -.11 -0.04 0.15 -.03
Dyad, emotional support 0.36 0.30 .13 0.38 0.28 .14
M’s sexual orientation -0.20 0.19 -.11 -0.32 0.18 -.17
M’s relationship status
.„ 'y ; ... _'
0.29 0.24 .14 0.50 0.20 .23 *
i =
Note. With dem ographics: R2 = . 1892 for Step 1 (g = .0006); A R2 .0274 for Step 2 (g > .05); A R2
.0228 for Step 3 (g >.05). W ithout dem ographics: R2 = .0337 for Step 2 (g > .05); A R2 " .0979 for Step 3
(g = .004). *g< .05, **g< .01.
Turning to the analysis of child well-being, in the analysis of child’s self-report
of depressive symptoms as the dependent variable, demographic variables (child’s
relationship status, child’s parental status and child’s age at divorce) were entered. (See
Table 33.) On step two, dyadic affectual solidarity, conflict and emotional support were
entered. In this analysis, at step one, the overall model significance test was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
nonsignificant (F (3, 101) = .63, g = .60). R-square for this step was .0183 indicating
that 1.8% of the variability in child depressive symptoms was accounted for by
demographic predictors entered. In this analysis, at step two, the overall model
significance test was also nonsignificant (F (6, 98) = 1.30, p = .26). R-square for this
step was .0739, indicating that 7.4% of the variability in child depressive symptoms was
accounted for by demographics and family relationship quality variables. The F-test for
the change in variance accounted for (R ) between step 1 and step 2 showed that family
relationship quality did not enhance the prediction of child depressive symptoms, after
controlling for relevant demographic predictors (F (3, 98) = 1.96, g = .12)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
Table 33
Summary o f Hierarchical Regression Models o f Transformed Child’ s Depressive
Symptoms on Demographic, Family Relationship Quality and Family Type with Beta
Weights.
With Demographics Without Demographics
Variable B SEB B B SEB B
Step 1
Child’s partner status -0.10 0.09 -.14 - - -
Child’s parental status 0.00 0.09 .00 - - -
Child’s age at divorce 0.00 0.01 .04 - - -
Step 2
Child’s partner status -0.06 0.09 -.08 - - -
Child’s parental status -0.01 0.09 -.01 - - -
Child’s age at divorce 0.00 0.00 .04 - - -
Dyad, affectual solidarity -0.21 0.13 -.20 -0.22 0.12 -.22
Dyad, conflict 0.04 0.07 .07 0.04 0.06 .07
Dyad, emotional support 0.02 0.11 .02 0.03 0.11 .03
Step 3
Child’s partner status -0.05 0.09 -.06 - - -
Child’s parental status -0.06 0.10 -.08 - - -
Child’s age at divorce 0.00 0.01 .02 - - -
Dyad, affectual solidarity -0.20 0.13 -.20 -0.21 0.12 -.21
Dyad, conflict 0.03 0.07 .05 0.04 0.06 .07
Dyad, emotional support 0.03 0.11 .03 0.04 0.11 .04
M’s sexual orientation 0.03 0.07 .05 0.01 0.07 .01
M’s relationship status
,_____ , • r,.'
-0.14 0.09 -.16
* ™ 2 =
-0.09 0.08 -.11
Note. W ith dem ographics: R 2 = .0183 for Step l(g > .05); A R2 .0556 for Step 2 (g > .05); A R2 .0264
for Step 3 (g > .05). W ithout dem ographics: R 2 = .0650 for Step 2 (g = .07); A R2 = .0116 for Step 3 (g >
.05).
In the analysis of child’s self-report of negative affect as the dependent variable,
demographic variables (child’s relationship status, child’s parental status, child’s age at
divorce and mother’s income) were entered. (See Table 34.) On step two, dyadic
affectual solidarity, conflict and emotional support were entered. In this analysis, at
step one, the overall model significance test was nonsignificant (F (3, 101) = 2.08, p =
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
.11). R-square for this step was .0581, indicating that 5.8% of the variability in child
negative affect was accounted for by demographic predictors entered. At step two,
however, the model was significant (F (6, 98) = 2.57, g = .024), and the R-square for
this step was .1358. This indicates that the combination of demographic and family
relationship quality variables accounted for 13.6% of the variability in child negative
affect. This indicates that 7.8% of the variability in child negative affect was uniquely
accounted for by family relationship quality variables. T-tests on individual predictor
parameter estimates showed that there was one parameter, dyadic affectual solidarity,
significantly different from zero (t (1) = -2.20, g = .029). The F-test for the change in
variance accounted for (R ) between step 1 and step 2 showed that family relationship
quality enhanced the prediction of child negative affect, after controlling for relevant
demographic predictors (F (3, 98) = 2.94, g = .037).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
Table 34
Summary o f Hierarchical Regression Models o f Transformed Child’ s Negative Affect on
Demographic, Family Relationship Quality and Family Type with Beta Weights
With Demographics Without Demographics
Variable B SEB 6 B SEB B
Step 1
Child’s partner status -0.43 0.38 .26 - - -
Child’s parental status -0.09 0.41 .82 - - -
Child’s age at divorce -0.03 0.02 .12 - - -
Step 2
Child’s partner status -0.19 0.38 -.06 - - -
Child’s parental status -0.13 0.39 -.04 - - -
Child’s age at divorce -0.03 0.02 -.16 - - -
Dyad, affectual solidarity -1.22 0.55 -.27
*
-1.16 0.51 -.26 *
Dyad, conflict 0.16 0.29 .06 0.25 0.25 .10
Dyad, emotional support 0.45 0.49 .10 0.50 0.46 .11
Step 3
Child’s partner status -0.18 0.38 -.06 - • - -
Child’s parental status -0.30 0.42 -.09 - - -
Child’s age at divorce -0.04 0.02 -.19 - - -
Dyad, affectual solidarity -1.22 0.55 -.27
*
-1.17 0.51 .03 *
Dyad, conflict 0.12 0.29 .05 0.25 0.25 .33
Dyad, emotional support 0.51 0.49 .11 0.53 0.47 .26
M’s sexual orientation 0.40 0.32 .12 0.19 0.30 .53
M’s relationship status
■ ■ • • • ,.. ....
-0.28 0.38 -.08
r\r\ * n 2 ~
0.00 0.34 .99
: a ........
N ote. W ith dem ographics: R“ = .0581 for Step 1 (g > .05); A R2 .0777 for Step 2 (g < .05); A R2 .0264
for Step 3 (g > .05). W ithout dem ographics: R2 = .0912 for Step 2 (g = .016); A R2' .0064 for Step 3 (g >
.05). *g< .05.
In the analysis of child’s self-report of positive affect as the dependent variable,
demographic variables (child’s relationship status, child’s parental status and child’s
age at divorce) were entered on step one. (See Table 35.) On step two, dyadic family
relationship quality variables were entered. In this analysis, at step one, the overall
model significance test was nonsignificant (F (3, 101) = 1.25, p = .30). R-square for
this step was .0357, indicating that 3.6% of the variability in child positive affect was
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
accounted for by demographic predictor entered. At step two, the model was also
nonsignificant (F (6, 98) = 1.34, p = .25), and the R-square for this step was .0757. This
indicates that the combination of demographic and family relationship quality variables
accounted for 7.6% of the variability in child positive affect. The F-test for the change
in variance accounted for (R2 ) between step 1 and step 2 showed that family
relationship quality did not enhance the prediction of child positive affect, after
controlling for relevant demographic predictors (F (3, 98) = 1.41, p = .24).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
Table 35
Summary o f Hierarchical Regression Models o f Transformed Child’ s Positive Affect on
Demographic, Family Relationship Quality and Family Type with Beta Weights.
With Demographics Without Demographics
Variable B SEB B B SEB B
Step 1
Child’s partner status 0.02 0.08 .04 - - -
Child’s parental status 0.08 0.08 .12 - - -
Child’s age at divorce -0.01 0.00 -.18 - - -
Step 2
Child’s partner status -0.01 0.08 -.01 - - -
Child’s parental status 0.09 0.08 .13 - - -
Child’s age at divorce -0.01 0.00 -.18 - - -
Dyad, affectual solidarity 0.14 0.11 .16 0.15 0.11 .17
Dyad, conflict -0.03 0.06 -.06 -0.02 0.05 -.04
Dyad, emotional support 0.00 0.10 .00 -0.02 0.10 -.03
Step 3
Child’s partner status -0.01 0.08 -.01 - - -
Child’s parental status 0.07 0.09 .10 - - -
Child’s age at divorce -0.01 0.00 -.19 - - -
Dyad, affectual solidarity 0.14 0.11 .16 0.16 0.11 .18
Dyad, conflict -0.04 0.06 -.07 -0.02 0.05 -.04
Dyad, emotional support 0.01 0.10 .01 -0.02 0.10 -.02
M’s sexual orientation 0.03 0.07 .05 0.03 0.06 .05
M’s relationship status
■ v r X T r • , t 1 1 • tt2
-0.04 0.08 -.05 -0.04 0.07 -.05
r,'2 '= "
Note. W ith dem ographics: R 2 = .0357 for Step 1 (2 > .05); A R2 .04 for Step 2 (g > .05); A R2 .0055 for
Step 3 (g > .05). W ithout dem ographics: R2 = .0346 for Step 2 (g > .05); A R2 ‘ .0064 for Step 3 (g > .05).
In the analysis of child’s self-report of self-esteem as the dependent variable,
demographic variables (child’s relationship status, child’s parental status and child’s
age at divorce) were entered on step one. On step two, dyadic family relationship
quality variables were entered with demographics. (See Table 36.) In this analysis, at
step one, the overall model significance test was nonsignificant (F (3, 99) = .24, p =
.87). R-square for this step was .0073, indicating that less than 1% of the variability in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
child positive affect was accounted for by demographic predictor entered. On step
two, dyadic affectual solidarity, conflict, and emotional support were entered with
demographics. In this analysis, at step two, the overall model significance test was also
nonsignificant (F (6, 96) = 1.37, p = .24). R-square for this step was .0786, indicating
that 7.9% of the variability in child self-esteem was accounted for by demographic and
family relationship quality variables entered. This indicates that 7.1% of the variability
in child self-esteem was uniquely accounted for by family relationship quality variables.
The F-test for the change in variance accounted for (R2 ) between step 1 and step 2
showed that family relationship quality approached enhancing the prediction of child
self-esteem, after controlling for relevant demographic predictors (F (3, 96) = 2.48, g =
.07) without any one specific variable being significant in the model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
Table 36
Summary o f Hierarchical Regression Models o f Transformed Child’ s Self-Esteem on
Demographic, Family Relationship Quality and Family Type with Beta Weights.
With Demographics Without Demographics
Variable B SEB fi B SEB B
Step 1
Child’s partner status 0.11 0.23 .06 - - -
Child’s parental status 0.03 0.24 .01 - - -
Child’s age at divorce 0.00 0.01 .04 - - -
Step 2
Child’s partner status 0.05 0.23 .02 - - -
Child’s parental status 0.05 0.23 .03 - - -
Child’s age at divorce 0.01 0.01 .04 - - -
Dyad, affectual solidarity 0.42 0.33 .16 0.28 0.31 .11
Dyad, conflict -0.06 0.17 -.04 -0.18 0.15 -.13
Dyad, emotional support 0.35 0.29 .13 0.38 0.28 .14
Step 3
Child’s partner status 0.05 0.23 .02 - - -
Child’s parental status 0.02 0.25 .01 - - -
Child’s age at divorce 0.00 0.01 .04 - - -
Dyad, affectual solidarity 0.42 0.33 .16 0.29 0.31 .11
Dyad, conflict -0.07 0.17 -.05 -0.18 0.15 -.13
Dyad, emotional support 0.36 0.29 .14 0.41 0.28 .15
M’s sexual orientation 0.08 0.19 .04 0.17 0.18 .09
M’s relationship status
............... . _ .
-0.02
a a - t a r _ _ _r w_
0.22 -.01
a r* \ a r \ 2 =
-0.09
A'T 1 A J?_ _ _O j. _ _ _ _ _ A
0.20 -.04
Note. With dem ographics: R2 = .0073 for Step 1 (p > .05); A R2 .0713 for Step 2 (p = 066); A R2 .0018
for Step 3 (p > .05). W ithout dem ographics: R2 = .0734 for Step 2 (p = .045); A R2^ .0115 for Step 3 (p >
.05).
Individual well-being and family relationship quality without demographics. Follow-up
analyses were run, using the same hierarchical regressions, but not controlling for
demographics. In step one (equivalent to step two of the original analyses) dyadic
relationship quality was entered into the equation with similar findings. (See Tables 29
through 36.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
In the analysis of mother’s self-report of depressive symptoms as the
dependent variable, without controlling for demographics, dyadic family relationship
quality was entered in step one. In this analysis, at step one, the overall model
significance test was significant at p <.05 (F (3, 106) = 2.81, p = .043). R-square for
this step was .0737, indicating that 7.4% of the variability in maternal depressive
symptoms was accounted for by dyadic family relationship quality predictors entered.
A t-test on the individual predictor parameter estimate showed that dyadic conflict was
significantly different from zero (t (1) = 2.81, p - .006).
In the analysis of mother’s self-report of negative affect as the dependent
variable, without controlling for demographic variables, dyadic family relationship
quality variables were entered in step one. The overall model significance for step one
was nonsignificant (F (3, 107) = 2.13, p - .10), and the R-square was .0563. This
indicates that the family relationship quality variables accounted for only 5.6% of the
variability in maternal negative affect. T-tests on individual predictor parameter
estimates showed that there was one parameter, dyadic conflict, was significantly
different from zero (t (1) = 2.12, p = .04).
In the analysis of mother’s self-report of positive affect as the dependent
variable, without controlling for demographic variables, dyadic family relationship
quality variables were entered in step one. At step one the model was nonsignificant (F
(3, 107) - 1.39, p = .25), and the R-square was .0376. This indicates that the family
relationship quality variables accounted for only 3.8% of the variability in maternal
positive affect.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
In the analysis of mother’s self-report of self-esteem as the dependent
variable, without controlling for demographic variables, dyadic family relationship
quality variables were entered in step one. At step one, the model was nonsignificant (F
(3, 106) = 1.23, p = .30), and the R-square was .0337. This indicates that the family
relationship quality variables accounted for only 3.4% of the variability in maternal
self-esteem.
In the analysis of child’s self-report of depressive symptoms as the dependent
variable, without controlling for demographics, dyadic family relationship quality was
entered in step one. In this analysis, at step one, the overall model significance test
approached significance (F (3,107) = 2.48, g = .07). R-square for this step was .0650,
indicating that 6.5% of the variability in child depressive symptoms was accounted for
by dyadic family relationship quality predictors entered.
In the analysis of child’s self-report of negative affect as the dependent variable,
without controlling for demographics, dyadic family relationship quality was entered in
step one. In this analysis, at step one, the overall model significance test was significant
(F (3,107) = 3.58, g = .02). R-square for this step was .0912, indicating that 9.1% of the
variability in child negative affect was accounted for by dyadic family relationship
quality predictors entered. T-tests on the individual predictor parameter estimates
showed that dyadic affectual solidarity was significantly different from zero (t (1) = -
2.29, g = .02).
In the analysis of child’s self-report of positive affect as the dependent variable,
without controlling for demographics, dyadic family relationship quality was entered in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
step one. In this analysis, at step one, the overall model significance test was
nonsignificant (F (3, 107) = 1.28, p = .29). R-square for this step was .0346, indicating
that only 3.5% of the variability in child positive affect was accounted for by dyadic
family relationship quality predictors entered.
In the analysis of child’s self-report of self-esteem as the dependent variable,
without controlling for demographics, dyadic family relationship quality was entered in
step one. In this analysis, at step one, the overall model significance test was significant
(F (3, 105) = 2.77, p = .045). R-square for this step was .0734, indicating that 7.3% of
the variability in child self-esteem was accounted for by dyadic family relationship
quality predictors entered without any one variable being significant in the model.
Summary for hypothesis III. In summary, for hypothesis III examining family
relationship quality as a predictor of well-being, family relationship quality related to
mother’s negative mood states, affectual solidarity and conflict for depressive
symptoms and conflict for negative affect. Higher levels of dyadic affectual solidarity
and conflict were related to higher levels of depressive symptoms. Higher levels of
conflict were also related to higher levels of negative affect. Dyadic affectual solidarity
was the one family relationship quality variable that predicted a child’s well-being
variable, negative affect. Higher levels of dyadic affectual solidarity were related to
lower levels of child’s negative affect, the reverse finding compared to the mothers.
These findings lend some support for the association among family relationship quality
and both mother’s and child’s well-being, with the findings not just a function of
demographics. There was a surprising lack of association among the variables,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
particularly for demographics, which consistently contributed nothing to the child’s
well-being model. At the same time, child’s parental status, child’s age at divorce and
mother’s income contributed to three of the four mother’s well-being models.
Individual well-being and family structure. Hypothesis IV stated that family structure
(lesbigay versus heterosexual headed families, and single versus married/partnered
mother headed families) would not make an independent contribution to the prediction
of well-being after controlling for family relationship quality (affectual solidarity,
conflict, and emotional support) and family background variables (demographics).
