University of Southern California Dissertations and Theses

The measurement, life course patterns, and outcomes of intergenerational ambivalence among parent-adult child dyads

(USC Thesis Other)

The measurement, life course patterns, and outcomes of intergenerational ambivalence among parent-adult child dyads

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content THE MEASUREMENT, LIFE COURSE PATTERNS, AND
OUTCOMES OF INTERGENERATIONAL AMBIVALENCE
AMONG PARENT-ADULT CHILD DYADS

by

Jessica Penn Lendon

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GERONTOLOGY)

May 2012

Copyright 2012 Jessica Penn Lendon
ii

Acknowledgements
I wish to thank my supportive committee members, Merril Silverstein, Roseann
Giarrusso, and Tim Biblarz, for transmitting their expertise, their accumulated wisdom,
and for their encouragement in completing this dissertation. I am honored to be a member
of the ―LSOG Family.‖ In addition, I want to recognize Jack McArdle, who was an
excellent model and fit for introducing me to longitudinal data analysis. I must also thank
my previous mentors, Angie Mertig and Ron Aday, whose encouragement fostered my
passion for survey research and gerontology, which significantly influenced my current
life trajectory.
I also greatly appreciate my colleagues and friends for their emotional and
instrumental support at all stages of this doctoral life course: Lindsey Baker, Trina Duke,
Zachary Gassoumis, Aaron Hagedorn, Shoshana Hindin, Martijn Hogerbrugge, Nichole
Kryla-Lighthall, Jeff Laguna, Naama Levitzki, Joohong Min, Adria Navarro, Hunhui Oh,
Nick Pisca, Cecilia Poon, Ricardo Reyes, and Maria Siciliano.
Without Danielle Zucker, my ability to manage and use the LSOG data would
have been very improbable. Her breadth of knowledge, skills, and support are beyond
measure.

iii

Table of Contents
Acknowledgements ............................................................................................................. ii

List of Tables ...................................................................................................................... iv

List of Figures ..................................................................................................................... v

Abstract .............................................................................................................................. vi

Chapter One: Introduction and Theory ............................................................................... 1

Chapter Two: Data Description: The Longitudinal Study of Generations ........................ 21

Chapter Three: Mixed Feelings, Mixed Measures: Measuring Ambivalence
among Older Parent-Adult Child Dyads ............................................................... 28

Chapter Four: A Decade of Love and Hate:
Patterns of Intergenerational Ambivalence Experienced by
Two Cohorts of Older Parent- Adult Child Dyads ................................................ 55

Chapter Five: Does Ambivalence Matter? The Relationship between Older
Parents‘ Feelings towards Their Children and Well-being over Time ................. 81

Chapter 6: Discussion ..................................................................................................... 104

References ........................................................................................................................ 111

iv

List of Tables
Table 0.1. Is Ambivalence a Concept, Model, or Theory? .................................................. 7

Table 0.2. Longitudinal Response for Parent-Child Dyads by Birth Cohort .................... 23

Table 1.1. Parent-Child Comparison of Means (Paired T-tests) for
Ambivalence, Positive, and Negative Measures ................................................... 43

Table 1.2. Correlations between Ambivalence, Positive, and Negative Measures .......... 44
Table 1.3. Multiple Regression Coefficients of the Effects of Characteristics
on Parents‘ and Children‘s Direct and Indirect Ambivalence ............................... 46

Table 2.1. Longitudinal Response for Parent-Child Dyads by Birth Cohort .................... 66
Table 2.2 Mean Comparisons and Correlations between Parent-Child
Ambivalence across Time and by Birth Cohort .................................................... 70

Table 2.3. Best-Fitting Multiple Group Latent Growth Model of Intergenerational
Ambivalence 1988-2005 ....................................................................................... 72

Table 2.4 Bivariate Latent Growth Model: Two Cohorts
of Child-Parent Ambivalence ................................................................................ 76

Table 3.1. Longitudinal Response for Parent-Child Dyads by Birth Cohort .................... 87

Table 3.2. Descriptive Statistics for Parents‘ Ambivalence,
Depression, and Covariates ................................................................................... 89

Table 3.3. Correlation Coefficients for Ambivalence and Depression across Time ......... 90

Table 3.4. Dual Latent Change Score Models for Ambivalence 1991-2005 ................... 94

Table 3.5. Dual Latent Change Score Models for Depression, 1991-2005 ..................... 95

Table 3.6. Bivariate Dual Latent Difference Score Model with Covariates .................... 99

v

List of Figures
Figure 0.1. Intergenerational Ambivalence Model ........................................................... 10

Figure 0.2. General Human Bioecological Model ............................................................ 14

Figure 2.1. Parents‘ Average Ambivalence by Birth Cohort (1988-2005) ........................ 73

Figure 2.2. Children‘s Average Ambivalence by Birth Cohort (1988-2005) .................... 74

Figure 3.1. SEM Results of Bivariate Latent Difference Score Model for
Depression and Ambivalence, 1991-2005 .......................................................... 103

Figure 4.1 Modified Bioecological Framework for Understanding
Intergenerational Relationships .......................................................................... 108

vi

Abstract

The overall goal of this dissertation is to describe the measurement, trajectories,
and consequences of intergenerational ambivalence among older parents and adult
children. Ambivalence is the coexistence of affection and conflict within the
intergenerational relationship and there are a variety of techniques in which to measure
ambivalence. However, there is little certainty about their validity or equivalence of the
measures within empirical research. Further, the majority of research on intergenerational
ambivalence is cross-sectional, which limits family researchers‘ knowledge to only
familial and individual characteristics associated with change in ambivalence.
Ambivalence has been related to well-being, but there is little evidence of a causal
relationship between ambivalence and psychological outcomes. Therefore, there is a need
to validate measurements and test causes and/or consequences of ambivalence over the
adult life course to advance the utility of the intergenerational ambivalence model. The
objectives are: (1) to compare and validate two methods of quantitatively measuring
intergenerational ambivalence within parent-child dyads; (2) to describe and explain
change in ambivalence over time among two cohorts of parent- adult child dyads; and (3)
to identify the effect of ambivalence over time on psychological well-being.
Analyses were performed using data from the USC Longitudinal Study of
Generations (LSOG)—a multi-wave, multi-level, four-generation survey of American
families. The first study analyzed 253 parent-child dyads from the 2005 wave of the
LSOG. Bivariate and multivariate analyses showed differences between parents and
children based on an indirect measure (but not a direct measure) of ambivalence and
vii

conceptual differences in the predictors of these two most commonly used measurement
strategies. I discuss applying the labels of implicit ambivalence to the indirect measure
and explicit ambivalence to the direct measure, as they each measure different
conceptualizations of ambivalence. The second study analyzed 848 parent-child dyads
from two birth cohorts across 10 years from the LSOG. The older cohort included parents
born between 1916 and 1931 and their children born between 1945 and 1955. The
younger cohort included parents born between 1945 and 1955 and their children born
between 1978 and 1983. Latent growth models showed an overall significant decline in
ambivalence over time, when both cohorts were combined. There was a positive
covariation between the latent level and negative slope for the younger dyad. And for the
older dyad, there was a nonlinear trend of decline from 1991 to 1997 and a slight increase
after 2000. Social and/or historical factors, along with life course stages, help to explain
the potential cohort differences observed in these varying trajectories. The final study
tested the causal relationship between ambivalence and depression in over 900 parents
using LSOG data from 1991 to 2005. Bivariate correlation analyses showed a significant
positive relationship between ambivalence and depression, as expected. However,
bivariate dual change-score analysis of the longitudinal data showed a significant
negative influence of parents‘ depression on ambivalence towards adult children, when
controlling for relationship characteristics, birth cohort of parent, gender, and health.
Contrary to expectations, ambivalence does not seem to influence levels of depression
over time. Although more depressed parents have greater levels of ambivalence, initially,
they actually experience a decline (or moderating effect on change) in ambivalence
viii

towards adult children over time. These results highlight the complexity of the impact
mental health has on older parents‘ relationships. The final discussion summarizes these
findings and discusses how they fit into the larger theoretical context.
1

Chapter One:
Introduction and Theory
A. Introduction
Importance of Ambivalence in Parent-Child Relationships in Later Life
The family, as a social institution, is experiencing a flux of changing norms in
response to social changes, demanding work lives, economic struggles, and lengthening
life spans (Lüscher, 2004). While families remain affectionate, supportive, and of
enduring importance, there are also many social forces placing burdens on the family as a
social institution (Lawton, Silverstein, & Bengtson, 1994). Feelings of ambivalence
within families have always been present, but mixed feelings between parents and adult
children may be more prevalent or of greater intensity because of these social changes
and complications in family life. It is important for family therapists, social workers,
policy makers, and individuals to understand strong contradictory feelings toward a loved
one, because the coexistence of affection and conflict may lead to detrimental outcomes
for individuals and late-life parent-child relationships (Lüscher, 2004).
Social changes relate to three primary aspects that directly impact
intergenerational relations: the lengthening life span, an increased need for informal care
giving of older parents, and increased diversity in family structures (Silverstein et. al,
2010). Lengthening life spans has allowed more generations within families to survive
each other, thus allowing more opportunity for both rewarding and burdensome
interactions. However, the existence of intergenerational ambivalence may inhibit mutual
reliance of family members across generations.
2

Families and adult children provide the bulk of informal care giving and often
take pride in their filial obligations; however, increasing demands from their many social
roles (child, parent, employee, spouse, etc.) can introduce ambivalence about taking on
the role of caregiver (Gans & Silverstein, 2006; Roberto & Jarrott, 2008). With longer
life spans, more illness and disability are experienced in later life. While the compression
or expansion of morbidity is still being debated, many older adults are living longer with
various acute and chronic illnesses, requiring both formal and informal long-term care
(Crimmins & Beltran-Sanchez, 2011). Care giving and dependency inherently produce
ambivalence, creating both costs and rewards within the relationship (Pillemer & Suitor,
2002; Connidis & McMullin, 2002).
The diversity in family structure and composition is also a major source of
conflict within otherwise affectionate and close parent-child ties. One key change is in
union formation: the increase in divorce, remarriage, delayed marriage, and decline in
marriage. Another source of family change is the decrease in fertility, increase in single-
parenthood, and choosing to have children outside of traditional marriage. Lastly,
families operate differently as women receive greater education and both partners pursue
careers (Bianchi & Casper, 2005). These changes have complicated the normative
expectations and obligations of the previous cohots‘ more traditional family structures.
With the inclusion of step-children and blended families, parent-child ties become less
certain, which is theorized to be potentially ambivalence-producing for older parents
(Pillemer & Suitor, 2005).
3

Illuminating both the good and bad aspects of parent-child relationships can be
informative for aging family scholars, but we should not overstate the case for
ambivalence or assume it always has a negative impact on family relationships. There are
likely some individuals with different levels of tolerance for ambivalence and some
individuals who experience stress, anxiety, confusion, or perhaps even depression over
unresolved mixed feelings for a loved one. Understanding ambivalence in contexts where
it presents detrimental consequences may be relevant for helping families through policy
or therapeutic settings.
Antecedents in Sociology and Psychology
Theorists from both psychology and sociology have wrestled with similar
conceptualizations of ambivalence. Sociological intergenerational ambivalence, similar to
role distance theory in which an individual experiences incompatible expectations within
one or multiple roles, is the experience of contradictory norms within a relationship. A
primary example of sociological ambivalence would be the contradictions experienced by
a frail older parent between retaining autonomy from and being dependent on a care
giving child (Coser, 1969; Connidis & McMullin, 2002). Psychological ambivalence
focuses on the contradictory feelings within the older adult about his or her child, and
vice versa (Lüscher, 2004). As such, psychological ambivalence can entail having
conflicted or mixed emotions, thoughts, or motivations about an individual (Fingerman et
al., 2008) and can be a consequence of sociological ambivalence. Social structures and
social roles are conventionally thought to produce the psychological mixed feelings;
4

however, in some cases, negotiations of ambivalent feelings can lead to personal agency
and social action (Connidis & McMullin, 2002).
Intergenerational Ambivalence within Context of Solidarity-Conflict
Intergenerational relationships have traditionally been theorized within the
solidarity-conflict paradigm, which was expanded to include ambivalence as the
intersection of solidarity and conflict (Bengtson et al., 2002; Silverstein et al., 2010). The
solidarity-conflict paradigm is multidimensional, including seven dimensions: affectual,
consensual, functional, associational, structural, normative solidarity, and conflict. The
dimensions of solidarity, for example, describe intergenerational relationships on a
continuum from most to least affectionate. Within one relationship, each of these
dimensions operate independently (e.g. having periodic contact with each other, but
maintaining high affection). The addition of conflict provides other possibilities to
measure how family relations can have, paradoxically, high solidarity in some
dimensions, but also experience great conflict (Bengtson et al., 2002). The measurements
for affectual solidarity developed by Bengtson et al. (1975) and conflict are similar to the
various items used by other researchers (e.g., Fingerman et al., 2006) to measure
intergenerational ambivalence as a mix of both affection and conflict.
Connidis and McMullin (2002) argue ambivalence is compatible with solidarity,
but is a distinct concept that the solidarity and conflict constructs cannot capture. They
claim it motivates action and negotiation in relationships as an attempt to reconcile
competing social expectations and emotions. From this perspective, the solidarity-conflict
paradigm could be viewed as measurements for ―outcomes of negotiating ambivalence‖
5

(Connidis & McMullin, 2002, p.595). Coser (1969) observed that, ―it is because of such
ambivalence-creating situations...that choices have to be made and that structural changes
become possible‖ (p.182). Ambivalence may not be ultimately resolved, especially when
stemming from larger social structures; however it is likely that family relationships
transition in and out of affectionate, conflictual, and ambivalent types over the life
course.
Luscher and Pillemer (1998) provide a helpful definition of intergenerational
ambivalence as the ―contradictions [in relationships] at the level of social structure,
evidenced in institutional resources and requirements, such as statuses, roles, and norms
and contradictions at the subjective level, in terms of cognitions, emotions, and
motivations‖ (p. 416).
B. Operational Definition of Intergenerational Ambivalence
The synthesis of both sociological and psychological conceptualizations of
ambivalence have led family scholars to assume that mixed emotions (psychological) are
the outcome of contradictory social structures and societal norms (sociological)
surrounding the intergenerational relationship, which may then lead to agency and
negotiations within the relationship (Lüscher, 2002; Curran, 2002; Connidis & McMullin,
2002). This tendency to define sociological contradictions as causes of psychological
ambivalence can be observed in much of the literature. Researchers theorize
intergenerational ambivalence arises from children‘s dependencies or problems and then
measure intergenerational ambivalence as the coexistence of positive and negative
emotions (Pillemer & Suitor, 2004; Luscher & Pillemer, 1998; Fingerman, 1996;
6

Lowenstein, 2007). For example, Willson et al. (2006) tested the association of
dependence within social structures (marriage and gender) on ―attitudinal ambivalence‖
(p. 241). This approach seems to bridge the divide between psychological and
sociological influences on family relationships, creating a useful and interdisciplinary
paradigm in which to understand aging families.
In this research, intergenerational ambivalence is conceptualized as the
coexistence of affection and conflict within individuals towards their family member,
which is influenced by life course changes and social structures. Ambivalence is
measured along a continuum as an inherently dynamic characteristic of intergenerational
relationships, rather than as a type of relationship in a classification scheme (Lüscher,
2002; Giarrusso, Silverstein, Gans, & Bengtson, 2005; van Gaalen & Dykstra, 2006).
Because the intensity of ambivalence is apt to change throughout the adult life course, a
continuous measure is more sensitive to this change than an observation of transitions
across discrete categories of relationship-types.
C. Concept, Framework, or Theory?
Intergenerational ambivalence has been referred to as a concept, a framework, and a
theory in the past decade. Table 0.1 shows published articles from 1998 to 2011 and the
language used by the authors to describe intergenerational ambivalence. Many of the
earlier articles called it a ―concept,‖ often qualifying it as a sensitizing or organizing
concept, but more recent articles referred to it as a framework, model, or a theory. Birditt,
Fingerman, and Zarit referred to ―ambivalence theory‖ in their 2010 article, but offer
little justification or theoretical development for calling it a theory. They test the idea that
7

children's problems and achievements may explain ambivalence, but not how
ambivalence can be applied as a general theoretical model.
Table 0.1. Is Ambivalence a Concept, Model, or Theory?

Authors

Year

Language

Blumer 1969 sensitizing concept
Lüscher and Pillemer 1998 general orientation; theoretical model
Luscher 1999 Model
Bengtson et al. 2002 Concept
Connidis and McMullin 2002 concept; critical view towards society
Curran 2002 Concept
Lüscher 2002 sensitizing concept
Pillemer and Suitor 2002 Concept
Willson, Shuey, and Elder 2003 Concept
Fingerman, Hay, Birditt 2004 sentiment; concept
Lüscher 2004 Concept
Pillemer & Suitor 2004 organizing concept; approach;
characteristic; perspective
Spitze and Gallant 2004 concept; perspective
Giarrusso, et al. 2005 Concept
Peters, Hooker, and
Zvonkovic
2006 theory of ambivalence; framework
van Gaalen and Dykstra 2006 Concept
Willson et al. 2006 Concept
Lownestein 2007 paradigm; model; conceptual perspective
Pillemer et al. 2006 organizing concept; perspective
Fingerman et al. 2008 Model
Steinbach 2008 model
Ha and Ingersoll-Dayton 2008 ambivalence theory; conceptual framework
Dolbin-Macnab et al. 2009 Perspective
vanGaalen et al. 2010 Concept
Birditt, Fingerman, Zarit 2010 ambivalence theory; framework
Silverstein et al. 2010 ambivalence paradigm

8

As a concept, intergenerational ambivalence is well suited as a tool for developing
existing theories of aging and human development (e.g., age stratification, life span
development, and social exchange theories) that explain intergenerational relations in
later life. However, if intergenerational ambivalence is labeled as a formal theory, then it
must be developed as a set of ideas and principles that explain something unique about
intergenerational relationships that other theories do a poorer job of explaining (e.g., why
certain children do or do not provide caregiving assistance to older parents).
It is perhaps premature to declare that ambivalence has become a formal theory;
however it is clearly a phenomenon within families that can be incorporated within
existing social-psychological theories. The language differences noted above are not
without consequences. Uncertainty about labeling ambivalence demonstrates that the
development of ambivalence-as-a-concept to ambivalence-as-a-theory has not come to
full fruition in the field of aging families. Lüscher (2002) noted that ambiguity in an idea
can drive insights, but he strongly advocated for explicit hypotheses and systematic
models to advance empirical research. The current state of intergenerational ambivalence
has relied heavily on empirical studies focused on measuring the prevalence of
ambivalence within families (using different measurement methods) and looking for
significant correlations with predictors, based on many theories and hypotheses. The
classifications in Table 0.1 reveal that family scholars, implicitly, seem to have accepted
that the concept of intergenerational ambivalence has evolved into a model, perspective,
or framework. The language used in research is important and the developing consensus
9

about what to call intergenerational ambivalence may help guide future research and
theory. For this reason, I describe it as the ―intergenerational ambivalence model.‖
The intermediary domain between the concept and theory, a model, is the best way to
describe the body of work surrounding intergenerational ambivalence. A model implies
multiple causes, consequences, and mechanisms for how ambivalence operates within
families that can be measured and tested. It goes beyond the level of concept, in that
understanding how ambivalence arises, is negotiated, and experienced may explain
certain family outcomes. Lüscher and Pillemer (1998) promoted a ―simple illustrative
model‖ of ambivalence in 1998 as the first step towards a formal conceptual model. It
shows how conflicting normative structures and social positions (defined as sociological
ambivalence by Connidis & McMullin, 2002) influence ambivalence (I label this as
psychological ambivalence), which impacts psychological outcomes for the individual
and leads to certain decisions to relieve or negotiate the ambivalence. Figure 0.1 is a
modified version of the original model, which includes the possible mutual influences
between psychological outcomes and personal agency, as well as, the feedback from
psychological outcomes to psychological ambivalence (represented by the dashed
arrows).
10