Hierarchical multiple regression analyses were used as described above to explore
whether family type makes an independent contribution to the prediction of mother and
child well-being after controlling for family relationship quality and demographics.
Results bearing on Hypothesis IV are shown in Tables 29 to 36. This hypothesis is
assessed in step 3 of each hierarchical regression analysis, in which family structure,
represented by two variables, maternal sexual orientation and maternal partner status, is
added to each model. One model was run for each of the four maternal well-being
variables, and one for each of the four child well-being variables, yielding eight models.
The contribution of family structure was assessed in two ways: First, we examined the
increment in variance accounted for uniquely by addition of maternal sexual orientation
(statistically assessed using the F-test for R-square change between the second and third
step). Second, we examined the significance of the t-test on the regression coefficient
for maternal sexual orientation and maternal relationship status in each final (step three)
model.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
Individual Family Member’s Well-Being and Family Structure
Hypothesis IV predicted that family structure would not make an independent
contribution to well-being after controlling for demographic and family relationship
variables. The third step of hierarchical regressions for mother’s and child’s well-being
(depressive symptoms, negative affect, positive affect and self-esteem) were reviewed
for this hypothesis (a total of eight regressions). The results were partially supported.
Mother’s sexual orientation did not make an independent contribution to either mother’s
or child’s well-being, however, in two cases (for mother’s negative well-being),
mother’s partner status did significantly add to the model. Follow-up analyses were
conducted by re-analyzing the data without controlling for the demographic predictors,
and a similar pattern emerged. Mother’s partner status significantly added to all the
mother’s models.
To examine hypothesis IV, the third step of each of the four hierarchical
regressions for mother’s well-being were reviewed. Examining step three of the model
for mother’s self-report of depressive symptoms, family structure (mother’s sexual
orientation and mother’s relationship status) was added to the model along with
demographic and dyadic family relationship quality. In this analysis, at step three, the
overall model significance test was significant at p < .015 (F (9,87) = 5.15,p < .0001)
with a Bonferroni correction. R-square for this step was .3475, indicating that 34.8% of
the variability in mother’s depressive symptoms was accounted for by demographic,
family relationship quality and family structure. This represents an increment of 6.0%
of the variance with the addition of mother’s sexual orientation and mother’s partner
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
status. T-tests on individual family structure predictor parameter estimates showed
that structure variable, mother’s partner status parameter estimate (b = -.28) was
significantly different from zero (t (1) = -2.76, p = .007). Significance testing for the
change in variance accounted for indicated that this increment was significant (F-
change (7, 87) = 3.97, p = .022).
Examining step three of the model for mother’s negative affect, family structure
(mother’s sexual orientation and mother’s relationship status) were added to the model
along with demographic and dyadic family relationship quality. In this analysis, at step
three, the overall model significance test was significant (F (9,88) = 2.55, g = .01). R-
square for this step was .2070, indicating that 20.7% of the variability in mother’s
negative affect was accounted for by demographic, family relationship quality and
family structure as predictors entered, with a change of 4.8% of the variance. T-tests on
individual predictor parameter estimates showed that for family structure variable,
mother’s partner status parameter estimate (b = -.26) was significantly different from
zero t (1) = -2.30, g = .024). Significance testing for the change in variance accounted
for indicated that this increment approached significance (F-change (7, 88) = 2.70, g =
.07).
Examining step three of the model for mother’s positive affect, mother’s sexual
orientation and mother’s relationship status was added to the model along with
demographic and dyadic family relationship quality. In this analysis, at step three, the
overall model significance test was significant at g <.05 (F (9,88) = 2.01, g = .047). R-
square for this step was .1705, indicating that 17.1% of the variability in mother’s
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
positive affect was accounted for by demographic, family relationship quality and
family structure as predictors entered, with a change of 5.2% of the variance. T-tests on
individual predictor parameter estimates showed no family structure variable was
significantly different from zero. Significance testing for the change in variance
accounted for indicated that this increment approached nonsignificance (F-change (7,
88) = 2.74, p = .07).
Examining step three of the model for mother’s self-report of self-esteem,
mother’s sexual orientation and mother’s relationship status (single or
married/partnered) were added to the model along with demographic and dyadic family
relationship quality. In this analysis, at step three, the overall model significance test
was significant at g = .015 (F (9,87) = 3.04, g = .003) with Bonferroni correction. R-
square for this step was .2394, indicating that 23.9% of the variability in mother’s self
esteem was accounted for by demographic, family relationship quality and family
structure as predictors entered, with a change of 2% of the variance. T-tests on
individual predictor parameter estimates showed that no family structure variable was
significantly different from zero. Significance testing for the change in variance
accounted for indicated that this increment was nonsignificant (F-change (7, 87) = 1.30,
g=.28).
Now we turn to children’s well-being models. Examining step three of the
model for child’s self-report of depressive symptoms, mother’s sexual orientation and
mother’s relationship status (single or married/partnered) were added to the model
along with demographic and dyadic family relationship quality. In this analysis, at step
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
three, the overall model significance test was nonsignificant (F (8, 96) = 1.34, g =
.24). R-square for this step was .1003, indicating that 10% of the variability in child’s
depressive symptoms was accounted for by demographic, family relationship quality
and family structure as predictors entered, with a change of 2.6% of the variance. The
F-test for the change in variance accounted for (R2 ) between step 1 and step 2 showed
that maternal partner status did not enhance the prediction of child’s depressive
symptoms, after controlling for relevant demographic predictors (F (6, 96) = 1.41, g =
.25).
Examining step three of the model for child’s self-report of negative affect,
family structure, measured by mother’s sexual orientation and mother’s relationship
status, was added to the model along with demographic and dyad family relationship
quality. In this analysis, at step three, the overall model significance test was significant
(F (8, 96) = 2.22, g = .03). R-square for this step was .1563 indicating that 15.6% of the
variability in child’s negative affect was accounted for by demographic, family
relationship quality and family structure as predictors entered, with a change of 2.1% of
the variance. The F-test for the change in variance accounted for (R ) between step 1
and step 2 showed that maternal partner status did not enhance the prediction of child’s
negative affect, after controlling for relevant demographic predictors (F (6, 96) = 1.17, g
= .32).
Examining step three of the model for child’s self-report of positive affect,
mother’s sexual orientation and mother’s relationship status were added to the model
along with demographic and dyadic family relationship quality. In this analysis, at step
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
three, the overall model significance test was nonsignificant (F (8, 96) = 1.06, 2 =
.40). R-square for this step was .0812, indicating that 8.1% of the variability in child’s
positive affect was accounted for by demographic, family relationship quality and
family structure as predictors entered, with a change of less than 1% of the variance.
The F-test for the change in variance accounted for (R ) between step 1 and step 2
showed that family structure did not enhance the prediction of child’s positive affect,
after controlling for relevant demographic predictors (F (6, 96) = .29, p = .75).
Finally, examining step three of the model for child’s self-report of self-esteem,
family structure, as measured by mother’s sexual orientation and mother’s relationship
status, was added to the model along with demographic and dyadic family relationship
quality. In this analysis, at step three, the overall model significance test was
nonsignificant (F (8, 94) = 1.03, p = .42). R-square for this step was .0804 indicating
that 8.0% of the variability in child’s self-esteem was accounted for by demographic,
family relationship quality and family structure as predictors entered, with a change of
less than 1% of the variance. The F-test for the change in variance accounted for (R2 )
between step 1 and step 2 showed that family structure did not enhance the prediction of
child’s self-esteem, after controlling for relevant demographic predictors (F (6, 94) =
.09, p = .91).
Individual well-being and family structure without demographics. Follow-up analyses
were conducted without controlling for demographics. In the first step, dyadic family
relationship variables (dyadic affectual solidarity, dyadic conflict and dyadic emotional
support) were entered. On the second step, family type variables (mother’s sexual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
orientation and mother’s partner status) were entered along with the family
relationship variables.
In the analysis of mother’s depressive symptoms as the dependent variable,
without controlling for demographics, but controlling for dyadic family relationship
quality, family structure was entered in step two along with dyadic family relationship
quality. In this analysis, at step two, the overall model significance test was significant
(F (5, 104) = 3,73, p = .004). R-square for this step was .1520, indicating that 15.2% of
the variability in mother’s depressive symptoms was accounted for by dyadic family
relationship quality and family structure predictors entered, with a change of 6.0%. T-
tests on individual family structure predictor parameter estimates showed that structure
variable, mother’s partner status parameter estimate (b = -.29) was significantly
different from zero (t (1) = -3.09, p = .003). The F-test for the change in variance
accounted for (R ) between step 2 and step 3 showed that family structure did enhance
the prediction of mother’s depressive symptoms, after controlling for relevant family
relationship quality predictors (F (7, 87) = 3.97, p = .022).
In the analysis of mother’s negative affect as the dependent variable, without
controlling for demographics, but controlling for dyadic family relationship quality,
family structure was entered in step two along with dyadic family relationship quality.
In this analysis, at step two, the overall model significance test was significant (F (5,
105) = 2.42, p = .04). R-square for this step was .1034, indicating that 10.3% of the
variability in mother’s negative affect was accounted for by dyadic family relationship
quality and family structure predictors entered, with a change of 4.9%. T-tests on
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
110
individual family structure predictor parameter estimates showed that structure
variable, mother’s partner status parameter estimate (b = -.20) was significantly
different from zero (t (1) = -2.13, p = .036). The F-test for the change in variance
accounted for (R ) between step 2 and step 3 showed that family structure did not
enhance the prediction of mother’s negative affect, after controlling for relevant family
relationship quality predictors (F (7, 88) = 2.70, p = .07).
In the analysis of mother’s positive affect as the dependent variable, without
controlling for demographics, but controlling for dyadic family relationship quality,
family structure was entered in step two along with dyadic family relationship quality.
In this analysis, at step two, the overall model significance test was significant (F (5,
105) = 3.41, p = .007). R-square for this step was .1396, indicating that 14.0% of the
variability in mother’s positive affect was accounted for by dyadic family relationship
quality and family structure predictors entered, with a change of 5.2%. T-tests on
individual family structure predictor parameter estimates showed that structure
variables, mother’s sexual orientation and mother’s partner status parameter estimates
(b = -.20, b = -.21, respectively) were significantly different from zero (t (1) = -2.14, p =
.03, t (1) = 2.28, p = .025, respectively). The F-test for the change in variance
'y
accounted for (R ) between step 2 and step 3 showed that family structure did not
enhance the prediction of mother’s positive affect, after controlling for relevant family
relationship quality predictors (F (7, 88) = 2.74, p - .07). Additionally, visual
inspection of the data did not suggest that there was an interaction affect between
mother’s sexual orientation and mother’s partner status.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
I l l
In the analysis of mother’s self-esteem as the dependent variable, without
controlling for demographics, but controlling for dyadic family relationship quality,
family structure was entered in step two along with dyadic family relationship quality.
In this analysis, at step two, the overall model significance test was significant (F (5,
104) = 3.15,p=.011). R-square for this step was .1316, indicating that 13.2% of the
variability in mother’s self-esteem was accounted for by dyadic family relationship
quality and family structure predictors entered, with a change of 2.3%. T-tests on
individual family structure predictor parameter estimates showed that structure variable,
mother’s partner status parameter estimate (b = .23) was significantly different from
zero (t (1) = 2.47, p = .015). The F-test for the change in variance accounted for (R2)
between step 2 and step 3 showed that family structure did not enhance the prediction of
mother’s self-esteem, after controlling for relevant family relationship quality predictors
(F (7, 87) = 1.30, p = .28).
With regard to child’s well-being, in the analysis of child’s depressive
symptoms as the dependent variable, without controlling for demographics, but
controlling for dyadic family relationship quality, family structure was entered in step
two along with dyadic family relationship quality. In this analysis, at step two, the
overall model significance test was nonsignificant (F (5, 105) = 1.74, p = .13). R-square
for this step was .0766, indicating that 7.7% of the variability in child’s depressive
symptoms was accounted for by dyadic family relationship quality and family structure
predictors entered, with a change of 1.2%. The F-test for the change in variance
accounted for (R ) between step 2 and step 3 showed that family structure did not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
112
enhance the prediction of child’s depressive symptoms, after controlling for relevant
family relationship quality predictors (F (3, 105) = .66, p = .52).
In the analysis of child’s negative affect as the dependent variable, without
controlling for demographics, but controlling for dyadic family relationship quality,
family structure was entered in step two along with dyadic family relationship quality.
In this analysis, at step two, the overall model significance test approached significance
(F (5, 105) = 2.20, p = .06), but was not significant after Bonferroni correction. R-
square for this step was .0949, indicating that 9.5% of the variability in child’s negative
affect was accounted for by dyadic family relationship quality and family structure
predictors entered, with a change of less than 1%. The F-test for the change in variance
accounted for (R~) between step 2 and step 3 showed that family structure did not
enhance the prediction of child’s negative affect, after controlling for relevant family
relationship quality predictors (F (3, 105) = .21,2 = -80).
In the analysis of child’s positive affect as the dependent variable, without
controlling for demographics, but controlling for dyadic family relationship quality,
family structure was entered in step two along with dyadic family relationship quality.
In this analysis, at step two, the overall model significance test was nonsignificant (F (5,
105) = .90, p = .49). R-square for this step was .0410, indicating that 4.1% of the
variability in child’s positive affect was accounted for by dyadic family relationship
quality and family structure predictors entered, with a change of less than 1%. The F-
"7
test for the change m variance accounted for (R ) between step 2 and step 3 showed that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
113
family structure did not enhance the prediction of child’s positive affect, after
controlling for relevant family relationship quality predictors (F (3, 105) = .35, p = .71).
In the analysis of child’s self-esteem as the dependent variable, without
controlling for demographics, but controlling for dyadic family relationship quality,
family structure was entered in step two along with dyadic family relationship quality.
In this analysis, at step two, the overall model significance test was nonsignificant (F (5,
103) = 1.91, p = .10). R-square for this step was .0849, indicating that 8.5% of the
variability in child’s self-esteem was accounted for by dyadic family relationship
quality and family structure predictors entered, with a change of 1.2%. The F-test for
the change in variance accounted for (R2) between step 2 and step 3 showed that family
structure did not enhance the prediction of child’s self-esteem, after controlling for
relevant family relationship quality predictors (F (3, 103) = .65, p = .53).
Summary for hypothesis IV. In summary, there was split support for hypothesis IV,
which stated family type (lesbigay versus heterosexual headed families, and single
versus dual parent families) would not make an independent contribution to the
prediction of well-being after controlling for family relationship quality (affectual
solidarity, conflict and emotional support) and family background variables
(demographics). Hierarchical multiple regression analyses for both mother’s and
child’s well-being were examined at the third step where family structure was added to
the model with demographics and family relationship quality. There was no evidence
for mother’s sexual orientation making a contribution to the variance of mother’s well
being or child’s well-being. At the same time, there was evidence for mother’s partner
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
114
status making an independent contribution to mother’s well-being. Mother’s partner
status contributed to mother’s negative mood states (depressive symptoms and negative
affect), after accounting for family relationship quality and demographics, so that a
mother being single suggested that she was more likely to report depressive symptoms
and negative affect than mothers who were married/partnered. There was no evidence
for mother’s partner status making an independent contribution to child’s well-being.
Age and Gender Differences
The third specific aim was exploratory in nature. The goal was to determine
whether there are age and gender differences for children’s perceptions of their
relationships with their mothers, as well as for children’s well-being. In particular, we
were interested in whether age and gender interact with maternal sexual orientation to
predict child well-being and perceived relationship quality. This aim was examined
using a series of hierarchical regression models: three for child’s perception of the
relationship with mother (affectual solidarity, conflict and emotional support) and four
for child’s well-being (depressive symptoms, negative affect, positive affect, and self
esteem). In each model, child’s age, child’s gender and maternal sexual orientation
were entered on step one. In addition to these variables, interaction terms representing
the product of child age and maternal sexual orientation, and child gender and maternal
sexual orientation were entered in step two. The variables used in the interaction terms
were centered as suggested by Pedhazur (1997). The contribution of the interaction
terms was assessed using the F-test of change in variance accounted for between step
one and step two. This was done in order to determine whether better well-being
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
115
outcomes were observed for older versus younger children or boys versus girls
raised by mothers of one sexual orientation or another. This is analogous to a treatment
aptitude interaction in randomized clinical trials. If we had been able to randomly
assign children to conditions, this term would be telling us whether lesbigay mothers
make girls healthier than boys, for example “happier” girls.
Examining the variables of interest in the first step of the analyses, there were
several findings of note. (See Appendix M.) For age effects on child’s perceived
family relationship quality, negative affect decreased with age and conflict decreased
with age. For gender effects on child’s perceived family quality, there were two
significant findings. Girls self-reported higher affectual solidarity and emotional
support than boys. There were no gender differences for well-being.
Several follow-up analyses examining mother’s perception and dyadic
perception of family relationship quality for age and gender contributions were
significant. (See Appendix N.) From the maternal perspective, child’s age significantly
predicted mother’s perception of conflict and emotional support with child, with both
conflict and support diminishing with increasing child’s age. The same findings held
for the dyadic perspective, child’s age significantly predicted dyadic perspective of
conflict and emotional support. For gender differences, maternal report and dyadic
report did not show the same findings as child report. There were no gender differences
for mother’s report of the relationship, or from the summed dyadic perspective.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
116
Child Age and Gender with Maternal Sexual Orientation
To test if there were interaction effects (age by mother’s sexual orientation, and
gender by mother’s sexual orientation) for child’s family relationship quality, step two
of the three hierarchical multiple regression analyses (child’s affectual solidarity,
conflict and emotional support) were used as to explore whether age or gender
interactions made an independent contribution to the prediction of child well-being.
The interaction terms did not contribute to the children’s family relationship quality
prediction models. (See Appendix M.) Similar analyses examining family relationship
quality from the mother’s perspective, as well as dyadic perspective, failed to uncover
any contribution from the interaction of child age and gender with maternal sexual
orientation. (See Appendix N.)
Discussion
Prior family research has not fully explored what aspects of family relate to an
individual’s well-being, particularly with regard to one’s family structure and family
relationships. This study set out to examine two theories regarding the ways in which
families influence their members’ well-being. We found that both family process and
family structure were associated with maternal well-being. Specifically, children do
just as well when raised by lesbigay mothers as by divorced heterosexual mothers.
Lesbigay mothers, as parents, are just as healthy as heterosexual divorced women.
Family process, in terms of affectual solidarity, conflict, and emotional support, is just
as functional in families headed by a lesbigay woman as by a heterosexual divorced
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
117
woman. Evidence suggests that mothers who remarry or repartner after divorce
show higher levels of well-being than mothers who do not. Furthermore, evidence
suggests that the quality of family interactions is associated with the well-being of
individuals in the family. Taken together, these findings suggest that family
relationship quality is a more closely linked to the well-being of individual family
members than the relationship status or sexual orientation of the head of the family. In
this section, we will first discuss the findings from each hypothesis in relationship to the
two family theories: family process theory and family structure theory. Next, we will
discuss the strengths and limitations of the present study. Finally, we will suggest
directions for future research.
Family Process Theory
Family process theory suggests that a strong association exists between family
relationships and the psychological well-being of family members. There was some
support for this theory. Golombok (2000), one of the first researchers in this field,
suggests that “what matters most for children’s psychological well-being is not family
type— it is the quality of family life” p. 104. In terms of family relationship quality,
higher levels of affectual solidarity were associated with less negative affect in
teenage/adult children. For mothers, higher levels of conflict were associated with
higher levels of depressive symptoms and higher levels of negative affect. And counter
to the children, higher levels of affectual solidarity for mothers were associated with
higher levels of depressive symptoms. (See Table 37.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table 37
Significant Variables in Hierarchical Regression Analyses for Transformed Data fo r Specific Aim II
M other Child
Transform ed data Transform ed data Transform ed data Transform ed data
with without with w ithout
dem ographics dem ographics dem ographics dem ographics
Depressive symptoms
Step 1 Dem ographics
C hild’s parental status,
C hild’s age at divorce,
M other’s income n.s.
Step 2 Fam ily process
Affectual solidarity,
Conflict Conflict n.s. n.s.
Step 3 Family_ structure M ’s relationship status M ’s relationship status n.s. n.s.
N egative affect
Step 1 Dem ographics
C hild’s parental status,
C hild’s age at divorce n.s.
Step 2 Fam ily process Conflict Conflict Affectual solidarity A ffectual solidarity
Step 3 Fam ily structure M ’s relationship status M ’s relationship status n.s. n.s.
Positive affect
Step 1 Dem ographics n.s. n.s. -
Step 2 Fam ily process n.s. n.s. n.s. n.s.
Step 3 Fam ily structure n.s.
M ’s relationship status
M ’s sexual orientation n.s. n.s.
Self-Esteem
Step 1 Dem ographics
C hild’s parental status,
M other’s income n.s.
.
Step 2 Fam ily process n.s. n.s. n.s. n.s.
Step 3 Fam ily structure n.s. M ’s relationship status n.s. n.s.
119
Findings related to family process theory were associated with small effect
sizes as compared to previous family studies comparing different groups of
heterosexual headed households (i.e., Lansford, et al., 2001; Demo & Acock, 1994). In
our study, family process accounted for only 2.7% to 11% of the variance for family
relationship quality and 11.9% to 28.8% for family relationship quality and family
demographics, as compared to up to 40% of the total variance in Acock & Demo, 1994.
There are a number of possible explanations. First, Acock and Demo’s (1994) research
measured family relationship quality from only one family member - the same member
who reported on well-being. Thus, the respondent’s well-being may have colored his or
her perception of the relationship. In this study, family relationship quality was
measured from both sides of the dyad, deflating the artificial source of correlation that
Demo and Acock might have exploited. Even exemplarly studies that have gathered
from more than one perspective may have been limited to mothers and teachers,
excluding the child’s perspective (Chan et al., 1998). Additionally, most of these
studies focused on much younger children, rather than young adults. Our finding of
family relationships and demographics accounting for less variance than other studies
could result if effects of family (whether it be relationship quality or structure) on
children dissipated with increasing age and/or with the adult child’s leaving the
household.
There were a few small- to medium-sized relationships in this study between
family relationship quality and well-being. Before correcting for multiple tests, dyadic
affectual solidarity was related to one maternal negative outcome (depressive
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
120
symptoms) and dyadic conflict was related to both maternal negative well-being
outcomes (depressive symptoms and negative affect) to a modest degree. At the same
time, dyadic affectual solidarity was related to one child negative well-being outcome
(negative affect was significant, and depressive symptoms approached significance).
After Bonferroni correction, maternal depressive symptoms were associated with dyadic
conflict. Increased maternal depressive symptoms were associated with increased
reports of dyadic conflict. After Bonferroni correction, child negative affect was
associated with dyadic affectual solidarity. Higher levels of child’s negative affect were
associated with lower levels of dyadic affectual solidarity. In general, family
relationship quality correlated with more child mental health variables (five of twelve
possible bivariate correlations were significant at p < .05), whereas family relationship
quality correlated with fewer maternal mental health variables (two of twelve possible
correlations were significant). However, in multiple regression analysis correcting for
demographic variables, family functioning was more consistently related to maternal
well-being than to child well-being.
Overall, these findings provide only modest support for the association between
family relationship quality and mental health. This may reflect a number of processes.
First, the fact that our observed associations are small (Rosenthal & Rosnow, 1991)
means that few correlations can survive a Bonferroni correction in a sample of this size.
With larger samples, it would be possible to classify some of our smaller effect sizes as
statistically significant. In fact, other much larger studies found more significant
relationships, with a sample in the thousands and medium effect sizes (Demo & Acock,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
121
1994). However, these studies did not examine lesbigay headed households.
(Studies with teenage and adult children have only compared well-being among family
types, heterosexual headed versus lesbigay headed, and looked for associations with
other aspects of family.) For effects as small as we observed, it may be necessary to
conduct larger studies, which creates sampling problems given the difficulties recruiting
lesbigay headed families.
Originally, we expected to find medium or medium-small effect sizes based on
previous estimates of multiple family variables accounting for up to 40% of the
variance. Estimates of power using the method suggested by Cohen (1977) were used
to assess the strength of the analyses. At the p_< .05 level, these correlations between
family relationship quality and psychological well-being had 80% power to detect
medium effect sizes, based on 130 subjects.
The size of our observed associations between well-being and relationship
quality may have been reduced by the fact that family functioning measures were taken
from both sides of the mother-child dyad. Most past research has examined the
association between family relationship functioning and well-being using only one
member of the family as a reporter. In some studies, the family member reported their
perceptions of the relationship as well as their perceptions of their own well-being, and
in other studies, a famil y member reported on their perceptions of the relationships as
well as their perceptions of someone else’s well-being (i.e., their child). This design
tends to inflate the observed associations between family functioning and well-being
since a reporter’s own moods and temperamental tendencies to focus on positive versus
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
122
negative aspects of experience may color all of their reports, leading to a common
underlying emotional factor detected by these studies (Garb, 1998). In this study,
family functioning was assessed by reports from both sides of the mother-child dyad, as
suggested by Bengtson and Roberts (1991). Reports by mothers and children were
averaged to yield a common measure of family functioning taken from two points of
view. In so doing, the common emotional factor that colors reports of family
functioning and well-being is reduced, which may explain why our findings were
modest compared with those found in other studies. The family process results were
similar for both NTF and LSOG families.
Family Structure Theory
Family structure theory suggests that family composition (specifically, mother’s
sexual orientation and mother’s partner status) is particularly important to the individual
family members’ well-being. There was some support for this theory. The major
finding for this hypothesis is that being a single mother was associated with reduced
well-being. Before Bonferroni correction, mother’s partner status was significantly
related to all mother’s well-being measures (reviewing the analyses both with and
without demographics in the models), and after correction, partner status was related to
maternal depressive symptoms and self-esteem. Single mothers showed higher
depressive symptoms and lower self-esteem than married or partnered mothers. With
regard to family structure, mother’s sexual orientation was significantly associated with
positive mood, with lesbigay mothers reporting higher levels of positive affect and self-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
123
esteem than heterosexual mothers. Neither mother’s sexual orientation nor mother’s
partner status were significantly related to child’s well-being.
A further aspect of family structure concerns child’s gender and age. Family
conflict decreased with increasing index child’s age (as did child negative affect).
Furthermore, girls (compared to boys) reported higher affectual solidarity and higher
emotional support with mother. These findings were not confirmed by analyses of
maternal reports of affectual solidarity and emotional support.
Thus, the most striking evidence for the association between family structure
and family process does not concern mother’s sexual orientation. Rather, the mother’s
relationship status and child’s age and gender are correlated with family process
variables.
The interactions between maternal sexual orientation and child gender and age
were also examined and found to make no significant contribution to prediction of
family relationship quality and child well-being, after controlling for main effect terms.
Not only does maternal sexual orientation fail to predict child well-being, but it is
further the case that maternal sexual orientation does not appear differentially to affect
boys and girls, or to differ in its effects with respect to child age. Furthermore, maternal
sexual orientation does not seem to differentially affect family relationship quality in
boys compared to girls, or younger compared to older children.
With regard to family structure and family relationship quality, before
Bonferroni correction, mother’s partner status was significantly related to dyadic
conflict, although this result did not survive Bonferroni correction. It appeared the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
124
result was coming from single mothers reporting more conflict with their children
than married/partnered mothers. In fact, previous studies of heterosexual divorced
families have found that when mothers remarry, conflict with their children increases
(Wallerstein, 2001).
Additionally, heterosexual women reported greater affectual solidarity with their
children (before controlling for demographics) than lesbigay mothers, while children of
lesbigay mothers reported a trend in the direction of greater affectual solidarity with
their lesbigay mothers than did children of heterosexual mothers. These results were
not significant after correcting for multiple tests, yet the interaction was significant.
Lesbigays and their children have greater agreement in rating relationship quality,
whereas heterosexual mothers, compared to their children, tend to overestimate the
quality of the relationship. These findings can be interpreted using the generational
stake hypothesis (Bengtson & Kuypers, 1971; Giarrusso, Stallings, & Bengtson, 1995).
We would expect there to be differences between the two generations, and this was
especially true for the heterosexual mothers and their children, but not the comparison
sample, lesbigay mothers and their children. What is it that may make these mothers
and children more closely mirror each other’s perspectives? It may be related to the
effects of society and environmental factors (Bengtson, 1995), such as the uncaptured
quality of how lesbigay mothers and children may scrutinize their own relationships and
family given the society’s evaluative nature of these families. Alternatively, it may be
that obvious, salient differences (such as sexual orientation) between two generations
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
125
allow the younger generation to experience greater emotional closeness without
threatening their identity and independence.
One could argue that these overall lack of differences were attenuated by
statistical control of demographic variables. In other words, it could be asserted that
differences across family type would be observed if demographic differences were not
statistically held constant. When t-tests were run to examine this question without
controlling for demographic differences, the conclusions did not change. Maternal
reports of affectual solidarity and help and support were significantly associated with
family type. Lesbigay mothers reported lower affectual solidarity and higher levels of
emotional support compared to heterosexual mothers. These differences were in the p
<. 05 but p > .01 range, did not survive a Bonferroni correction, and represent very
small differences in absolute terms and as assessed by delta. Furthermore, the two
findings are more or less contradictory, thus failing to reveal a consistent tendency to
favor one family structure over the other. In summary, we generally found support for
the hypothesis that family structure, evidenced by maternal sexual orientation, has only
small and inconsistent associations with family relationship functioning. These findings
are consistent with other studies, Lansford, et al. (2001) found the most support in their
study for the importance of family processes that occur in all types of families, rather
than family structure.
This lack of striking differences associated with family structure could be
understood in several ways. First, it may be the case, in reality, that parental sexual
orientation is not a significant predictor of family relationship quality, subjectively
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
126
rated. Second, it may be the case, in reality, that differences exist, but that they are
complex and multifaceted and do not clearly favor one type of family over the other
(lesbigay versus heterosexual headed). We are, at least, able to rule out the possibility
that child gender or age interacts with maternal sexual orientation to determine child
well-being. Third, it may be the case that differences in family functioning across the
two types of families are characterized by small effect sizes such that a much larger
study would be needed to examine the issue. If so, such findings might have purely
epidemiologic implications, without clinically significant implications in individual
cases. At the same time, it is important to consider that if, in general, lesbigay mothers
are more likely to divorce from heterosexual marriages than heterosexual mothers, then
a comparison of lesbigay headed households might reach an erroneous conclusion
unless analyses stratify for divorce status.
However, a second finding with respect to this study’s first aim was that there
were too few single mothers to look at any interaction effects of mother’s sexual
orientation and mother’s partner status, because there were very few mothers in the
single cells for lesbigay mothers and heterosexual mothers. This could result from a
tendency for over-representation of partnered mothers. Partnered mothers might be
over-represented because they have more time to participate in research or because they
feel better about themselves and tend to be more willing to think about their functioning
and report it on a survey. It may also be that single, lesbigay mothers have such a high
level of stress that they are less likely to volunteer for research than coupled women.
Thus, our study has systematically excluded stressed single mothers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
127
In analyses examining differences in mental health and family functioning
across family type (lesbigay versus heterosexual headed), we anticipated that we would
have adequate power to detect moderate effect sizes. However, this study did not
produce moderate effect sizes. In fact, those effect sizes observed are small. We
estimated that our analyses should have 80% power to detect moderate effect sizes
(delta=.5) using 65 subjects in each group (Rosenthal & Rosnow, 1991). In order to
detect small effects (.20), a sample of 400 subjects would be needed for each group.
Much larger studies would be required to detect such small effects, but the usefulness of
any findings would be questionable.
The fact that neither family process nor family structure was dramatically
supported raises a question: what does predict well-being? A recent review by Harris
(2000) argues that the importance of family relationships for child functioning has been
systematically over-rated by theoreticians. Harris reviewed the developmental literature
and suggests that researchers have systematically failed to adequately attend to the
importance of children’s extra-family social relationships on their overall well-being
and development. By extending her argument to parents, one may also expect that
parental non-familial relationships may be important predictors of well-being not often
addressed in the literature on family functioning and individual well-being. Future
research should pay more attention to these extra-familial factors and how they are
associated with well-being of all members of the family.
Although this study observed more significant associations between family
functioning and child well-being compared to the number of associations between
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
128
family functioning and maternal well-being (which argues in favor of Harris’s
thesis), the opposite pattern held after demographics were statistically controlled. The
examination of the pattern of associations between family functioning and the
functioning of parental figures as compared with children would be a fruitful direction
for future study.
Furthermore, Putney and Bengtson (2001) emphasize the heterogeneity of
intergenerational relationship in the many variations of family, and the findings from
this study suggest heterogeneity across family type as well. Additional exploration of
this heterogeneity and its effects on family life would be another fruitful direction for
additional inquiry.
This study hypothesized that family structure would not be related to well-being
after controlling for family background and family relationship quality. This hypothesis
was supported for children’s well-being, but only partially supported for mothers’ well
being, since maternal partner status but not sexual orientation was important in the
analyses. Family type did not significantly predict any child well-being measure after
controlling for demographic factors and family relationship quality. Family structure-
embodied by maternal partnership status-was related to maternal well-being. This
could result either from a tendency for partnerships to have a beneficial effect on
mothers, or from a tendency for more euthymic mothers to have more interest in or
better luck finding partners. Finally, there could be an underlying third cause for the
association.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
129
Theoretical Limitations
In contradiction to the family structure hypothesis, children’s well-being in
lesbigay headed families did not differ from that of children in heterosexual headed
families on any of the four well-being measures examined, even when examined using
t-tests (i.e., not controlling for demographic differences, and, thus, having greater power
to detect differences). In terms of the well-being of lesbigay parents, the study found
greater well-being among lesbigays in two of the measures examined: self-esteem and
positive affect, findings which did not hold after correcting for multiple tests. These
two measures were the positive dimensions of well-being in this study, as compared to
the negative dimension (depressive symptoms and negative affect) which showed no
difference between the two groups.