Figure 0.1. Intergenerational Ambivalence Model

D. Theoretical and Empirical Developments
The theoretical basis for understanding intergenerational relationships is drawn from
many sociological and psychosocial theoretical traditions. Most importantly is the life
course perspective and human bioecological framework. The following section reviews
each of these perspectives, as applied to the intergenerational ambivalence model, and the
empirical findings related to these theories.
Life Course Perspective
The lifespan developmental theory (Fingerman, 1996) and the life course perspective
(Pillemer & Suitor, 2004) can help researchers understand the diversity and dynamics
within aging families, especially in terms of ambivalence. The life course perspective
includes several principles important to theorizing aging families, such as, linked lives,
historical influences, and personal agency, which interact to influence parent-child
relationships (Elder, 1994; Bengtson, Elder, & Putney, 2005). One of the most important
Relationship
Decisions/
Personal Agency
Psychological
outcomes
Psychological
Ambivalence
Sociological
Ambivalence:

Conflicting
Normative &
Social Structures

11

contributions of this perspective to the development of the intergenerational ambivalence
model is the recognition that relationships are dynamic and apt to change throughout the
life course. The current state of research on ambivalence has not yet investigated whether
(or how) ambivalence changes or remains stable over the life course of intergenerational
relationships. Lüscher (2004) discussed the temporal aspects of ambivalence and
ambivalence as a potentially enduring, oscillating, or temporary characteristic of
intergenerational relationships. Ambivalence could be only a temporary state for a
daughter caring for her mother after a brief hospital stay or may be an enduring
characteristic in a parent's relationship with his or her child, stemming from the child's
adverse decisions in young adulthood. The life course theory would predict that given
different situations and changes throughout the life course, ambivalence between parents
and children also changes. The life course perspective also posits that historical and
social trends impact family relationships. Likely, baby boomer children have different
relationships with their parents than their parents had with their parents, as a result of
sociodemographic and family changes. The timing and types of transitions in children‘s
lives, whether they follow a ―normative life course‖ or not, can impact the feelings of
both generations about each other. When timing or transitions are ―off,‖ unexpected, or in
direct conflict with each other, ambivalent situations and feelings may arise.
The lifespan developmental theory, coinciding with the generational stake hypothesis
(Giarrusso, Feng, & Bengtson, 2004), is closely related to the life course perspective, and
would predict that individuals at certain developmental stages (e.g., adolescence or frailty
in old age) may experience greater ambivalence. Additionally, parents and children are
12

each in different life stages and the ―linking‖ of parent‘s and children‘s lifespans have
reciprocal impacts on each other.
The life course and lifespan developmental theories offer broad explanations and
ways to understand aging families. However, few specific predictions about
intergenerational ambivalence can be made. The human ecological theory, while similar
to the life course, offers additional expectations about ambivalence in aging families.
Human Bioecological Framework
The human bioecological perspective is a multi-level and multidisciplinary
perspective that lends itself to a better understanding of aging individuals within families
and within society (Bronfenbrenner, 2001). While this specific dissertation research
depicts only a slice of the experience of human aging; it is embedded within the
bioecological perspective (Moen, Elder, Luscher, 1995). The point of addressing the
ambivalence model within this more integrated model is not to study the entire system
around an individual, but to better formulate problems and studies and make conclusions
that take into account the whole social system (Magnusson, 1995). The bioecological
theory is a useful multidisciplinary and critical theory for applying to ambivalence in
parent-child relationships that opens up new ideas, possibilities, and innovations that
might not be realized without a multi-disciplinary viewpoint. In fact, Moen (1995, p. 6)
states, ―Bronfenbrenner‘s ecology of human development paradigm not only furnishes a
theoretical model, but also becomes a catalyst, inspiring a redefinition of the problems to
be studied. In framing their research designs, he reminds researchers to explicitly
acknowledge that individuals and their environments are in constant, reciprocal
13

interplay.‖ Parents and children not only function and develop continuously in a
reciprocal dyad, but also individually through interactions between socioemotional,
physiological, interpersonal, and situational factors (Magnusson, 1995).
The human bioecological paradigm combines four primary elements:
characteristics of the person, developmental processes, social context, and time (PPCT),
within which individuals can experience continuity, change, and different rates of change
dependent on other social/individual elements (Moen, 1995). These factors include
situational demands, available opportunities, and potential barriers, which shape
trajectories and transitions across the life course (Moen, 1995). The cumulative advantage
and disadvantage theory can also be subsumed by this larger framework, as some
individuals have more situational opportunities and fewer barriers in their lives, based on
earlier life course advantages (Dannefer, 2003). There are strong associations between
behavior in childhood and psychosocial environmental risk factors in adult life through
multiple mechanisms and indirect links (Rutter et al., 1995). Additionally, macro social
changes impact individual experiences, development, and opportunities throughout the
life course (Elder, 1995).
The theoretical model of intergenerational ambivalence, which implicitly borrows
heavily from Bronfenbrenner‘s human bioecological framework and the life course
theory, posits that when contradictory macro and micro level forces impact the individual
and relationship dyad, ambivalent situations and feelings may be produced. Figure 0.2
shows how the parent-child dyad is embedded within the multiple levels of the
14

bioecological paradigm, providing a rich, but complex theoretical model for
understanding aging families.
Figure 0.2. General Human Bioecological Model

Macro-Level: Political Economy
Social Structures: Gender, Race, Class, Age
Policy: FMLA, NFCSP, Social Security, LTC
Historical Period: Norms, Ideology, Opportunity

Meso/Micro-Level: Lifecourse Perspective
Linked Lives: Work-family balance
Timing & Transitions
Social Networks & Exchange
Individual-Level:
Lifespan Development & Personality
Health & Biological Aging
Agency & Personal Experiences

Macro-level Social Structures Associated with Ambivalence
The macro level of this framework includes culture, social structures (age, gender,
class, race), the political economy, the medical-industrial complex (Estes, 2000), and
overall ideology. These macro-level forces impact individuals in terms of normative
expectations for filial obligation, work and family norms, socioeconomic status, social
inequalities, and health and lifespan, all of which directly impact the parent-child
15

relationship (Connidis & McMullin, 2002). Social structural inequalities (e.g. based on
gender, race, age stratification, and class) and social conflicts can create ambivalence in
families where little control can be exerted to negotiate the ambivalence (Connidis &
McMullin, 2002; Silverstein et al., 2010; Coser, 1969; Pillemer & Suitor, 2004). Previous
research has focused on some of the possible social structural characteristics and
demonstrated associations between ambivalence and competing norms and values,
gender, and racial differences. Ambivalence is particularly related to parents‘ tensions
between valuing both ideologies of autonomy and dependency. In terms of gender
differences, mothers have been shown to experience more ambivalence when children are
more dependent or do not live up to expectations (Pillemer & Suitor, 2004), suggesting
an influence of social norms and ideology about the normative life course and the role of
parents in raising successful children. Racial differences in intergenerational ambivalence
have also been briefly explored in the research, comparing two generations of black and
white families (Pillemer et al., 2007). These findings revealed greater ambivalence in
black mothers than white mothers, which was accounted for by their children's education,
marital status, employment, and behavioral problems.
Meso/ Micro-level Relationship Characteristics Associated with Ambivalence
At the meso-level, conflicts between various interpersonal roles (daughter, wife,
employee, mother), which are directly impacted by the macro-level social structures, can
lead to mixed feelings. The meso-level is conceptualized as the interaction between
multiple micro-level systems, such as family and work. The life course perspective helps
explain how timing, transitions, and linked lives, which operate at this level, impact the
16

parent-child relationship. Ambivalence has also been associated with overall similarity
between the generations and the child's characteristics and status attainment (health,
education, marital status, and neuroticism) (Fingerman et al., 2008; Pillemer & Suitor,
2002; Willson, Shuey, & Elder, 2003). Interestingly, the more social roles either
generation is invested in is related to lower levels of ambivalence (Fingerman et al.,
2006). Social support, instrumental support, informal caregiving, and emotional support
are all a fundamental part of family and intergenerational relationships. Families are a
huge source of support and solidarity for individuals. The social exchange theory would
suggest that situations where a parent (or a child) is providing inequitable amounts of
support would be more ambivalent (Ekeh, 1974; Lee & Ellithorpe, 1982). Research has
found that unequal exchanges, where a child is caregiving and feels like they are
receiving little support in return have a lower quality relationship with their parent, which
results in poorer well-being (Merz, Schuengel, & Schulze, 2009).
Individual-level Characteristics Associated with Ambivalence
The individual - level within the bioecological model, when applied to
ambivalence, includes factors such as personality, lifespan development, health, and
personal achievements. Previous research has focused on some of the possible individual
characteristics associated with intergenerational ambivalence. Older parents' ambivalence
towards their children has typically been associated with their children‘s poor health,
neuroticism, emotional problems, career success, and children who are unmarried or
divorced (Birditt et al., 2010; Pillemer & Suitor, 2005). Research has found adult children
were more ambivalent towards parents with poorer physical and psychological health,
17

towards mothers, and when providing assistance to their parent (Fingerman et al., 2006;
Fingerman et al., 2008; Wilson, Shuey, & Elder, 2003). Limited studies have also
examined the outcomes of ambivalent relationships in terms of well-being for the older
generation. Fingerman et al. (2008) found that greater ambivalence was associated with
more depression and less life satisfaction. While affection has a stronger association with
better quality of life, ambivalence was slightly associated with worse quality of life in
cross-country research (Lowenstein, 2007).
Applying the life course and bioecological perspectives to parent-child
relationships expands our understanding of the model of ambivalence developed by
Luscher and Pillemer (1998). Their original model (Figure 0.1) highlights the external
nature of macro-level social factors (conflicting norms and social structural positions) on
producing sociological ambivalence. However, it is limited by the directional
relationships between the meso-micro-individual level influences on the parent-child
dyad. Their model benefits from the addition of time and understanding that each of these
levels in an individual‘s life can be dynamic and reciprocal over the life course.
E. Objectives of Dissertation
This dissertation addresses important gaps in the empirical research on
intergenerational relationships, focusing on three main questions, (1) how is
intergenerational ambivalence best measured, (2) how does ambivalence change over the
life course for parents and children, and (3) what are the causes and outcomes of parents‘
ambivalent feelings towards their children?
18

The overall goal of my dissertation is to investigate the measurement, causes, and
consequences of intergenerational ambivalence among older parents and adult children.
There are a variety of techniques in which to measure this concept, but little certainty of
their validity or equivalence within the empirical research. All of the research on
ambivalence is cross-sectional, which limits family researchers‘ knowledge to familial
and individual characteristics associated with ambivalence. Therefore, there is a need to
validate measurements and test causes and consequences of ambivalence over the adult
life course to advance the utility of the concept within research of intergenerational
relationships. The specific aims for my dissertation are:
Aim 1 (Chapter 3). Will compare and validate two methods of quantitatively
measuring intergenerational ambivalence within a parent-child dyad, drawing on previous
research and theory. The purpose of the first aim is to compare two common strategies for
measuring intergenerational ambivalence. The comparison attempts to validate the two
measurements and assess the extent to which they are comparable. Researchers tend to
use either the indirect or the direct measurement of ambivalence, without
acknowledgment of the possible conceptual differences between the two strategies. This
aim assesses the face, predictive, and convergent validity of the measurement types from
both the parent and child perspectives.
Aim 2 (Chapter 4). Will describe and explain change in ambivalence over time
among two cohorts of parent- adult child dyads. The current state of research on
ambivalence has not yet investigated whether (or how) ambivalence changes or remains
stable over the life course of intergenerational relationships. Researchers have only
19

looked at ambivalence within cross-sections because this concept in family research is
new and measures have only recently been incorporated in studies. This research will
utilize multigenerational, longitudinal data to further develop the intergenerational
ambivalence model by identifying life course situations that cause ambivalence and
patterns or changes in ambivalence over the life course.
There also has been little investigation of the interplay or congruence between
generations' assessment of ambivalence and outcomes of ambivalence. The generational
stake hypothesis would suggest that parents may have a positivity bias in their reports of
relationships with their children, thus showing less ambivalence (Giarrusso, Feng, &
Bengtson, 2004). Fingerman et al. (2006) found that mothers' and children's assessments
were more similar than fathers' and children's, suggesting gender differences in the
congruence of intergenerational attitudes. It may be important to know if parents‘
assessments of ambivalence influence their childrens‘, and vice versa.
Aim 3 (Chapter 5). Will identify the effect of ambivalence over time on
psychological well-being, specifically depression, and identify potential factors that
explain varying outcomes. Researchers theorize that ambivalence may be associated with
negative well-being, quality of life, solidarity, and conflict, but studies have not directly
addressed potential causal linkages between ambivalence and outcomes (Luscher &
Pillemer, 1998). Although ambivalence is not necessarily associated with positive or
negative outcomes, it has been shown to be correlated with poor well-being (Pillemer &
Suitor, 2004; Fingerman et al., 2008). The existence of ambivalence in family
relationships may cause stress and tension within both sides of the intergenerational
20

relationship, leading to increased conflict or motivation for conflict resolutions. The
longitudinal outcomes of intergenerational ambivalence have not yet been empirically
tested. The third aim will investigate the long-term causal effects of parents‘
intergenerational ambivalence on depression.

21

Chapter Two:
Data Description: The Longitudinal Study of Generations
A. General Description of Sample
The data for the analyses come from the Longitudinal Study of Generations (LSOG),
which began in 1971 with 2,044 original respondents who were members of three-
generation families. Grandparents (G1) were selected using a multi-stage stratified
random sampling procedure from a population of 840,000 individuals enrolled in an
HMO in southern California. Adult children (G2) and grandchildren (G3) of the G1
grandparents were also invited to participate in the survey. Follow-up mail surveys were
administered to original respondents and additional spouses in 1985, 1988, 1991, 1994,
1997, 2000, 2005. The great-grandchildren (G4s) of the G1s, who were 16 or older, were
added in 1991 and in subsequent waves (Bengtson & Schrader, 1982).
This data has been used by family and aging researchers to answer a variety of
questions. Primarily, it is used to describe intergenerational relationships, explain filial
obligations and intergenerational support, and the transmission of social attitudes
between three-generation families.
B. Unique Features of the LSOG
One of the most obvious features is the LSOG‘s long period of data collection,
spanning over 30 years. The data for this research will span up to 15 years. There are
many psycho-social variables to test hypotheses that are comparable to measures used in
other social surveys, especially those on family relationships. Taking advantage of the
LSOG‘s multigenerational perspectives, we do not have to rely on second-hand
22

information about family members. Being a multigenerational study, we have first-person
accounts from both the parents and children, with the ability to compare generational
similarity and potential response biases that may occur.
C. Parent-Child Dyad Construction
In these analyses, I will be using reciprocal parent-child dyads from up to two
birth cohorts. Parents from the second generation (G2) and third generation (G3) birth
cohorts and children from the third generation (G3) and fourth generation (G4) birth
cohorts will be used to construct the two cohorts of parent-child dyads. The G3s will be
analyzed as both adult children and parents. Parents and children were randomly matched
in the beginning of the LSOG data collection, so that parents consistently responded
about one specific ―study‖ child and that child responded about his/her ―study‖ parents.
Given this information, I was able to match reciprocal dyads and include two
generational perspectives and two birth cohort perspectives.
The older dyad is composed of parents (G2) born between 1916 and 1931 and
children (G3) born between 1945 and 1955. The younger dyad is composed of parents
(G3) born between 1945 and 1955 and children (G4) born between 1978 and 1983.
Including incomplete cases, there are 238 unique older and 263 younger mother-child
dyads and 185 older and 162 younger father-child dyads. Table 0.1 lists the number of
dyads available for the ambivalence measure at each time point from 1988 to 2005. Only
respondents who had a matching parent-child dyad during at least one of the time periods
were included in the analyses.

23

Table 0.2. Longitudinal Response for Parent-Child Dyads by Birth Cohort

Total

1988

1991

1994

1997

2000

2005
Older (G2-G3) Cohort 408 355 335 373 375 319 253

Younger (G3-G4)
Cohort
495 0 145 263 336 303 319

D. Variable Construction used in Analyses
Intergenerational Ambivalence
There are various ways to measure intergenerational ambivalence and this
research will use two methods. The direct measurement was developed by Pillemer and
Suitor (2007) through focus group interviews. Their scale asks respondents to rate the
degree to which they have ―mixed feelings,‖ ―get on each other‘s nerves, but nonetheless
feel close,‖ or feel ―torn in two directions‖ toward a parent or child (Pillemer et al., 2007,
p. 782). This direct measure of ambivalence has been utilized in several studies
(Lowenstein, 2007; Pillemer et al., 2007; Pillemer & Suitor, 2002) and shows the ability
to tap into subjective ambivalence or mixed feelings about the target individual.
Several techniques have been used to indirectly capture ambivalence, including
additive scales of discordant (positive and negative) measures that describe the intensity
of opposing feelings (e.g., Willson et al., 2006), and classification procedures to group
relationships into ambivalent and non-ambivalent types (Giarrusso, Silverstein, Gans, &
Bengtson, 2005; Steinbach, 2008; van Gaalen & Dysktra, 2006). One common indirect
measure of ambivalence uses a set of statements about positive and negative feelings, or
relationship qualities, and calculates a score that reflects the balance between positive and
24

negative feelings (Fingerman et al., 2008; Thompson, Zanna, & Griffin, 1995; Willson,
Shuey, & Elder, 2003). Noticeably, the positive and negative items typically used are
similar to those used to measure affectual solidarity and conflict within the solidarity
paradigm. Fingerman and colleagues (2006) use the following two items to assess
positive qualities of the relationship: ―how much does he or she make you feel loved and
cared for?‖ and ―how much does he or she understand you?‖ They used these two items
to assess negative feelings: ―how much does he or she criticize you?‖ and ―how much
does he or she make demands on you?‖ This measurement strategy combines two
necessary aspects of ambivalence: similar magnitude of the negative and positive
components and at least moderate intensity of both components. For instance, an
individual who reports the highest positive and negative feelings would be considered
very ambivalent, while an individual with high positive and low negative (or vice versa)
feelings would be considered to have low ambivalence. Someone with both lower
positive and negative feelings, while similar in magnitude, does not possess the intensity
of feelings to be considered ambivalent (Thompson, Zanna, & Griffin, 1995).
For this research, I used slightly different, but similar items to measure positive
aspects of the intergenerational relationship: ―How close do you feel is the relationship
between you and your study child (or mother/father)?,‖ ―How well do you get along?,‖
and ―How good is communication with your study child?‖ The following three negative
items about the relationship, or conflict, were also averaged: ―How much conflict do you
feel there is between you and your study child (or mother/father)?,‖ ―How much do you
feel this child is critical of you or what you do?,‖ and ―How much does your study child
25

argue with you?‖ The positive scale had a Chronbach‘s alpha of greater than 0.85, and the
negative scale had a Chronbach‘s alpha of greater than 0.65 for each generation. These
items were rated from 1 to 6, with 6 indicating greater positive or negative aspects. The
commonly used ―Griffin‖ formula (Thompson, Zanna, & Griffin, 1995) was then applied
using the two scales:
Ambivalence = [ (Positive + Negative) / 2 - | Positive – Negative | ] + 1.5
This formula takes the average of both positive and negative aspects to evaluate
the strength of both emotions, which is then subtracted by the difference to evaluate the
similarity of both emotions. The addition of 1.5 to the original formula eliminates
negative scores, so the range is from 0 to 5.25, with a higher score indicating more
ambivalence. Individuals with a high degree of both positive and negative feelings have a
higher score and are conceptualized to be more ambivalent.
Both the direct and indirect measures reflect ―attitudinal‖ or psychological
ambivalence within the intergenerational relationship and fit the concept of a mixture of
good and bad aspects about the relationship.
Psychological Well-Being
Depression and self-esteem are used to measure ―psychological well-being.‖ The
depression scale includes 20 statements relating to depression symptoms, rated by
frequency of experience within the past week, from rarely or none (0) to most or all of the
time (3). Each item is averaged to create a scale from 0 to 3. Self-esteem will be
measured as the average of 12 items, which were rated from strongly agree (0) to strongly
26

disagree (3), where a higher score indicates greater self-esteem. These variables serve as
independent and dependent variables, depending on the model.
Relationship Characteristics
Relationship quality is measured by one question asking ―as parents and children
get older, relationships sometimes change. Compared to four years ago, would you say
that your relationship with your ―study‖ child (or parent) has: declined considerably,
declined somewhat, stayed about the same, improved somewhat, or improved
considerably?‖ The responses range from 0 to 4, with a high score indicating
improvement. This variable is taken from both parent and child perspectives.
Proximity to each other is measured by an ordinal item, taken from the children's
perspectives, asking ―how far from you do your mother and father live?‖ The response
options include: we live together, less than 5 miles, 5-50 miles, 51-150 miles, 151-250
miles, 251-500 miles, and more than 500 miles. A dummy variable will be constructed to
identify children who live less than or greater than 50 miles from their parents.
Frequency of contact includes the frequency of both in-person and contact by
phone. Both types of contact are measured ordinally, from not at all (0), once or twice a
year, several times a year, every month, every week, and daily or more often (5).
Frequency of contact is measured from both the parent and child perspectives.
Instrumental support will be measured by parents‘ and children‘s perceptions of
what they receive from the other by asking respondents whether they received any
support in the following domains: household chores, transportation, financial assistance,
help when sick, and assistance with personal care.
27

Emotional support will be indexed by these four domains: information/advice,
emotional support, discussing life decisions, and visiting/sharing leisure time. The
measurement of receipt of support will be taken from the individual‘s perception (not
what parents or children say they provide to one another) to account more for perception
of support received.
Similarity between both generations was measured by one item asking: ―in
general, how similar are your opinions and values about life to those of your child (or
parent) at this point in time?‖ The responses include not too similar (0) to extremely
similar (5). Responses were taken from both perspectives in the dyad.
Individual Characteristics
Individual characteristics such as age, birth cohort, race (white and non-white),
gender of child and parent (if mothers and fathers are analyzed together), years of
education or a dummy variable measuring graduation from college, marital status,
parental status of the child, and work status are included.
28

Chapter Three:
Mixed Feelings, Mixed Measures: Measuring Ambivalence among
Older Parent-Adult Child Dyads
A. Abstract

This research compares two commonly used measurements of intergenerational
ambivalence between older parents and their adult child. I compared direct and indirect
measures of ambivalence among 253 older parent-adult child dyads in the Longitudinal
Study of Generations. Children had greater average scores for indirect ambivalence, but
there were no differences between parents and children in average direct ambivalence
scores. Both measures were moderately correlated and multivariate regression analyses
examined the relationship between the types of ambivalence and related individual and
relationship characteristics (demographics, health, dependency, similarity, filial norms,
and provision of support). Results show significant differences, suggesting that indirect
and direct measurements may be tapping into related, but overlapping constructs. Future
research should consider possible generational differences when measuring ambivalence
and the conceptual differences between both ambivalent feelings measured directly and
ambivalent relationship qualities measured indirectly. I propose the underlying concept of
the indirect variable is ―implicit‖ ambivalence and the direct measure relates to ―explicit‖
ambivalence.
B. Introduction
Great strides in the past decade to investigate intergenerational ambivalence have
led to a variety of theories, methods, and results from both older parents‘ and adult
29

children‘s perspectives. The purpose of this research is to integrate some of these diverse
studies and compare two common strategies for measuring intergenerational
ambivalence. Previous studies raise questions about the ability to compare studies that
use these two different direct and indirect measurements. Despite the interest in
ambivalent relationships between parents and children, few studies have shown whether
or not various measures are tapping into the same construct. This research will answer the
following questions: 1) How similar are the two measures within each generation? 2) Do
parents and children have similar levels of ambivalence, within each type of ambivalence
measure? 3) Are direct and indirect measurements equivalent in terms of the theoretical
factors associated with ambivalence?
History of the Concept of Ambivalence

The idea that forces of attraction and repulsion are simultaneously present in the
closest personal relationships, such as parent-child relationships, formed the basis for the
psychoanalytic concept of ambivalence. According to this perspective, ambivalence is the
mix of positive and negative emotions toward the same relational object, and has long
been considered a source of neurosis deriving from early childhood attachment problems
(Freud, 1913). Sociologists moved the concept of ambivalence beyond its pathological
implications by determining it to be an intrinsic property of human relationships
structured by irreconcilable demands for opposing behaviors toward another (Merton &
Barber, 1963). Coser (1956) elaborated that ―converging and diverging motivations may
be so comingled in the actual relationship that they can be separated only for
classificatory or analytical purposes, while the relationship actually has a unitary
30

character sui generis.‖ (p. 64). Luescher and Pillemer (1998) extended the application of
ambivalence to intergenerational relations, critiquing the ―either/or approach‖ to
affectionate and conflictual aspects of mature parent-child relationships. Given that
family life has its basis in the tension between the desire for autonomy and the need for
interdependence, it is not surprising that intergenerational relations—throughout the
family life-cycle—are among the most ambivalent of social relationships (Fingerman,
Hay, & Birditt, 2004). Luscher and Pillemer (1998) defined intergenerational
ambivalence as ―contradictions [in relationships] at the level of social structure,
evidenced in institutional resources and requirements, such as statuses, roles, and norms
and contradictions at the subjective level, in terms of cognitions, emotions, and
motivations‖ (p. 416).
Intergenerational relationships have typically been theorized and researched
within the solidarity-conflict paradigm. This paradigm is multidimensional, including
seven dimensions: affectual, consensual, functional, associational, structural, normative
solidarity, and the most recently added dimension, conflict (Bengtson et al., 2002). The
dimensions of solidarity describe intergenerational relationships in terms of opposites
(intimacy versus emotional distance, for example). The addition of conflict allowed
researchers to more completely capture the complexities in family life, where,
paradoxically, some parents experience high solidarity on some dimensions (even
affection), but still experience conflict with their child. Research demonstrates the
application of the solidarity-conflict paradigm to intergenerational ambivalence
(Bengtrson et al., 2002). Empirical research has explored different ways of measuring and
31

explaining this complex psychological and sociological concept, using both quantitative
and qualitative strategies. This paper, however, will focus on the differences between two
commonly used quantitative measurements of psychological intergenerational
ambivalence.
Indirect Measure of Ambivalence
Various techniques have been used to indirectly capture the potential for
psychological ambivalence, including additive scales of discordant (positive and
negative) measures that describe the intensity of opposing feelings (Willson et al., 2006),
and classification procedures to group relationships into ambivalent and non-ambivalent
types (Giarrusso, Silverstein, Gans, & Bengtson, 2005; Hogan, Eggebeen, & Clogg,
1993; Lawton , Silverstein,& Bengtson, 1997; Steinbach, 2008; Van Gaalen & Dysktra,
2006). One such indirect measure of ambivalence uses a set of statements about positive
and negative feelings, or relationship qualities, and calculates a score that reflects the
balance between positive and negative feelings (Fingerman et al., 2008; Thompson,
Zanna, & Griffin, 1995; Willson, Shuey, & Elder, 2003).
Noticeably, the positive and negative items typically used are similar to
those used to measure affectual solidarity and conflict within the solidarity paradigm.
Fingerman and colleagues (2006) use the following two items to assess positive qualities
of the relationship: ―How much does he or she make you feel loved and cared for?‖ and
―How much does he or she understand you?‖ They used these two items to assess
negative feelings: ―How much does he or she criticize you?‖ and ―How much does he or
she make demands on you?‖ This specific strategy of combining two independent
32

dialectical assessments of an attitude is known as the Griffin formula, derived from work
by Thompson, Zanna, and Griffin (1995) .This formula combines two necessary aspects
of ambivalence: similar magnitude of the negative and positive components and at least
moderate intensity of both components. For instance, an individual who reports the
highest positive and negative feelings would be considered very ambivalent, while an
individual with high positive and low negative (or vice versa) feelings would be
considered to have low ambivalence. Someone with lower levels of both positive and
negative feelings (while similar in magnitude) does not the intensity of mixed feelings to
be considered ambivalent (Thompson, Zanna, & Griffin, 1995).
Direct Measure of Ambivalence
Pillemer and Suitor (2002) developed a quantitative survey-based strategy to
directly measure ambivalence. Their measure asks respondents to rate the degree to
which they have ―mixed feelings,‖ ―get on each other‘s nerves, but nonetheless feel
close.‖ or feel ―torn in two directions‖ toward a parent or child (Pillemer et al., 2007, p.
782). These items are then averaged or summed to create an ordinal-interval level scale.
The direct measure of ambivalence has been utilized in several studies (Lowenstein,
2007; Pillemer et al., 2007; Pillemer & Suitor, 2002) and shows high reliability to tap into
ambivalence or mixed feelings about the target individual. To date, these direct measures
have not been utilized as often as the indirect formula, but published research shows
promising results for better understanding families using a more direct approach to
ambivalence.

33

Direct Measure versus Indirect Measure
The direct measurement of ambivalence developed by Pillemer and Suitor (2002)
requires respondents to evaluate their contradictory feelings and be aware of having
―mixed feelings‖ for these items to accurately reflect ambivalence. Are individuals who
recognize their ambivalence towards a loved parent or child different from individuals
who may not be cognizant of ambivalence? The direct measurement method taps into a
more explicit and expressed experience of ambivalence within the intergenerational
relationship.
In contrast, the indirect measure may be able to tap into both conscious and
subconscious ambivalent feelings. While the respondent would need to confirm
experiencing conflict in the relationship, he or she might not necessarily be aware of
experiencing both positive and negative feelings. This indirect method allows
respondents to report the co-occurrence of positive and negative feelings independently,
which then can be evaluated by researchers as potentially ambivalent. This method may
be tapping into a potentially ambivalent relationship of which the individual may be
unaware; whereas, the individual must be able to express their ambivalence when using
the direct measurement.
Theoretical Factors Associated with Ambivalence
The theoretical model of intergenerational ambivalence posits that macro- and
micro- level forces impact the parent-child relationship, creating ambivalent situations
and feelings. Previous research has focused on some of the possible social structural,
34

relationship, and personal characteristics associated with ambivalence, using both direct
and indirect measurements.
At the macro- and meso-levels, social structures such as age, gender, and race
may lead to ambivalence. Additionally, lengthening lifespan, workforce demands, and
competing norms of filial obligations and autonomy may lead to ambivalent situations.
The empirical work using the direct measurement has demonstrated associations between
ambivalence and competing norms and values, women‘s roles within the family, and
tensions between autonomy and dependency (Pillemer & Suitor, 2005).
At the micro-level, shared familial histories, inequitable exchanges of support,
personality, and poor health are potential factors contributing to intergenerational
ambivalence (Fingerman, Hay, & Birditt, 2004; Fingerman et al., 2006; Lüscher, 2004).
Ambivalence has also been associated with overall similarity between the generations
and the child's characteristics and status attainment (health, education, marital status, and
neuroticism) (Fingerman et al., 2008; Pillemer & Suitor, 2002; Willson, Shuey, & Elder,
2003). These characteristics are commonly found to relate to ambivalence for both
parents‘ feelings about their children and adult children‘s feelings about their parents.
Several studies have also examined the outcomes of ambivalent relationships in terms
of well-being for the older generation. Fingerman et al. (2008) found that greater
ambivalence was associated with more depression and less life satisfaction. While
affection has a stronger association with better quality of life, ambivalence was slightly
associated with worse quality of life (Lowenstein, 2007). Outcomes of ambivalence
should be evaluated with caution, since the literature has only employed cross-sectional
35

data. However, the different types of ambivalence measures, if capturing overlapping
concepts of ambivalence (manifest vs. latent) has implications for the consequences of
having these mixed feelings.
Generational Stake Phenomenon
The generational stake phenomenon refers to the tendency of parents to over-
report more positive attitudes about their relationships with their children (Giarrusso,
Feng, & Bengtson, 2004). This positivity bias in the parental generation is in contrast to
an underreporting of affection by the younger child generation. This phenomenon or
incongruency in parent-child reports about intergenerational relations has been observed
since the earliest work on the solidarity paradigm (Bengtson & Kuypers, 1971). Parents
invest emotional and physical work in raising children, taking pleasure and pain in their
relationship. Over-emphasizing the pleasurable, affectionate qualities of the relationship
with their children is a potentially beneficial coping mechanism.
Researchers have tested several psycho-social and social structural theoretical
explanations for these discrepancies with mixed findings. Giarrusso, Feng, and Bengtson
(2004) reviewed four possible theories regarding developmental stages, status-
inheritance, stress and life-events, and gender that could predict parents‘ positivity bias
(and conversely, children‘s negativity bias). Developmentally, parents have an interest in
closer relationships with their children to feel successful in their role as ―parent,‖ while
children are more interested in establishing independence, thus under-emphasizing
parental closeness. Additionally, stressful life events may elicit differences in emotional
qualities between the parent and adult child. For instance, where a child‘s divorce may
36

increase her feelings of closeness towards her mother, it may decrease closeness from the
mother‘s perspective. Similarity of values and social status is consistently found to be
associated with agreement in emotional aspects between the generations.
Interdependence, similarity, and common interests (or equal stakes within the
relationship) lend themselves to a greater level of perceptual agreement about the quality
of the relationship (Shapiro, 2004).
While this bias has been observed in terms of affectual solidarity (Giarrusso,
Stallings, & Bengtson, 1995) and for perceptions of conflict (Fingerman, 1996) between
older parents and adult children, it has not been observed in reports of ambivalence for
either type of measurement. The generational stake phenomenon has been found to work
differently, depending on the dimension of the relationship being measured (such as
contact, provision of support and conflict) (Shapiro, 2004). Thus it is likely that
ambivalence may be perceived differently by the parent and child. The generational stake
hypothesis would predict that older parents, who have a stake in positive relationships
with their adult children, are less likely to be aware of or report mixed feelings and
ambivalence than their children. However, utilizing an indirect measurement of
ambivalence, including reports of some conflict and affection, may be more likely to tap
into implicit feelings of ambivalence among older parents and their children.
C. Research Questions
Drawing from questions raised about previous studies utilizing direct and indirect
measurements, in addition to the potential influence of the generational stake on
ambivalence, this investigation has three primary research questions: 1) How similar are
37

the two measures within each generation? If they are both measuring the same concept,
we would expect the two measures to be nearly perfectly correlated. 2) Do parents and
children have similar levels of ambivalence, within each type of ambivalence measure?
The generational stake phenomenon suggests parents would be less likely to report
conflict and mixed feelings, focusing more on the positive aspects of their relationship.
Thus, we predict parents to have less ambivalence within each measurement, compared to
their child. 3) Are direct and indirect measurements equivalent in terms of the theoretical
factors associated with ambivalence? If both measures tap into the same construct of
intergenerational ambivalence, we would expect the same individual and relationship
characteristics, used in previous research, to be significantly associated with both
measures.
This research is important for investigating whether the concept of ―ambivalence‖
is comparable, regardless of whether it is implicitly or explicitly measured. Additionally,
it is important to know whether these measures are valid for both the parent and child
generations.
D. Method
Sample

In this analysis, the sample is composed of 253 parent-child reciprocal dyads
where parents report about the child and that same child reports about that parent. We
used the latest wave of data from the University of Southern California‘s Longitudinal
Study of Generations (LSOG), collected in 2004. The LSOG began in 1971 with 2,044
original respondents who were members of three-generation families. Grandparents (G1)
38

were selected via a multi-stage stratified random sampling procedure from a population
of 840,000 individuals enrolled in southern California's first large HMO (see Bengtson &
Schrader, 1982 for further details). Adult children (G2) and grandchildren (G3) of the G1
grandparents were also invited to participate in the survey. Follow-up surveys were
administered seven times from 1985 to 2004. In the LSOG survey, the G2 mothers and
fathers were matched with a randomly designated G3 ―study child,‖ for whom they
consistently answered about throughout the waves of the study, making possible the
creation of reciprocal dyads. In 2004, all parents were at least 60, but averaged 75 years
of age. The sample includes 153 mothers and 100 fathers. The ―study children‖ were
between the ages of 38 and 60, and on average, 53 years old. Over 60% of the adult
children were daughters, and over 90% of all the respondents were white. Almost 72% of
parents and 70% of children were married at the time of the survey. The older generation
had four children, and the younger generation had three children, on average. Over 85%
of the adult children were still working, at least part time, and both generations had above
a high school education. About half of the parents lived within an hour of their adult
child.
Measures
Ambivalence. The dependent variables included two measurements of
ambivalence, taken from both the parent and child perspectives. The direct measurement
is an index of three questions asked of both generations about the other. These questions
were: ―How often do you feel torn in two directions about your study child at this point in
your life (never, seldom, now and then, often, or very often)?;‖ ―I have mixed feelings
39

about this daughter or son;‖ and ―My study child and I often get on each other‘s nerves,
but nevertheless we feel close (strongly disagree, disagree, agree, strongly agree).‖ The
scores for each were averaged and range from 1 to 4.5, with 4.5 indicating greater
ambivalence. This scale had high reliability for each dyad type (father-child, child-father,
mother-child, and child-mother); Chronbach‘s α ranged between .58 and .72.
Ambivalence was also assessed using an indirect measurement for each of the
perspectives. We averaged three items measuring positive aspects of the intergenerational
relationship: ―How close do you feel is the relationship between you and your study child
(or mother/father)?,‖ ―How well do you get along?,‖ and ―How good is communication
with your study child?‖ The following three negative items about the relationship, or
conflict, were also averaged: ―How much conflict do you feel there is between you and
your study child (or mother/father)?,‖ ―How much do you feel this child is critical of you
or what you do?,‖ and ―How much does study child argue with you?‖ The positive scale
had a Chronbach‘s alpha of greater than 0.85, and negative scale had a Chronbach‘s alpha
of greater than 0.65 for each generation. These items were rated from 1 to 6, with 6
indicating greater positive or negative aspects. The commonly used ―Griffin‖ formula
(Thompson, Zanna, & Griffin, 1995) was then applied using the two scales:
Ambivalence = [ (Positive + Negative) / 2 - | Positive – Negative | ] + 1.5.
This formula takes the average of both positive and negative aspects to evaluate
the strength of both emotions, which is then subtracted by the difference to evaluate the
similarity of both emotions. The addition of 1.5 to the original formula eliminates
negative scores, so the range is from 0 to 5.25, with a higher score indicating more
40

ambivalence Individuals with a high degree of both positive and negative feelings have a
higher score and are conceptualized to be more ambivalent.
Demographic and Health Characteristics. The following characteristics of both
the parents and children were included in the analyses: gender, age, marital status
(currently married), number of years of education, number of living children, work status
(currently working vs. not working), self-rated health (scale of 1 to 4 from poor to
excellent), depression (20 averaged items from the CES-D scale, where a higher score
indicates more severe depressive symptoms), and self-esteem (average of 12 items,
indicating greater self-esteem). These independent variables were included in the
analyses as potential correlates of ambivalence, which have been used in previous studies
that utilized both direct and indirect ambivalence measures (Fingerman et al., 2006;
Fingerman et al., 2008; Ha & Ingersoll-Dayton, 2008; Pillemer & Suitor, 2002).
Parent-Child Relationship Characteristics. Additionally, the following variables
were also included to assess aspects of the relationship perceived from both parent and
child perspectives. The provision of support was measured with a dichotomous variable
indicating whether the respondent provided at least one type of support. The possible
types of support exchanged were information and advice, emotional support, discussing
important life decisions, visiting or sharing leisure activities, giving help when sick,
financial support, help with household chores, transportation, and help with personal care
or hygiene. This measurement was taken from both the parents‘ perspective of support
provided to children and children‘s perspective of support provided to parents.
41

Both the parents' and children‘s filial norms, or expectations about whether adult
children should care for aging parents, were also included from both perspectives in the
dyads. Respondents were asked to indicate how much responsibility children should
have in providing the following: companionship, household chores and transportation,
advice and guidance, personal and health care needs, financial support, and housing. The
scale ranged from 1 to 5, where a score of 5 indicated children should take total
responsibility.
Consensual solidarity, or similarity, was measured by one item on a scale of 1 to
6 (―How similar are your opinions and attitudes about life to those of your child (or
parent) at this point in time?‖).
Analyses
Correlational analyses compared the direct and indirect ambivalence
measurements within each generation, and paired-samples T-tests compared the two types
of ambivalence across generations. Multivariate regression analyses were estimated to
determine the relationship between the independent variables (parents‘, children's, and
relationship characteristics) with both of the ambivalence measurements for each
generation. The regression analyses were run using Mplus Version 4.0 to simultaneously
predict both direct and indirect ambivalence, accounting for the covariance between the
dependent variables for both the child and parent perspectives. Incomplete cases were
analyzed using maximum likelihood estimations to account for data missing at random
and reduce response bias. Additionally, standard errors were adjusted based on family
membership (using the cluster function in Mplus) because the cases were not independent
42

(fathers and mothers could be spouses; when spouses or former spouses were included,
the same child was used in both mother and father dyads).
E. Results
Between-Generation Differences
Firstly, how prevalent is direct and indirect ambivalence? Approximately one
third of both children and parents agreed with the three direct ambivalence items. This
finding is consistent with previous prevalence reported for the direct measure (Pillemer &
Suitor, 2002). Secondly, we compared means of both types of ambivalence between the
generations to answer the first research question: Do parents and children have similar
levels of ambivalence, within each type of ambivalence? The mean for the direct
ambivalence for fathers about children (m = 1.92) and for children about fathers (m =
2.03) was not statistically significantly different (t = 0.877). Similarly, there was no mean
difference for mothers‘ direct ambivalence about children (m = 2.02) and children about
mothers (m = 2.07).
However, there was a significant difference between children‘s and parents‘
indirect ambivalence. Children had a significantly greater mean indirect ambivalence
score towards both mothers and fathers (see Table 1). To further explore differences
within the indirect ambivalence measure, we analyzed differences between the
components of indirect ambivalence: positive and negative relationship qualities. Both
mothers and fathers were significantly more likely to report greater positive feelings
about their child. Children also reported greater negative aspects about their mothers.
43

The differences in indirect ambivalence scores between children and parents may be due
to the greater ratings of positive aspects reported by the parents.