Why would we observe greater well-being among lesbigays mothers compared
to heterosexual mothers? One important consideration is that we examined only
lesbigay mothers who had come out to their children at least two years earlier. Thus,
we missed people who might otherwise qualify but who have not decided to identify
themselves to their children, as lesbian, gay or bisexual. As a result our sample may
favor those with greater comfort with their own sexual orientation, self-confidence,
basic trust in one’s security in the world, and trust in investigators. These factors may
be associated with greater positive affect and well-being. A number of unidentifiable
women probably exist who related to other women as identified lesbians do, but whose
psychological functioning does not lead them to incorporate their orientation into their
identity and public behaviors. Furthermore, the tendency to identify one’s orientation,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
130
incorporate it into one’s identity, then make public commitments and actions
consistent with that identity suggests an active, direct problem solving style. The
tendency to ignore one’s orientation, or to gratify part of it privately without
incorporating these actions, suggests an avoidant coping style and one associated with
cognitive dissonance, both of which are found in previous research to be related to
higher levels of negative affect (Taylor, 1999). Thus, by selecting identified lesbigay
women (the only group that we could sample, as a practical matter), we are also
selecting a group that survived a generally difficult process of coming to terms with a
minority identity and making public commitments to it. People who have come through
that process, as opposed to people who have stopped short of the end of it, would almost
certainly tend to be characterized by greater ego strength.
Children’s well-being was not related to family structure. In exploratory
analyses, children’s gender and age did not interact with mother’s sexual orientation to
predict family relationship quality or well-being. So there were no differences in
children’s well-being based on family structure, but relationships were observed
between single versus partnered maternal status and maternal well-being.
An explanation to account for family structure’s association with maternal (but
not child) well-being, is that while both mothers’ and children’s well-being can directly
influence their own actions, mothers’ actions can influence their relationship or partner
status, while children’s actions do not have as much influence on mother’s partner
status. (See Figure 2.) Thus, there are two plausible mechanisms to account for a
correlation between mother’s well-being and relationship status (well-being is causal
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
131
versus relationship status is causal), while there is only one plausible mechanism to
explain a correlation between child well-being and maternal relationship status
(maternal relationship is causal).
For example, the finding could result either from a tendency for partnerships to
have a beneficial effect on mothers, or from a tendency for more euthymic mothers to
have more interest in or better luck finding partners. Therefore, the correlation between
maternal well-being and maternal relationship status does not necessarily support the
family structure theory. It may be that mothers who are depressed make decisions that
lead them to be single, either by choice or because their partners leave them. And so we
cannot conclude that family structure (single versus partnered) is making these mothers
depressed.
Mother’ s
Children’ s
W ell-being
W ell-being
Fam ily
J I
Structure
J I
(Mother’ s
Mother’ s —> Relationship cz...£ > Child's
Actions Status) Actions
Figure 2. Alternative Theoretical Explanation to Family Structure Theory for the Relationships
Between Well-Being and Family Structure____________________
If family structure theory were correct, it would predict a different pattern:
maternal relationship status would be as strongly related to children’s well-being as it is
to maternal well-being. Furthermore, it would predict significant associations between
maternal sexual orientation and child well-being; no such association was observed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
132
Mother’s sexual orientation was not related to child’s well-being, even in t-tests.
So, even before controlling for demographics on t-tests (an analysis with maximum
power) there was no support for the family structure hypothesis in precisely the place
where the family structure proponents make their most passionate case. In fact, family
processes theory can explain the findings of structure and its relationship to well-being
when it comes to maternal partner status and maternal well-being.
Another component of family to consider is whether the effect of the child’s age
at the time of the divorce is offsetting the findings on mother’s sexual orientation.
Although we do not have sufficient numbers to ask that question, it is a worthwhile
question for future research.
Strengths and Limitations of the Study
Several limitations of the study are recognized. First, questions of causal
linkages in family samples are of great interest, and this dataset cannot address those
linkages. The present data are correlational in nature and do not support causal
inference. Future work using longitudinal research designs will be needed to
disentangle causes and consequences in these domains, given the impossibility of
randomly assigning children to mothers of different sexual orientations or of assigning
sexual orientations to randomly selected mothers. It may be that the ongoing National
Lesbian Family Study may answer these questions over the next 15 years as the children
(currently 8 years old) become young adults (Gartrell, Banks, Hamilton, Reed, Bishop,
& Rodas, 1999; Gartrell, Banks, Reed, Hamilton, Rodas, & Deck, 2000).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
133
Second, this study relied on mother and child reports of family relationship.
Assessments of mother’s and child’s well-being completed by observers who were
unaware of the family background of participants would have been valuable. This
method was used in Golombok’s et al. 1983 original study of children of lesbian
mothers using other measures such as blind clinical interviews. At the same time, our
well-being measures are widely used and well-known instruments. The measures for
family relationship quality have been developed and used by Bengtson and others for
almost 30 years, and are established in the field of family studies, allowing the results of
this study to be compared with those of other researchers (e.g., Bengtson, Schaie, &
Burton, 1995).
Furthermore, sampling issues posed significant problems. Gay, lesbian and
bisexual parents are a difficult population to sample representatively as participants for
research. However, some of the sampling biases, such as the healthy volunteer bias,
also affects the LSOG heterosexual sample. Legal discrimination against lesbians,
especially against those who choose to raise children, may create a population of
invisible lesbian mothers who are virtually impossible to research with standard
designs. To leam about these lesbians, it may be necessary to use other methods, such
as clinical case studies. Similarly, our findings might not generalize beyond the United
States to populations with very different attitudes toward lesbians who are mothers or
with different social structures affecting single mothers.
We sampled women who conceived children with men, and then ended the
relationship. Concerns about legal rights and risks may have kept them from speaking
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
134
about their families in the past. Although the overall political climate may be less
threatening now and their children are older, they may still face other discriminatory
acts against them, even if they are no longer threatened with the possibility that their
children could be removed from their home. As a result, the first limitation is the
representativeness of the sample. Who were we able to locate and solicit for
participation in the study? As noted in the results section, we sampled people who have
a fair amount of education, are mostly partnered, white and relatively affluent. They are
also a group that chose to use social support systems, from either COLAGE or family
pride coalition.
Participants in this sample may systematically exclude those who are
particularly worried or suspicious, especially those who have previously experienced
legal difficulties related to their sexual orientation. In the future, research might use
other sampling methodologies, such as snowball sampling or examining lesbians in
cultures with more legal protections for lesbian relationships, such as Denmark or
Holland, as well as cultures in which lesbian motherhood is even more taboo.
Although it could be argued that this study falsely claims that lesbigay mothers
are not detrimental to children on the basis of sampling a particularly healthy group of
lesbigay mothers and their children, this conclusion is not supported when one
considers the sampling limitations affecting the LSOG study. Both study samples are
affected by a nearly unavoidable healthy volunteer bias.
Another limitation of the study lies in its cross-sectional design. All of the
family members have been exposed to divorce. However, the degree that divorce
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
135
disrupted the family and affected perceived family relationship quality and
psychological well-being are not known. Longitudinal data would be ideal for
understanding the influence of divorce on family members before examining the
predictive quality of family structure and perceived relationship quality in these
families. These data could be used to control for relationship quality.
The sample is relatively small for family research. However, most family
research does not inquire about individuals’ sexual orientation. This study is large
compared to other studies that have identified lesbigay headed households.
While the present study is somewhat limited by factors just listed, it has several
major strengths. First, it tests empirical ideas grounded in various theories by
comparing four types of non-traditional mothers: single-parent lesbigay mothers, dual
parent lesbigay mothers, single-parent heterosexual mothers, and dual-parent
heterosexual mothers and their children. Only by examining both family structure and
family process variables, can we test the effects of both simultaneously. In addition,
this study (unlike most) examined family relationship variables and family structure
differences in lesbigay parented families.
This study demonstrates a method of data collection that is useful for studying
nontraditional families at the dyadic level. As noted by Bengtson and Mangen (1988),
designs involving data collected from multiple family members have the potential to
yield rich new information above and beyond that provided by individuals. These data
support the view that it is informative to study coping strategies using data provided not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
136
only by individuals but also by family members because both contribute to our
understanding of family relationships and psychological well-being.
This project extends previous research in several ways. First, it examines
multiple non-traditional families. Second, it extends the research on well-being and
family relationships in families with lesbigay parents. Researchers have recommended
that family member’s well-being and predictors of well-being need to be explored in
families with lesbigay mothers (Patterson, 1994). In order to explore mother and child
well-being, the study controls for potential demographic confounders. Through these
methods, the study explores various theories’ whether it is way family affects individual
well-being.
Implications and Further Research
In terms of child well-being, as noted in the introduction, some writers (e.g.,
Cameron & Cameron, 1997; Cameron et al., 1996; Wardle, 1997) have suggested that
studies with small samples and inadequate methodology justify restrictions on the
parental rights of lesbian mothers. This study fails to provide support for the contention
that children of lesbian mothers have lower subjective well-being compared to children
of divorced heterosexual mothers.
In contradiction to family structure theory, we found that the well-being of
children was modestly related to family relationship quality, especially affectual
solidarity. Affectual solidarity appears to protect children from depressive symptoms,
although the direction of causation cannot be firmly established. This appears to
suggest that efforts to assist children of lesbigay mothers should focus more on
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
137
enhancing the non-specific supportive aspects of family members’ relationships
with each other. Further research to understand the reasons why some lesbigay headed
households are more supportive than others would be of clinical and theoretical interest.
However, given the findings in this study, one may wonder whether it is
important to examine lesbigay headed households separately from heterosexual headed
households. This study did not find striking differences between the two types of
families, at least in terms of child outcomes. Given the sampling difficulties inherent in
studying lesbigay headed households in the United States and the intense politicization
of the issue, future researchers in this area must take great care to provide sufficient
justification for looking separately at lesbigay headed households and sufficient
methodological rigor to allow reliable conclusions to emerge.
Family structure was related to maternal well-being: mothers who were in
relationships had higher well-being than those who weren’t. Further research should
examine the process of repartnering after divorce, especially examining the pattern of
changes in family relationship quality as one parent leaves and a new substitute parent
enters. Intensive examination of a small number of cases prospectively over time, from
the moment of a lesbigay mother’s decision to leave the father of her children to the
time of repartnering or a decision not to repartner, would yield rich hypotheses for
further quantitative survey research. Many aspects of repartnering-such as the timing
and pattern of dating in relation to the developmental phase of the child/children in the
house-are inherently complex and difficult to handle adequately in quantitative survey
methods.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
138
Clinically, efforts should be directed at supporting the mother/child dyad in
order to facilitate the optimal adjustment and development of both parties. Golombok
(2000) suggests that “what matters most for children’s psychological well-being is not
family type-it is the quality of family life” (p. 104). The overwhelming trend of the
evidence suggests that sexual orientation is not amenable to change in psychotherapy,
and since this study does not suggest that the gender of the mother’s subsequent
partners affects child well-being, efforts to convince a lesbian mother to change her
orientation are not supported. Furthermore, no evidence suggests that partnered
mothers are inherently better for child well-being than unpartnered mothers, so
clinicians should avoid rushing clients into relationships for the good of their children.
However, the effects of some legal sanctions implicitly force lesbian mothers to
choose between retaining their children and entering lesbian relationships. Since
mothers in this study had better well-being when partnered, and non-significant
statistical differences between children of partnered versus unpartnered mothers favored
children of partnered mothers, legal sanctions preventing lesbian mothers from
repartnering may be detrimental to mothers as well as children.
An additional point raised in this study is that sampling lesbigay households is
impeded by culture’s longstanding stigmatization and devaluation lesbian relationships-
especially the mother-child relationship of a lesbian and her son or daughter. This leads
to sampling difficulties, which in turn make comparisons between lesbigay mothers and
heterosexual mothers problematic. A possible solution to this difficulty would be to
conduct this research in a culture with less entrenched prejudice against lesbian
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
139
parenthood, preferably a culture in which community registers can be used to
identify a more representative sample of lesbigay families than can be completed using
social organization register sampling. Thus, a health or marriage registry sampling in
Scandinavian countries could much more readily address the fundamental structural
question of whether a man and woman do a better job raising a child than a woman and
another woman. Further, the use of Scandinavian samples could allow us to include
outcomes such as treatment for clinical depression, which would allow a more objective
assessment of problems with well-being of a clinical significant level.
Longitudinal studies, with larger and different samples of lesbigay mothers and
their children, would greatly benefit our understanding of family relationships and well
being, as long as these studies take care to sample at least a small number of individuals
who are relatively “closeted”. Nonetheless, structural comparisons such as those
presented in this study help move the literature beyond theory to empirically testing
these longstanding hypotheses of how a family member’s well-being is affected by
family structure and family process.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
140
References
Acock, A. C., & Hurlbert, J. S. (1990). The effects of marital status on the form and
composition of social networks. Social Science Quarterly, 71(1). 163-182.
Amato, P. (2001). Children of divorce in the 1990s: An update of the Amato and Keith
(1991) meta-analysis. Journal of Family Psychology, 15(3), 355-370.
Amato, P. R., & Booth, A. (1997). A generation at risk: Growing up in an era of family
upheaval. Cambridge, MA: Harvard University Press.
Aquilino, W. S., & Supple, A. J. (2001). Long-term effects of parenting practices during
adolescence on well-being outcomes in young adulthood. Journal of Family
Issues. 22(3), 289-308.
Belsely, D. A., Kuh, E., & Welsch, R. E. (1980). Regression diagnostics: Identifying
influential data and sources of colinearity. New York: John Wiley and Sons.
Bengtson, V. L., & Kuypers, J. A. (1971). Generational differences and the
developmental stake. Aging and Human Development, 2(2), 249-260.
Bengtson, V. L., & Mangen, D. J. (1988). Family intergenerational solidarity revisited:
Suggestions for Future Management. In D. Mangen, V. L. Bengtson, & P. Landry
(Eds.), Measurement of Intergenerational Relations (pp. 222-238). Newbury Park,
CA: Sage.
Bengtson, V. L., & Roberts, R. E. L. (1991). Intergenerational solidarity in aging
families: An example of formal theory construction. Journal of Marriage and the
Family. 53, 856-870.
Bengtson, V. L., Schaie, K. W., & Burton, L. M. (1995). Adult intergenerational
relations: Effects of social change. New York: Springer.
Blechman, E. A. (1982). Are children with one parent at psychological risk? A
methodological review. Journal of Marriage and the Family, 44, 179-195.
Bradbum, N. (1969). The structure of psychological well being. Chicago: Aldine.
Brewaeys, A., Ponjaert, I., Van Hall, E. V., & Golombok, S. (1997a). Lesbian
motherhood: The impact on child development and family functioning. Journal of
Psychosomatic Obstetrics and Gynecology, 18, 1-16.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
141
Brewaeys, A., Ponjaert, I., Van Hall, E. V., & Golombok, S. (1997b). Donor
insemination: Child development and family functioning in lesbian mother
families. Human Reproduction, 12(6), 1349-1359.
Bronfenbrenner, U., & Crouter, A. C. (1983). The evolution of environmental models in
developmental research. In W. Kessen (Ed.), Handbook of child psychology: Vol.
1 History. Theory and Methods (pp. 357-414). New York: John Wiley and Sons.
Bronfenbrenner, U., (1960). The ecology of human development: Experiments by
nature and design. Cambridge, MA: Harvard University Press.
Bryner, C. L. (2001). Children of divorce. The Journal of American Board of Family
Practice. 14(3), 201-210.
Burkhauser, R. V., & Duncan, G. J. (1989). Economic risks of gender roles: Income and
loss and life events over the life course. Social Science Quarterly. 70. 3-23.
Cameron, P., & Cameron, K. (1998). Homosexual parents: A comparative forensic
study of character and harms to children. Psychological Reports. 82, 1155-1191.
Cameron, P., & Cameron, K. (1997). Did the APA misrepresent the scientific literature
to courts in support of homosexual custody? The Journal of Psychology. 131(3).
313-332.
Cameron, P., Cameron, K, & Landess, T. (1996). Errors by the American Psychiatric
Association, the American Psychological Association, and the National
Association in representing homosexuality in amicus briefs about Amendment 2
to the U.S. Supreme Court. Psychological Reports, 79, 383-404.
Carter, B., & McGoldrick, M. (1989). The changing family life cycle: A framework for
family therapy. Boston: Allyn & Bacon.
Chan, R. W., Raboy, B., & Patterson, C. J. (1998). Psychosocial adjustment among
children conceived via donor insemination by lesbian and heterosexual mothers.
Child Development, 69(2). 443-457.
Chodorow, N. (1980). The reproduction of mothering: Psychoanalysis and the
sociology of gender. Berkeley, CA: University of California Press.
Cochran, S. D., & Mays, V. M. (2000). Are lesbians more at risk for psychiatric
disorders? Evidence from the 1996 National Household Survey of drug abuse. In
Proceedings from the National Conference on Health Statistics. Washington, D.C:
U.S. Department of Health and Human Services.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
142
Cohen, J. (1977). Statistical power in the social sciences (rev. ed.). New York:
Academic Press.
Cowan, P. A., Cowan, C. P., & Kerig, P. K. (1993). Mothers, fathers, sons, and
daughters: Gender differences in family formation and parenting style. In P. A.
Cowan, & D. Field (Eds.), Family, self, and society: Toward a new agenda for
family research (pp. 165-195).
Demo, D. H., & Acock, A. C. (1994). Family diversity and well-being. Thousand Oaks,
CA: Sage.
Demo, D. H. (1992). Parent-child relations: Assessing recent changes. Journal of
Marriage and the Family. 54, 104-117.
Dundas, S., & Kaufman, M. (2000). The Toronto lesbian family study. Journal of
Homosexuality, 40(2). 65-77.
Emery, R. E. (1988). Marriage, divorce, and children’s adjustment. Newbury Park, CA:
Sage.
Emery, R. E. (1982). Interparental conflict and the children of discord and divorce.
Psychological Bulletin. 92, 310-330.
Emery, R. E., & O’Leary, K. D. (1984). Marital discord and child behavior problems in
a nonclinical sample. Journal of Abnormal Child Psychology. 12(3), 411-420.
Erel, O., & Burman, B. (1995). Interrelatedness of marital relations and parent-child
relations: A meta-analytic review. Psychological Bulletin, 118(1), 108-132.
Fustenberg, F. F., Jr., Morgan, P., & Allison, P. D. (1987). Paternal participation and
children’s well-being after marital dissolution. American Sociological Review, 52,
695-701.
Gartrell, N., Banks, A., Reed, N., Hamilton, J., Rodas, C., & Deck, A. (2000). The
national lesbian family study: 3. Interviews with mothers of five-year-olds.
American Journal of Orthopsychiatry. 70(4), 542-548.
Gartrell, N. Banks, A., Hamilton, J., Reed, N., Bishop, H., & Rodas, C. (1999). The
national lesbian family study: 2. Interviews with mothers of toddlers. American
Journal of Orthopsychiatry. 69(3). 362-369.
Garb, H. N. (1998). Studying the clinician: Judgment research and psychological
assessment. Washington, DC: American Psychological Association.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
143
Gershon, T. D., Tschann, J. M., & Jemerin, J. M. (1999). Stigmatization, self-esteem,
and coping among adolescent children of lesbian mothers. Journal of Adolescent
Health. 24, 437-445.
Giarrusso, R., Stallings, M., & Bengtson, V. L. (1995). The intergenerational stake
hypothesis revisited: Parent-child differences in perceptions of relationships 20
years later. In V. L. Bengtson, K. W. Schaie, & L. M. Burton (Eds.), Adult
Intergenerational Relations: Effects of Social Change (pp. 227-296). New York:
Springer.
Golombok, S. (2000). Parenting: What really counts. London: Routledge.
Golombok, S., Spencer, A., & Rutter, M. (1983). Children in lesbian and single-parent
households: Psychosexual and psychiatric appraisal. Journal of Child Psychology
and Psychiatry. 24(4), 551-572.
Golombok, S., Tasker, F. L., & Murray, C. (1997). Children raised in fatherless families
from infancy: Family relationships and the socioeconomic development of
children of lesbian and single heterosexual mothers. Journal of Child Psychology
and Psychiatry. 38. 783-791.
Green, R., Mandel, J. B., Hotvedt, M. E., Gray, J., & Smith, L. (1986). Lesbian mothers
and their children: A comparison with solo parent heterosexual mothers and their
children. Archives of Sexual Behavior, 7, 175-184.
Gringlas, M., Weinraub, M. (1995). More things change ...: Single parenting revisited.
Journal of Family Issues. 16(1), 29-42.
Grove, W. R., Hughes, M., & Style, C. B. (1983). Does marriage have positive effects
on the psychological well-being of the individual? Journal of Health and Social
Behavior, 24, 122-131.
Harris, J. R. (2000). The outcome of parenting: What do we really know? Journal of
Personality, 68(3), 625-637.
Heady, B., & Wearing, A. (1991). Subjective well-being: A stocks and flows
framework. In F. Strack, M. Argyle, & N. Schwarz (Eds.), Subjective Well-
Being: An Interdisciplinary Perspective (pp.49-72). Oxford: Pergamon Press.
Herek, G. M. (1991). Stigma, prejudice, and violence against lesbians and gay men. In
J. C. Gonsiorek & J. D. Weinrich (Eds.), Homosexuality: Research implications
for public policy (pp.60-80). Newbury Park, CA: Sage.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
144
Herzog, A. R., Rodgers, W. L., & Woodworth, J. (1982). Subjective well-being among
different age groups. Ann Arbor, MI: Institute for Social Research.
Hess, R. D., & Camera, K. A. (1979). Post-divorce family relationships as mediating
factors in the consequences of divorce for children. Journal of Social Issues, 35,
79-98.
Hetherington, E. M. (Ed.). (1999). Coping with divorce, single parenting, and remarriage:
A risk and resiliency perspective. Mahway, NJ: Erlbaum.
Hetherington, E. M., Cox M., & Cox, R. (1982). Effects of divorce on parents and
young children. In M. Lamb (Ed.), Nontraditional families: Parenting and child
development (pp.233-288). Hillsdale, NJ: Lawrence Erlbaum.
House, J. S., Kahn, R. L., McLeod, J. D., & Wiliams, D. (1985). Measures and
concepts of social support. In S. Cohen, & S. L. Syme (Eds.), Social support and
health (pp. 83-108). San Diego, CA: Academic Press, Inc.
Howes, P., & Markman, H. J. (1989). Marital quality and child functioning: A
longitudinal investigation. Child Development, 60, 1044-1051.
Huggins, S. L. (1989). A comparative study of self-esteem of adolescent children of
divorced lesbian mothers and divorced heterosexual mothers. In F. W. Bozett
(Ed.), Homosexuality and the Family (pp. 123-135), New York: Harrington Park
Press.
Hughes, T. L., Pollinger Haas, A., Avery, L. (1997). Lesbians and mental health:
Preliminary results from the Chicago Women’s Health Survey. Journal of the Gay
and Lesbian Medical Association, 1(3), 137-148.
Hughes, T. L., Pollinger Haas, A., Razzano, L., Cassidy, R., Matthews, A. (2000).
Comparing lesbians’ and heterosexual women’s mental health: A multi-site
survey. Journal of Gay and Lesbian Social Services, 11(1), 57-76.
Huston, A. (1983). Sex typing. In E. M. Heterington (Ed.), Handbook of child
psychology: Socialization, personality and social development, Vol 4 (pp.387-
467). New York: John Wiley and Sons.
Kahn, R., & Antonnucci, T. (1981). Convoys of social support: A life-course approach.
In S. Keisler, J. Moragn, & V. Oppenheimer (Eds.), Aging: Social Change (pp.
383-403).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
145
Kelly, J. B. (2000). Children’s adjustment in conflicted marriage and divorce: A
decade review of research. Journal of the American Academy of Child and
Adolescent Psychiatry. 39(8), 963-973.
Kingsbury, N., & Scanzoni, J. (1993). Structural-functionalism. In P. G. Boss, W. J.
Doherty, R. LaRossa, W. R., Shumm, & S. K. Steinmetz (Eds.), Sourcebook of
family theories and methods: A contextual approach (pp. 195-217). New York:
Plenum.
Kirkpatrick, M., Smith, C., & Roy, R. (1981). Lesbian mothers and their children: A
comparative survey. American Journal of Orthopsychiatry, 51(3), 545-551.
Lansford, J. E., Ceballo, R., Abbey, A., Stewart, A. J. (2001). Does family structure
matter? A comparison of adoptive, two-parent biological, single-mother,
stepfather, and stepmother households. Journal of Marriage and Family, 63(3),
840-851.
Lemer, R., & Nagai, A. K. (2000). Out of nothing comes nothing: Homosexual and
heterosexual marriage not shown to be equivalent for raising children. Paper
presented at the Revitalizing the Institution of marriage for the 21st Century
conference, Brigham Young University, March, Provo, UT.
Lewinsohn, P., Redner, J., & Seeley, J. (1991). The relationship between life
satisfaction and psychological variables: New perspectives. In F. Strack, M.
Argyle, & N. Schwartz (Eds.), Subjective well-being: An interdisciplinary
perspective (pp.141-169). Elmsford, New York: Pergamon Press.
Lucas, R. E., & Diener, E., & Suh, E. (1996). Discriminant validity of well-being
measures. Journal of Personality and Social Psychology, 71(3), 616-628.
Mancini, J. A., & Blieszner, R. (1989). Aging parents and adult children: Research
themes in intergenerational relations. Journal of Marriage and the Family, 51, 275-
290.
Moos, R. (1974). Family Environment Scale: Preliminary manual. Palo Alto, CA:
Consulting Psychologists Press.
Mucklow, B. M., & Phelan, G. K. (1979). Lesbian and traditional mothers’ responses to
child behavior and self concept. Psychological Reports, 44(3), 880-882.
Parks, C. A. (1998). Lesbian parenthood: A review of the literature. American Journal
of Orthopsvchiatrv. 68(3), 376-389.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
146
Patterson, C. J. (1995). Lesbian mothers, gay fathers, and their children. In A. R.
D’Augelli & C. J. Patterson (Eds.), Lesbian, gay and bisexual identities across the
lifespan: Psychological perspectives (pp. 262-290). New York: Oxford University
Press.
Patterson, C. J. (1994a). Lesbian and gay families, Current Directions in Psychological
Science, 3, 62-64.
Patterson, C. J. (1994b). Children of the lesbian baby boom: Behavioral adjustment,
self-concepts, and sex-role identity. In B. Greene & G. Herek (Eds.),
Contemporary Perspectives on Lesbian and Gay Psychology: Theory, Research
and Applications (pp. 156-175). Beverly Hills, CA: Sage.
Patterson, C. J. (1992). Children of lesbian and gay parents. Child Development, 63,
1025-1042.
Patterson, C. J., & Redding, R. E. (1996). Lesbian and gay families with children:
Implications of social science research policy. Journal of Social Issues. 52(3), 29-
50.
Pedhazur, E. J. (1997). Multiple regression in behavioral research: Explanation and
prediction. Third Edition, London: Wadsworth Thomson Learning.
Perlin, L. (1983). Problems in everyday life survey. Unpublished Manual.
Perrin, E. C., & the Committee on Psychosocial Aspects of Child and Family Health.
(2002). Technical report: Coparent or second-parent adoption by same-sex
parents. Pediatrics, 109(2), 341-344.
Putney, N. M., & Bengtson, V. L. (2001). Families, intergenerational relationships, and
kinkeeping in midlife. In M. Lachman, (Ed). Handbook on Midlife Development,
(pp. 528-570). New York: John Wiley & Sons.
Huggins, S. L. (1989). A comparative study of self-esteem of adolescent children of
divorced lesbian mothers and divorced heterosexual mothers. In F. W. Bozett
(Ed.), Homosexuality and the Family (pp.123-135), New York: Harrington Park
Press.
Radloff, L. (1977). The CES-D scale: A self-report depression scale of research in the
general population. Applied Psychological Measurement. I, 3 85-401.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
147
Rand, C., Graham, D. L. R., & Rawlings, E. I. (1982). Psychological health and
factors the court seeks to control in lesbian mother custody trials. Journal of
Homosexuality, 8(1). 27-38.
Roberts, E. L., & Bengtson, V. L. (1996). Affective ties to parents in early adulthood
and self-esteem across 20 years. Social Psychology Quarterly. 59, 96-106.
Rosenberg, M. (1965). Society and the adolescent self image. Princeton, NJ: Princeton
University Press.
Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods and
data analysis. New York, NY: McGraw-Hill, Inc.
Ross, C. E., & Mirowsky, J. (1999). Parental divorce, life-course disruption, and adult
depression.
Ross, C. E., Mirowsky, J., & Goldsteen, K. (1991). The impact of the family on health:
The decade in review. In A. Booth (Ed.), Contemporary families: Looking
forward, looking back (pp. 341-360). Minneapolis, MN: National Council on
Family Relations.
Rothblum, E. D., & Factor, R. (2001). Lesbians and their sisters as a control group:
Demographic and mental health factors. Psychological Science. 12(1), 63-69.
Shavelson, E., Biaggio, M. K., Cross, H. H., & Lehman, R. E. (1980). Lesbian women’s
perceptions of their parent-child relationships. Journal of Homosexuality, 5(3),
205-215.
Shek, D. T. L. (1998). A longitudinal study of the relations between parent-adolescent
conflict and adolescent psychological well-being. The Journal of Genetic
Psychology, 159(1), 53-67.
Stacey, J., & Biblarz, T. J. (2001). (How) Does the sexual orientation of parents matter?
American Sociological Review, 66, 159-183.
Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3rd Edition),
New York: Harper Collins.
Tasker, F., & Golombok, S. (1998). Growing up in a lesbian family: Effects on child
development. New York: Guilford Press.
Tasker, F., & Golombok, S. (1995). Adults raised as children in lesbian families.
American Journal of Orthopsychiatry, 65, 203-215.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
148
Tasker, F., & Golombok, S. (1994). Children in lesbian and gay families: Theories
and evidence. Annual Review of Sex Research, 5, 73-100.
Taylor, S. E. (1999). Health psvchology(4th Ed.), Boston: McGraw-Hill.
Thompson, N., McCandless, B., Strickland, B. (1971). Personal adjustment of male and
female homosexuals and heterosexuals. Journal of Abnormal Psychology, 78,
237-240.
Voydanoff, P. (1990). Economic distress and family relations: A review of the eighties.
Journal of Marriage and the Family, 52, 1099-1115.
Wallerstein, J. S. (2001). Unexpected legacy of divorce. Saint Paul, MN: High Bridge.
Wardle, L. D. (1997). The potential impact of homosexual parenting on children.
University of Illinois Law Review. 833-919.
Weiss, R. (1984). The impact of marital dissolution on income and consumption in
single-parent households. Journal of Marriage and the Family, 46, 115-127.
Weitzman, L. (1985). The divorce revolution: The unexpected social and economic
consequences for women and children in America. New York: Free Press.
Wortman, C., & Dunkel-Shetter, C. (1987). Conceptual and methodological issues in
the study of social support. In J. E. Singer, & A. Baum (Eds.), Handbook of
psychology and health (pp.63-108). Hillsdale, NJ: Laurence Erlbaum.
White, L. (1992). The effects of parental divorce and remarriage on parental support for
adult children. Journal of Family Issues, 13(2), 234-250.
Zill, N., Morrison, D., & Coiro, M. (1993). Long term effects of parental divorce on
parent-child relationships, adjustment, and achievement in young adulthood.
Journal of Family Psychology, 7, 91-103.
Zweig, R. (2000, August). Androgyny and well-being in adults from non-traditional and
traditional families of origin. Paper presented at the Annual American
Psychological Association, Washington, D.C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
149
Appendix A
Demographics
Background Variables / Demographics
1 .What is your CURRENT marital status or living arrangement? (Check only one.)
Year This
Occurred
□ Married.................................................................................... ..............
□ Separate................................................................................... ..............
□ D ivorced................................................................................. ..............
□ W idow ed................................................................................ _ _ _ _
□ Living with a partner as though married........................ _ _ _ _ _
□ Stopped living with a partner as though married _ _ _ _ _
□ Single, never married...........................................................
□ Other: ......................................
2. Who lives with you? (Check as many as apply.)
□ N o one else— live alone
□ Your parentl:__________________________
□ Your parent2:_____________ ____________
□ Your parent:____________________________(first name)
□ Your parent:___________________________ (first name)
□ Your brother(s) or sister(s)
□ Your grandchild(ren) (How many?_________ )
□ Your partner/spouse
□ Your child(ren) (How many?_______ )
□ Friend(s) or roommate(s) (How many? )
□ Other relative (Who? )
□ Other (Who? )
3. Do you have any children?
□ No
□ Yes -» If yes, how many children do you have? (Please include biological,
guardianship, adopted, and step-children, etc.)_______
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
150
4. How old were you when your parents divorced? ____ (Question for children)
5. What is your date o f birth? ____ / /
Mo. Day Year
6. What is your sex?
□ M aleD Female
7. What is your racial or ethnic background?
8. What was YOUR TOTAL HOUSEHOLD INCOME
for last year? (Check one box).
□ Less than $10,000 □ $80,000 - $89,999
□ $10,000- $19,999 □ $90,000 - $99,999
□ $20,000 - $29,999 □ $100,000-- $109,999
□ $30,000 - $39,999 □ $110,000--$119,999
□ $40,000 - $49,999 □ $120,000--$129,999
□ $50,000 - $59,999 □ $130,000--$139,999
□ $60,000 - $69,999 □ $140,000-- $149,999
9. What is the highest level o f education you have
attained?
□ 8th grade or less
□ Some high school (9th - 11th grade)
□ GED certificate
□ High school or vocational school graduate
□ Specialized technical, business, or,other training after high school
□ Some college (1-3 years)
□ College or university graduate
□ One or more academic years beyond college, including M.A.
□ Post-graduate degree (Ph.D., M.D., J.D.)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
151
Appendix B
Lesbigay Mother’s Age at Coming Out
1. When did you first feel an attraction to a member of the same sex?
(age)
2. What age did you first come out to yourself?
(age)
3. When did you have your first experience/relationship with a person of
the same sex? (age)
4. When did you first tell another person about your GLB orientation?
(age)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
152
Appendix C
Family Relationship Quality
Mother’ s Perception o f Affectual Solidarity
1. Taking everything into consideration, HOW CLOSE do you feel the relationship
between you and your “study” child is, at this point in your life?
□ N ot at all close
□ Not too close
□ Somewhat close
□ Pretty close
□ Very close
□ Extremely close
2. How is COMMUNICATION between you and your “study” child — exchanging ideas
or talking about things that really concern you at this point in your life?