Table 1.1. Parent-Child Comparison of Means (Paired T-tests) for Ambivalence,
Positive, and Negative Measures (n = 253)

Indirect
Ambivalence

Direct
Ambivalence

Positive Negative
Father-Child Dyad (n = 100)

Child 2.032 2.032 4.2 1.8
Father 1.647 1.916 4.6 1.7
Paired t-test 2.66** 0.877 -4.1** 1.3

Mother-Child Dyad (n = 153)

Child 1.802 2.067 4.43 1.98
Mother 1.538 2.012 4.6 1.8
Paired t-test 2.39* 0.536 -2.62* .2.0*

* * p < 0.01; * p < 0.05

44

Between-Measurement Differences
Next, we compared the two types of ambivalence measures to answer the second
question: How similar are the two measures within each generation? For each
generation, the direct and indirect measurements were significantly correlated (see Table
2). The correlations between indirect and direct ambivalence in the father-child dyads
was .49 and .59 for mother-child dyads (p < .001). All correlations indicated moderate
convergent validity between the two types of ambivalence, within each generation. To
more specifically test discriminant validity, we correlated direct ambivalence with the
individual positive and negative components comprising indirect ambivalence with the
assumption that each is measuring different concepts. All were statistically significantly
correlated, although they indicated weaker associations than that between the two types
of ambivalence. These findings suggest direct ambivalence is also conceptually related to
independent assessments of affection and particularly, conflict.

Table 1.2. Correlations between Ambivalence, Positive, and Negative Measures

Indirect and
Direct
Direct and
Positive
Direct and
Negative
Fathers about Children .488** -.387** .362**
Children about Fathers .494** -.293** .477**

Mothers about Children .589** -.572** .613**
Children about Mothers .587** -.379** .603**

Parent about Child .548** -.500** .528**
Child about Parent .551** -.356** .544**
** p < .001

45

Parents’ Intergenerational Ambivalence
Multivariate regression analyses, reported in Table 3, were conducted to examine
whether both measurements were associated with the same characteristics. To establish
construct validity, we would expect both measurements to be related to the same
variables. The analyses included the parents‘ and children‘s ambivalence for both
measures as four dependent variables, regressed on the selected individual characteristics
for the parent and child and relationship characteristics. For the parents‘ indirect
ambivalence, the parents‘ age (b = - 0.038), children‘s marital status (b = - 0.445),
parents‘ reported similarity with children (b = - 0.372), and frequency of contact (b = -
0.299) were statistically significant. Parents had greater levels of indirect ambivalence
when they were younger, their children were not married (including divorced, widowed,
or never married), when they felt less similar to their children, and when they had less
frequent contact with children. The r-squared for this outcome was .325.
46

Table 1.3. Multiple Regression Coefficients of the Effects of Characteristics on
Parents‘ and Children‘s Direct and Indirect Ambivalence

Parents' Perspective Children's Perspective
Indirect Direct Indirect Direct
Parents' Demographics

Female -1.87 0.185 -.358* 0.024
Age -.038* 0.006 -0.025 0.007
Currently Employed -0.109 -0.006 -.468* 0.061
Number of children -0.033 -0.013 -0.069 -.078*
Married 0.116 0.17 0.109 -0.031
Parents' Health

Good Self-Rated Health 0.108 0.204 -0.366 0.064
Depression 0.382 .296* -0.053 0.165
Self-esteem -0.013 -0.021 0.052 0.098
Children's Demographics

Female -0.022 0.036 -0.08 -0.103
Currently Employed 0.12 0.089 -0.102 -0.061
Years of Education -0.008 -0.002 -0.015 -0.001
Parent 0.062 0.037 0.118 0.061
Married -.445* -.348* 0.037 0.082
Children's Health

Good Self-Rated Health -0.691 0.054 0.039 -0.12
Depression -0.141 0.023 -0.24 -0.002
Self-esteem -0.2 -0.053 -.491* -.467*
Relationship Characteristics

Parents' Filial Norms -0.005 0.091 0.013 -0.022
Childs' Filial Norms -0.117 -.174* 0.034 .189*
Parent similarity w/ child -.372* -.208* -0.037 0.001
Child similarity w/ parent -0.123 -0.015 -.504* -.152*
Support from parent 0.071 0.026 0.297 .270*
Support from child -0.225 -0.328 0.532 0.366
Live over 50 miles apart 0.14 0.005 -.720* -0.197
Frequency of contact -.299* -0.111 0.251* .166*
r-square 0.325 0.309 0.326 0.235

*p <.05
47

For the parents‘ direct ambivalence, parents‘ depression score (b = 0.296), the
children‘s marital status (b = - 0.348), children‘s filial norms (b = - 0.174), and parents‘
reported similarity with the child (b = - 0.208) were statistically significant (see Table 3).
Parents had greater levels of direct ambivalence if they experienced greater levels of
depression. Parents also had greater direct ambivalence when their children were not
married (including divorced, widowed, or never married), had weaker filial norms, and
when they felt less similar to their children. The r-squared for this outcome was 0.309.
As measured from the parents‘ perspectives, the two types of ambivalence
measures were primarily related to different variables, except for the children‘s marital
status and similarity. Additionally, slightly more variance in indirect ambivalence was
explained by the selected variables.
Children’ s Intergenerational Ambivalence
Table 3 also shows the multivariate regression results to estimate the children‘s
indirect and direct ambivalence. The gender (b = - 0.358) and employment status (b = -
0.468) of the parents were significantly associated with children‘s indirect ambivalence.
Additionally, children‘s self-esteem (b = - 0.491), similarity with parents (b = - 0.504),
proximity (b = - 0.72), and frequency of contact (b = 0.251) were also statistically
significant. Children had greater indirect ambivalence towards fathers and parents who
were not currently employed. Children with lower self-esteem also had greater indirect
ambivalence. Greater indirect ambivalence was associated with children who felt less
similar to parents, lived within 50 miles of their parents, and had more contact with
48

parents. The r-square was .326, indicating similar explanatory power of the variables for
indirect ambivalence of parents and children.
Greater levels of direct ambivalence towards parents was significantly associated
with the children having fewer siblings (b = - 0.078), lower self-esteem (b = - 0.467),
stronger filial norms (b = 0.189), less similarity with parents (b = - 0.152), and greater
frequency of contact with parents (b = 0.166). As opposed to the association between
filial norms and parents‘ direct ambivalence, children experienced greater direct
ambivalence when they had stronger filial norms. Additionally, when the parents reported
providing at least one form of support to children, the children experienced greater direct
ambivalence (b = 0.270). The r-square for children‘s direct ambivalence was 0.235,
which is 10% lower than for indirect ambivalence.
For children, there was overlap between the types of ambivalence in terms of self-
esteem, similarity to parents, and frequency of contact. As observed for parents, there are
mixed results about the construct validity between indirect and direct ambivalence.
F. Discussion
The goal of this research was to compare two commonly used measurements of
psychological intergenerational ambivalence. We also used both the parent and child
perspectives to investigate generational differences within each measurement. These two
methods of quantitatively measuring ambivalence are widely used in the aging families‘
literature, but likely capture different aspects of the concept. Given their conceptual
differences discussed previously and confirmed by the empirical findings, I propose
49

calling ambivalence measured using the indirect method implicit ambivalence. The
underlying concept measured by the direct method will be labeled explicit ambivalence.
Implicit Ambivalence: A Generational Stake?
This study suggests both similarities and differences between the two types of
measurements across two generational perspectives. Parents and children seem to report
similar levels of explicit ambivalence, but children have greater levels of implicit
ambivalence as measured indirectly. The greater implicit ambivalence for children about
their parents may be due to parents reporting more solidarity than children. In the implicit
ambivalence formula, similar and moderate to high levels of both the positive and
negative scales indicate ambivalence. However, if parents are more likely to report
greater positive aspects (and mothers are less likely to report negative), their implicit
ambivalence score will be smaller than children who are reporting more equal levels of
positive and negative aspects. The differences in the component parts of implicit
ambivalence may explain the generational differences, which were not observed for
explicit ambivalence. The generational stake phenomenon explains how measuring
positive and negative aspects independently may not provide an accurate picture about
parents‘ feelings because they over emphasize positive and underemphasize negative
feelings. The multivariate findings also suggest, as expected, that implicit ambivalence
for parents is distinct from ambivalence for children.
Conceptual Overlap between Latent and Manifest Ambivalence
The indirect and direct measurements were compared to test convergent and
construct validity using many variables theorized to relate to ambivalence. The literature
50

cites interdependency, children‘s problems, social structures, and personality
characteristics as being associated with ambivalence. Both measures were moderately,
but not perfectly, correlated with each other, suggesting that similar, but not equivalent,
underlying concepts are being captured by the two measurements. A multivariate
regression further investigated the possible different factors associated with the two
related measures. While difficult to discern and explain the conceptual differences and
overlap, we discuss several of the patterns interpreted from the analysis. As discussed
previously, the indirect measure is tapping into implicit or potential ambivalence. The
parent or child is not necessarily aware of their mixed feelings, but has a relationship with
both conflict and affection, potentially leading to ambivalent feelings. The direct measure
is tapping into the conscious recognition of holding contradicting emotions, thus I label
this explicit ambivalence.
Interdependency
For both generations, explicit and implicit ambivalence we associated with
variables related to interdependency, or the tensions between autonomy and dependence.
Children‘s implicit ambivalence was related to living closer to their parents and having
more frequent contact. The explicit measurement of ambivalence was also associated
with more frequent contact, indicating overlap in some aspects of interdependency.
Children who had stronger feelings about providing care to older parents were also more
ambivalent when measured explicitly. Also highlighting the importance of autonomy for
the adult children, children experienced more mixed feelings when receiving support
from their parents. These relationships indicate that physical closeness and contact, which
51

encourages more interdependency between parents and their children, may increase
opportunities for conflict, in the midst of affection. However, normative expectations and
dependency on parents for support does not relate to active conflict (as measured in the
implicit measurement), but to an internal uncertainty (―torn in two directions‖) about
their relationship with their parents. Despite having stronger filial piety, these children
may feel torn about their desires to care for their parents and also desiring their own
independent lifestyles without parental caregiving responsibilities. This interesting
relationship between filial norms and ambivalent feelings should be explored further, as it
seems to represent the competition between social norms and personal desires that will
likely influence supportive behaviors in the future. The pull of strong filial norms to
provide care and the push of ambivalent feelings to withdraw create a structurally
ambivalent context for adult children, as theorized by Connidis and McMullin (2002).
However, this relationship was only found for the explicit ambivalence measure, raising
questions about the validity of implicitly measuring ambivalence.
For parents, the relationship between interdependency and ambivalence seems
different from that of children. Parents who report having less contact with their child are
more ambivalent on the implicit scale. As opposed to the relationship between contact
and conflict from the child‘s perspective (where more contact is associated with more
conflict), the older parent‘s lack of contact may suggest more disagreements and tensions
in the relationship. Interestingly, the children‘s reported filial norms were also associated
with parents‘ explicit ambivalence: when children reported weaker filial piety, the parent
reported more ambivalence. This may reflect the parent‘s knowledge of his child‘s
52

attitudes and a reluctance to rely on the child, producing more mixed feelings about that
child. Future research should consider older adults in more intensely dependent,
caregiving situations.
Structural and Social Forces
Children seem to have more implicit ambivalent feelings about fathers than
mothers and when a parent is not working. Being retired or unemployed may be an
indicator to an adult child of her parents‘ possible future dependency and aging, which
leads to more conflicted feelings that are not explicitly acknowledged by the child
through the explicit measurement.
Children with more siblings experienced less explicit ambivalence, indicating that
knowledge of more resources to help care for aging parents may reduce mixed feelings.
The relationship between parents and children with many siblings may also be more
diffused and less susceptible to mixed feelings. The number of siblings as a future
caregiving resource is an interesting topic for further exploration since caregiving
research typically shows that even when there are many siblings; one or two children
(usually daughters) end up providing the bulk of care. The current expectations for
siblings to step up to the plate may be driving the lower levels of ambivalence seen in this
study.
Children’ s Successes and Problems
Our analyses did not include indicators of more serious children‘s problems or
failures (such as drug abuse or attitudinal differences). However, parents were more
ambivalent on both the implicit and explicit scales when they felt less similar to their
53

child and when their child was never married or divorced. Both measures seem to be
tapping into similar concepts, in that consensus between the generations is related to
more ―certain‖ relationships, where each generation may have more clear expectations
about the other and fewer circumstances to experience ambivalence. Additionally, parents
may have more satisfaction and feelings of success as a parent when their child is similar
to them.
Personality and Health
Explicit ambivalence was associated with parents‘ depression. The relationship
between depression and explicit ambivalence is not causal and could indicate a unique
tendency for more depressed parents to experience or perceive ―mixed feelings‖ due to
poorer mental health.
Both measurements of ambivalence suggest a relationship between mental health
and ambivalence for the child generation. Children with lower self-esteem had greater
levels of both types of ambivalence towards parents. Arguably, self-esteem could be a
predictor or an outcome of a very ambivalent relationship to a close relative. Previous
studies have found that psychological well-being, quality of life, and neuroticism are
associated with intergenerational ambivalence. Further longitudinal studies should
disentangle the causal relationships between these aspects.
Limitations
The sample is not representative of all parent-child relationships, as the sample is
based on a regional population and does not represent a racially or ethnically diverse
population. Our results and conclusions should be generalized only to white middle-aged
54

children and their young-old parents. The research literature utilizes other measurements
of intergenerational ambivalence, including latent classes and qualitative strategies. We
do not think this paper presents a definitive statement on all measures or conceptions of
ambivalence, only the specific chosen measures.
Conclusion
This research suggests that the two common ways of empirically measuring
intergenerational ambivalence, though related, may represent different constructs. We do
not conclude that one measurement type is better or ―more valid,‖ but rather that
researchers should be aware of these possible differences when interpreting results and
comparing studies with different measurement types. We propose using the terms explicit
ambivalence (the conscious recognition of mixed feelings identified by the explicit
measurement) and implicit ambivalence (the coexistence of conflict and affection not
necessarily cognitively experienced) to distinguish between the different ―strains‖ of
intergenerational ambivalence. Specifically, implicit ambivalence is more associated with
circumstances where there is more opportunity for conflict in an already affectionate
relationship (frequency of contact, proximity, gender differences) and explicit
ambivalence is related to mental health and social norms (depression, filial norms). We
think explicit acknowledgement of the likely conceptual differences between these two
measurements (as well as the overlap) in the research literature will greatly benefit the
theoretical development of ambivalence and fine-tune our understanding of family
relationships.

55

Chapter Four:
A Decade of Love and Hate: Patterns of Intergenerational Ambivalence
Experienced by Two Cohorts of Older Parent- Adult Children Dyads
A. Abstract
The coexistence of affection and conflict towards an older parent or adult child, known as
intergenerational ambivalence, has sparked great interest in the study of aging families.
Researchers have demonstrated that families experience more intimate than conflicted
relationships, but as a result of the increase in more complex family structures, it is also
important to understand the contradictions and problems inherent within these important
family ties. Studies have investigated many social, familial, and individual characteristics
associated with parent-child ambivalence; however, we know little about how families
experience implicit ambivalence over the life course. This research used five waves of the
Longitudinal Study of Generations from 1991 to 2005 to assess the patterns of
ambivalence for both parents and children. Intergenerational ambivalence was assessed
using an indirect measurement combining both positive and negative aspects of the
relationships. The analyses included two cohorts of parent-child dyads (N = 900) to
compare ambivalence by familial perspective (parent versus child) and birth cohort
(younger versus older generation) using latent growth models. Average scores indicated
low to moderate levels of ambivalence for parents and children, which declined over the
life course. Regardless of birth cohort, children experienced greater ambivalence than
their parents. Cohort differences suggest greater ambivalence among baby-boomer
parents (born 1945-1955) and their children (born 1978-1983) than among the older
56

dyads. Differences in trajectories suggest that younger parents have greatest ambivalence
prior to launching their children, which sharply declines as their children enter adulthood.
The older cohort of parents experienced more stable levels of ambivalence that indicate
an increase with older age. These results are discussed in terms of potential historical
forces shaping cohort differences versus the differences in developmental life stage. This
research supports current ambivalence theory and presents new empirical evidence of the
dynamic presence of implicit ambivalence in aging families.
B. Introduction
The coexistence of affection and conflict towards an older parent or adult child,
known as intergenerational ambivalence, has sparked great interest in the study of aging
families. Researchers assert that ambivalence arises from contradictions in social
institutions and structures, competing social norms, and interdependency between parents
and children (Connidis & McMullin, 2002). Proponents of ambivalence claim it
motivates action and negotiation in relationships as an attempt to reconcile competing
social expectations and emotions. Coser (1969) observed that ―it is because of such
ambivalence-creating situations...that choices have to be made and that structural changes
become possible‖ (p.182). For example, a daughter‘s mixed feelings about her mother
may never be ultimately resolved, especially if it stems from larger social structures, but
understanding the contradictions within these family ties may help her (and all children
and parents) deal with her ambivalence to achieve more positive outcomes for herself and
the relationship with her parents.
57