□ Not at all good
□ Not too good
□ Somewhat good
□ Pretty good
□ Very good
3. Overall, how well do you GET ALONG WITH your “study” child at this point in
your life?
□ Not at all well
□ Not too well
□ Somewhat well
□ Pretty well
□ Very well
□ Extremely well
4. How well do you feel that YOU UNDERSTAND your “study” child?
□ Not at all well
□ Not too well
□ Somewhat well
□ Pretty well
□ Very well
□ Extremely well
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
153
5. How much AFFECTION do you feel toward your “study” child?
□ None at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
Children’ s Perception o f Affectual Solidarity
1. Taking everything into consideration, HOW CLOSE do you feel the relationship
between you and your “study” parents is, at this point in your life?
Parent
□ Not at all close
□ Not too close
□ Somewhat close
□ Pretty close
□ Very close
□ Extremely close
2. How is COMMUNICATION between you and your parents— exchanging ideas or
talking about things that really concern you at this point in your life?
Parent
□ Not at all good
□ Somewhat good
□ Pretty good
□ Very good
□ Extremely good
□ Extremely good
3. Overall, how well do you GET ALONG WITH your parent(s) at this point in your
life? Parent
□ Not at all well
□ Not too well
□ Somewhat well
□ Pretty well
□ Very well
□ Extremely well
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4. How w ell do you feel that YOU UNDERSTAND your parent(s)?
Parent
□ Not at all well
□ Not too well
□ Somewhat well
□ Pretty well
□ Very well
□ Extremely well
5. How much AFFECTION do you feel toward your parent(s)?
Parent
□ None at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
155
M other’ s Perception o f Conflict
1. Taking everything into
consideration, how much
CONFLICT, TENSION, or
DISAGREEMENT do you feel there
is between you and this child
at this point in your life?
□ None at ail
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
2. How much do you feel this
child is CRITICAL of
you or what you do?
□ Not at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
3. How much does this child
ARGUE with you?
□ Not at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
Child’ s Perception o f Conflict
1. Taking everything into
consideration, how much
CONFLICT, TENSION, or
DISAGREEMENT do you feel there
is between you and your parent(s)
at this point in your life?
Parent
□ None at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
2. How much do you feel your
parent(s) are CRITICAL of
you or what you do?
Parent
□ Not at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3. How much do your parent(s)
ARGUE with you?
Parent
□ Not at all
□ A little
□ Some
□ Pretty much
□ Quite a bit
□ A great deal
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Mother’ s Perception o f Emotional Support
157
Now we want to talk about friends and relatives other than parents. For each type of
help and support listed below, put a check in the box beneath each person to whom
YOU PROVIDE that kind of assistance or support. (Check as many as apply.)
Partner/ "Study" Other Other
"Study"
grand-
"Study"
grand-
Other
relatives/ Friends/
Profes
sionals
spouse child Son
hi
daughter parentl
hi
parent2 Brother Sister
hi hi
in-laws Neighbors and
paid
helpers
Inform ation
and advice ...
□ □ □ □ □ □ □ □ □ □
□
E m otional
su p p o rt........
□ □ □ □ □ □ □ □ □ □
□
Discussing
im portant
life
decision s......
□ □ □ □ □ □ □ □ □ □ □
2. Turning things around the other way, we're interested in learning about the kinds of
help and support YOU RECEIVE from family, friends, and others. For each type of
help and support listed below, put a check in the box beneath each person who gives
you that kind of assistance or support. (Check as many as apply.)
Partner/
spouse
Inform ation
and a d v ic e .. q
Em otional
Support q
D iscussing
im portant q
life
d ecisio n s......
’Study"
child
□
□
□
Profes-
"Study" "Study" Other sionals
Other Other grand- grand- relatives/ Friends/ and
Son Daughter parentl parent2 Brother Sister in-laws neighbors paid
£sl h i jsl {si helpers
□
□ □
□ □
□
□
□
□ □ □ □ □ □
□ □ □ □ □ □
□ □ □ □ □ □
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
158
C hild’ s Perception o f Emotional Support
1. We're interested in learning about the kinds o f help and support exchanged between
adult children and their parents. In the first two columns, check the boxes indicating
the types o f help and support you PROVIDE TO your parentl and parent2. In the
second two columns, check the boxes indicating the types of help and support you
RECEIVE FROM your parentl and parent2.
Information and advice......................
Emotional support................................
Discussing important life decisions .
YOU PROVIDE YOU RECEIVE
TO FROM
"Study" "Study” "Study" "Study"
Parentl Parent2 Parentl Parent2
□ □ □ □
□ □ □ □
□ □ □ □
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
159
Appendix D
Well-Being Measures
Depressive Symptoms (CES-D)
For each o f the following statements, check the box that best describes HOW OFTEN YOU
HAVE FELT THIS WAY DURING THE PAST WEEK.
Rarely A A M ost
or none little moderate or all
DURING THE PAST WEEK: of the o f the amount of o f the
time time the time time
a. I was bothered by things that don’t usually bother me □ □ □ □
b. I did not feel like eating; my appetite was poor □ □ □ □
c. I felt that I could not shake the blues even with help
from my family or friends □ □ □ □
d. I felt that I was just as good as other people □ □ □ □
e. I had trouble keeping my mind on what I was doing □ □ □ □
f. I felt depressed □ □ □ □
g-
I felt that everything I did was an effort □ □ □ □
h. I felt hopeful about the future □ □ □ □
I. I thought my life had been a failure □ □ □ □
j-
I felt fearful □ □ □ □
k. My sleep was restless □ □ □ □
1 . I was happy □ □ □ □
m. I talked less than usual □ □ □ □
n. I felt lonely □ □ □ □
0. People were unfriendly □ □ □ □
P-
I enjoyed life □ □ □ □
q -
I had crying spells □ □ □ □
r. I felt sad □ □ □ □
s. I felt that people disliked me □ □ □ □
t. I could not “get going” □ □ □ □
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
160
Negative and Positive Affect (BAB)
6. The following sentences describe how people sometimes feel about their lives. DURING
THE PAST FEW WEEKS, did you ever feel: (C heck one box for each question.)
Y es No
a. Particularly excited or interested in something? □ □
b. So restless that you couldn't sit long in a chair? □ □
c. Proud because someone complimented you on
something you had done? □ □
d. Very lonely or remote from other people? □ □
e. Pleased about having accomplished something? □ □
f. Bored? □ □
g. On top o f the world? □ □
h. Depressed or very unhappy? □ □
i. That things were really going your way? □ □
j. Upset because someone criticized you? □ □
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
161
Self-Esteem (Rosenburg Self-Esteem Scale)
AT THIS POINT IN YOUR LIFE. Please indicate how much YOU AGREE or DISAGREE
with each statement.
Strongly
agree Aaree Disagree
Strong
disaar
I feel that I'm a person o f worth, at least □ □ □ □
on an equal basis with others..............................
I feel that I have a number o f good qualities... □ □ □ □
At times I think I am no good at all................... □ □ □ □
I wish I could have more respect for m yself... □ □ □ □
I am able to do things as well as most other □ □ □ □
p eople........................................................................
I certainly feel useless at tim es........................... □ □ □ □
I feel I do not have much to be proud o f .......... □ □ □ □
I take a positive attitude toward m yself............ □ □ □ □
On the whole, I am satisfied with m yself........ □ □ □ □
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Appendix E
Correlation Matrices of All Main Variables for Full Sample
Table E-a
Transformed Family Relationship Quality for the Full Sample.
1 2 3 4 5 6 7 8 9
tel -0.46**** 0.45**** 0.74**** -0.35*** 0.39**** 0.81**** -0.44** * * 0.40****
(112) (113) (113) (113) (113) (113) (112) (113)
ten -0.09 -0.45**** 0.80**** -0.06 -0.26** 0.82**** -0.10
(112) (112) (112) (112) (112) (112) (112)
tes 0.32*** 0.00 0.74**** 0.37**** -0.09 0.80****
(114) (114) (114) (113) (112) (114)
tmocl -0.47**** 0.22* 0.37**** -0.36*** 0.28**
(134) (134) (113) (112) (114)
tmocn 0.02 -0.14 0.44* * * * 0.01
(134) (113) (112) (114)
tmoes 0.33*** -0.08 0.37****
(113) (112) (114)
tchcl -0.35**** 0.31***
(121) (122)
tchcn -0.08
(121)
tches --
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support
tmocl = transformed mother’s closeness tmocn = transformed mother’s conflict tmoes = transformed mother’s emotional support tchcl “ transformed child’s closeness
tchcn = transformed child’s conflict tches = transformed child’s emotional support *p< 05 **p< 01 ***j j<.001 ****£<.0001.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table E-b
Transformed Family Relationship Quality and Well-Being Variables for the Full Sample.
tel ten tehs tmocl tmocn tmoes tchcl tchcn
tches
tmocesd -0 . 0 2 0.24* -0.06 -0.14 0.17 -0 . 0 2 0.07 0.17
-0.05
( 1 1 2 ) ( 1 1 1 ) (113) (133) (133) (133) ( 1 1 2 ) (1 1 1 ) (113)
tmonegaff -0 . 1 1 0 .2 2 * -0.03 -0.19* 0.16 -0.09 -0 . 0 1 0.04
-0 . 0 0
( 1 1 2 ) ( 1 1 2 ) (113) (133) (133) (133) ( 1 1 2 ) ( 1 1 2 ) (113)
tmoposaff 0.13 -0.04 0.18 0.13 -0.06 0 .2 2 * 0.07 0 . 0 0
0.09
( 1 1 2 ) ( 1 1 2 ) (113) (133) (133) (133) ( 1 1 2 ) ( 1 1 2 ) (113)
tmose 0.08 -0.04 0.17 0.05 -0 . 0 1 0.19* -0 . 0 1 0.06
0.18
( 1 1 2 ) ( 1 1 1 ) (113) (133) (133) (133) ( 1 1 2 ) ( 1 1 1 ) (113)
tchcesd -0 .2 2 * 0.18 -0.07 -0.18 0.18 -0.14 -0.14 0.13
-0.06
(113) (1 1 2 ) (114) (114) (114) (114) (1 2 1 ) ( 1 2 0 )
( 1 2 2 )
chnegaff -0.24** 0.23* 0 . 0 1 -0.18 0.15 -0.09 -0.24** 0 .2 0 *
-0.05
(113) ( 1 1 2 ) (114) (114) (114) (114) ( 1 2 2 ) ( 1 2 1 )
(123)
tchposaff 0.17 -0 . 1 2 0.05 0.18 -0.13 0 . 0 2 0.06 -0.05
0 . 1 0
(113) ( 1 1 2 ) (114) (114) (114) (114) ( 1 2 2 ) ( 1 2 1 )
(123)
tchse 0 .2 0 * -0.19 0.19* 0.15 -0.19* 0.13 0.15 -0 . 1 1
0.16
( 1 1 1 ) ( 1 1 0 ) ( 1 1 2 ) ( 1 1 2 ) (1 1 2 ) ( 1 1 2 ) ( 1 2 0 ) (119) ( 1 2 1 )
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support
tmocl = transformed m other’s closeness tmocn = transformed m other’s conflict tmoes = transformed m other’s emotional support tchcl “ transformed child’s closeness
tchcn = transformed child’s conflict tches ■ = transformed child’s emotional support tmocesd = transformed m other’s depressive symptoms tm onegaff= transformed m other’s negative affect
tm oposaff = transformed m other’s positive affect tmose = transformed m other’s self-esteem tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect
tm oposaff = transformed child’s positive affect tchse = transformed child’s self-esteem *£<.05 **£<.01 ***£<.001 ****£<.0001.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table E-c
Transformed Well-Being Variables for the Full Sample.
1 2 3 4 5 6 7 8
tmocesd
57**** _ jg****
- 5 6 ***** .22* .15 -.17 -.08
(132) (132) (132) (113) (113) (113) (111)
tmonegaff
-31***
.13 .23* .02 -.04
(133) (132)
(113) (113) (113) (111)
tmoposaff
—
5 2****
-.13 -.16 .09 . 0 0
(132) (113) (113) (113)
(HI)
tmose - . 1 1 -.07 .15 . 0 0
(113) (113) (113)
( 1 1 1 )
tchcesd
—
5 ^****
38****
_ 5 4 ****
( 1 2 2 ) ( 1 2 2 ) ( 1 2 0 )
chnegaff
— -.15
_ 4 4 ****
(123) ( 1 2 1 )
tchposaff
—
^5****
( 1 2 1 )
tchse
—
Note, tmocesd * transformed mother's depressive symptoms tmonegaff = transformed mother’s negative affect tmoposaff = transformed mother’s positive affect
tmose = transformed mother’s self-esteem tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tchposaff = transformed child’s
positive affect tchse = transformed child’s self-esteem *g<.05 **p<.01 ’ ***£<.001 ****£<.0001.
O n
4= -
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Appendix F
Correlation Matrices of All Main Variables for Lesbian Mothers and Their Children
Table F-a
Transformed Family Relationship Q uality for the Lesbian Mothers and Their Children.
1 2 3 4 5 6 7 8 9
tel -0.48*** 0.56**** 0.75**** -0.36* 0.45** 0.88**** -0.44** 0.42**
(47) (48) (48) (48) (48) (48) (47) (48)
ten -0.05 -0.49*** 0.81**** -0.12 -0.28a 0.81**** 0.09
(47) (47) (47) (47) (47) (47) (47)
tes 0.45** 0.03 0.63**** 0.49*** -0.10 0.75****
(48) (48) (48) (48) (47) (48)
tmocl -0.55**** 0.26* 0.46*** -0.37* 0.29*
(68) (68) (48) (47) (48)
tmocn -0.03 -0.12 0.40** 0.12
(68) (48) (47) (48)
tmoes 0.40** -0.16 0.13
(48) (47) (48)
tchcl -0.39** 0.38**
(56) (57)
tchcn 0.07
(56)
tches
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support
tmocl = transformed mother’s closeness tmocn = transformed mother’s conflict tmoes = transformed mother’s emotional support tchcl “transformed child’s closeness
tchcn = transformed child’s conflict tches = transformed child’s emotional support *p<.05 **p<.01 ***p<.001 ****p<.0001.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table F-b
Transformed Family Relationship Q uality with Well-Being Variables for the Lesbian Mothers and Their Children.
tel ten tes tmocl tmocn tmoes tchcl tchcn tches
tmocesd 0.07 0.07 -0.03 -0.09 -0 . 0 2 0.05 0.15 0 . 1 2 -0.09
(47) (46) (47) (67) (67) (67) (47) (46) (47)
tmonegaff 0.05 0.13 -0.14 -0.05 0.06 -0.13 0 . 1 1 0.09 -0.08
(47) (47) (47) (67) (67) (67) (47) (47) (47)
tmoposaff 0.18 -0.04 0.09 0,19 -0.05 0 . 1 0 0.05 -0.04 0 . 1 0
(47) (47) (47) (67) (67) (67) (47) (47) (47)
tmose -0 . 0 2 0.06 0 . 1 0 0.09 0.04 0 . 0 2 -0 . 1 1 0 . 1 1 0.13
(48) (47) (48) (6 8 ) (6 8 ) (6 8 ) (48) (47) (48)
tchcesd -0.32* 0.26 -0.08 -0 . 1 2 0.17 -0.27 -0.33* 0.24 -0.07
(48) (47) (48) (48) (48) (48) (56) (55) (56)
chnegaff -0.25 0.26 -0.08 -0 . 0 1 0.06 -0.29* -0.27* 0.27* 0.07
(48) (47) (48) (48) (48) (48) (57) (56) (57)
tchposaff 0.23 0.04 0.14 0 . 1 2 0 . 0 0 0.16 0 . 2 0 0 . 1 0 0.17
(48) (47) . (48)
(48) (48) (48) (57) (56) (57)
tchse 0.40** -0.23 0.30* 0 . 2 0 -0.26 0.30* 0.32* -0.13 0.19
(47) (46) (47) (47) (47) (47) (56) (55) (56)
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support tmocl = transformed mother’s closeness tmocn = transformed m other’s
conflict tmoes = transformed m other’s emotional support tchcl tra n sfo rm e d child’s closeness tchcn = transformed child’s conflict tches = transformed child’s emotional support tmocesd =
transformed m other’s depressive sym ptom s tm onegaff = transformed m other’s negative affect tm oposaff= transformed m other’s positive affect tmose = transformed m other’s self-esteem
tchcesd = transformed child’s depressive symptoms chnegaff= child’s negative affect tm oposaff= transformed child’s positive affect tchse = transformed child’s self-esteem ^
*p<05 **e<.01. O N
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table F-c
Transformed Well-Being Variables for the Lesbian Mothers and Their Children.
1 2 3 4 5 6 7 8
tmocesd .63**** -.34**
m 5 5 ****
.13 .09 . 0 0 .09
(6 6 ) (6 6 ) (67) (47) (47) (47) (46)
tmonegaff
— -.24
_ 42***
.07 .14 .14 .04
(67) (67) (47) (47) (47) (46)
tmoposaff
—
4g****
.03 -.08 - . 0 1 - . 1 1
(67) (47) (47) (47) (46)
tmose -.05 - . 1 1 .13 -.03
- (48) (48) (48) (47)
tchcesd
—
gg****
- . 2 2 - 59****
(56) (56) (55)
chnegaff
— -.18
_ 5 4 ****
(57) (56)
tchposaff
~ .31*
(56)
tchse -
Note, tmocesd = transformed mother’s depressive symptoms tmonegaff = transformed mother’s negative affect tmoposaff= transformed mother’s positive affect
tmose = transformed mother’s self-esteem tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tchposaff= transformed child's
positive affect tchse = transformed child’s self-esteem *g<05 **g<,01 ***g<.001 ****£<.0001. -.j
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Appendix G
Correlation Matrices of All Main Variables for Heterosexual Mothers and Their Children
Table G a
Transformed Family Relationship Quality for the Heterosexual Mothers and Their Children
1 2 3 4 5 6 7 8 9
tel -0.46***
(65)
0.38**
(65)
0.78****
(65)
-0.35**
(65)
0.35**
(65)
0.79****
(65)
-0.44***
(65)
0.38**
(65)
ten -- -0.10
(65)
-0.45***
(65)
0.81****
(65)
-0.02
(65)
-0.25*
(65)
0.82****
(65)
-0.20
(65)
tes -- 0.30*
(66)
-0.02
(66)
0.81****
(66)
0.25*
(65)
-0.09
(65)
0.83****
(66)
tmocl -- -0.41***
(66)
0.27*
(66)
0.41***
(65)
-0.36**
(65)
0.31*
(66)
tmocn 0.05
(66)
-0.17
(65)
0.47****
(65)
-0.07
(66)
tmoes
'
0.23a
(65)
-0.05
(65)
0.51****
(66)
tchcl -- -0.33**
(65)
0.23a
(65)
tchcn -- -0.17
(65)
tches --
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support
tmocl = transformed mother’s closeness tmocn = transformed mother’s conflict tmoes = transformed mother’s emotional support tchcl =transformed child’s closeness ^
tchcn = transformed child’s conflict tches = transformed child’s emotional support *p<.05 **p<.01 ***p<.001 ****p<.0001. 00
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table G-b
Transformed Family Relationship Quality with Well-Being Variables for the Heterosexual Mothers and Their Children.
tel ten tes tmocl tmocn tmoes tchcl tchcn tches
tmocesd -0.06 0.31* -0.07 -0.18 0.30* -0.05 0.03 0.19 -0 . 0 2
(65) (65) (6 6 ) (6 6 ) (6 6 ) (6 6 ) (65) (65) (6 6 )
tmonegaff -0 . 2 1 0.27* 0.06 -0.35** 0.27* 0 . 0 0 -0.08 0 . 0 2 0.07
(65) (65) (6 6 ) (6 6 ) (6 6 ) (6 6 ) (65) (65) (6 6 )
tmoposaff 0 . 1 0 -0.04 0 . 2 1 0.19 -0.09 0.26* 0 . 0 2 0 . 0 2 0.06
(65) (65) (6 6 ) (6 6 ) (6 6 ) (6 6 ) (65) (65) (6 6 )
tmose 0.14 -0 . 1 0 0 . 2 1 0.09 -0 . 1 0 0.29* 0 . 0 2 0 . 0 2 0 . 2 0
(64) (64) (65) (65) (65) (65) (64) (64) (65)
tchcesd -0.16 0.14 -0.05 -0.23 0.18 -0.05 0.03 0.06 -0.05
(65) (65) (6 6 ) (6 6 ) (6 6 ) (6 6 ) (65) (65) (6 6 )
chnegaff -0.23 0 . 2 1 0.08 -0.29* 0 . 2 2 0.08 -0.19 0.15 0.05
(65) (65) (6 6 ) (6 6 ) (6 6 ) (6 6 ) (65) (65) (6 6 )
tchposaff 0.14 -0 . 2 1 0 . 0 1 0 . 2 0 -0 . 2 2 -0.05 -0.06 -0.15 0.06
(65) (65) (6 6 ) (6 6 ) (6 6 ) (6 6 ) (65) (65) (6 6 )
tchse 0.07 -0.17 0 . 1 2 0 . 1 1 -0.14 0.05 0 . 0 0 -0.09 0.14
(64) (64) (65) (65) (65) (65) (64) (64) (65)
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support tmocl = transformed m other’s closeness tmocn = transformed m other’s
conflict tmoes = transformed m other’s emotional support tchcl = transformed child’s closeness tchcn = transformed child’s conflict tches = transformed child’s emotional support tmocesd =
transformed m other’s depressive symptoms tm onegaff = transformed m other’s negative affect tm oposaff = transformed m other’s positive affect tmose = transformed m other’s self-esteem
tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tchposaff= transformed child’s positive affect tchse = transformed child’s self-esteem *p< 05 ,_,
**o<.01. . O j
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table G-c
Transformed W ell-Being Variables fo r the H eterosexual Mothers and Their Children.
1 2 3 4 5 6 7 8
tmocesd — .53**** -.37** - 60**** .25* .18 -.26* -.18
(6 6 ) (6 6 ) (65) (6 6 ) (6 6 ) (6 6 ) (65)
-taaovieaaff -.34** .16 .30* -.06 - . 1 2
(6 6 ) (65) (6 6 ) (6 6 ) (6 6 ) (65)
tmoposaff
—
.48**** - . 2 1 -.19 .18 .09
(65) (6 6 ) (6 6 ) (6 6 ) (65)
tmose —
-.13 - . 0 1 .19 .05
(65) (65) (65) (64)
tchcesd
___
5 3 ****
- 48**** _ 5 2 ****
(6 6 ) (6 6 ) (65)
chnegaff —
-.13
_ ^y**
(6 6 ) (65)
tchposaff — .38**
(65)
tchse
—
Note, tmocesd = transformed mother’s depressive symptoms tmonegaff = transformed mother’s negative affect tmoposaff = transformed mother’s positive affect
tmose = transformed mother’s self-esteem tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tchposaff = transformed child’s positive affect tchse =
transformed child’s self-esteem *p<.05 **p<.01 ***p<.001 ****j><.0001.
O
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Appendix H
Correlation Matrices for Single Mothers and their children
Table H-a
Transform ed Fam ily Relationship Q uality fo r Single M others and Their Children
1 2 3 4 5 6 7 8 9
tel - -0.53** 0.23 0.80**** -0.48** 0.23
Q
-0.53** 0.26
(29) (29) (29) (29) (29) (29) (29) (29)
ten — -0.03 -0.48** 0.85**** -0.15 -0.21 0.89**** -0.16
(29) (29) (29) (29) (29) (29) (29)
tes 0.28 0.07
Q
0.13 -0.12
q 7 9 * * * *
(29) (29) (29) (29) (29) (29)
tmocl
— -0.49** 0.30 0.36 -0.42* 0.26
(34) (34) (29) (29) (29)
tmocn — -0.17 -0.23 0.61*** 0.03
(34) (29) (29) (29)
tmoes — 0.07
(29)
-0.19
(29)
0.43*
(29)
tchcl - -0.29
(29)
0.15
(29)
tchcn - -0.29
(29)
tches
—
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support tmocl = transformed mother’s closeness tmocn = transformed mother’s
conflict tmoes = transformed mother’s emotional support tchcl = transformed child’s closeness tchcn = transformed child’s conflict tches = transformed child’s emotional support
*£<•05 **£<.01 ***£<.001.
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table H- b
Correlation M atrix o f Transform ed Fam ily Relationship Q uality and W ell-Being fo r Single M others and Their Children
tel ten tes tmocl tmocn tmoes tchcl tchcn tches
tm ocesd -0.22 0.20 -0.18 -0.30 0.25 -0.36* -0.04 0.15 0.04
(29) (29) (29) (34) (34) (34) (29) (29) (29)
tm onegaff -0.29 0.23 0.23 -0.25 0.27 -0.05 -0.19 0.05 0.27
(29) (29) (29) (34) (34) (34) (29) (29) (29)
tm oposaff -0.03 0.12 0.22 0.09 0.00 0.45** 0.03 0.07 0.02
(29) (29) (29) (34) (34) (34) (29) (29) (29)
tm ose -0.12 0.05 0.12 -0.09 -0.07 0.35* -0.08 0.07 -0.01
(29) (29) (29) (34) (34) (34) (29) (29) (29)
tchcesd 0.10 0.17 0.15 -0.14 0.21 -0.17 0.26 0.15 0.27
(29) (29) (29) (29) (29) (29) (29) (29) (29)
chnegaff 0.02 0.18 0.34 0.06 0.24 0.10 -0.07 0.15 0.32
(29) (29) (29) (29) (29) (29) (29) (29) (29)
tchposaff -0.07 -0.03 0.03 0.09 -0.12 0.07 -0.28 0.07 0.02
(29) (29) (29) (29) (29) (29) (29) (29) (29)
tchse 0.07 -0.21 0.16 0.12 -0.14 0.25 0.07 -0.21 0.04
(29) (29) (29) (29) (29) (29) (29) (29) (29)
Note, tel— transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support tmocl = transformed m other’s closeness tmocn = transformed m other’s
conflict tmoes = transformed m other’s emotional support tchcl = transformed child’s closeness tchcn = transformed child’s conflict tches = transformed child’s emotional support tmocesd =
transformed m other’s depressive symptoms tm onegaff= transformed m other’s negative affect tm oposaff = transformed m other’s positive affect tmose = transformed m other’s self-esteem
tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tm oposaff = transformed child’s positive affect tchse = transformed child’s self-esteem *p<.05
* * p < 0 1 - K>
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table H -c
Correlation M atrix o f Transform ed W ell-Being Variables fo r Single M others and Their Children
1 2 3 4 5 6 7 8
tm ocesd — 0.55*** -0.33 -0.67**** 0.36 0.31 -0.24 -0.23
(34) (34) (34) (29) (29) (29) (29)
tm onegaff -0.38* -0.63**** 0.17 0.38* 0.24 -0.09
(34) (34) (29) (29) (29) (29)
tm oposaff 0 62**** -0.24 -0.36 0.05 0.24
(34) (29) (29) (29) (29)
tmose
_
-0.08 -0.34 0.07 0.12
(29) (29) (29) (29)
tchcesd
_
0.43* -0.42* -0.53**
(29) (29) (29)
chnegaff
— 0.08 -0.33
(29) (29)
tchposaff — 0.31
(29)
tchse
_
Note, tmocesd = transformed mother’s depressive symptoms tmonegaff = transformed mother’s negative affect tmoposaff = transformed mother’s positive affect
tmose = transformed mother’s self-esteem tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tchposaff = transformed child’s
positive affect tchse = transformed child's self-esteem *p<.05 **£<.01 ***£<.001 ****£<.0001.
■ - J
O J
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Appendix I
Correlations Matrices for Married/Partnered Mothers and Their Children
Table I-a
Correlation Matrices ofFamily Relationship Quality fo r Married/Partnered Mothers and Their Children
1 2 3 4 5 6 7 8 9
Tel - -0.43****
0 **** Q 7| ****
-0.29**
0 45**** 0 gj****
-0.40***
0 43****
(82) (83) (83) (83) (83) (83) (82) (83)
Ten — -0.10 -0 43****
0 7g****
-0.03 -0.27*
0 jg****
-0.07
(82) (82) (82) (82) (82) (82) (82)
Tes
—
0.33*** -0.01 0 73**** 0.44****
-0.08 0.80****
(84) (84) (84) (83) (82) (84)
Tmocl — -0.45**** 0.19 0.37*** -0.33** 0.28*
(99) (99) (83) (82) (84)
Tmocn
— 0.08 -0.10 0.37*** 0.02
(99) (83) (82) (84)
Tmoes
—
0 43***
-0.06 0.36***
(83) (82) (84)
Tchcl
— -0.34** 0.34**
(82) (83)
Tchcn - 0.00
(82)
Tches
-
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support mod = transformed mother’s closeness tmocn = transformed mother’s
conflict tmoes = transformed mother's emotional support tchcl =transformed child’s closeness chcn = transformed child’s conflict tches = transformed child’s emotional support *p< 05 i — i
***E<.001 **p<.0001. ^
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table I-b
Correlation M atrix o f Fam ily Relationship Q uality and W ell-Being fo r M arried/Partnered M others and Their Children
tel ten tes tm ocl tmocn tmoes tchcl tchcn tches
tm ocesd 0.10 0.25* 0.02 -0.07 0.14 0.15 0.13 0.19 -0.06
(82) (81) (83) (98) (98) (98) (82) (81) (83)
tm onegaff -0.04 0.23* -0.08 -0.18 0.13 -0.05 0.05 0.06 -0.07
(82) (82) (83) (98) (98) (98) (82) (82) (83)
tm oposaff 0.16 -0.10 0.15 0.14 -0.08 0.11 0.08 -0.04 0.11
(82) (82) (83) (98) (98) (98) (82) (82) (83)
tmose 0.12 -0.07 0.16 0.09 0.00 0.08 0.00 0.04 0.24*
(82) (81) (83) (98) (98) (98) (82) (81) (83)
tchcesd -0.29** 0.17 -0.10 -0.17 0.15 -0.13 -0.20 0.10 -0.10
(83) (82) (84) (84) (84) (84) (83) (82) (84)
chnegaff -0.31** 0.23* -0.07 -0.24* 0.11 -0.16 -0.27* 0.20 0.00
(83) (82) (84) (84) (84) (84) (83) (82) (84)
tchposaff 0.27* -0.16 0.07 0.21 -0.13 0.03 0.15 -0.10 0.13
(83) (82) (84) (84) (84) (84) (83) (82) (84)
tchse 0.25* -0.17 0.20 0.16 -0.21 0.11 0.14 -0.04 0.18
(81) (80) (82) (82) (82) (82) (81) (80) (82)
Note. tcl= transformed dyadic closeness ten = transformed dyadic conflict tes = transformed dyadic emotional support tmocl = transformed m other’s closeness tmocn = transformed m other’s
conflict tmoes = transformed m other’s emotional support tchcl = transformed child’s closeness tchcn = transformed child's conflict tches = transformed child’s emotional support tmocesd =
transformed m other’s depressive symptoms tm onegaff = transformed m other's negative affect tm oposaff = transformed m other’s positive affect tmose = transformed m other’s self-esteem
tchcesd = transformed child’s depressive symptoms chnegaff = child’s negative affect tchposaff = transformed child’s positive affect tchse = transformed child’s self-esteem *p< 05
* *E < .01 * * *jj< .001 * * * *j><.0001
Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission.
Table I-c
Correlation M atrix o f W ell-being Variables fo r M arried/Partnered M others and Their Children
1 2 3 4 5 6 7 8
tm ocesd - 0.55**** -0.30** -0.48**** 0.15 0.11 -0.18 -0.05
(97) (97) (97) (83) (83) (83) (81)
tm onegaff -0.24* -0.33** 0.12 0.23* -0.08 -0.05
(98) (97) (83) (83) (83) (81)
tm oposaff
— 0 42**** -0.09 -0.13 0.15 -0.06
(97) (83) (83) (83) (81)
tmose
- -
-0.12 -0.02 0 .2 2 * -0.01
(83) (83) (83) (81)
tchcesd
__
0.62**** -0.39*** -0 5 7 ****
(84) (84) (82)
chnegaff
— -0.18 -0 4 7 ****
(84) (82)
tchposaff
— 0.35**
(82)
tchse
Note, tmocesd = transformed mother’s depressive symptoms tmonegaff= transformed mother’s negative affect tmoposaff = transformed mother's positive affect
tmose = transformed mother’s self-esteem tchcesd = transformed child's depressive symptoms chnegaff = child’s negative affect tchposaff = transformed child’s
positive affect tchse = transformed child’s self-esteem *p<.05 **p<.01 ***p<001 ****g<0001.
177
Appendix K
Follow-up Models for Hypothesis I
Table K-a.
Summary o f Hierarchical Regression Models o f Transformed Mother’ s Affectual
Solidarity With Child on Demographic, Mother’ s Sexual Orientation with Beta Weights.
Variable ' B SE B B "
Step 1
Child’s partner status 0.13 0.08 .21
Child’s parental status -0.02 0.08 -.02
Child’s age at divorce 0.00 0.00 .11
Mother’s income -0.01 0.04 -.01
Step 2
Child’s partner status 0.13 0.08 .19
Child’s parental status -0.03 0.08 -.05
Child’s age at divorce 0.00 0.00 .10
Mother’s income 0.00 0.05 .01
Mother’s sexual 0.07 0.07 .11
orientation
Note. R2 = .0562 for Step 1; A R2 = .0092 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table K-b.
Summary o f Hierarchical Regression Models o f Transformed M other’ s Conflict With
Child on Demographic, M other’ s Sexual Orientation with Beta Weights.
Variable B SEB fi
Step 1
Child’s partner status -0.02 0.02 -.12
Child’s parental status -0.01 0.02 -.07
Child’s age at divorce 0.00 0.00 -.12
Mother’s income 0.00 0.01 .03
Step 2
Child’s partner status -0.02 0.02 -.13
Child’s parental status -0.02 0.02 -.10
Child’s age at divorce 0.00 0.00 -.13
Mother’s income 0.01 0.01 .06
Mother’s sexual 0.02 0.02 .13
orientation
Note R2 = .0557for Step 1; A R2= 0131 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
179
Table K-c.