Intergenerational ambivalence is a promising paradigm within the
intergenerational relations literature, which has helped researchers gain more in-depth
insight into the complexities of family life. Studies have investigated many social,
familial, and individual characteristics associated with ambivalence; however, the field
has yet to investigate the ―oscillating‖ (Luscher, 2004) nature of ambivalence and test
hypotheses about ambivalence over the life course. Luscher (2004, p. 37) observed that
ambivalence ―… can be experienced in [specific] situations …and it also can be seen in
regard to the entire biographical history of the relationships between parents and their
children.‖ This research begins to fill this gap by investigating intergenerational implicit
ambivalence over more than 10 years of the life course of both parents and adult children.
C. Literature Review
Operational Definition of Intergenerational Ambivalence
Intergenerational ambivalence has emerged as a valuable concept to describe and
understand relationships among aging families. Ambivalence between older parents and
their adult children typically refers to mixed emotions, simultaneous feelings of love and
hate, or having both affection and conflict towards each other. Theorists from both
psychology and sociology have wrestled with similar, but distinct, conceptions of
ambivalence.
Sociological intergenerational ambivalence, similar to role distance theory where
an individual experiences incompatible expectations within one or multiple roles, is the
experience of contradictory norms within a relationship (Coser, 1969; Connidis &
58

McMullin, 2002). Psychological ambivalence focuses on the contradictory feelings
within the older adult about their child, and vice versa (Lüscher, 2004).
A synthesis of both sociological and psychological conceptions of ambivalence
have lead family researchers to assume that mixed emotions (psychological) are the
outcome of contradictory social structures and societal norms (sociological) surrounding
the intergenerational relationship, which may then lead to agency and negotiations within
the relationship (Lüscher, 2002; Curran, 2002; Connidis & McMullin, 2002). Luscher and
Pillemer‘s earliest theorizing about ambivalence included a multidisciplinary approach to
intergenerational relationships, but few current articles explicitly distinguish between the
psychological and sociological. In terms of psychological ambivalence, there are also two
conceptions of intergenerational ambivalence measured in the previous literature: implicit
and explicit. Implicit ambivalence is the coexistence of affection and conflict, which is
not necessarily acknowledged by the parent or child. This form of ‗potential‘ ambivalence
is measured indirectly. Explicit ambivalence is mixed feelings that the parent or child is
aware of and able to express, measured by directly asking about their mixed feelings.
This paper aims to empirically describe the psychological aspect of intergenerational
implicit ambivalence and test two of the sociological structural factors, gender and birth
cohort that may influence ambivalent feelings over the life course.
Life Course Perspective
The lifespan developmental theory (Fingerman, 1996) and the life-course perspective
(Pillemer & Suitor, 2004) can help researchers understand the diversity and dynamics
within aging families, especially in terms of ambivalence. The life course perspective
59

includes several principles important to theorizing aging families, such as, linked lives,
historical influences, and personal agency, which interact to influence parent-child
relationships (Elder, 1994; Bengtson, Elder, & Putney, 2005). One of the most important
contributions of this perspective to the development of the intergenerational ambivalence
model is the recognition that relationships are dynamic and apt to change throughout the
life course. The current state of research on ambivalence has not yet investigated whether
(or how) ambivalence changes or remains stable over the life course of intergenerational
relationships. Lüscher (2004) discussed the temporal aspects of ambivalence and
ambivalence as a potentially enduring, oscillating, or temporary characteristic of
intergenerational relationships. Ambivalence could be only a temporary state for a
daughter caring for her mother after a brief hospital stay or may be an enduring
characteristic in a parent's relationship with their child, stemming from the child's adverse
decisions in young adulthood. The life course theory would predict that given different
situations and changes throughout the life course, ambivalence between parents and
children also changes. The life course perspective also posits that historical and social
trends impact family relationships. Likely, babyboomer children have different
relationships with their parents than their parents had with their parents, due to
sociodemographic and family changes. The timing and types of transitions in children‘s
lives, whether they follow a ―normative life course‖ or not, can impact the feelings of
both generations about each other. When timing or transitions are ―off,‖ unexpected, or in
direct conflict with each other, ambivalent situations may arise.
60

The life course perspective predicts changes and transitions over the life course,
simultaneously experienced by parents and children, interact and impact their
relationships. However, parents and children are experiencing different life course
transitions, situations, and norms, dependent on their own life stage. For example,
children transitioning out of the home may illicit more mixed feelings towards their
parents, while experiencing stability and maturity of their children might actually
increase feelings of affection for their parents. Of course, as evidenced by the
bioecological framework, all of these changes are constrained by earlier life experiences,
social contexts, shared histories, and personal characteristics. The macro level of this
framework includes culture, social structures (age, gender, class, race), the political
economy, medical-industrial complex (Estes, 2000) and overall ideology. These macro-
level forces impact individuals in terms of normative expectations for filial obligation,
work and family norms, socioeconomic status, social inequalities, and health and
lifespan, all of which directly impact the parent-child relationship (Connidis &
McMullin, 2002). Social structural inequalities and social conflicts can create
ambivalence in families where little control can be exerted to negotiate the ambivalence.
Inequalities in terms of gender, race, age stratification, and class that create contradictions
within family relationships are important examples (Connidis & McMullin, 2002;
Silverstein et al., 2010; Coser, 1969; Pillemer & Suitor, 2004).
The empirical work has demonstrated associations between implicit and explicit
ambivalence and competing norms and values, gender, and racial differences.
Ambivalence is particularly related to parents‘ tensions between valuing both autonomy
61

and dependency (Pillemer & Suitor, 2005). In terms of gender differences, mothers have
been shown to experience more ambivalence when children are more dependent or do not
live up to expectations (Pillemer & Suitor, 2004), suggesting an influence of social norms
and ideology about the normative life course and the role of parents‘ in raising successful
children. Racial differences in intergenerational ambivalence have also been briefly
explored in the research, comparing two generations of black and white families
(Pillemer et al., 2007). These findings revealed greater ambivalence in black mothers than
white mothers, which was accounted for by their children's education, marital status,
employment, and behavioral problems.
Most of the current research on ambivalence has used only cross-sectional data,
limiting our knowledge to only snapshots of the experience of ambivalence at certain life
stages. Kiecolt, Blieszner, Savla (2011) have pioneered the longitudinal research on
ambivalence, by examining the long-term effects of intergenerational ambivalence on
psychological well-being. Their study found that parents‘ ambivalence decreases over
time and that greater ambivalence is associated with the child being unmarried or having
more behavioral problems earlier in their life. As expected by the researchers,
ambivalence towards their child also increased levels of depression and decreased levels
of happiness in older parents.
The findings of previous cross-sectional and the newest longitudinal study lead to
several expectations about the nature of implicit ambivalence over an extended period of
the life course. These findings suggest that parents‘ implicit ambivalence may be greatest
during the life course when children‘s lives are more volatile or uncertain, but decreases
62

as children become more settled and achieve normative social statuses. Children‘s
implicit ambivalence is likely to be greater when the parents are older, have poorer
health, and may require some care giving.
Intergenerational Congruency in Ambivalence
Research has shown discrepancies in parents‘ and children‘s reports of support for
each other, levels of contact, solidarity, and conflict, but little is known about how
ambivalent attitudes may be shared within intergenerational relationships over time
(Willson et al., 2006; Giarrusso, Stallings, & Bengtson, 1995; Fingerman, 1996; Shapiro,
2004). Parents are expected to report less conflict and more solidarity, leading to lower
levels of ambivalence, according to the generational stake hypothesis. The generational
stake phenomenon refers to the tendency of parents to over-report more positive attitudes
about their relationships with their children (Giarrusso, Feng, & Bengtson, 2004). This
positivity bias in the parental generation is in contrast to an underreporting of affection by
the younger child generation. This phenomenon or incongruency in parent-child reports
about intergenerational relations has been observed since the earliest work on the
solidarity paradigm (Bengtson & Kuypers, 1971; Giarrusso, Feng, & Bengtson, 2004).
Developmentally, parents have an interest in closer relationships with their children to
feel successful in their role as ―parent,‖ while children are more interested in establishing
independence, thus under-emphasizing parental closeness. Additionally, stressful life
events may elicit differences in emotional qualities, where a child‘s divorce may increase
her feelings of closeness towards her mother, but decrease closeness from the mother‘s
perspective. Ambivalence may be both outcome and catalyst for change in
63

intergenerational relationships, where the child may attempt to negotiate ambivalent
feelings, which may alter their parents‘ feelings towards the child.
It is likely that the level of parent-child correspondence in ambivalence changes
over the life course, due to changes in similarity and social status and changes in
dependency. Similarity of values and social status between parents and children is
consistently found to be associated with intergenerational agreement, thus as children
develop and take on similar social status as their parents, the level of agreement in
ambivalence is likely to be greater. Interdependency, similarity, and common interests (or
equal stakes within the relationship) lend themselves to a greater level of perceptual
agreement about the quality of relationship and this level of agreement may change over
time (Shapiro, 2004). Parents, especially in later life, may depend on their children and
experience great satisfaction in their roles as parents, generating more positive feelings
about their children. Children, on the other hand, might experience more ambivalence,
depending on filial obligations. They might hold positive attitudes and desire for
closeness with their parents, but also conflict and internal conflict about supporting them
and maintaining independence and their own life.
While families are remaining affectionate, supportive, and of enduring
importance, there are also many social forces placing burdens on families (Lawton,
Silverstein, & Bengtson, 1994). Ambivalence has always been present in society, but it
may be becoming more prevalent, leading to detrimental effects on individuals and late-
life parent-child relationships (Lüscher, 2004). Lüscher (2004) discussed the temporal
aspects of ambivalence and ambivalence as a potentially enduring, oscillating, or
64

temporary characteristic of intergenerational relationships. Ambivalence could be only a
temporary state for a daughter caring for her mother after a brief hospital stay or may be
an enduring characteristic in a parent's relationship with their child, stemming from the
child's adverse decisions in young adulthood. This research will utilize longitudinal data
to further develop the intergenerational ambivalence model by identifying and describing
patterns or changes in ambivalence over the life course. This research is expected to help
researchers generate new and exciting hypotheses about the potential life stage and life
course circumstances leading to intergenerational ambivalence for both parents and
children.
D. Research Questions
This research will address the following exploratory research questions:
1. How does intergenerational ambivalence change for both parents and children over
the life course?
2. How do historical and/or social changes impact the trajectory of ambivalence for two
birth cohorts parents and children?
3. Is there similarity in parents‘ and children‘s levels of ambivalence over the life
course?
I expect ambivalence to be greater when children are entering young adulthood and
both generations are experiencing life-changing role transitions and increasing
individuation of the child from their parents. However, ambivalence should decline for
both parents and children over the life course, but perhaps begin to increase as the parents
enter old age and experience age-related declines and losses.
65

In terms of the two cohort groups, the younger dyad is expected to have more
ambivalence as their relationship is exposed to more social upheaval and changes within
family structures (divorce, step-parents, siblings, and children, poverty, fewer children,
etc).
I expect parents‘ and children‘s changes in ambivalence to have a moderate, positive
impact on each other, where children‘s stronger levels of ambivalence lead to an increase
in parents‘ ambivalence.
E. Methods
Sample
This research uses five waves of the Longitudinal Study of Generations (LSOG)
from 1991 to 2005. The LSOG began in 1971 with 2,044 original respondents who were
members of three-generation families. Grandparents (G1) were selected via a multi-stage
stratified random sampling procedure from a population of 840,000 individuals enrolled
in southern California's first large HMO (see Bengtson & Schrader, 1982 for further
details). Adult children (G2) and grandchildren (G3) of the G1 grandparents were also
invited to participate in the survey. Follow-up surveys were administered to original
respondents in 1985, 1988, 1991, 1994, 1997, 2000, and 2005. All data have been
collected by mail-back surveys.
The sample includes two sets of parent-adult child dyads from two birth cohorts.
Including incomplete cases, there are 408 older parent-child dyads and 495 younger
parent-child dyads (903 total dyads). In the LSOG, the parents were matched with a
randomly designated ―study child,‖ for whom they consistently answered about
66

throughout the waves of the study, and whom answered about their designated ―study
parents.‖ Table 2.1 shows the number of dyads by gender of the parent and cohort
membership available from 1991 to 2004.

Table 2.1. Longitudinal Response for Parent-Child Dyads by Birth Cohort

Total 1988 1991 1994 1997 2000 2005

Older (G2-G3) Cohort 408 355 335 373 375 319 253

Younger (G3-G4)
Cohort
495 0 145 263 336 303 319

Measures
In this research, intergenerational implicit ambivalence is conceptualized as the
coexistence of affection and conflict within individuals towards their family member,
which is susceptible to life course changes and social structures. Implicit Ambivalence is
measured along a continuum as an inherent characteristic of intergenerational
relationships, rather than as a classification of relationship types (Lüscher, 2002;
Giarrusso, Silverstein, Gans, & Bengtson, 2005; van Gaalen & Dykstra, 2006). At any
given time, some relationships could be classified as ambivalent; however, because levels
of ambivalence are apt to change throughout the adult life course, a continuous
measurement may be more sensitive to changes in ambivalence, rather than changes from
discrete categories of relationship-types.
Intergenerational implicit ambivalence was assessed using an indirect
measurement combining both positive (closeness and getting along) and negative
67

(conflict and criticism) aspects of the relationships. I averaged three items measuring
positive aspects of the intergenerational relationship: ―how close do you feel is the
relationship between you and your study child (or mother);‖ ―how well do you get
along;‖ and ―how good is communication with your study child?‖ The following three
negative items about the relationship, or conflict, were also averaged: ―how much conflict
do you feel there is between you and your study child (or mother),‖ ―how much do you
feel this child is critical of you or what you do,‖ and ―how much does study child argue
with you?‖ The positive scale had a Chronbach‘s alpha of greater than 0.8 and the
negative scale had a Chronbach‘s alpha of greater than 0.6 for each generation at each
time wave. These items were rated from 1 to 6, indicating greater positive or negative
aspects. The commonly used formula for measuring the similarity and strength of
opposing feelings (Thompson, Zanna, & Griffin, 1995) was then applied using the two
scales:
Ambivalence = [ (Positive + Negative) / 2 - | Positive – Negative | ] + 1.5.
The addition of 1.5 to the original formula eliminates negative scores, so the
range is from 0 to 7.5, with a higher score indicating more ambivalence. This formula
takes the average of both positive and negative aspects to evaluate the strength of both
emotions, which is then subtracted by the difference to evaluate the similarity of both
emotions. Individuals with a high degree of both positive and negative feelings have a
higher score and are conceptualized to be more ambivalent. Ambivalence was assessed
from both the parents‘ perspectives about their child and children‘s perspectives about
their parent.
68

Birth cohort membership and gender were used as control variables to test for
variations in change over time. For parents, they either belonged to the older (G2) cohort,
born between 1916 and 1931 or the younger (G3) cohort, born between 1945 and 1955.
The children either belonged to the older (G3) cohort or the younger (G4) cohort, born
between 1978 and 1983. Gender was assessed for both parents and children.
Analysis
These analyses used latent growth models to examine change in ambivalence over
time. Specifically, two separate models were estimated, first one model from the
children‘s perspectives, controlling for cohort and gender. The second model estimated
implicit ambivalence from the parents‘ perspectives, also controlling for gender and birth
cohorts. Secondly two-group latent growth models were then estimated for children‘s and
parents‘ ambivalence to directly compare the differences between the older and younger
birth cohorts. Thirdly, a two-group bivariate latent growth model was estimated to test the
effects of parents‘ and children‘s implicit ambivalence on each other, over time,
comparing the birth cohorts. Latent slopes, intercepts, regression coefficients, and
indices of fit are reported.
Incomplete cases were analyzed using full information maximum likelihood
estimations in Mplus 4.0 to account for data missing at random and reduce response bias.
A cluster analysis was also applied to adjust standard errors and compensate for the
interdependency between families in the sample.

69

F. Results
Overall, average implicit ambivalence scores indicated little to moderate levels of
ambivalence at each time wave. For both cohorts, children seemed to experience slightly
greater average levels of ambivalence towards parent than the parents did towards
children. Table 2.2 displays these means and paired t-tests. The t-tests showed significant
differences at each time point between children and parents; although there were greater
differences between the younger parent- child dyads. These initial results indicate
different ambivalence trajectories between parents and children and by birth cohort.
70

Table 2.2 Mean Comparisons and Correlations between Parent-Child
Implicit Ambivalence across Time and by Birth Cohort

Children's
Ambivalence
Parents'
Ambivalence
Mean
Difference
(paired t-test)
Paired
Correlation n

Younger Birth Cohorts (G3-G4)

1988 ---- ---- ---- ---- ----
1991
2.68 2.62
.059 .428* 121
1994
2.75 2.34
.413* .358* 216
1997
2.49 2.18
.313* .468* 290
2000
2.23 2.12
.107 .183* 263
2005
2.16 1.94
.215* .480* 319

Older Birth Cohorts (G2-G3)

1988 2.12 1.85 .269* .387* 313
1991
2.03 1.54
.492* .323* 300
1994
1.97 1.50
.467* .487* 323
1997
1.81 1.46
.352* .429* 332
2000
1.85 1.54
.311* .391* 286
2004
1.89 1.58
.31* .384* 250

* p < .001

Trajectory by Generational Perspectives
After examining the means, I estimated two latent growth models to further
describe the overall changes in ambivalence for parents and children. The latent growth
model for children‘s ambivalence towards their parents had a significant fit (χ
2
= 16, df =
14) and confirmed an average decline in ambivalence over time (slope µ = -.111) with
significant variation in both the intercept and slope. A negative correlation (σ
is
= -.086)
between the intercept and slope indicates that children with the greatest levels of
ambivalence in 1991, when they are younger, experience a greater decrease in mixed
71

feelings over time. The parents‘ ambivalence trend was similar, with a smaller, and
nonlinear decline in ambivalence (slope µ = -.094) and negative correlation (σ
is
= -.247).
Differences between Birth Cohorts of Parents
The differences between the two cohort groups were tested by estimating and
comparing the fit statistics of an invariant multiple-group growth model to a model in
which the two groups were allowed to vary. Table 2.3 shows these results for parents and
children. The invariant model was not a good fit (χ
2
= 207, df = 24, RMSEA = .134), but
the variant model proved to be a significantly better fit (χ
2
= 143, df = 41, RMSEA =
.07).
72

Table 2.3. Best-Fitting Multiple Group Latent Growth Model of Intergenerational
Ambivalence 1991-2005

Children Parents
Older (G3)
Younger
(G4)
lder (G2)
Younger
(G3)
Slope Loadings [1,2,3,4,5] [1,2,3,4,5] [1,*,*,*,5] [1,*,*,*,5]
Intercept mean 2.276* 3.260* 1.625* 3.541*
Slope mean
-.077*
-.281* -.032 -.332*
Intercept variance .92* 1.26* 1.682* 1.303*
Slope variance
0.01
0.04 .054* .048*
Intercept/Slope cov
-.009
-.138 -.203* -.153*
Daughter->int. -.437* 0= 0= -.587*
Daughter -->Slope 0.06 0= 0= .140*
Mother-->Intercept 0.08 0= 0= -.358
Mother-->Slope -.018 0= 0= 0.07
Goodness of fit

χ
2
49.50 143.00
df 44.00 41.00
RMSEA 0.02 0.07
CFI 0.99 0.92

* p < .05

The best fitting model showed significant non-linear change for the younger and
older parents. However, the average latent intercept for the older cohort of parents was
1.63 and for the younger cohort was 3.54, indicating the younger group of parents have a
higher initial level of ambivalence. On average, the older parents did not have significant
change in ambivalence over time, but the younger parents had a steep decline (slope µ = -
.33, p < .05) in ambivalence scores. The younger parents experienced higher levels of
ambivalence when children were younger, which declined as their children matured.
There was significant variation in initial levels and change in ambivalence over the life
73

course and in the covariation between them for both cohorts. This variation indicates
there are individual differences in the trajectories, dependent on other important factors
not yet tested in these analyses. The parent‘s gender did not affect their levels of
ambivalence, but younger parents experienced lower initial levels of ambivalence
towards daughters that steadily decreased over time, compared to sons who they were
more ambivalent about in 1991. Figure 2.1 shows the latent growth model of parents‘
intergenerational ambivalence by birth cohort.

Figure 2.1. Parents‘ Average Ambivalence by Birth Cohort (1988-2005)

Differences between Birth Cohorts of Children
A two-group latent growth model was also estimated for the children‘s
ambivalence about their parents. The invariant model was not a good fit (χ
2
= 226, df =
26, RMSEA = .129), but the variant model proved to be a significantly better fit (χ
2
= 49,
74

df= 44, RMSEA = .02). The best fitting model showed significant linear change in
ambivalence for both of the birth cohorts. The average latent intercept for the older
children was 2.27 and for the younger children was 3.26, indicating the younger cohort
had a higher initial level of ambivalence. The average latent slope for the older children
was -.077 and -.281 for the younger, indicating the younger children had a sharper
decline in ambivalence levels over time. There was significant variation in initial levels,
but no variation in the change and covariation for either group. Figure 2.2 shows the
latent growth model of children‘s intergenerational ambivalence by birth cohort.

Figure 2.2. Children‘s Average Ambivalence by Birth Cohort (1988-2005)

Relationship between Parents’ and Children’ s Ambivalence
A bivariate latent growth model was estimated to test the effect of parents‘
children‘s initial levels on each other‘s latent slopes, and vice versa. This model was run
as a two-group bivariate latent growth model, in order to compare the younger and older
75

birth cohorts. The first model constrained the bivariate relationship between the parents‘
and children‘s intercepts and slopes to be equal across the two groups. However, the
variant model (results shown in Table 2.4) was a statistically significantly better fit (χ
2
=
300, df = 119, RMSEA = .05).

76

Table 2.4 Bivariate Latent Growth Model: Two Cohorts of Child-Parent
Ambivalence

Children Parents
Older (G3) Younger (G4) Older (G2) Younger (G3)
Slope Loadings [1,2,3,4,5] [1,2,3,4,5] [1,*,*,*,5] [1,*,*,*,5]
Intercept mean 2.3* 3.518* 1.6* 3.597*
Slope mean
0.03
.55* 0.13 .379*
Intercept variance .922* 1.373* 1.66* 1.358*
Slope variance
-.002
-.024 .053* -.007
Intercept <->Slope cov
0.05
.158* -.16* 0.13
Daughter -->Intercept -.471* -.346 0.04 -.656*
Daughter -->Slope .067* -.1 -.053 0.09
Mother-->Intercept 0.08 -.012 -.036 -.34
Mother-->Slope -.023 -.061 -.018 0.06
Parent-Child Bivariate Effects
Older
Younger
Covariation Intercepts

.798*
1.439*

Covariation Slopes

.014*
.087*

Regressions from Intercepts

Parent Intercept -->Child Slope
-.069*
-.227*

Child Intercept -->Parent Slope
-.057*
-.206*

Goodness of fit
χ
2
300.00
Df 119.00
RMSEA .058 (.05-.06)
CFI 0.93

* p < .05

In this bivariate latent growth model (Table 2.4), I examined the covariation
between the intercepts for parents and children and between the slopes of parents‘ and
children‘s ambivalence. Parents‘ and children‘s initial levels of ambivalence (younger =
1.43; older = .798) and their rates of change (younger = .087; older = .014) positively
covaried, with a stronger relationship between parents and children‘s ambivalence in the
younger cohort.
77

Next, I examined the regression of children‘s latent slope on parents‘ latent
intercept. The older cohort of parents‘ latent level of ambivalence significantly predicted
the children‘s change (-.069), but the younger cohort of parents had a much stronger
influence on children‘s rate of change (-.227). These results suggest when parents who
have a greater level of ambivalence towards their children in 1991, the positive rate of
change in children‘s ambivalence is attenuated.
Thirdly, I examined the regression of parents‘ latent slope on children‘s initial
levels of ambivalence. The older cohort of children‘s latent level of ambivalence
significantly predicted the parents‘ change (-.057), but the younger cohort of children had
a much stronger influence on parents‘ rate of change (-.206).
Lastly, I also examined the effects of gender of both parents and children.
Daughters in the older birth cohort had a significantly lower initial level of ambivalence
toward parents (-.471), which increased over time (slope = .067). Interestingly, the
younger cohort of parents had higher levels of ambivalence towards son than towards
their daughters.
G. Discussion
This paper presents several analyses to examine overall trajectories of
ambivalence over time, comparing familial generations (parents versus children) and two
birth cohorts to answer three research questions. The first questions asks how
intergenerational ambivalence change for both parents and children over the life course?
Overall, parents had less ambivalence over their life course than their children. Both
children and parents‘ levels of ambivalence declined over the 15 years of the study.
78

Interestingly, daughters had less intergenerational ambivalence than sons. These findings
are consistent with a previous study using longitudinal data of parents‘ ambivalence
(Kiecolt, Blieszner, & Savlo, 2010).
The second research question and analysis was designed to investigate how
historical and/or social changes may impact the trajectory of ambivalence. To answer this
question, I compared two birth cohorts of parents-child dyads to explore how family
relationships differ, depending on the historical period in which they developed. These
results showed conclusively that more recent cohorts of parents and children have more
ambivalent relationships. Baby boomer parents have more mixed feelings towards their
children than their parents had towards them. The younger parents‘ relationships with
their children were more volatile, on average they had much higher levels of ambivalence
which declined more rapidly. Interestingly, the older parents had greater individual
variation in their ambivalence, suggesting further research into individual causes of
ambivalence for parents in later life.
The third research question asked whether there was similarity or influences
between parents‘ and children‘s levels of ambivalence over the life course. On average,
children were more ambivalent than their parents, at each time of the survey. While the
levels of ambivalence did not show congruency, there is a positive relationship between
parents and children‘s ambivalence. The influence of the ambivalence of one family
member on the trajectory of ambivalence of another suggests a dynamic and reciprocal
nature of parent-child ties. These findings also confirm possible agency and negotiations
79

within intergenerational relationships, where feelings of parents may lead to changes in
feelings of children.
These findings offer a more detailed description of ambivalence over the life
course, but are limited because they do no empirically test factors that could explain why
there are generational and cohort differences. However, the life course perspective and
generational stake phenomenon offer potential explanations that can be further explored.
These descriptive results lend themselves to two possible hypotheses about the
trajectories of ambivalent feelings for parents and adult children. Because the two birth
cohorts of parent-child dyads in this dataset are not age-matched, the potential cohort and
developmental effects are confounded.
The potential explanations for the cohort effect may be higher divorce rates, more
uncertainty in children‘s jobs and career options, and increased stressors in everyday life
from a more chaotic political, financial, media-driven culture. The younger group of
parents and children may not be any less close or affectionate than the older group, but
experience more stress and conflict in their relationships (Pillemer & Suitor, 2005).
The different life stage or age-based hypothesis, however, seems more likely
supported by these results, prior research, and theories about ambivalence. I could re-
conceptualize the two birth cohorts as a ―synthetic cohort‖ of parent-child relationships at
two different life stages. The older dyad represents parents who are in old age with
middle-aged children. However, the younger dyad represents parents of middle-aged with
young adult children. The dyad of parents and children at younger life-stages had greater
levels of ambivalence than the dyad at later life stages. shows a theoretical decrease in
80

ambivalence for both parents and children, where ambivalence is greatest when children
are in their 20‘s, declines through mid-life, and then begins to increase when parents enter
late life. Significant individual variations in ambivalent relationships, with cohort
differences, suggest different influences on ambivalence within aging families over the
life course. One possible explanation is that ambivalence is greater when children are
younger and in earlier life stages (as seen in the younger dyad), which declines as
children become more independent. As parents age into older adult, with possible health
problems and increasing dependency, ambivalence begins to increase at the later stages of
parents' lives.
This paper confirms Luscher‘s (2004) original hypotheses about the temporal and
―oscillating‖ nature of ambivalence, which is susceptible to changes over the life course
and dynamic between parents and children. The findings also support research on older
parents‘ ambivalence over 14 years, which also found a decline in mixed feelings over
time, as mothers and fathers age (Kiecolt, Blieszner, & Savlo, 2011).
As with any study, there are some limitations. The sample is not representative of all
parent-child relationships, as the sample is based on a regional population and does not
represent a racially or ethnically diverse population. The research literature utilizes other
measurements of intergenerational ambivalence, including latent classes and qualitative
strategies, and direct measurements, which have been found to differ conceptually
(Suitor, Gilligan, & Pillemer, 2011). Using different measurements of ambivalence,
researchers may find different trajectories or differences between parents and children.
81

Chapter Five:
Does Ambivalence Matter? The Relationship between Older Parents’
Feelings towards Their Children and Well-being over Time
A. Abstract

Previous research has examined the prevalence and correlates of ambivalence, or mixed
feelings between adult children and their older parents, but very few studies of examined
the long-term trends and consequences of ambivalence. In this paper I investigated the
mutual interplay between parents‘ ambivalent feelings for older children and
psychological well-being over fifteen years of their life course. Relationship strain as
manifest by ambivalence may negatively affect psychological wellbeing, but the effects
of being more depressed may also impact the relationships between parents and children.
In this paper, I ask (1) is having implicit ambivalent feelings toward children were
correlated with depression at different points in the life course? (2) Does implicit
ambivalence influence elements of psychological well-being or vice versa? I used data
from the Longitudinal Study of Generation to analyze 900 parent-child dyads in five
waves between 1991 and 2005. Dual change score models of both depression and
ambivalence showed that depression and ambivalence are significantly associated, with
depression leading to declines in ambivalence over the time frame of the study. This
research provides evidence for a dynamic link between well-being and the quality of
relationships between aging parents and their adult children. These results are discussed
in terms of risk factors for ambivalence and how successful negotiation of ambivalence
may produce more positive outcomes for intergenerational relationships.
82

B. Introduction
Why do aging family researchers study ambivalence, specifically? Does
experiencing intergenerational ambivalence matter for older parents and adult children?
Until very recently, most of the research on intergenerational ambivalence in aging
families have been cross-sectional and focused on potential influence of ambivalence on
well-being. In previous research on intergenerational relationships, researchers have
demonstrated that some aspects of psychological well-being (depression, self-esteem, and
quality of life) and exchange of social support is indeed associated with ambivalence for
both parents and children. The variations over time due to life course transitions, the
emergence of health problems, and competing social norms or cohort differences may all
interact to affect the behaviors and feelings of older parents.
C. Literature Review
Social Relationships and Psychological Well-Being
Research in psychology and sociology has examined the positive influence of
social networks and support on well-being. For older adults, we know that social
networks are often reduced, but retain the most important and salient relationships
(Carstensen, 1992). Despite the potential decrease in social relationships, the quality of
social relationships has been found to positively relate to both psychological and physical
health (Umberson et. al, 1996; Uchino, 2009; Umberson & Montez, 2010). However, the
negative relationships or conflict in older adults‘ lives may matter more than positive
relationships. For example, research has found that for older adults, more negative social
interactions are more salient than positive social relationships, relating more to
83

psychological distress (Rook, 1984; Newsom, et. al, 2005). In regards to intergenerational
relationships, the parent-child relationship is usually maintained over their lifetime,
despite possible conflicts and tensions. Even if older adults hone down their social
networks to the most positive and important relationships, parents are involuntarily and
permanently ―linked‖ to their children and the existence of conflict in parent-child
relationships has a stronger effect on parents‘ psychological well-being than having
positive relationships (Umberson, 1992). Even though family researchers show positivity
and high solidarity in parent-child ties, these relationships do experience conflicts and
little is understood about the potential negative influences of parental ambivalence.
Intergenerational Ambivalence and Well-Being
The intergenerational ambivalence model posits that macro and micro level social
forces impact the parent-child relationship, creating ambivalent situations and feelings.
For example, power imbalances, lengthening lifespans, gender, workforce demands, and
interdependency are macro-level forces affecting ambivalence. Research has
demonstrated associations between ambivalence and competing norms and values,
women‘s roles within the family, and tensions between autonomy and dependency
(Pillemer & Suitor, 2005). On the micro-level, shared familial histories, inequitable
exchanges of support, personality, and poor health are potential factors contributing to
intergenerational ambivalence (Fingerman, Hay, & Birditt, 2004; Fingerman et al., 2006;
Lüscher, 2004; Fingerman et al., 2008; Pillemer & Suitor, 2002; Willson, Shuey, & Elder,
2003). In fact, much of the research in intergenerational ambivalence has focused on the
association between health and psychological well-being on parent-child ambivalence.
84

Previous research utilizing both direct and indirect measurements has found adult
children were more ambivalent towards parents with poorer physical and psychological
health, and that parents‘ ambivalence is related to worse life satisfaction, depression, and
less happiness (Fingerman et al., 2006; Fingerman et al., 2008; Wilson, Shuey, & Elder,
2003; Ward, 2008). However, Pillemer and Suitor (2002) using a direct measurement did
not find relationship to mothers‘ health when considering factors such as mothers age,
child‘s statuses and dependency, similarity, and proximity. Additionally, Birditt,
Fingerman and Zarit (2010) found that self-rated health was associated with ambivalence
until considering the problems (psychical, lifestyle, career, and relationship) of their adult
child. Research on the impact of intergenerational relationships on quality of life found
interesting and different relationships between solidarity, conflict, and ambivalence
(Lowenstein, 2007). The quality of life of older parents was assessed using the World
Health Organizations‘ Quality of Life-BREF inventory to cover many dimensions of life
(health, psychological well-being, satisfaction with relationships, etc) and found that
affection and similarity to an adult child has a stronger association with better quality of
life, but that ambivalence had an independent relationship with worse quality of life.
While all of this previous research has been cross-sectional and cannot conclude
that ambivalence leads to worse psychological well-being, Kiecolt, Blieszner, and Savla‘s
(2011) research is an exception. This longitudinal study on parental ambivalence
examined the long-term effects of intergenerational ambivalence on psychological well-
being. Their study found that parents‘ ambivalence decreases over time and that greater
ambivalence is associated with the child being unmarried or having more behavioral
85

problems earlier in their life. Parents with greater ambivalence towards their child
experienced an increase in depression and decreased levels of happiness. However, a
major limitation in this research is that the analyses assume ambivalence is the predictor
of depression and happiness, without testing the opposite or bi-directional effects of
parental well-being and depression.
Since Luscher and Pillemer‘s (1998) first conceptual model of ambivalence (see
Figure 0.1), researchers have assumed intergenerational ambivalence leads to a decrease
in psychological well-being (less happiness, more depressed, lower quality of life, and
lower life satisfaction). However, this causal mechanism has not been supported by the
research, nor does the theoretical model of ambivalence support the idea that
ambivalence is always detrimental (Connidis & McMullin, 2002). The current study
addresses this gap in the research and provides a better test of the causal relationships
between intergenerational relationships and well-being for older adults.
D. Research Questions
Given the previous research and evidence for the influence of ambivalence on
psychological health, the current paper is focused on the relationship between
ambivalence and depression in older parents. (1) Is having ambivalent feelings toward
children correlated with depression at different points in the life course? (2) Does
ambivalence influence older parents‘ levels of depression or does depression influence
ambivalent feelings? I expect the mixed feelings of older parents to influence an increase
in depression, which will also reinforce their levels of ambivalence. However, controlling
86

for time-varying factors, such as similarity, contact, and self-rated health will weaken the
relationships between depression and ambivalence.
E. Method
Sample
This research used four waves of the Longitudinal Study of Generations (LSOG)
from 1991 to 2005. The LSOG began in 1971 with 2,044 original respondents who were
members of three-generation families. Grandparents (G1) were selected via a multi-stage
stratified random sampling procedure from a population of 840,000 individuals enrolled
in southern California's first large healthcare management organization. Adult children
(G2) and grandchildren (G3) of the G1 grandparents were also invited to participate in
the survey. Follow-up surveys were administered to original respondents in 1985, 1988,
1991, 1994, 1997, 2000, and 2005. All data have been collected by mail-back surveys or
web survey for the 2005 data collection.
The sample includes two sets of parent-adult child dyads from two birth cohorts.
Including incomplete cases, there were 900 total dyads. The G2 mothers and fathers were
matched with a randomly designated G3 ―study child,‖ for whom they consistently
answered about throughout the waves of the study. Table 3.1 shows the number of dyads
by gender of the parent and cohort membership available from 1991 to 2005. The
youngest cohort (G4) was introduced to the sample as they turned 16 beginning in 1991,
which explains the initial increase in G4 respondents and then attrition over time.

87

Table 3.1. Longitudinal Response for Parent-Child Dyads by Birth Cohort

Total 1988 1991 1994 1997 2000 2005
Older (G2-G3) Cohort 408 355 335 373 375 319 253
Younger (G3-G4)
Cohort
495 0 145 263 336 303 319

Measurements
Intergenerational ambivalence was assessed using the indirect measurement of
implicit psychological ambivalence, combining both positive (closeness and getting
along) and negative (conflict and criticism) aspects of the relationships. The following
three items measuring positive aspects of the intergenerational relationship were
averaged: ―how close do you feel is the relationship between you and your study child (or
mother);‖ ―how well do you get along;‖ and ―how good is communication with your
study child?‖ The following three negative items about the relationship, or conflict, were
also averaged: ―how much conflict do you feel there is between you and your study child
(or mother),‖ ―how much do you feel this child is critical of you or what you do,‖ and
―how much does study child argue with you?‖ The positive scale had a Chronbach‘s
alpha greater than 0.8 and negative scale had a Chronbach‘s alpha greater than 0.6 for
each generation at each time wave. These items were rated from 1 to 6, indicating either
greater positive or greater negative aspects. The commonly used Griffin formula for
ambivalence (Thompson, Zanna, & Griffin, 1995) was then calculated using the two
scales:

88

Ambivalence = [ (Positive + Negative) / 2 - | Positive – Negative | ] + 1.5.