Summary o f Hierarchical Regression Models o f Transformed M other’ s Emotional
Support on Demographic, M other’ s Sexual Orientation with Beta Weights.
_ _ _ _ _ _ _ _ _ _ _ _ _ £
Step 1
Child’s partner status -0.10 0.08 -.15
Child’s parental status 0.01 0.09 .01
Child’s age at divorce 0.00 0.00 -.03
Mother’s income 0.12 0.05 .25
Step 2
Child’s partner status -0.09 0.08 -.12
Child’s parental status 0.04 0.09 .05
Child’s age at divorce 0.00 0.00 .00
Mother’s income 0.10 0.05 .21
Mother’s sexual -0.14 0.07 -.20
orientation
Note R2 = .0897 for Step 1; A R2 = .0321 (ps > .05) * p < .05
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table K-d.
Summary o f Hierarchical Regression Models o f Transformed Child’ s Ajfectual
Solidarity on Demographic, M other’ s Sexual Orientation with Beta Weights.
Variable B SEB B
Step 1
Child’s partner status 0.07 0.06 .12
Child’s parental status -0.06 0.07 -.10
Child’s age at divorce 0.00 0.00 -.02
Step 2
Child’s partner status 0.07 0.06 .13
Child’s parental status -0.04 0.07 -.07
Child’s age at divorce 0.00 0.00 -.01
Mother’s sexual -0.08 0.06 -.15
orientation
Note R2 = .0134for Step 1; A R2= ,0201(ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table K-e.
Summary o f Hierarchical Regression Models o f Transformed Child’ s Conflict on
Demographic, M other’ s Sexual Orientation with Beta Weights.
Variable B SEB fi
Step 1
Child’s partner status -0.03 0.02 -.19
Child’s parental status 0.01 0.02 .03
Child’s age at divorce 0.00 0.00 -.10
Step 2
Child’s partner status -0.03 0.02 -.19
Child’s parental status 0.00 0.02 .02
Child’s age at divorce 0.00 0.00 -.11
Mother’s sexual 0.01 0.02 .06
orientation
Note. R2 - .0415 for Step 1; A R2 = .0036 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
182
Table K-f.
Summary o f Hierarchical Regression Models o f Transformed Child’ s Emotional
Support on Demographic, M other’ s Sexual Orientation with Beta Weights.
Variable B SE B B
Step 1
Child’s partner status 0.05 0.08 .07
Child’s parental status -0.07 0.09 -.10
Child’s age at divorce 0.00 0.00 -.04
Step 2
Child’s partner status 0.05 0.08 .07
Child’s parental status -0.06 0.09 -.09
Child’s age at divorce 0.00 0.00 -.03
Mother’s sexual -0.03 0.07 -.05
orientation
Note. R2 = .0098 for Step 1; A R2= .0018 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Appendix L
Follow-up Models for Hypothesis II
183
Table L-a.
Summary o f Hierarchical Regression Models o f the Second Step o f Transformed
Mother’ s Ajfectual Solidarity on Demographic, M other’ s Relationship Status with Beta
Weights.
' Variable B SE B — fi
Step 2
Child’s partner status 0.13 0.08 .19
Child’s parental status 0.02 0.09 .03
Child’s age at divorce 0.00 0.00 .10
Mother’s income -0.03 0.05 -.08
Mother’s relationship status 0.12 0.09 .17
Note. A R2 = .0205 (p > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
184
Table L-b.
Summary o f Hierarchical Regression Models o f Transformed M other’ s Emotional
Support on Demographic, M other’ s Relationship Status with Beta Weights.
Variable B SE B 1 3
Step 2
Child’s partner status -0.12 0.08 -.16
Child’s parental status 0.01 0.09 .02
Child’s age at divorce 0.00 0.00 .00
Mother’s income 0.11 0.05 .23
Mother’s relationship status 0.06 0.09 .08
Note.A R2= .0048 (2 >.05) * p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
185
Table L-c.
Summary o f Hierarchical Regression Models o f Transformed Child’ s Affectual
Solidarity on Demographic, Mother’ s Relationship Status with Beta Weights.
Variable B SEB 1 3
Step 2
Child’s partner status 0.08 0.07 .13
Child’s parental status -0.04 0.08 -.07
Child’s age at divorce 0.00 0.00 -.10
Mother’s relationship status 0.00 0.07 .00
Note. A R2 = .0000 (2 > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
186
Table L-d.
Summary o f Hierarchical Regression Models o f Transformed Child ’ s Conflict on
Demographic, Mother’ s Relationship Status with Beta Weights.
Variable B SEB B
Step 2
Child’s partner status -0.03 0.02 -.19
Child’s parental status -0.01 0.02 -.04
Child’s age at divorce -0.00 0.00 -.04
Mother’s relationship status -0.01 0.02 -.03
Note. R2 = .0467 for Step 1; A R2- .001 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
187
Table L-e.
Summary o f Hierarchical Regression Models o f Transformed Child’ s Emotional
Support on Demographic, Mother’ s Relationship Status with Beta Weights.
Variable B SEB 1 3
Step 2
Child’s partner status 0.04 0.09 .06
Child’s parental status -0.07 0.10 -.10
Child’s age at divorce -0.00 0.00 -.05
Mother’s relationship status 0.01 0.09 .01
Note. A R2 = .0002 (p> .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
188
Appendix M
Analyses for the Third Specific Aim
Table M-a.
Summary o f Hierarchical Regression Models o f Child’ s Affectual Solidarity on Age,
Gender, and Interactions with Beta Weights (N = 122)
Variable B SE B fi
Step 1
Child’s age 0.00 0.00 .02
Child’s gender 0.12 0.06 .20*
Mother’s sexual orientation -0.08 0.05 -.15
Step 2
Child’s age -0.00 0.00 -.06
Child’s gender 0.13 0.06 .21*
Mother’s sexual orientation -0.07 0.05 -.14
Ch. age * m’s s.o. 0.04 0.03 .16
Ch. gender * m ’s s.o. -0.02 0.02 -.08
Note. R2 = .0662 for Step 1 (p = .04); A R 2 = .0234 ( p > .05). * P < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
189
Table M-b.
Summary o f Hierarchical Regression Models o f Child’ s Conflict
Interactions with Beta Weights (N = 121)
on Age, Gender, and
Variable B SEB B
Step 1
Child’s age -0.00 0.00 -.19*
Child’s gender 0.01 0.02 .08
Mother’s sexual orientation 0.00 0.01 .03
Step 2
Child’s age -0.00 0.00 -.19
Child’s gender 0.01 0.02 .05
Mother’s sexual orientation 0.00 0.01 .03
Ch. age * m’s s.o. -0.00 0.01 -.03
Ch. gender * m’s s.o. 0.01 0.01 .13
Note. R2 = .0374 for Step 1; A R2= .0155 (ps > .05). * p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
190
Table M-c.
Summary o f Hierarchical Regression Models o f Child’ s Emotional Support on Age,
Gender, and Interactions with Beta Weights (N = 123)
Variable B SEB £
Step 1
Child’s age -0.01 0.00 -.13
Child’s gender 0.16 0.07 .20*
Mother’s sexual orientation -0.05 0.06 -.07
Step 2
Child’s age -0.01 0.01 -.13
Child’s gender 0.15 0.07 .19*
Mother’s sexual orientation -0.05 0.07 -.07
Ch. age * m’s s.o. -0.00 0.04 -.01
Ch. gender * m ’s s.o. 0.02 0.03 .05
Note. R2 = .0580 for Step 1; A R2 = .0024 (ps > .05). * p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table M-d.
Summary o f Hierarchical Regression Models o f Child’ s Depressive Symptoms on Age,
Gender, and Interactions with Beta Weights (N = 122)
Variable B SEB B
Step 1
Child’s age -0.00 0.00 -.00
Child’s gender 0.06 0.07 .07
Mother’s sexual orientation 0.03 0.07 .04
Step 2
Child’s age 0.00 0.01 .06
Child’s gender 0.06 0.07 .08
Mother’s sexual orientation 0.02 0.07 .03
Ch. age * m’s s.o. -0.03 0.04 -.10
Ch. gender * m’s s.o. 0.00 0.03 .00
Note. R2 = .0066 for Step 1; A R2 = .0075 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
192
Table M-e.
Summary o f Hierarchical Regression Models o f Child’ s Negative Affect on Age,
Gender, and Interactions with Beta Weights (N = 123)
Variable B SE B 6
Step 1
Child’s age -0.04 0.02 -.19*
Child’s gender 0.09 0.31 .03
Mother’s sexual orientation 0.27 0.28 .09
Step 2
Child’s age -0.02 0.02 -.10
Child’s gender 0.07 0.32 .02
Mother’s sexual orientation 0.24 0.28 .08
Ch. age * m’s s.o. -0.25 0.16 -.17
Ch. gender * m’s s.o. 0.10 0.14 .07
Note. R2 = .0374 for Step 1; A R2= .0235 (ps > .05). * g < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
193
Table M-f.
Summary o f Hierarchical Regression Models o f Child’ s Positive Affect on Age, Gender,
and Interactions with Beta Weights (N = 123)
Variable B SE B fi
Step 1
Child’s age -0.00 0.00 -.06
Child’s gender -0.08 0.06 -.11
Mother’s sexual orientation 0.02 0.06 .04
Step 2
Child’s age -0.00 0.00 -.10
Child’s gender -0.07 0.07 -.10
Mother’s sexual orientation 0.03 0.06 .05
Ch. age * m’s s.o. 0.03 0.03 .11
Ch. gender * m’s s.o. -0.03 0.03 -.11
Note. R2 = .0198 for Step 1;AR2 = .0171 (ps> .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
194
Table M-g.
Summary o f Hierarchical Regression Models o f Child’ s Self-Esteem on Age, Gender,
and Interactions with Beta Weights (N - 121)
Variable B SE B B
Step 1
Child’s age -0.01 0.01 -.04
Child’s gender -0.34 0.19 -.16
Mother’s sexual orientation 0.09 0.17 .05
Step 2
Child’s age 0.00 0.01 .00
Child’s gender -0.34 0.20 -.17
Mother’s sexual orientation 0.08 0.17 .04
Ch. age * m’s s.o. -0.07 0.10 -.08
Ch. gender * m’s s.o. 0.01 0.09 .02
Note. R2 = .0431 for Step 1; A R2= .004 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
195
Appendix N
Summary of Hierarchical Regression Models for Mother and Dyadic Family
Relationship Quality on Age, Gender and Interactions.
Table N-a.
Summary o f Hierarchical Regression Models o f Mother’ s Affectual Solidarity on Age,
Gender, and Interactions with Beta Weights (N — 114)
Variable B SE B B
Step 1
Child’s age 0.00 0.00 .11
Child’s gender 0.01 0.07 .01
Mother’s sexual orientation 0.11 0.06 .17
Step 2
Child’s age 0.00 0.01 ,11
Child’s gender 0.01 0.07 .01
Mother’s sexual orientation 0.11 0.06 .17
Ch. age * m ’s s.o. -0.00 0.04 -.00
Ch. gender * m’s s.o. 0.00 0.03 .01
Note. R2 = .0443 for Step 1; A R2= .00 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table N-b.
Summary o f Hierarchical Regression Models o f M other’ s Conflict on Age, Gender, and
Interactions with Beta Weights (N = 114)
Variable B SEB fi
Step 1
Child’s age -0.00 0.00 -.36***
Child’s gender 0.03 0.02 .16
Mother’s sexual orientation 0.01 0.01 .05
Step 2
Child’s age -0.00 0.00
_ 42***
Child’s gender 0.03 0.02 .14
Mother’s sexual orientation 0.01 0.02 .06
Ch. age * m ’s s.o. 0.01 0.01 .09
Ch. gender * m’s s.o. 0.00 0.01 .04
Note. R2 = .1373 for Step 1 (p = .001); A R 2 = .0068 (e > .05). ***'£<.001
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table N-c.
Summary o f Hierarchical Regression Models o f M other’ s Emotional Support on Age,
Gender, and Interactions with Beta Weights (N = 114)
Variable B SEB B
Step 1
Child’s age -0.01 0.00
. 27**
Child’s gender 0.12 0.07 .16
Mother’s sexual orientation -0.12 0.06 -.18
Step 2
Child’s age -0.01 0.01 -.27*
Child’s gender 0.10 0.07 .13
Mother’s sexual orientation -0.13 0.06 -.18
Ch. age * m ’s s.o. -0.01 0.04 -.02
Ch. gender * m’s s.o. 0.03 0.03 .10
Note. R2 = . 1351 for Step 1 (p = 001); A R2- .0084 (p > .05). * P < .05 ** p < .01.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
198
Table N-d.
Summary o f Hierarchical Regression Models o f Dyadic Affectual Solidarity on Age,
Gender, and Interactions with Beta Weights (N = 113)
Variable B SE B 1 3
Step 1
Child’s age 0.00 0.00 .05
Child’s gender 0.12 0.08 .15
Mother’s sexual orientation -0.04 0.07 -.06
Step 2
Child’s age -0.00 0.01 -.04
Child’s gender 0.12 0.08 .15
Mother’s sexual orientation -0.03 0.07 -.04
Ch. age * m ’s s.o. 0.05 0.04 .15
Ch. gender * m’s s.o. -0.01 0.04 -.02
Note. R2 - .0311 for Step 1; A R2 = .0139 (ps > .05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
199
Table N-e.
Summary of Hierarchical Regression Models of Dyadic Conflict
Interactions with Beta Weights (N = 112)
on Age, Gender, and
Variable B SEB 1 3
Step 1
Child’s age -0.03 0.01 -.30**
Child’s gender 0.15 0.13 .11
Mother’s sexual orientation 0.08 0.12 .06
Step 2
Child’s age -0.03 0.01
_ 32**
Child’s gender 0.11 0.14 .08
Mother’s sexual orientation 0.09 0.12 .06
Ch. age * m’s s.o. 0.01 0.07 .02
Ch. gender * m’s s.o. 0.06 0.06 .09
Note. R2 = .0195 for Step 1 (p := .015); A R2 = = .0076 (p > .05). ** p < .01.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 0 0
Table N-f.
Summary o f Hierarchical Regression Model o f Dyadic Emotional Support on Age,
Gender, and Interactions with Beta Weights (N = 113)
Variable B SE B B
Step 1
Child’s age -0.01 0.00 -.23*
Child’s gender 0.13 0.07 .16
Mother’s sexual orientation -0.05 0.07 -.06
Step 2
Child’s age -0.01 0.01 -.23
Child’s gender 0.11 0.08 .14
Mother’s sexual orientation -0.05 0.07 -.07
Ch. age * m’s s.o. -0.01 0.04 -.02
Ch. gender * m ’s s.o. 0.04 0.03 .11
Note. R2 = .0778 for Step 1 (p = .03); A R 2 =
trT
o
A
3
o
* p < .05.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract (if available)
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Psychological well-being, religious affiliation, and cultural identification in adult children of Jewish-Gentile interfaith marriages
PDF
Parents' psychological distress and children's temperament as moderators of the effects of daily work stress on family interactions
PDF
Hostile family environments and children's perceptions of social support
PDF
The relationship between changes in the type and frequency of contact in open and closed adoptions and behavioral outcomes of adopted children
PDF
Children's and parents' exposure to community violence: Links with neighborhood disadvantage and parenting
PDF
Parents as harbingers: A new model for differentially predicting life outcomes for African-Americans and Caucasian Hispanics
PDF
Comparative analysis of the stress process in AIDS caregivers: Gay male partners, Anglo women, and Latinas
PDF
Emotion regulation as a mediator between family-of-origin aggression and marital aggression
PDF
Social support, social change, and psychological well-being of the elderly in China: Does the type and source of support matter?
PDF
The associations among family stress, friendship quality, and romantic quality
PDF
Psychological reactions to racial hate crime victimization: An investigation utilizing the articulated thoughts in simulated situations paradigm
PDF
The flow of children through the foster care system
PDF
Latino and African American mothers' involvement in their children's early education
PDF
Family functioning, intergenerational conflict, and psychological symptomology of Asian American college students
PDF
Intra-familial child sexual abuse: The experience and effects on non-offending mothers
PDF
Black urban non-family households: Their socioeconomic position and spatial buffering
PDF
Emergent literacy: A sense of becoming and being literate in Latino families
PDF
The role of affect in the marital satisfaction of monoethnic Latino, biethnic Latino, and European -American newlywed couples
PDF
Female parenting styles: associations with psychological distress, interpersonal relations, conscientious-assertiveness, and general deviance
PDF
Organizational development and coalition building among domestic violence agencies in California: conflict and compromise between grassroots groups and established institutions
Asset Metadata
Creator
Rose, Tara
(author)
Core Title
Psychological well -being in non -traditional families
Degree
Doctor of Philosophy
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,psychology, social,sociology, individual and family studies
Language
English
Contributor
Digitized by ProQuest
(provenance)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-264305
Unique identifier
UC11339331
Identifier
3093813.pdf (filename),usctheses-c16-264305 (legacy record id)
Legacy Identifier
3093813.pdf
Dmrecord
264305
Document Type
Dissertation
Rights
Rose, Tara
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
psychology, social
sociology, individual and family studies