The addition of 1.5 to the original formula eliminates negative scores, so the range is
from 0 to 5.25, with a higher score indicating more ambivalence. This formula takes the
average of both positive and negative aspects to evaluate the strength of both emotions,
which is then subtracted by the difference to evaluate the similarity of both emotions.
Individuals with a high degree of both positive and negative feelings have a higher score
and are conceptualized to be more implicitly ambivalent. Ambivalence was assessed
separately from both the parents‘ perspectives about their child and the children‘s
perspectives about their parent.
Depression is used as one factor representing ―psychological well-being.‖ The
depression scale includes 20 statements relating to depression symptoms, rated by
frequency of experience within the past week, from rarely or none (0) to most or all of the
time (3). Each item is averaged to create a scale from 0 to 3.
Birth cohort membership, age, gender, contact with their child, similarity to their
child, and self-rated health were used as control variables to test for variations in change
over time. The children in the older birth cohort were born between 1945 and 1955 and
the younger cohort was born between 1978 and 1983. Gender was assessed for both
parents and children.
Summary Statistics
A description of the summary statistics for both variables at all time waves
appears in Table 3.2. This includes a list of all the univariate statistics for depression,
89

ambivalence, and the covariates (age, gender, contact, similarity, self-rated health, etc).
The raw means showed a slight decrease in ambivalence and an increase in depression
over time.
Table 3.2. Descriptive Statistics for Parents‘ Ambivalence, Depression, and
Covariates

N M SD Min Max
Dependent Variables

Ambivalence 1991 480 1.89 1.39 0 5.5
Ambivalence 1994
582 2.3 1.43 0 5.67
Ambivalence 1997
660 2.11 1.39 0 5.83
Ambivalence 2000
729 2.17 1.41 0 6.17
Ambivalence 2005
569 2.04 1.34 0 6.33
Depression 1991
715 1.48 0.43 1 3.85
Depression 1994
772 1.48 0.43 1 3.35
Depression 1997
799 1.48 0.41 1 3.5
Depression 2000
793 1.51 0.43 1 3.5
Depression 2005
570 1.5 0.4 1 3.3
Time-Invariant Covariates

Mother
528 58% 0.49 0 1
Daughter
497 55% 0.5 0 1
Older Birth Cohort
408 54% 0.49 0 1
Age
746 54.6 12.4 34 82

Table 3.3 displays the correlations between all pairs of variables at each of the
five time points. In general, cross-sectional correlations between ambivalence and
depression at each time was 0.02 and statistically significant (p <.05). The longitudinal
or lead-lag relationships within ambivalence and depression (the test-retest correlations
90

between ambivalence 1991 and 1994) show a significant and very strong correlation
between ambivalence levels over time (0.6 – 0.75) and a strong, stable association in
depression over time (.55). The lead-lag relationships between the two variables (the
correlation between ambivalence in 1991 and depression in 1994 and vice versa) shows a
similar strength of relationship, regardless of whether ambivalence or depression comes
first.
Table 3.3. Correlation Coefficients for Ambivalence and Depression across Time

1 2 3 4 5 6 7 8 9
1. Amb 1991 1

2. Amb1994 0.71 1

3. Amb 1997 0.59 0.62 1

4. Amb 2000 0.57 0.58 0.75 1

5. Amb 2005 0.42 0.48 0.51 0.52 1

6. Dep 1991 0.17 0.16 0.21 0.19 0.17 1

7. Dep 1994 0.14 0.22 0.19 0.19 0.18 0.58 1

8. Dep 1997 0.17 0.21 0.24 0.23 0.14 0.54 0.63 1

9. Dep 2000 0.11 0.14 0.18 0.22 0.09 0.52 0.58 0.64 1
10. Dep 2005 0.06 0.13 0.13 0.15 0.15 0.51 0.53 0.56 0.63

Analyses
The analysis strategy used to determine the relationship between change in
ambivalence and change in depression is the dual (or dynamic) latent difference score
structural equation model (McArdle, et al, 2004; Jajodia, 2011). First, two univariate dual
difference score models were estimated for ambivalence and depression overtime,
separately. Then a bivariate dual difference score model, based on results of the best-
fitting univariate models, was estimated to test the mutual effects of ambivalence on
91

depression. Lastly, the covariates (e.g., gender and age) were added to the bivariate
model.
This structural equation modeling approach combines features of the latent
growth and the crossed-lagged latent difference score models. With a latent growth
model, we can fit the data to an average trajectory over time based on the scores of
ambivalence (or depression) at each time point. This growth equation is written as
X[t]= xi + B[t]xs + e[t]
In this equation, the average trajectory of ambivalence (X[t]) is a function of the
individual‘s initial latent level of ambivalence in 1991 (xi), the latent slopes (xs) or
individual changes, multiplied by the time points (B[t]), and the latent errors (e[t]) at each
measurement.
The cross-lagged latent difference score model measures the effect of a previous
time point to the next (auto-regression), and the effects of one variable x on change in a
second variable y (and vice versa) (the cross-lagged effect).
The equations for the latent changes are written as:
Delta X[t]= alpha + beta x[t] + lambda y[t] + e[t]
Delta Y[t]= alpha + beta y[t] + lambda x[t] + e[t]
The dual (or dynamic) latent difference score model combines the constant rate of
change of a latent growth model and the proportional accumulation of changes over time
of a cross-lagged difference score model. The bivariate relationships between depression
and ambivalence are assessed by the cross-lagged or coupling parameters (effects of one
92

variable on the change in the other) and by the covariation between the latent intercepts
and slopes of both variables.
In the bivariate dual latent difference score model (Figure 3.1), both variables
have latent variables at each time point, which are defined as latent difference scores.
Each variable has a latent intercept and latent slope, which are allowed to covary. In
addition, the autocorrelations (proportional change) between the latent variable and
difference scores are estimated, but constrained as equal over time. The useful feature of
the bivariate model is the addition of coupling parameters, indicating the lead-lag effects
of ambivalence at each time point on change in depression and, simultaneously, the
effects of depression at each time point on change in ambivalence. The strength and
statistical significance of these parameters can be tested to determine which variable
influences the other.
I used nested models to test each univariate difference score model for both
constant and proportional change, relying on chi-square difference testing to choose the
best-fitting model. The bivariate difference score model combines the best-fitting
univariate models of depression and ambivalence and the addition of coupling parameters
was tested using nested models. Finally, the addition of the covariates were added in
stages, also comparing nested models, to determine the unique effects the control
variables may have on the relationship between ambivalence and depression. The chi-
square, CFI, and RMSEA model fit statistics are reported to assess the models.
All analyses were conducted in Mplus version 4.0, using Full Information
Maximum Likelihood estimation for incomplete data. The number of incomplete cases
93

used in the analyses is 902; however, the number of complete cases is 286. Each model
was run using only the complete cases. These models had a worse fit, but individual
parameters were more likely to be statistically significant than in the FIML models.
Given the FIML theory and superiority over other missing data treatments for MAR and
MCAR (Enders & Bandalos, 2001) I chose to use the incomplete cases to estimate the
models.
F. Results
Latent Difference Score Results for Ambivalence
The four nested models for ambivalence are displayed in Table 3.4. The best-
fitting and parsimonious model shows ambivalence as having a constant rate of change
with a proportional change at each time point (χ2 = 99, df = 13). There is not a
significant constant rate of decline in ambivalence when including the proportional
change. Both the latent intercept and slope have significant individual variation and a
positive covariance (σ
is
=

-.094).

94

Table 3.4. Dual Latent Change Score Models for Ambivalence 1991-2005

No Change
Constant
Change
Proportional
Change Dual Change
Fixed Effects

Slope Loading 0= 1= 0= 1=
Proportional change 0= 0= -.045* -.216*
Intercept mean 1.878* 2.076* 2.08* 2.108*
Slope mean 0= -.087* 0= 0.328
Random Effects

Intercept variance .962* 1.434* 1.21* 1.506*
Slope variance 0= .054* 0= .075*
Residual error
variance .734* .601* .709* .599*
Int <--> Slope cov 0= -.155* 0= .094*
Fit Statistics

Chi-square/df 217.4/17 103.4/14 161.2/16 99/13
CFI 0.847 0.932 0.889 0.934
RMSEA 0.118 0.087 0.103 0.088

* p < .05

Latent Difference Score Results for Depression
The four nested models for depression are displayed in Table 3.5. The best-fitting
model also shows depression as also having a constant rate of change and proportional
change at each time point (χ
2
= 28, df = 13). There is a significant average increase in
depression over time, but a proportional decline in ambivalence between each time point.
The contradictory parameters suggest the overall increase in depression levels is
tempered by the proportional decrease based on previous levels of depression. There is
significant individual variation in the intercept, in that parents start out at varying levels
of depression.
95

Table 3.5. Dual Latent Change Score Models for Depression, 1991-2005

No Change
Constant
Change
Proportional
Change
Dual
Change
Fixed Effects

Slope Loading 0= 1= 0= 1=
Proportional change 0= 0= 0.003 -.383*
Intercept mean 1.499* 1.487* 1.49* 1.486*
Slope mean 0= 0.006 0= .579*
Random Effects

Intercept variance .104* .121* .103* .126*
Slope variance 0= .003* 0= 0.017
Residual error variance .077* .07* .077* .069*
Int <--> Slope cov 0= -.007* 0= .034*
Fit Statistics

χ
2
/df 60/17 33/14 58/16 28/13
CFI 0.97 0.986 0.97 0.989
RMSEA 0.053 0.039 0.054 0.036

* p < .05

Bivariate Latent Difference Score Results
The bivariate models were also tested hierarchically, testing the fit differences
between bivariate dual change models, comparing coupling parameters and adding
covariates to subsequent models. Of all the models, as expected, the bivariate dual change
score model with selected time variant and invariant control variables fit best to the data
(χ
2
= 370, df = 177, RMSEA = .035). Table 3.6 compares two of these bivariate dual
change score models, comparing the hierarchical addition of the control variables. Model
A tested a bivariate dual change (including constant and proportional change) model on
the mutual influence of ambivalence and depression over time, without any control
variables. The fit was poor (χ
2
= 1260, df=225, RMSEA= .071). Model B included
additional controls for gender and birth cohort on the effect of the fixed effects (slopes
96

and intercepts) and Model C included gender, birth cohort, similarity, contact, and self-
rated health at the fixed effects level. Models B and C (not shown) were better fits (χ
2
=
1088, df=213, RMSEA= .067; χ
2
= 779, df=201, RMSEA= .056) compared to Model A.
However, the best fitting model included time invariant and time-varying control
variables. Figure 3.1 illustrates Model D as a structural equation model. Although there
was a slight proportional decline in ambivalence over time, with no significant changes in
depression or relationship between the changes in depression and ambivalence in Model
A (with no covariates), the relationship between changes in ambivalence and depression
is altered by including the important control variables, that are significantly correlated
with either ambivalence or depression. Model D (χ
2
= 370, df = 177, RMSEA = .035)
suggests a different story to the relationship between ambivalence and depression, once
including gender, generation, self-rated health, similarity between parent and child, and
the level of contact between parent and child.
There are statistically significant negative proportional changes for ambivalence
(λ
a
= -.287) and depression (λ
d
= -.597). There is also a statistically significant positive
rate of change in the latent slope for both ambivalence (s
a
= 2.48) and depression (s
d
=
.515). The negative proportional change reduces the overall increase of the slope for both
ambivalence and depression, leading to a simultaneous acceleration and deceleration that
leads to a smaller overall increase in both constructs, on average.
Perhaps most importantly, the results show a statistically significant negative
influence of depression on the latent difference in ambivalence (γ
a
= -1.07), which
97

indicates that parents with greater levels of depression actually experience a decline in
ambivalence over time.
Effects of Covariates on Implicit Ambivalence
The covariates helped explain some of the significant variation in the initial levels
and rate of change for ambivalence. The younger cohort of parents had greater initial
levels of ambivalence and parents with poorer self-rated health in 1991 experienced a
greater overall increase in ambivalence. As expected, the frequency of contact in 1991
had a negative effect on the intercept (β = -.126), indicating that parents who had more
contact with their child were also less ambivalent. Similarity to their child in 1991 also
had a negative effect on initial levels of ambivalence (β = -.508), such that parents who
felt more similar were less ambivalent.
Self-rated health, contact, and similarity were also being controlled for at each
time point. Of note, having more contact is related to lower ambivalence in 1994 (time
2)and 2005 (time 4), suggesting that the relationship between contact with children and
ambivalence may only be important in certain points in the life course. However, as
expected, at each point, parents who are more similar to their child report lower levels of
ambivalence.
Effects of Covariates on Depression
The covariates helped explain some of the significant variation in the initial levels
and rate of change for depression. Mothers had greater initial levels of depression (β =
.104) and were more likely to experience an overall increase in depression, as compared
to fathers (β = .048). As expected, parents with poorer self-rated health had greater initial
98

levels of depression and experienced an overall increase in depression over time (β =
.044). Self-rated health, contact, and similarity were also being controlled for at each
time point. Only self-rated health was positively associated with depression at each time
point, as expected.

99

Table 3.6. Bivariate Dual Latent Difference Score Model with Covariates
Model A Model D
Ambivalence Depression Ambivalence Depression
Fixed Effects

Loading alpha 1= 1= 1= 1=
Proportion beta -.220* -.255 -.287* -.597*
Coupling lambda 0.487 0.02 -1.07* 0.046
Intercept mean 2.112* 1.485* 4.07* 1.38*
Slope mean -.391 0.348 2.48* .515*
Random Effects

Intercept variance 1.5* .125* .86* .094*
Slope variance 0.081 0.009 .213* .028*
Error variance .596* .069* .53* .07*
Latent Variable Covariance

Intercept--Slope 0.048 0.018 .117* .04*
Amb i--Dep i .121* ___ 0.012 ___
Amb i--Dep s -.014 ___ -.013 ___
Amb s--Dep s -.008 ___ .053* ___
Amb s--Dep i -.03 ___ .105* ___
Time In-Variant Covariates [regression coefficient for effect on Intercept, Slope]
Birth Cohort 0= 0= .887*, .071 .06, '-.01
Mother 0= 0= .038, .099 .104*, .048*
Daughter 0= 0= .005, .085 .07*, .017
Self-Rated Health 0= 0= .081, .166* .174*, .044*
Contact 0= 0= -.126*, -.016 -.014, -.003
Similarity 0= 0= -.508*, -.041 -.077*, .006
Time Variant Covariates [t1, t2, t3, t4]
Self-Rated Health 0= 0= .08, .00, -.02, -.18*
.08*, .11*, .13*,
.14*
Contact 0= 0= -.02, -.07*, -.05, -.09* .00, -.00, .01, .00
Similarity 0= 0= -.21*, -.29*, -.39*, -.36* -.00, -.00, -.03, -.03
Fit Statistics

chi-square/df 1260/225

370/177

CFI 0.722

0.948

RMSEA 0.071 0.035

100

G. Discussion
The aim of this paper was to determine the causal relationship between
intergenerational ambivalence and depression for parents of adult children. This study is
one of the first to employ longitudinal data to examine parents‘ ambivalence towards their
children over 15 years. Additionally, this is the first study that can actually distinguish the
independent and simultaneous effects of ambivalence and depression on each other.
Within the bioecological and life course perspectives, the causes and consequences for
parent-child ambivalence are changing and dependent on many macro-micro interactions
that impact the life of the parent and, independently, the child. This presents a complex
picture of intergenerational ambivalence that is susceptible to the diversity in families.
The literature and previous empirical evidence points to a negative influence of
ambivalence on older parents‘ well-being, but to date, only one study has enlisted
longitudinal data to support these findings (Kiecolt, Blieszner, & Savla, 2010). This
previous study found that having greater levels of ambivalence does lead to an increase in
depression for parents. Kiecolt, Blieszner, and Savla find this relationship even when
controlling for many characteristics of the child and parent. However, their analysis
differs from the current research because it does not take into account the dynamic effects
of ambivalence and depression on each other over time. They employ longitudinal data
and establish a temporal relationship between earlier ambivalence on long term
depression and happiness trajectories, but do not also consider mutual influences of
psychological well-being and intergenerational ambivalence.
101

The first research question asked was whether parents‘ ambivalent feelings toward
children correlated with depression at different points in the life course? The correlations
between ambivalence and depression are significant at each time point of the study.
Implicit ambivalence and depression are positively related, as predicted and shown in
previous research (e.g. Kiecolt, Blieszner, & Savla, 2010).
The second research question was whether ambivalence influences older parents‘
levels of depression or does depression influence ambivalent feelings? The findings from
this analysis seem contrary to expectations that ambivalence and depression mutually
influence changes in each other. Intergenerational ambivalence does not seem to be the
cause of poor psychological health, at least in terms of depression. In fact, it seems that
having greater levels of depression leads to declines in intergenerational ambivalence for
older parents. This finding likely demonstrates a moderating effect of depression on
emotional intensity (conflict and affection) towards the focal child. However, it is
important to note, that while the cross-section bivariate correlations show a positive
relationship between ambivalence and depression, the dual change model does not refute
this. Rather, ambivalence levels are greater among parents who are more depressed, but
those levels of ambivalence do not actually increase, but decline and regress toward the
mean.
Most theory and research has focused on the unidirectional relationship of
ambivalence on elements of well-being, without considering the impact individual
psychological health can have on their relationships with adult children (Luscher &
Pillemer, 1998; Fingerman et al., 2006; Lownestein, 2007; Kiecolt, Blieszner, & Savlos,
102

2010). This study is one of the first to utilize a novel methodological approach to assess
the mutual interplay between parent-child relationships and mental health. Once
accounting for the bidirectional effects, I find that mental health actually impacts the
parent-child relationship, rather than vice versa. To address the title question of the
chapter, changes in intergenerational ambivalence over time does not seem to matter for
parents‘ changes in depression levels.
There are some limitations to this analysis in terms of the sample, missing data,
measurement of ambivalence, and data analysis strategy. The sample is not representative
of all parent-child relationships, as the sample is based on a regional population and does
not represent a racially or ethnically diverse population. The research literature utilizes
other measurements of intergenerational ambivalence, including latent classes and
qualitative strategies, and direct measurements, which have been found to differ
conceptually (Suitor, Gilligan, & Pillemer, 2011). Using different measurements of
ambivalence, researchers may find different trajectories or differences between parents
and children. Additionally, using other related factors as time variant and invariant
controls may change the relationship between ambivalence and depression, so it‘s
important to recognize that this analysis is only controlling for contact, self-rated health,
similarity, and gender. Other characteristics of the parent or child, health measures, or
other psychologically well-being indicators could impact the results of future studies.
103

Figure 3.1. SEM Results of Bivariate Latent Difference Score Model D for
Depression and Ambivalence, 1991-2005

*p < .05

Note: Not shown in the figure: the observed depression and ambivalence measures at
each time point are also regressed on the time invariant covariates (self-rated health,
similarity, and contact).

104

Chapter 6:
Discussion
The three studies in this dissertation are unique theoretical and methodological
contributions to gerontology and the study of aging families. Intergenerational
ambivalence is a growing topic in aging families‘ research and represents an
interdisciplinary approach to understanding families in later life. Scholars from both
sociological and psychological backgrounds have provided evidence for the impact of
contradictory social structures on family relationships, particularly in terms of the conflict
between dependency and autonomy. Additionally, researchers have demonstrated the
existence of and potential negative outcomes for mixed feelings within both parents and
children. This dissertation aims to extend the empirical and theoretical development of
the ambivalence model, particularly by emphasizing the importance of refining the
measurement of ambivalence, including reciprocal dyads, and using longitudinal data to
capture family-life over the life course.
Summary
The overall goal of this dissertation is to describe the measurement, trajectories,
and consequences of intergenerational ambivalence, coexistence of affection and conflict,
among older parents and adult children. While there is a growing body of research on
intergenerational ambivalence, I identified several gaps in the literature to address in this
dissertation. There is a need to validate measurements and test causes and/or
consequences of ambivalence over the adult life course to advance the utility of the
intergenerational ambivalence model. Therefore, the objectives were: (1) to compare and
105

validate two methods of quantitatively measuring intergenerational ambivalence within
parent-child dyads; (2) to describe and explain change in ambivalence over time among
two cohorts of parent- adult child dyads; and (3) to identify the effect of ambivalence
over time on psychological well-being.
Analyses were performed using data from the USC Longitudinal Study of
Generations (LSOG)—a multi-wave, multi-level, four-generation survey of American
families. The first study analyzed 253 parent-child dyads from the 2005 wave of the
LSOG. Parents and children report similar levels of direct ambivalence, but children had
greater levels of ambivalence when measured using the indirect scale. The levels of the
component parts of the indirect scale, affectual solidarity and conflict, are different
according to the familial generation. The generational stake phenomenon, where parents
over emphasize positive and under emphasized negative feelings may help explain this
incongruency. The multivariate analyses show support for previous research, which
shows that tensions between autonomy and dependence, associational solidarity (contact
and physical closeness), having poorer health, and children‘s problems are related to
ambivalence. My analysis uniquely demonstrates; however, that some of these factors
may not be statistically related to different measures of ambivalence. Additionally, the
macro and micro-level variables related to ambivalence are different for parents than for
children.
The second study analyzed 848 parent-child dyads from two birth cohorts across
10 years from the LSOG. Overall, there was significant decline in ambivalence over time,
when both cohorts were combined. However, there was a positive covariation between
106

the latent level and negative slope for the younger dyad. And for the older dyad, there
was a nonlinear trend of decline from 1991 to 1997 and a slight increase after 2000.
These results show conclusively that more recent cohorts of the baby boomer parents and
their children have more ambivalent relationships than the previous cohort. The younger
parents‘ relationships with their children seem more volatile. Additionally, children,
across both birth cohorts, experience more ambivalence than their parents. As seen in the
previous study, this finding may be due to the measure of ambivalence used; however, the
generational stake hypothesis supports the idea that children would report more conflict
and have a more negative perspective on their relationship.
The final study tested the causal relationship between ambivalence and depression
in over 900 parents using LSOG data from 1991 to 2005. Contrary to expectations and
previous research, I found that parents‘ depression influences changes in their
ambivalence towards adult children, even when controlling for relationship
characteristics, birth cohort of parent, gender, and health. Parents who are more
depressed experience an increase in intergenerational ambivalence, but these changes in
ambivalence do not seem to influence levels of depression over time. These results
highlight the importance of mental health and well-being on family relationships, in
addition to the traditional understanding of the importance of social relationships for
well-being.
Bioecological Framework for Studying Family Relationships
In the introduction I discuss the theoretical and empirical developments of the
intergenerational ambivalence model, and argue for the utility of considering the life
107

course and bioecological perspectives to further understand intergenerational
ambivalence. I discuss how the bioecological model can also incorporate many other
social gerontological perspectives (e.g., political economy, life course, social exchange,
personality and lifespan development) seen in Figure 0.2, that are applicable to
understanding the intergenerational relationship. Ambivalence arises when there are
conflicts or contradictions within the macro and micro-level domains of social life, but
these contradictions are likely different for parents than for children. Each familial
generation needs to be included in the theoretical development, as parents and children
interact with each other. Over the process of writing this dissertation, I developed Figure
4.1 to illustrate the incorporation of a lifecourse and dyadic bioecological model of
intergenerational relationships. In this diagram, the parent and child, individually, are
embedded within macro (represented by the large oval), micro, and individual level
forces (the smaller circles), which independently impact their life course trajectories over
time. There is likely overlap in their macro-level social positions. Additionally, parents
and children are interacting with each other at multiple points over time (represented by
the bold double-headed arrows), which impacts their feelings towards each other (the
change in conflict, affection, and/or ambivalence is represented by the triangles). These
changes in feelings, in turn, impact the individual and meso/micro-level factors for both
child and parent, which also affect their future interactions. In terms of the human
bioecological model, Figure 4.1 includes characteristics of the person, developmental
processes, social context, and time (Moen, 1995; Magnusson, 1995). However, it also
108

adds an added dimension important for studying family relationships, which is the
reciprocal dyadic component.
Figure 4.1 Modified Bioecological Framework for Understanding
Intergenerational Relationships

In these analyses, I tried to incorporate many elements of the above framework.
Chapters 3 and 4 analyzed data using reciprocal dyads of both parents and their children.
The first study, in Chapter 3, demonstrated that there is not a clear, distinct,
operationalization of ambivalence, but likely many strategies to capture different aspects
of this complex mix of love and hate. Additionally, there are different macro and micro-
level factors related to ambivalence for both the parent and child generations, suggesting
different theoretical processes for producing ambivalence. The second study, Chapter 4,
109

demonstrates the differences between familial generations in ambivalence: children are
more ambivalence toward their parents. Another important aspect of the bioecological
and life course perspectives is change over time. Chapters 4 and 5 analyzed data across
time in order to describe and explain changes in ambivalence. The second study shows
different trajectories of ambivalence based on cohort membership and familial
generation, in addition to significant individual variation. The third study, Chapter 5,
shows that psychological health and well-being of parents influences their relationships
with their child over time, suggesting dynamic effects of macro, micro, and individual-
level variables over the life course.
Conclusion
Taken together, these studies of intergenerational ambivalence show a complex
picture of family life that is susceptible to many social forces and changes over time.
Ambivalence is still a newer and elusive concept, but seems well-suited to help
understand how families, in spite of affection and closeness, negotiate conflict and
tensions with each other and in their social roles and obligations. Future research on
differences in the measures of ambivalence is needed, especially applying these
conceptual differences over time. Additionally, the relationship between ambivalence and
the various macro and micro-level social and psychological factors included in the first
study (Chapter 3) should be further explored using longitudinal analyses. The surprising
relationship between depression and ambivalence over time, which represents a unique
contribution to the literature, suggests that theories about the negative outcomes of
ambivalence are not yet supported. Does ambivalence matter? It is likely that
110

experiencing ambivalence toward a loved one does matter, but in different ways, in
different relationships, and for different individuals. Understanding the contexts where
ambivalence does matter for older parents and adult children is, I think, the most
important next step in developing the intergenerational ambivalence model.

111

References
Bengtson, V. L. & Kuypers, J.A. (1971). Generational differences and the developmental
stake. Aging and Human Development, 2, 249-260.
Bengtson, V . L., Elder, G. H., & Putney, N. M. (2005). The life course perspective on
ageing: Linked lives, timing, and history. In M. L. Johnson (ed.) in Association
with Bengtson, V . L., Coleman, P., and Kirkwood, T., The cambridge handbook of
age and ageing. (pp. 493–501). Cambridge, MA: Cambridge University Press.
Bengtson, V . L., & Schrader, S. S. (1982). Parent-child relations. In D. Mangen & W.
Peterson (Eds.), Handbook of research instruments in social gerontology (pp. 115-
185). Minneapolis, MN: University of Minnesota Press.
Bengtson, V . L., Giarrusso, R., Mabry, J. B., & Silverstein, M. (2002). Solidarity, conflict,
and ambivalence: Complementary or competing perspectives on intergenerational
relationships? Journal of Marriage and the Family, 64, 568-576.
Bianchi, S.M., & Casper, L.M. ( 2005). Explanations of family change: A family
demographic perspective . In V. L. Bengston, A. A. Acock, K. R. Allen, P.
Dilworth-Anderson, & D. M. Klein (Eds.), Sourcebook of family theory &
research (pp. 93-117). Thousand Oaks, CA: Sage.
Birditt, K. S., Fingerman, K. L., & Zarit, S. (2010). Adult children‘s problems and
successes: Implications for intergenerational ambivalence. Journal of
Gerontology: Psychological Science, 65, 145-153.
112

Bronfenbrenner. U. (2001). The bioecological theory of human development. In N.J.
Smelser, and P.B. Baltes (Eds.). International encyclopaedia of the social and
behavioural sciences (volume 10, pp. 6963-6970). New York, NY: Elsevier.
Carstensen, L.L. (1992). Social and emotional patterns in adulthood: Support for
socioemotional selectivity theory. Psychology and Aging, 7, 331-338.
Connidis, I.A. & McMullin, J.A. (2002). Ambivalence, family ties, and doing sociology.
Journal of Marriage and Family, 64, 594-601.
Coser, L. A. (1956). The Functions of social conflict. Glencoe, Ill: Free Press.

Coser, R. L. (1969). Role distance, sociological ambivalence, and transitional
status systems. The American Journal of Sociology, 72, 173-187.
Crimmins, E. & Beltran-Sanchez, H. (2011). Mortality and morbidity trends: Is there
compression of morbidity? Journals of Gerontology: Social Sciences, 66, 75-86.
Curran, S. R. (2002). Agency, accountability, and embedded relations: What‘s love got to
do with it? Journal of Marriage and Family, 64, 577-584.
Dannefer, D. (2003). Cumulative advantage, and the life course: Cross-fertilizing age and
social science knowledge. Journal of Gerontology: Social Sciences, 58, 327-337.
Elder, G.H. (1994). Time, human agency, and social change: Perspectives on the life
course. Social Psychology Quarterly, 57, 4-15.
Elder, G.H. (1995). The life course paradigm: Social change and individual development.
In P. Moen, G.H. Elder, and K. Luscher (Eds.), Examining lives in context:
Perspectives on the ecology of human development. APA Books.
113

Ekeh, Peter P. (1974). Social exchange theory: The two traditions. Cambridge, MA:
Harvard University Press.
Enders, C.K. & Bandalos, D.L. (2001). The relative performance of full information
maximum likelihood estimation for missing data in structural equation models.
Structural Equation Modeling: A Multidisciplinary Journal, 8, 430-457.
Estes, C.L. (2001). The political economy of aging: A critical perspective on social
gerontology" In Biggs, S., Lowenstein, A. and Hendricks, J. (Eds.) The need for
theory: Social gerontology for the 21st century. Amityville, NY: Baywood.
Fingerman, K. L. (1996). Sources of tension in the aging mother and adult daughter
relationship. Psychology and Aging, 11, 591-606.
Fingerman, K. L., Chen, P. C., Hay, E. L., Cichy, K. E., & Lefkowitz, E. S. (2006).
Ambivalent reactions in the parent and offspring relationship. Journals of
Gerontology: Psychological Sciences, 61, 152-160.
Fingerman, K. L., Hay, E., and Birditt, K. S. (2004). The best of ties, the worst of ties:
Close, problematic, and ambivalent social relationships. Journal of Marriage and
the Family, 66, 792-808.
Fingerman, K.L., Pitzer, L., Lefkowitz, E.S., Birditt, K.S., and Mroczek,, D. (2008).
Ambivalent relationship qualities between adults and their parents: implications
for the well-being of both parties. Journal of Gerontology, 63B, 362-371.
Freud, S. (1913). Totem and taboo. (A. A. Brill, Trans.). New York, NY: New Republic.
Gans, D. & Silverstein, M. (2006). Norms of filial responsibility for aging parents
across time and generations. Journal of Marriage and the Family, 68, 961- 976.
114

Giarrusso, R., Feng, D., & Bengtson, V . L. (2004). The intergenerational-stake
phenomenon over 20 years. In M. Silverstein (Ed.), Annual review of gerontology
and geriatrics (V ol. 24, pp. 55 – 76). New York, NY: Springer.
Giarrusso, R., Silverstein, M., Gans, D., and Bengtson, V . L. (2005). Ageing
parents and adult children: New perspectives on intergenerational
relationships. In M. L. Johnson, V . L. Bengtson, P. G. Coleman and T. B. L.
Kirkwood, (Eds.), Cambridge handbook of age and ageing (pp. 413-421).
London: Cambridge University Press.
Hogan, D. P., Eggebeen, D J., & Clogg, C. C. (1993). The structure of intergenerational
exchanges in American families. American Journal of Sociology, 98, 1428-1458.
Ha, J. & Ingersoll-Dayton, B. (2008). The effect of widowhood on
intergenerational ambivalence. The Journals of Gerontology Series B:
Psychological Sciences, 63, 49-58.
Lawton, L., Siverstein, M., and Bengtson, V .L. (1994). Solidarity between
generations in families. In V .L. Bengtson and R.A. Harootyan (Eds.),
Intergenerational linkages: Hidden connections in American Society (pp. 19-42).
New York, NY: Springer.
Jajodia, A. (2011). Dynamic structural equation models of change. In Newsom, J.T.,
Jones, R.N., & Hofer, S.M. (Eds.), Longitudinal data analysis: A practical guide
for researchers in aging, health, and social sciences. New York, NY: Routledge.

115

Lee, G., and E. Ellithorpe. 1982. Intergenerational exchange and subjective well-
being among the elderly. Journal of Marriage and the Family, 44, 217-224.
Lowenstein, A. (2007). Solidarity-conflict and ambivalence: testing two
conceptual frameworks and their impact on quality of life for older family
members. Journal of Gerontology: Social Sciences, 62, 100-107.
Lüscher, K., & Pillemer, K. (1998). Intergenerational ambivalence: A new
approach to the study of parent-child relations in later life. Journal of Marriage
and the Family, 60, 413-425.
Lüscher, K. (2002). Intergenerational ambivalence: Further steps in theory and
research. Journal of Marriage and Family, 64, 585-593.
Lüscher, K. (2004). Conceptualizing and uncovering intergenerational
ambivalence. In Pillemer, K. and Lüscher, K. (Eds.), Intergenerational
Ambivalences: New Perspectives on Parent-child Relations in Later Life (pp. 23-
62). Amsterdam: Elsevier.
Magnusson, D. (1995). Individual development: A holistic, integrated model. In P. Moen,
G.H. Elder, and K. Luscher (Eds.), Examining lives in context: Perspectives on the
ecology of human development. APA Books.
McArdle J.J. (2009). Latent variable modeling of differences and changes with
longitudinal data. Annual Review of Psychology, 60, 677-605.
Merton, R. K., & Barber, E. (1963). Sociological ambivalence. In E. Tiryakian (Ed.),
Sociological theory: Values and sociocultural change (pp. 91-120). New York,
NY: Free Press.
116

Merz, E.M., Schuengel, C. & Schulze, H.J. (2009). Intergenerational relations across
four years: wellbeing is affected by quality not by support exchange. The
Gerontologist, 49, 536-548.
Moen, P., Elder, G.H., & Luscher, K. (1995). Preface. In P. Moen, G.H. Elder, and K.
Luscher (Eds.), Examining lives in context: Perspectives on the ecology of human
development. APA Books.
Moen. P. (1995). Introduction. In P. Moen, G.H. Elder, and K. Luscher (Eds.), Examining
lives in context: Perspectives on the ecology of human development. APA Books.
Newsom, J. T., Rook, K.S., Nishishiba, M., Sorkin, D.H., & Mahan, T.L. (2005).
Understanding the relative importance of positive and negative social exchanges:
examining specific domains and appraisals. Journals of Gerontology:
Psychological Sciences, 60, 304-312.
Peters, C. L., Hooker, K., & Zvonkovic, A. M. (2006). Older parents‘ perceptions
of ambivalence in relationships with their children. Family Relations, 55, 539-
551.
Pillemer, K. & Suitor, J. J. (2002). Explaining mothers' ambivalence toward their adult
children. Journal of Marriage and Family, 64, 602-613.
Pillemer, K. & Suitor, J. J. (2004). Ambivalence and the study of intergenerational
relations. In Silverstein, M. & Schaie, W. (Eds.), Annual Review of Gerontology
and Geriatrics, (pp. 3-23). Springer Publishing.
117

Pillemer, K., Suitor, J. J., Mock, S.E., Sabir, M., Pardo, T.B., & Sechrist, J. (2007).
Capturing the complexity of intergenerational relations: Exploring ambivalence
within later-life families. Journal of Social Issues, 63, 775-791.
Rutter, M., Champion, L., Quinton, D., Maughan, B., & Pickles, A. (1995).
Understanding individual differences in environmental-risk exposure. In P. Moen,
G.H. Elder, and K. Luscher (Eds.), Examining lives in context: Perspectives on the
ecology of human development. APA Books.
Roberto, K. A., & Jarrott, S. E. (2008). Family caregivers of older adults: A life-span
perspective. Family Relations, 57, 100-111.
Rook, K. (1984). The negative side of social interaction: Impact on psychological well-
being. Journal of Personality and Social Psychology, 46, 1097-1108.
Silverstein, M., Gans, D., Lowenstein, A., Giarrusso, R., & Bengtson, V .L. (2010). Older
parent-child relationships in six developed nations: comparisons at the
intersection of affection and conflict. Journal of Marriage and Family, 72, 1006-
1021.
Silverstein, M., & Litwak, E. (1993). A task specific typology of intergenerational family
structure in later life. The Gerontologist, 33, 258-264.
Shapiro, A. (2004). Revisiting the generation gap: Exploring the relationships of
parent/adult-child dyads. The International Journal of Aging and Human
Development, 58, 127-146.
Suitor, Gilligan, Pillemer (2011). Conceptualizing and measuring intergenerational
ambivalence in later life. Journal of Gerontology: Social Sciences, 66, 769-781.
118

Spitze, G. & Gallant, M. P. (2004). The bitter with the sweet: Older adults‘
strategies for handling ambivalence in relations with their adult children.
Research on Aging, 26, 387-412.
Steinbach, A. (2008). Intergenerational solidarity and ambivalence: Types of
relationships in German families. Journal of Comparative Family Studies, 39,
115-127.
Thompson, M. M., Zanna, M. P., Griffin, D. W. (1995). Let's not be indifferent about
(attitudinal) ambivalence. In R. E. Petty & J. A. Krosnick (Eds.), Attitude
strength: Antecedents and consequences (pp. 361–385). Mahwah: NJ: Erlbaum.
Uchino, B. N. (2009). Understanding the links between social support and physical
health: A lifespan perspective with emphasis on the separability of perceived and
received support. Perspectives in Psychological Science, 4, 236-255.
Umberson, D. (1992). Relationships between adult children and their parents:
psychological consequences for both generations. Journal of Marriage and
Family. 54.
Umberson, D., Chen, M.D., House, J.S., Hopkins, K., & Slaten, E. (1996). The effects of
social relationships on psychological well-being: Are men and women really so
different?‖ American Sociological Review, 61, 837-857.
Umberson, D. and Montez, J.K. (2010). Social relationships and health: A flashpoint for
health policy. Journal of Health and Social Behavior, 51, 54-66.
119

van Gaalen, R. I. & Dykstra, P.A. (2006). Solidarity and conflict between adult
children and parents: a latent class analysis. Journal of Marriage and Family, 68,
947-960.
van Gaalen, R. I., Dykstra, P. A. & Komfter, A. E. (2010). Where is the exit?
Intergenerational ambivalence and relationship quality in high contact ties.
Journal of Aging Studies, 24, 105-114.
Ward, R.A. (2008). Multiple parent-adult child relations and well-being in middle and
later life. Journal of Gerontology, 63B, 239-247.
Willson, A. E., Shuey, K. M., & Elder, G. H. (2003). Ambivalence in the relationship of
adult to aging parents and in-laws. Journal of Marriage and Family, 65, 1055-
1072.
Willson, A. E., Shuey, K. M., Elder, G. H., & Wickarama, K. A. S. (2006). Ambivalence
in mother-adult child relations: A dyadic analysis. Social Psychology Quarterly,
69, 235-252.

Abstract (if available)

Abstract The overall goal of this dissertation is to describe the measurement, trajectories, and consequences of intergenerational ambivalence among older parents and adult children. Ambivalence is the coexistence of affection and conflict within the intergenerational relationship and there are a variety of techniques in which to measure ambivalence. However, there is little certainty about their validity or equivalence of the measures within empirical research. Further, the majority of research on intergenerational ambivalence is cross-sectional, which limits family researchers’ knowledge to only familial and individual characteristics associated with change in ambivalence. Ambivalence has been related to well-being, but there is little evidence of a causal relationship between ambivalence and psychological outcomes. Therefore, there is a need to validate measurements and test causes and/or consequences of ambivalence over the adult life course to advance the utility of the intergenerational ambivalence model. The objectives are: (1) to compare and validate two methods of quantitatively measuring intergenerational ambivalence within parent-child dyads

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Intergenerational transmission of values and behaviors over the family life course

PDF

Childlessness and psychological well-being across life course as manifested in significant life events

PDF

Normative obligations and parental care in social context

PDF

Transmission and transitions: young adults' beliefs, values and life course transitions in familial context

PDF

Family relationships and their influence on health outcomes over time among older adults in rural China

PDF

Filial expectations and social exchange patterns among older Taiwanese parents and their adult children

PDF

Intergenerational coresidence of older adults in contemporary Japan: traditional cultural norms in divergent styles

PDF

The immediate and long term legacy of relationships with grandparents for the well-being of grandchildren

PDF

Racial/ethnic variation in care preferences and care outcomes among United States hospice enrollees

PDF

Age integration in late life: sociodemographic & psychosocial correlates of intergenerational-only, peer-only, and age-integrated social networks

PDF

Children 's migration and the financial, social, and psychological well-being of older adults in rural China

PDF

Intergenerational social support and the psychological well-being of older parents in China

PDF

Improvements in women's status, decision-making and child nutrition: evidence from Bangladesh and Indonesia

PDF

Internet communication use, psychological functioning and social connectedness at older ages

PDF

The effect of body weight on mortality: different countries and age groups

PDF

Mechanisms of stress effects on learning and decision making in younger and older adults

PDF

The effects of wisdom-related personality traits on caregivers’ health: an application of the resilience model

PDF

How emotional arousal influences memory and learning in younger and older adults

PDF

Intergenerational support between grandparents and grandchildren in rural China and its effect on the psychological well-being of older adults

PDF

Palliative care: breaking the knowledge barrier among community-based older adults

Asset Metadata

Creator Lendon, Jessica Penn (author)

Core Title The measurement, life course patterns, and outcomes of intergenerational ambivalence among parent-adult child dyads

School Leonard Davis School of Gerontology

Degree Doctor of Philosophy

Degree Program Gerontology

Publication Date 05/02/2012

Defense Date 01/13/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag aging families,Gerontology,intergenerational ambivalence,life course,OAI-PMH Harvest

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Silverstein, Merril (committee chair), Biblarz, Timothy (committee member), Giarrusso, Roseann (committee member)

Creator Email jessicapenn@gmail.com,lendon@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-23765

Unique identifier UC11290091

Identifier usctheses-c3-23765 (legacy record id)

Legacy Identifier etd-LendonJess-721.pdf

Dmrecord 23765

Document Type Dissertation

Rights Lendon, Jessica Penn

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 a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA