AN ABSTRACT OF THE DISSERTATION OF
Lorina S. Evans for the degree of Doctor of Philosophy in Human Development
and Family Studies on July 10, 1996. Title: Amount of Care Given by Daughters
and Perceived Relationship Quality with Care-Receiving Mothers
Abstract approved:
Redacted for Privacy
Alexi J. Walker
The relationship between mothers and daughters has been shown to be
generally positive and strong throughout the life cycle. The relationship may
become strained, however, as mothers become dependent due to deteriorating
health or chronic health conditions. It has long been recognized in both the
gerontological and intergenerational literatures that caring for a dependent
elderly parent can be burdensome and stressful for some caregiving adult
daughters. Using the elaborated wear-and-tear hypothesis, this study examined
longitudinally the impact of the amount of care given on 140 caregiving
daughters' perception of the quality of their relationship with their elderly
dependent mothers. It was hypothesized that increases in the amount of care
given by caregiving daughters over time would lead to a corresponding decrease
in relationship quality. It was also hypothesized that relationship quality and
amount of care given would be stable over time. Results indicated that providing
more care does decrease relationship quality between caregiving daughters and
care-receiving mothers over time although it does not do so initially. Further, the
results of study showed that relationship quality and amount of care given, on
average, are stable over time.
Amount of Care Given by Daughters and Perceived Relationship Quality with
Care-Receiving Mothers
By
Lorina S. Evans
A DISSERTATION
Submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Doctor of Philosophy
Completed July 10, 1996
Commencement June 1997
Doctor of Philosophy dissertation of Lorina S. Evans presented on July 10, 1996
APPROVED:
Redacted for Privacy
Major Profes or, representing Human Development and Family Studies
Redacted for Privacy
Head of Department o Human Development and Family Sciences
Redacted for Privacy
Dean of Gr
uate School
I understand that my thesis will become part of the permanent collection of
Oregon State University libraries. My signature below authorizes release of my
thesis to any reader upon request.
Redacted for Privacy
Lorina S. Evans, Author
ACKNOWLEDGMENTS
I wish to express my sincere appreciation to my major advisor, Dr. Alexis
J. Walker, for her valuable time, advice, and guidance in this project. I also wish
to thank my other committee members, Dr. Alan Acock, Dr. Mina Carson,
Dr. Brooke Collison, Dr. Patricia Moran, and Dr. Anisa Zvonkovic for their
encouragement and their support in bringing this project to completion. I am also
indebted to both Dr. Alan Acock and Dr. Fuzhong Li for their assistance and
guidance in the editorial process of the statistical analyses section.
Additionally, I would like to give a special thank you to my friend Shari
Sakashita, for her support over a difficult journey, encouraging me to pursue my
dreams, and to hang on when things were tough.
Lastly, I wish to express my gratitude to my parents and friends for their
patience and support along this journey.
This research was supported by grant R01-AG-06766 from the National
Institute on Aging.
TABLE OF CONTENTS
Chapter
Chapter One: Introduction
Page
1
Purpose of the Study
4
Definition of Terms
6
Chapter Two: Literature Review
7
Wear-and-Tear Hypothesis
7
Research on the Wear-and-Tear Hypothesis
8
Overview of Elderly Caregiving
11
Outcomes for Caregiving Daughters
13
Intergenerational Relationships
14
Elderly Caregiving and Mother-Daughter Relationships
16
Impact of Amount of Care Given and Relationship Quality
17
Covariates
18
Modeling of Change in Relationship Quality
20
Control Variables
20
Hypotheses
23
TABLE OF CONTENTS (Continued)
Chapter
Page
Chapter Three: Method
24
Sample
24
Attrition
28
Procedures
30
Measures
31
Care given
Perceived relationship quality
Control variables
31
32
33
Data Analysis
34
Missing Data
36
Chapter Four: Results
38
Preliminary Data Analyses
38
Structural Equation Model
40
Measurement Model
41
Structural Model
45
Measurement Invariance
48
Loadings on the latent variables
Path coefficients
Effects of the exogenous variables
49
51
53
TABLE OF CONTENTS (Continued)
Chapter
Page
Chapter Five: Discussion and Conclusion
Primary Findings
Stability of the endogenous latent variables
Relationship between amount of care given and relationship
quality
Relationship between covariates and endogenous latent
variables
55
56
56
58
60
Limitations and Strengths
62
Limitations
Strengths
62
63
Implications
64
Directions for Future Research
64
Bibliography
66
Appendices
80
Appendix A: Caregiving Tasks and Activities
Appendix B: Intimacy Scale
Appendix C: Attachment Scale
81
84
85
LIST OF FIGURES
Figure
Page
1
Conceptual Model
2
Structural and Measurement Model
22
3
Measurement Model of Latent Constructs
42
5
LIST OF TABLES
Table
Page
1
Selected Sociodemographic Characteristics of Caregiving Daughters
26
2
Comparisons of Daughters in the Sample with Those Lost to Attrition
29
3
Correlations Among Observed Indicators and Exogenous Variables
39
4
Loadings of Measurement Model
44
5a
Goodness-of-Fit Indices of Structural Models
46
5b
Model Comparisons
47
6
Loadings for Structural Model 4
50
7
Parameter Estimates of the Hypothesized Model (Beta)
52
8
Parameter Estimates of the Exogenous Variables to Latent Variables
54
AMOUNT OF CARE GIVEN BY DAUGHTERS AND PERCEIVED
RELATIONSHIP QUALITY WITH CARE-RECEIVING MOTHERS
Chapter One
Introduction
Most elderly persons continue to be in good health despite advancing
age. Some become dependent due to deteriorating health or chronic
conditions, however. Kunkel and Applebaum (1992) projected that by the
year 2020, 4.9 million elderly persons aged 65 and over will experience some
type of chronic problems. When chronic conditions limit daily activities, the
need for assistance is usually provided by a family member.
Caregiving to impaired elderly persons usually begins when they are
unable to perform such tasks as housekeeping, shopping, or meal
preparation without some assistance (Brody, 1985; Brody & Schoonover,
1986; Cicirelli, 1981; Lang & Brody, 1983; Stephens & Christianson, 1986;
Stone, Cafferata, & Sang!, 1987; Walker, Pratt, & Eddy, 1995). Spouses
usually provide care to their dependent partner. In the absence of a spouse,
adult children, particularly adult daughters, assume the responsibility of
providing care needed by their aging parent(s) (Brody, 1986; Cicirelli, 1983;
Mancini & Blieszner, 1989; Shanas, 1979; Stoller, 1983).
2
Providing care to an impaired parent has been shown to be costly,
particularly for adult daughters. Caregivers often give up their personal and
social activities (Abel, 1989) and are confined to their homes (Robinson &
Thurnher, 1979). Additionally, caregivers report experiencing emotional costs
such as frustration, resentment, and feeling overwhelmed (e.g., Archbold,
1983; Cicirelli, 1983; Robinson & Thurnher, 1979). Increasingly, it appears
that the amount of care given is a critical factor in caregiving costs. That is,
those who give more care, have more costs (Brody, 1985; Walker, Acock,
Bowman, & Li, 1996). Indeed, the amount of care given is far more important
in predicting the outcomes of caregiving than is caregiving duration.
Despite the myth that aging parents are socially isolated and
abandoned by their adult children, researchers have shown that both
continue to interact and depend on each other throughout life (Brody,
Johnsen, Fulcomer, & Lang, 1984; Cicirelli, 1983; Rossi & Rossi, 1990;
Shanas, 1979). On average, adult children have frequent contact with their
parents, live within a 30-mile radius of them, and provide emotional and
instrumental assistance as they grow older (Shanas, 1979). As aging parents
receive emotional support and assistance from their adult children in tasks
they no longer are able to perform independently, aging parents may
continue to provide monetary support as well as love and affection.
Mother-daughter relationships in middle and later life have been
shown to be strong and enduring ties, characterized by affection and by the
exchange of aid (Bretherton, Biringen, & Ridgeway, 1991; Boyd, 1989;
3
Bromberg, 1982-1983; Cicirelli, 1983, 1991; Himes, 1994; Lawton,
Silverstein, & Bengtson, 1994; Mancini & Blieszner, 1989; Rossi & Rossi,
1990; Stueve & O'Donnell, 1984). Relationships between mothers and
daughters in middle and later life have also been shown to be positive and
rewarding to both parties (Baruch & Barnett, 1983; Barnett, Baruch, Kibria, &
Pleck, 1991). These relationships may be strained however, by experiences
such as daughter's marital dissolution or even by the experience of
caregiving (Johnson & Catalano, 1983; Umberson 1989a, 1989b, 1992).
Caregiving has produced both positive and negative consequences for
mothers and daughters. Some researchers have found that adult daughters
generally have a good relationship with their parent(s) despite the strain and
burden that caregiving often brings (e.g., Abel, 1986; Miller, 1989; Walker,
Shin, & Bird, 1990). Other researchers report opposing results. Scharlach
(1987), for example, found that adult daughters reported a deterioration in
their relationship with their parent(s) as the need for care increased. Overall,
little research has examined the potential detrimental effect on caregiving of
elderly mother-adult daughter relationships. Furthermore, only a handful of
longitudinal studies have examined the impact of caregiving (e.g., Johnson &
Catalano, 1983; Townsend, Noelker, Deimling, & Bass, 1989). Most studies
have been cross-sectional in nature and have looked at adult children caring
for elderly parents with Alzhiemer's Disease or other types of dementia (e.g.,
Pratt, Schmall, Wright, & Cleland, 1985; Pruchno & Resch, 1989; Reed,
Stone, & Nee le, 1990). Research to understand the impact of caregiving on
4
relationships between caregivers and care receivers is needed, particularly
longitudinal research.
Purpose of the Study
The purpose of the study was to examine the impact of the amount of
care given on daughters' perception of their relationship with their carereceiving mothers. The study investigated longitudinally the effect of the
amount of care given over time on the perceived quality of the relationship
between caregiving adult daughters and their mothers, thus filling in a gap in
the existing literature (See Figure 1). The study builds on findings that the
amount of care given and not duration is the important factor in determining
caregiving outcomes.
5
Figure 1. Conceptual Model
6
Definition of Terms
Caregivers.
Caregivers refers to individuals, such as spouses and adult children,
who assist dependent elderly individuals with activities of daily living (ADL)
(Stone et al., 1987).
Care recipients.
Care recipients refers to elderly individuals who need assistance with
activities of daily living (ADL) (Stone et al., 1987).
7
Chapter Two
Literature Review
Wear-and-Tear Hypothesis
The wear-and-tear hypothesis, as applied to caregiving, proposes that as
duration of caregiving increases, caregivers will experience increasing strain and
burden (e.g., Townsend, Noelker, Deimling, & Bass, 1989). The wear-and-tear
hypothesis is derived from social exchange theory. Social exchange theory has
emerged as a major theoretical framework by which to explain social
interactional processes in dyadic and small-group relationships (McDonald,
1981). The basic premise of social exchange theory is that individuals involved in
social interaction attempt to maximize rewards and minimize costs to obtain the
most profitable outcome (Thibaut & Kelley, 1959). That is, individuals will seek
what is most beneficial to them at the lowest cost. Costs are defined as both
unpleasant experiences and rewards foregone. Rewards are defined as
experiences that lead to some type of gratification such as relationship
satisfaction (Thibaut & Kelley, 1959).
Relationship changes in later life as a consequence of care given by adult
children to their aging parents can be explained by social exchange theory.
Thibaut and Kelley (1959) posit that in order for a relationship to be satisfying,
both members of a dyad must find interaction more rewarding than costly. When
8
one or both individuals is unable to reciprocate, an imbalance may be created
that is costly to individuals in the relationship. Dowd (1975) applied social
exchange theory to aging suggesting that as individuals age, their resources are
diminished, thus creating an imbalance in the exchange of aid. In other words,
as aging parents become more dependent on their adult children, their level of
independence diminishes. This decline may result in their feeling helpless and
useless. For caregiving adult children, the increasing amount of care given
coupled with declines in aid received may lead to burden, strain, and relationship
deterioration (Gerstel & Gallagher, 1993; Miller, McFall, & Montgomery, 1991;
Montgomery, Gonyea, & Hooyman, 1985; Stoller, 1983; Young & Kahana,
1985).
Research on the Wear-and-Tear Hypothesis
According to the wear-and-tear hypothesis, caregivers experience
increasing burden and stress as duration of caregiving increases. Research on
caregiving generally has focused on negative outcomes for caregivers (Archbold,
1983; Baruch & Speid, 1989; Cicirelli, 1983; Montgomery & Kamo, 1989).
Mathews (1988), Stull, Bowman, and Smerglia (1994), and Zarit, Reeves, and
Bach-Peterson (1989) suggested, however, that the majority of caregivers report
low levels of strain, depression, and burden, thus questioning the validity of the
wear-and-tear hypothesis. Stephens and Zarit (1989) proposed an adaptation
9
hypothesis in which caregivers adjust to the caregiving situation over time
despite stressors (see Pear lin, Mullen, Semple, & Skaff, 1990). In other words,
stress should level off as caregivers adapt and adjust to meeting the elderly
person's needs.
Only a few studies have looked at the wear-and-tear hypothesis
longitudinally. Johnson and Catalano (1983) followed 115 elderly individuals
recently discharged from the hospital and their caregivers over an 8-month
period. They found that on average, caregivers reported experiencing stress and
burden as the amount of care given increased. They also found that the
caregiver-care recipient relationship deteriorated due to the increased care
provided. Townsend, Noelker, Deimling, and Bass (1989) examined
longitudinally the impact of caregiving to impaired elderly parents on adult
children's mental health. They found substantial individual variability in adult
children's level of stress and depression. For example, 36% (n = 40) of the adult
child caregivers reported increased stress while 52% n = 57) reported decreased
stress over time. Approximately 12% (n = 13) reported no change in their stress
levels and 8% (n = 9) of the caregivers reported no change in depression.
Depression decreased for 58% (n = 65) of the caregivers but increased for 34%
(n = 38) The overall changes were small, however. More importantly, they
.
found that duration of caregiving did not influence adult children's mental health.
Longitudinal studies of caregiving for persons with Alzheimer's Disease or
dementia have produced mixed results (Haley & Pardo, 1989; Schulz &
Williamson, 1991; Zarit, Todd, & Zarit, 1986). Zarit et al. (1986) followed
10
caregivers of family members with dementia for a 2-year period. They found that
women (wives) initially reported being more burdened than men (husbands) at
Time 1. At Time 2, women were able to adapt and adjust to the caregiving
situation despite the progression of dementia. Schulz and Williamson (1991) also
found considerable variability in depression scores for men and women giving
care to persons with Alzheimer's Disease over a 2-year period. For example,
only 27.3% (n = 6) of the men reported being depressed over the 2-year period.
Approximately 47% (n = 24) of the women (both wives and daughters) reported
higher levels of depression over time, thus, supporting the wear-and-tear
hypothesis.
Walker, Acock, Bowman, and Li (1996) extended the wear-and-tear
hypothesis by explaining individual variability in the adaptation to caregiving over
time. The researchers examined the impact of change on the amount of care
given by 130 White caregiving daughters over a 4-year period. They found that
on average, adult daughters gave increasing amounts of care over the 4-year
period, and they reported high caregiving satisfaction. They also found that, as
the amount of care increased over time, caregiving satisfaction decreased even
though daughters' caregiving satisfaction scores were high. Additionally, they
found that duration of caregiving (wear-and-tear hypothesis) was unrelated to
change in daughters' caregiving satisfaction. The small decrease in caregiving
satisfaction over the 4-year period was related to the increasing amount of care
given, rather than duration or time. It appears then, that the wear-and-tear
hypothesis must be modified to include change in the amount of care. The
11
increased amount of care that caregivers provide over time appears to be related
to caregiving stress and burden. Therefore, researchers need to focus more on
the amount of care given than on duration of caregiving to explain caregiving
outcomes.
Overview of Elderly Careaivinq
Persons aged 65 and over comprise approximately 13% of the population
(Harris, 1990). Increased life expectancy, the advancement of life-saving medical
technology, low fertility rates, and low mortality rates have contributed to the
increased need for elderly caregiving (Kunkel & Applebaum, 1992; Mancini &
Blieszner, 1989; Treas, 1977).
Although most elderly persons continue to be in good health and are selfsufficient, nearly 10% of those aged 65 and older, and one quarter of those over
age 84 become dependent due to deteriorating health or chronic conditions such
as arthritis, loss of hearing, and vision impairment (Nimes, 1994). Under such
conditions, family members provide assistance to the elderly (Archbold, 1983;
Brody, 1986; Cicirelli, 1983; Nimes, 1994; Lee, 1987; Mancini & Blieszner, 1989;
Stoller, 1983). Shanas (1979) noted that spouses usually provide care for the
dependent elderly who are married. In the absence of a spouse, however, adult
children typically provide care to their aging parent(s). Increased longevity
12
means that such care is provided for longer periods of time (Brody, 1985;
Shanas, 1979; Stoller, 1983).
Stone, Cafferata, and Sang! (1987) provided a national profile of
caregivers and care recipients using the 1982 National Long-Term Care Survey
(also known as the channeling demonstration) and the matched Informal
Caregivers Survey. These surveys were the first databases where national
estimates of family members, friends, and other unpaid caregivers to the
uninstitutionalized elderly were provided. Stone et al. (1987) found that the
majority (72%) of caregivers are women, with adult daughters comprising 29% of
those caring for an impaired aging person. Wives comprised 23% of this
population. The average age for caregiving wives was 69 years with 48%
between the ages 65 to 74 years. The average age for caregiving adult
daughters was 52 years with 63% between the ages 45 to 64 years. Women
caregivers provide assistance in personal care, medical care, and other in-home
care (e.g., meal preparation, shopping, transportation).
Stone et al. (1987) found that 60% of the care recipients are women and
51% are married. The mean age for care recipients was 78 years with 78% aged
65 and above. Approximately 40% lived with a spouse, 36% lived with a spouse
and/or children, and 11% lived alone. Approximately 23% of the care recipients
reported being in good health despite increasing chronic conditions. Elderly care
recipients reported needing help performing activities of daily living (ADL), such
as bathing, dressing, toileting, getting in and out of bed or chairs, and eating, and
instrumental activities of daily living (IADL), such as shopping, cleaning house,
13
and laundry. Stone et al.'s (1987) study substantiated earlier findings that women
are actively involved as caregivers and as care receivers (Archbold, 1983; Brody,
1985; Cicirelli, 1981; Lang & Brody, 1983; Stoller, 1983).
Outcomes for Caregiving Daughters
The impact of elderly caregiving on adult daughters continues to be a
major focus in gerontological research (Abel, 1986; Aronson, 1992; Brody, Litvin,
Albert, & Hoffman, 1994; Cicirelli, 1981; Dwyer & Coward, 1991; Dwyer &
Seccombe, 1991; Nimes, 1994; Horowitz, 1985; Mancini & Blieszner, 1989;
Stoller, 1983; Walker et al., 1992). With caregiving to their aging parent(s), adult
daughters report being distressed and burdened (Barusch & Spaid, 1989;
Gerstel & Gallagher, 1993; Montgomery et al., 1985; Montgomery & Kamo,
1989; Young & Kahana, 1986), depressed (Archbold, 1983; Pruchno & Resch,
1989), frustrated (Circirelli, 1983; Motenko, 1989) and having a restricted outlook
for the future (Rakowski & Clark, 1985). Not all outcomes are negative, however.
Abel (1986) suggested that the mother-daughter relationship can be
strengthened as a consequence of caregiving. Archbold (1983) found that
caregivers reported being satisfied when they managed rather than gave care to
their dependent elderly parent(s). She also found that caregivers learned much
about themselves and about aging in the caregiving process, and reported that
their relationship with their dependent elderly parent(s) improved as a result of
14
caregiving. Walker, Martin, and Jones (1992) found that daughters who have a
good relationship with their dependent elderly mothers reported less caregiving
costs (i.e., insufficient time, frustration, anxiety).
Intergenerational Relationships
Most intergenerational relationships in later life are continuous in that
these relationships are shaped by past intimate experiences, histories, patterns,
and behaviors. Research has indicated that family members live in close
proximity to each other, maintain contact, share important emotional ties, and
exchange aid (Cicirelli, 1983; DeWitt, blister, & Burch, 1988; Frankel & DeWitt,
1988; Lee, 1988; Lee & Ellithorpe, 1982; Mancini & Blieszner, 1989; Stoller,
1985; Walker & Pratt, 1991; Walker, Pratt, & Oppy, 1992). For example, parents
and adult children provide all types of services (e.g., babysitting, major help with
housework, financial assistance) to each other throughout life (Hogan,
Eggebeen, & Clogg, 1993; Wellman & Wortley, 1990). Barnett, Baruch, Kibria,
and Pleck (1991) studied the effect of relationship quality on mental health (i.e.,
subjective well-being and psychological distress) in 403 adult daughters. They
found that adult daughters generally reported having positive relationships with
their aging parent(s). They also found that relationship quality was positively
related to daughters' well-being. In other words, daughters who reported having
15
a good relationship with their aging parent(s) had higher psychological wellbeing.
In addition to examining intergenerational solidarity, such as type and
frequency of interactions and activities, and affectional sentiments (see
Harootyan & Bengtson, 1994; Bengtson & Roberts, 1991 for a review),
researchers have focused their attention on the effect of early family
relationships in later life (Rossi & Rossi, 1990; Whitbeck, Simons, and Conger,
1991). For example, Rossi and Rossi (1990) examined parent-child relationships
across the life course. They found that parents who were affectionate,
accessible, and had a close relationship with their children in adolescence were
more intimate with their adult children later in the life course. They also found
that early family cohesiveness influenced intimacy and closeness between adult
children and their parents in later life. Whitbeck, Simons, and Conger (1991) also
examined the effects of early parent-child relationships on current relationships.
They found that adult children's perceptions of early relationships with their
elderly parents did impact their relationships with them in later life. Adult children
who perceived their parents as rejecting and hostile and as harsh disciplinarians
early in life reported having poorer relationships with them currently. They also
found depression to be negatively related to current relationships. Their study
substantiates Rossi and Rossi's (1990) findings that parents who had poor
relationships with their children earlier in life were more likely to have poor
relationships with them in later life.
16
Elderly Caregiving and Mother-Daughter Relationships
The relationships between elderly parents and adult daughters in later life
have been studied in both the gerontological and intergenerational literatures
(Abel, 1989; Barer & Johnson, 1990; Brody & Schoonover, 1986; Cicirelli, 1983;
Nimes, 1994; Lang & Brody, 1981; Sheehan & Nuttall, 1988; Walker & Allen,
1991). Research has shown that on average, adult daughters live near their
elderly parents, have frequent contact with them, and generally view their
relationships as positive (Abel, 1986; Bengtson, Cutler, Mangen, & Marshall,
1985; Lawton, Silverstein, & Bengtson, 1994; Mancini & Blieszner, 1989;
Silverstein, Lawton, & Bengtson, 1994). Baruch and Barnett (1983) found that
adult daughters perceived their relationship with their mothers as rewarding and
positive. Walker, Shin, and Bird (1990) studied perceived relationship changes
as a consequence of caregiving in 133 elderly mother-adult daughter pairs. The
researchers found that 50.4% of daughters reported a positive impact of
caregiving on their relationship, but 44.2% reported no change. Additionally,
daughters who reported a positive impact or no change in their relationship with
their mothers were satisfied with caregiving.
On the negative side, Boyd (1987) and Walker and Allen (1991) found
that cargiving adult daughters who were in conflict with their elderly mothers
reported their relationship as unrewarding. Additionally, caregiving daughters
experienced greater strain or costs (e.g., time constraints, burden) when their
relationships with their mothers were poor (Scharlach, 1987; Sheehan & Nuttall,
17
1988). Very few researchers, however, have examined the impact of caregiving
on relationship quality.
Impact of Amount of Care Given and Relationship Quality
In order to understand factors that contribute to relationship quality,
researchers have measured this latent construct as intimacy (Walker &
Thompson, 1983), attachment (Cicirelli, 1983; Thompson & Walker, 1984), and
affection (Jarrett, 1985; Lawton, Silverstein, & Bengtson, 1994). Although adult
children and in particular, adult daughters, do not have to have strong emotional
ties to provide care to an elderly parent(s), some research has shown that
caregiving brings the mother-daughter relationship closer (Abel, 1989; Baruch &
Barnett, 1983; Cicirelli, 1983; Lang & Brody, 1983; Walker, Acock, Bowman, &
Li, 1996). For example, Cicirelli (1983) examined 148 adult children's helping
behavior to their elderly parents. The researcher found that adult children
provided more care to their elderly parent when they felt a strong attachment.
Lawton, Silverstein, and Bengtson (1994) also found that affectional ties bring
the mother-daughter relationship closer. Other research has shown that
relationship deterioration and strain are consequences of care given (Scharlach,
1987; Sheehan & Nuttall, 1988). Mancini and Blieszner (1989) and Walker, Pratt,
Shin, and Jones (1990) noted that more research is needed on the impact of
care given on the quality of mother-daughter relationships in later life. It is
18
noteworthy that there is little research on relationship quality as it applies to
mother-daughter relationships in middle and later life.
Covariates
A number of variables may influence the amount of care given. It is
important to control for these influences in any examination of the effect of the
amount of care given on relationship quality.
Age. Age is often used as a demographic variable in many gerontological
studies because it is often assumed that individuals change in predictable
patterns as they age (e.g., Kart, 1985; Markides & Cooper, 1989). Age is
multidimensional, however, and encompasses biological, psychological, and
chronological age. Most research on caregiving has shown that a person's age is
related to care received (see Cicirelli, 1981; Hareven, 1995; Kinsella, 1995). As
individuals live longer, they face an increased likelihood of becoming dependent
and of requiring care over time. Age, however, has not been found to be
associated with relationship quality (Walker & Allen, 1991).
Health. The health of an elderly parent is associated with the amount of
care given. Some researchers have found that elderly parents' deteriorating
health is related to an increase in helping behaviors by caregiving daughters
(Abel, 1986; Archbold, 1983; Lang & Brody, 1983; Markides & Cooper, 1989;
19
Walker et al., 1992). Walker, Shin, and Bird (1990), however, found no
association between mother's health and intimacy.
Coresidence. The majority of elderly individuals do not live alone. Of those
aged 65 and older, 67% live with a family member, predominately a spouse.
Approximately 13%, most of them widowed, live with other family members, such
as an adult daughter (U.S. Bureau of the Census, 1994). When adult children
and their elderly parents live together, they so do for a variety of reasons, such
as high housing costs, unemployment, and divorce (Aquilino, 1990, Aquilino &
Supple, 1991; Ward & Spitze, 1992). Chappell (1991), Lang and Brody (1983),
and Walker and Pratt (1991) found that coresidence with adult children is
associated with higher rates of assistance from them. Regarding the association
between coresidence and relationship quality, Aquilino and Supple (1991) found
that coresidential parents generally reported having as positive a relationship
with their adult children as noncoresidential parents.
Past Relationship Quality. As stated previously, past relationship history
influences current relationships. Research has shown that parents who had poor
relationships earlier in life are likely to have poor relationships in the present
(Rossi & Rossi, 1990; Whitbeck et al., 1991; Whitbeck et al., 1994). Whitbeck et
al. (1991) found that adult children's past or current relationship with their
parents was not related to giving assistance to their parents.
20
Modeling of Change in Relationship Quality
Past research has suggested either stable levels of perceived relationship
quality (Baruch & Barnett, 1983; Lang & Brody, 1983) or decreasing levels of
perceived relationship quality (Johnson & Catalano, 1983; Scharlach, 1987;
Sheehan & Nuttall, 1988) with caregiving. Walker et al. (1996) found that
increased amount of care given is associated with negative caregiving outcomes;
that is, the more care given, the poorer the outcomes.
Structural equation modeling is used to examine the impact of the amount
of care given on perceived relationship quality over a two-year period (see Figure
2). In accordance with the mothers' chronic rather than acute conditions, it was
predicted that on average, amount of care given would remain stable over time.
Perceived relationship quality too, would remain stable over time, on average. In
accordance with the elaborated wear-and-tear hypothesis, it was predicted that
caregiving daughters would report a deterioration in relationship quality when the
amount of care given increases over time.
Control Variables
Age and deteriorating health are associated with the amount of care given
by caregiving adult daughters (e.g., Lang & Brody, 1983); that is, older persons
and persons in poorer health receive more care. In addition, Chappell (1989) and
21
Walker et al. (1996) found that caregivers give more care to care recipients when
they coreside. Furthermore, age, health, and coresidence have been found to be
correlated. Therefore, these variables were controlled in the model. Because
Whitbeck et al. (1991, 1994) and others (e.g., Caspi & Elder, 1988; Rossi &
Rossi, 1990) found an association between past and present relationship quality,
it also was important to control for prior relationship quality in examining the
impact of amount of care given on relationship quality.
Relationship
Quality (Wave 1)
12
Figure 3. Measurement Model of Latent Constructs
23
Hypotheses
Drawing specifically on the elaborated wear-and-tear hypothesis and
findings from the literature, the following hypotheses were derived:
1. On average, amount of care given will remain stable over
time.
2. On average, relationship quality will remain stable over time.
3. The initial amount of care given in Wave 1 will be negatively
related to the initial level of relationship quality in Wave 1;
that is, respondents who initially give higher levels of care will
have lower initial levels of relationship quality.
4. For any increase in amount of care over time, there will be a
corresponding decrease in relationship quality.
5. Mothers' age will be positively associated with the initial
amount of care given.
6. Daughters' perception of mothers' health will be negatively
related to the initial amount of care given; that is, the better
the perceived health, the less care given.
7. Daughters who reside with their mothers will give more care
than daughters who live separately from their mothers.
8. Daughters' perceived past relationship quality with their
mothers will be positively related to their perceptions of
relationship quality at Wave 1.
24
CHAPTER THREE
Method
Sample
The sample consists of 140 (136 White, 2 African American, 1 American
Indian, 1 Other) caregiving daughters, who, with their mothers, participated in a
two-year longitudinal study on mother-daughter relationships in middle and later
life between 1986 and 1989 (see Walker et al., 1996). This sample was drawn
from a larger sample of 173 elderly mother-caregiving adult daughter pairs from
Western Oregon. Participants were recruited primarily through local newspaper
articles describing a study of mother-daughter relationships wherein caregiving
sometimes occurred. To participate in the study, the following criteria had to be
met: (a) mothers were aged 65 and older and unmarried (i.e., widowed, divorced,
or deserted); (b) mothers had no cognitive impairment as reported by daughters
or by project interviewers; (c) daughters lived with or within 45 miles of their
mothers; (d) dependent mothers needed assistance in one or more of the
following areas: transportation, housekeeping, meal preparation, laundry,
personal care, or financial aid; and (e) daughters reported providing more than
half of the assistance needed by their dependent mothers.
The mean age of caregiving daughters was 52.7 years (Mdn = 53.0 years,
SD = 10.3) with ages ranging from 29 to 73 years at Wave 1. The majority
(62.9%) were married (n = 88) although 21.4% (n = 30) were divorced. The
25
average length of marriage was 21.7 years (Mdn = 19.0 years, SD = 15.2).
Approximately 24.3% completed high school while the majority (75.7%) had
some college. The mean years of education was 14.1 (Med. = 14.0 years, SD =
2.2). (See Table 1). The median annual family income was $28,000 (M =
$30,181, SD = $20,458) at Wave 1 and 59.3% (n = 83) reported being
employed.
Caregiving daughters reported giving care to their mothers for a median of
5 years (M = 7.3 years, SD = 7.1) with a range from 0 months to 51 years. The
estimate of 0 months came from daughters who did not describe their activities
as caregiving, even though they met the study criteria. The estimate of 51 years
of caregiving came from a daughter who felt she had been taking care of her
mother since her father had abandoned the family years earlier. Regarding living
arrangements, 80% (n = 112) of the caregiving daughters did not live with their
mothers while 20% (n = 28) did. Mothers' median age was 84 years (M = 81.5,
SD = 8.8) and ranged from 64 years to 101 years. Daughters reported that their
mothers had, on average, 4.1 (Mdn = 4.0, SD = 1.8) of 14 health problems
common among the elderly. The four most common health problems
experienced by their mothers, as reported by caregiving daughters, were arthritis
(n = 94, 67.1%), loss of vision (n = 90, 64.3%), loss of hearing (n = 55.0%), and
high blood pressure (n = 53, 37.9%). Despite the chronic health problems,
daughters perceived their mothers to be in good health (M = 2.8, SD = .9).
26
Table 1: Selected Sociodemooraphic Characteristics of Caregiving Daughters
Characteristic
n
%
18
13.0
23.7
34.5
23.7
5.0
Age
29-39
40-49
50-59
60-69
70-79
33
48
33
7
Marital Status
Never married
Married
52
88
37.1
8
5.7
24.3
32.9
25.7
11.4
62.9
Education
< High school
High school
Some college
4 years college
> 4 years college
34
46
36
16
Income
< $10,000
$10,000-$19,999
$20,000-$29,999
$30,000-$39,999
$40,000 and
above
15
32
36
24
29
11.0
23.5
26.5
17.7
21.3
57
83
40.7
59.3
Employment Status
Not employed
Employed
27
Table 1 (Continued)
Characteristic
n
ok
Coresidence
Lived with mother
Other
28
112
20.0
80.0
44
31.4
22.1
12.9
21.4
Years in marital
status
0-10
11-20
21-30
31-40
> 41
31
18
30
17
12.1
Parental Status
Mothers
No children
12
91.4
8.6
106
75.7
127
Children Under 18
0
1
2
3
4
5
17
10
5
1
1
Note. Cases with missing data are excluded.
12.1
7.1
3.6
.7
.7
28
Attrition
Of the original 173 mother-daughter pairs, 33 daughters were lost to the
study over two years resulting in a total sample of 140. The 33 who dropped out
of the study did so for the following reasons: (a) mother's death ( = 19, 57.6%),
(b) lack of interest (n = 9, 27.3%), (c) mother's physical illness (n = 3, 6.1%),
daughter's cognitive decline (n = 1, 3.0%) and daughter's physical illness (n = 1,
3.0%). To determine if differences existed between these women and those who
remained in the sample over the two waves, I compared them on the following
Wave 1 measures: marital status, education, income, employment status,
number of children under age 18, and all the variables in the model. As shown in
Table 2, the two groups did not differ significantly on marital status (married or
not married), education, employment status (employed or not employed),
number of children under age 18, mothers' age, coresidence, past relationship
quality, or on the indicators that make up the relationship quality construct (i.e.,
closeness, good relationship quality, attachment, and intimacy). Daughters'
perceptions of their mother's health, however, did differ between the two groups
such that daughters who dropped out of the study reported their mothers being
in poorer health, x2 (3, n = 173) = 17.97, p. = .001. The two groups also differed
in the amount of care given as measured by their IADL and ADL scores.
Daughters who were lost to the study gave more IADL and ADL care to their
mothers than those who remained. Additionally, daughters who stayed in the
study had more income than those who left the study (See Table 2).
29
Table 2: Comparisons of Daughters in the Sample with Those Lost to Attrition
Study Sample
Variable
Respondents lost due
to attrition
M or %
SD or n
M or %
SD or n
Daughters' age
52.68
11.17
51.30
10.26
.50
Mothers' age
81.48
8.79
79.69
1.55
.24
Education
14.11
2.28
13.42
2.22
.11
% Married
62.9
140
45.5
33
.37
Years married
21.74
15.22
22.24
14.40
.86
$30.18
$20.46
$22.26
$12.98
.01
33
.87
Income in thousands
% Employed
No. of Children < 18
59.3
140
63.6
2
.44
.92
.30
.64
.87
IADL help
2.17
.71
2.60
.78
.003
ADL help
1.40
.72
1.88
1.07
.02
Attachment
3.27
.70
3.25
.87
.91
Intimacy
4.04
.67
3.92
.83
.40
Daughters' perception
of mothers' health
2.82
.88
2.09
1.01
Closeness
3.46
.60
3.42
.66
.78
Good relationship
quality
3.49
.62
3.25
.80
.13
Past relationship
quality
3.12
.76
3.18
.85
.71
.001
Note. n = 140 in the study sample; n = 33 respondents lost to attrition.
30
Procedures
The data were drawn from a preexisting data set in which data were
collected during face-to-face interviews with daughters conducted primarily in
their homes by trained graduate research assistants and the principal
investigator. Interviews lasted approximately 1.5 hours. Daughters received $20
in Wave 1 and $15 in Wave 2 to thank them for their participation. Daughters
were asked a series of questions about background characteristics, caregiving
tasks and activities, mother's health, coresidence, and relationship quality.
Respondents also completed paper-and-pencil measures assessing perceived
relationship quality.
To help maintain interest in the study, daughters and mothers received
birthday cards each year, and a general report about the study was distributed to
respondents at the end of the second wave of data collection. At the end of each
interview, the interviewers asked respondents if they would be willing to
participate in the study in the following year. Before the next interview was
scheduled, respondents were sent a postcard indicating that someone would call
them to set up a time for the next interview. When possible, respondents had the
same interviewer across waves.
31
Measures
Two latent variables were of central interest in this study: amount of care
given and perceived relationship quality. Four additional variables, coresidence,
age, health, and past relationship quality were included as controls.
Care given.
Daughters were asked about the specific tasks with which they helped
their mothers in seven IADL categories: (a) shopping/errands, (b) indoor
maintenance, (c) financial tasks, (d) food preparation/clean up, (e) financial aid,
(f) outdoor maintenance, and (g) bureaucratic mediation. In addition, they were
asked about personal care (i.e., bathing, dressing) or ADLs. For each activity in
every category, daughters were asked: "Does your mother do
herself or do you help her to do it?" For each category, responses were coded:
(1) no help at all. (2) help with half or less than half of the activities in the
category, (3) help with more than half of the activities in the category, or (4) help
with all of the activities in the category. Categories not applicable (e.g., outdoor
maintenance for mothers who lived in retirement homes or apartments) were
coded (5). A mean level of IADL care given was calculated by adding across the
seven categories any score from 1 to 4, and dividing by the total number of
categories in which scores from 1 to 4 were obtained. The level of ADL care
given was the code given for that category (i.e., 1, 2, 3, or 4) (Appendix A).
32
Perceived relationship quality.
Perceived relationship quality consisted of four measures: (a) daughter's
intimacy score, (b) daughter's attachment score, (c) daughter's perception of
overall relationship quality with her mother, and (d) daughter's perception of the
closeness or distance of the present relationship with her mother. The intimacy
scale (Walker & Thompson, 1983), reflecting generalized affection for the
intergenerational partner, consisted of 17 items. Examples of the items are 'We
want to spend time together," "She shows that she loves me," "Our lives are
better because of each other," and "We love each other." Daughters rated each
item from (1) not true to (5) always true. A mean score was computed for each
participant with higher scores indicating greater intimacy. Cronbach's alphas for
this measure have ranged from .91 to .97 across mother-daughter relationships
(Walker & Thompson, 1983) (Appendix B).
The attachment scale (Thompson & Walker, 1984) consisted of nine items
that ranged from (1) not true to (5) always true. Examples of the items include:
'We are dependent on each other," 'We anticipate each other's moods," "Our
best times are with each other," and "We are emotionally dependent on each
other." A mean score was computed for each participant with higher scores
indicating greater attachment. Cronbach's alphas have ranged from .86 to .91
across relationship reports (Thompson & Walker, 1984) (Appendix C).
Daughters were asked the following question to describe their current
relationship with their mother: "How would you describe your relationship with
33
fMother's name] at the present time?" Daughters responded from (1) poor to (4)
excellent. In addition, daughters were asked the following question: "How
(emotionally) distant or close would you say you are?" Responses ranged from
(1) very distant to (4) very close.
Control variables.
Mothers' age was measured in years. Coresidence was measured by
asking the daughter the following question: "What is your current living
arrangement?" Coresidence was dichotomized into two categories (0) other
(e.g., live alone, live with other relatives but not mother) or (1) live with mother.
Daughters' perception of mothers' health was measured by the following
question: "In general, how would you describe your mother's health right now?"
(1) poor, (2) fair, (3) good, or (4) excellent. Daughter's perception of past
relationship quality was measured by the following question: "How would you
describe your relationship with your mother in the past?" (1) poor, (2) fair, (3)
good, or (4) excellent.
34
Data Analysis
Data on the 140 participants were obtained from the measures described
above. The SAS statistical program was used for descriptive analyses.
Descriptive statistics such as the mean, median, standard deviation, range,
frequency, and percent were generated. Intercorrelations of independent and
dependent variables were computed for caregiving daughters.
The method used to analyze change in the amount of care given by
daughters to care-receiving mothers and change in perceived relationship quality
from Wave 1 to Wave 2 was the LISREL (Linear Structural Relations) computing
program (Version 8.12a; J6reskog & Sorbom, 1993). All analyses were
conducted using maximum likelihood estimation on covariance matrices,
although the major emphasis was on parameter estimates for the completely
standardized solution.
Structural equation modeling (SEM) takes into account multiple indicators
allowing for more valid measurements of latent constructs (i.e., amount of care
given and relationship quality) (Bent ler, 1988). As Byrne (1994) and others
(Bent ler, 1988; Bollen, 1989; JOreskog & SOrbom, 1993) noted, SEM allows for
testing and modeling visually the adequacy of fit of a theoretical model to data by
using a covariance matrix (see Bent ler, 1988; Joreskog & SOrbom, 1993 for
further explanations). SEM consists of two conceptually distinct submodels: a
measurement model that deals with the relationship between observed
measures and their underlying latent constructs, and the structural model that
35
specifies the relationship between latent and observed variables that are not
indicators of the latent construct (see Bent ler, 1988; Hoyle, 1995 for a review).
The model, which included both the measurement and structural models,
is presented in Figure 2. There were four exogenous variables (mothers' age,
daughters' perception of mothers' health, coresidence, and past relationship
quality) and four endogenous variables (amount of care given and relationship
quality measured at Waves 1 and 2). In structural equation terminology, the four
exogenous variables are called Ksi's, (4)'s and the four endogenous variables
are called Eta's, (Ti)'s. The structural parameters from the exogenous variables
to the endogenous variables are called the Gamma's, (y)'s. The structural
parameters from one endogenous construct to the next are called Beta's, (3)'s.
The Zeta (0 coefficients are the unexplained variance for each endogenous
latent variable. The measurement model includes paths from the latent
constructs to the measurement indicators. These are called Lamda's, (A)'s. The
Epsilon's, (E)'s represent the error (residuals) for the indicators of the latent
endogenous variables.
There are several advantages to using structural equation modeling in the
present study. First, SEM allows for the joint specification and estimation of the
hypothesized measurement and structural models to account for the observed
data (Long, 1983). Second, SEM takes into account measurement errors and
uses multiple indicators to improve the construct validity of measurements
(Bent ler, 1980; Hayuduk, 1987). Third, SEM allows for "restricted" models that
36
include systematic constraints on relationships among theoretical constructs and
between observed variables and theoretical constructs. A key implication of such
restrictions is that models can be tested that include only relationships that are
hypothesized a priori. Fourth, SEM provides a systematic basis for evaluating the
"fit" of the hypothesized model to the data. Thus, SEM allows for theory
development and construct validation (see Anderson & Gerbing, 1988 for further
explanation).
Missing data
Missing data are often a major problem in longitudinal research. Standard
analysis of missing data such as pairwise deletion, listwise deletion, mean
substitution, and imputation assume that data are missing at random (Little &
Rubin, 1987). Due to sample bias, inefficient estimates, and discarding of
potentially useful data, these analyses often produce biased results (Muthen,
Kaplan, & Hollis, 1987). Structural equation modeling was applied to test whether
missing data were problematic.
Of the 140 sample daughters, 7 have missing data on at least one of the
variables in the model. The Expectation Maximization (EM) Algorithm was used
to account for missing data patterns (Dempster, Laird, & Rubin, 1977; Little &
Rubin, 1987). EM Algorithm uses multiple-regression-based imputation to predict
the missing variable using all other relevant variables. EM Algorithm assigns the
37
missing variables an imputed value that includes the error component (see
Dempster et al., 1977; Little & Rubin, 1987 for further explanation).
38
Chapter Four
Results
This study examined the effect of the amount of care given on the quality
of mother-daughter relationships in middle and later life. Using SEM
methodology, a model describing mother-daughter relationships was empirically
tested. This chapter is organized into three sections. First, univariate descriptive
statistics as well as correlations among observed variables are presented.
Second, results of the structural equation modeling analyses are presented.
Finally, the results are discussed in relation to the hypotheses presented in
Chapter 2.
Preliminary Data Analyses
Means, standard deviations, and correlations for all observed variables
are presented in Table 3. An inspection of Table 3 shows that the two observed
measurement indicators, IADL and ADL, for the latent construct of amount of
care given, were highly correlated at Wave 1 (r = .67, p < .001) and Wave 2 (r =
.74, p. < .001). The correlations of the four observed measurement indicators,
closeness, good relationship quality, attachment, and intimacy, for the latent
construct of relationship quality, were all positively related to each other.
Table 3: Correlations Among Observed Indicators and Exogenous Variables
Indicators
V1
1. IADL1
2. ADL1
3. Close1
4. Good1
1.00
.67
.02
-.07
.16
.00
.77
.60
.00
-.12
.04
-.04
.22
-.27
5. Attachment)
6. Intimacy1
7. IADL2
8. ADL2
9. Close2
10. Good2
11. Attachment2
12. Intimacy2
13. MAGE
14. Health
15. Coresidence
16. Past Rel.Q1ty.
MEANS
SD
V2
1.00
-.05
-.09
.12
.04
.53
.51
-.07
-.14
-.02
-.10
.22
V3
V5
V6
V7
V8
V9
V10
1.00
.65
.59
.68
1.00
.46
.68
V13
V14
V15
V16
.31
-.11
.43
1.00
.23
.08
.15
1.00
.00
.09
1.00
-.06
1.00
3.06
.72
4.01
.69
81.48
8.79
2.82
.88
.20
.40
3.12
.76
V11
V12
1.00
.76
-.06
-.08
-.04
1.00
-.01
.00
1.00
.61
.45
.47
.06
-.02
.54
.39
.48
.45
-.03
-.02
.31
-.30
.24
.05
.02
-.21
.28
2.17
1.40
.72
3.46
.60
.71
V4
1.00
.44
.64
-.12
-.09
.56
.45
.47
.61
.04
.07
-.12
.39
3.49
.62
1.00
.58
.14
.09
.54
.28
.75
.56
-.05
-.15
-.04
.36
3.27
.70
1.00
.00
.01
.64
.52
.54
.74
.04
-.05
-.15
.36
4.04
.67
1.00
.74
.00
-.10
-.05
-.12
.27
-.29
.15
.02
1.00
.00
-.10
-.10
-.14
.22
-.22
.13
-.01
2.53
1.68
.81
.91
-.10
-.02
-.20
.37
3.44
.64
.01
.03
-.17
.38
3.45
.72
Note. IADL1, (ADL2 - Instrumental Activities of Daily Living in waves 1 & 2; ADL1, ADL2 - Activities of Daily Living in Waves 1 & 2; Closel, Close2
Closeness in Waves 1 & 2; Good1, Good2 - Good relationship quality in Waves 1 & 2; MAGE - Mothers' age; HEALTH - Daughters' perception of
mothers' health.
40
The correlations ranged from .44 to .61 at Wave 1 and .46 to .76 at Wave 2,
respectively. All correlations were statistically significant.
Table 3 also provides the correlations between exogenous variables
(mothers' age, daughters' perceptions of mothers' health, coresidence, and past
relationship quality). The correlations ranged from -.06 to .23. Results indicated
that mother's age and daughter's perception of mother's health were positively
associated with each other (r = .23, p < .01). Mother's age was not associated
with past relationship quality or coresidence. Finally, coresidence was not
associated with past relationship quality or daughter's perception of mother's
health.
Structural Equation Model
A two-step process was used in analyzing the structural model (See
Figure 2). Step 1 involved assessing the adequacy of the measurement model
and step 2 verified the structural model. The two-step process has the
advantage of minimizing the interpretational confounding of both the structural
and measurement model covariances (Anderson & Gerbing, 1988).
41
Measurement Model
A preliminary confirmatory factor analysis was conducted to test the
measurement indicators of the four latent endogenous constructsamount of
care given, measured at Waves 1 and 2, and relationship quality, measured at
Waves 1 and 2. A structural equation modeling program, LISREL8.12a
(Joreskog & Sombom,1993), was employed using the maximum likelihood (ML)
estimation method. The ML estimation assumes that the data being analyzed
have multivariate normal distribution. When the data are not normally distributed,
the weighted least squares (WLS) estimation method is usually recommended.
Although the WLS estimation method provides unbiased estimates (e.g., chisquare, standard errors), it requires a very large sample (see Bollen, 1989;
Joreskog & Sombom,1993). The sample size used in this study precludes the
use of WLS.
Two observed measurement indicators (IADL and ADL) were loaded on
the latent construct, amount of care given. Four observed measurement
indicators (closeness, good relationship quality, intimacy, and attachment) were
loaded on the latent construct, relationship quality. Because the same items
were used at both waves (i.e., closeness, Wave 1 and closeness, Wave 2),
corresponding error terms were allowed to be correlated. These correlations are
not shown in Figure 2 for simplicity. Finally, the four latent variables were allowed
to be correlated in confirmatory factor analyses (see Figure 3).
ci
Mothers'
Age
E2
IADL
ADL
W1
y11
kl
Mothers'
Health
IADL
W2
W1
X2
C3
1331
Amount of Care
Coresidence
Amount of Care
(Wave 2)
(Wave 1)
E3
rI3
71
[343
c4
Past Relation.
Quality
y 24
Relationship
Quality (Wave 2)
Relationship
Quality (Wave 1)
r14
112
19
x3
W1
63
'Attach
I nt.
W1
W1
64
e6
rial ood I 17 IW2
Frri
W2
I
69
(G W2
I
610
W2
I
61 I
Figure 2. Structural and Measurement Models
Note. Correlated error terms among observed measurement indicators are not shown
in order to simplify the presentation.
1
el2
43
The initial analysis of the measurement model yielded a statistically
significant chi-square, x2 (42, N = 140) = 96.26, p < .001. Contrary to the
conventional chi-square test, a nonsignificant chi-square indicates a good fit of
the model to the data. The chi-square statistic, however, has been criticized for
its practical use in model evaluation due to its dependency on large sample sizes
and its sensitivity to multivariate normality violation (Bent ler & Bonett, 1980).
Other measures of fit indices such as the Non-Normed Fit Index (NNFI) = .92,
the Comparative Fit Index (CFI) = .95 and the Incremental Fit Index (IFI) = .95,
indicated an adequate fit of the model to the data. While this model does not fit
the data perfectly, the fit indices were all above the recommended .90 minimum
criterion (JOreskog & Sorbom, 1993). The root mean square error of
approximation (RMSEA) is another goodness-of-fit index that corrects for
degrees of freedom. The RMSEA for the measurement model is .10. Although
the RMSEA exceeds the recommended level of .08 (Browne & Cudeck, 1993),
inspection of the loadings for the measurement model indicated that all loadings
were statistically significant and strong, ranging from .63 to .94 (completely
standardized estimates) (See Table 4). The results indicated that all observed
variables were good indicators of their underlying constructs.
44
Table 4: Loadings of Measurement Model
Parameter
estimates
Unstandardized
estimates
Completely
standardized
estimates
t-ratio
IADL, Wave 1<== Amount of care given,
Wave 1
.56
.79
6.37
ADL, Wave 1<== Amount of care given,
Wave 1
.62
.86
6.02
Closeness, Wave 1<== Relationship
quality, Wave 1
.38
.63
7.78
Good relationship quality, Wave 1
<== Relationship quality, Wave 1
.47
.77
10.01
Attachment, Wave 1<==
Relationship quality, Wave 1
.48
.67
8.75
Intimacy, Wave 1<== Relationship
Quality, Wave 1
.56
.78
11.48
IADL, Wave 2<== Amout of care given,
Wave 2
.64
.94
6.52
ADL, Wave 2<== Amout of care given,
Wave 2
.89
.75
7.10
Closeness, Wave 2<== Relationship
quality, Wave 2
.48
.73
10.09
Good relationship quality, Wave 2
<== Relationship quality, Wave 2
.52
.73
9.58
Attachment, Wave 2<== Relationship
quality, Wave 2
.57
.79
11.06
Intimacy Wave 2,<== Relationship
quality, Wave 2
.65
.93
14.03
Note. IADL Instrumental Activities of Daily Living; ADL - Activities of Daily
Living.
t-ratios above 3.21 are significant at p < .001
45
Structural Model
Table 5a shows the goodness of fit indices of the structural models. The
initial analysis of the structural model, Model 1a structural model where
corresponding error terms were allowed to covary because the same items were
used in both waves (i.e., IADL, Wave 1 and IADL, Wave 2), yielded a statistically
significant chi-square, x2 (88, N = 140) = 146.80, 2 < .001; NNFI = .93; CFI = .95;
IFI = .95; RMSEA = .07. Upon inspection of the residuals and the LISREL
modification indices (Joreskog & SOrbom, 1993), however, the hypothesized
model was respecified to include correlations between additional errors as
specified below.
Model 2 allowed closeness and good relationship quality errors to covary
within each wave because the two items contain similar wordings. The resulting
model yielded the following fit: x2 (86, N = 140) = 122.18, 2 < .01; NNFI = .96;
CFI = .97; IFI = .97; RMSEA = .06. The final model, Model 3, allowed attachment
and intimacy error terms to covary within each wave at both time periods due to
the conceptual similarity between the two scales. Model 3 resulted in the
following fit: x2 (84, N = 140) = 114.11, 2 < .05; NNFI = .96; CFI = .97; IFI = .98;
RMSEA = .05 (See Table 5a). With these respecifications, the structural models
(Models 1 and 2) were improved, showing a better fit to the data. As shown in
Table 5b, Model 2 was significantly better than Model 1 and Model 3 was
significantly better than models 1 and 2.
Table 5a: Goodness-of-Fit Indices of Structural Models
Model
Equality Constraints
df
NNFI
/2
CFI
IFI
RMSEA
Model 1
None
88
146.80
.00
.93
.95
.95
.07
Model 2
Closet error term
correlated to Good1,
Close2 error term
correlated to Good2
86
122.18
.01
.96
.97
.97
.06
Model 3
Model 2 plus correlated
attachment) and
intimacy) error terms,
attachment2 error term
correlated intimacy2
84
114.11
.02
.96
.97
.98
.05
Model 4
Model 3 with equality
constraints
88
118.76
.02
.96
.97
.97
.05
Note. NNFI Non-Normed Fit Index; CFI Comparative Fit Index; IFI Incremental Fit Index; RMSEA Root Mean
Square of Approximation. Close 1, 2 -Closeness in Waves 1 & 2; Good 1, 2 Good relationship quality in Waves 1 & 2;
attach - Attachment in Waves 1 & 2; int. Intimacy in Waves 1 & 2.
47
Table 5b: Model Comparisons
Zdiff
Model
df if
Model 3 - Model 1
32.69***
2
Model 2
24.62***
2
Model 3 - Model 2
8.07*
2
Model 4 - Model 3
4.65
4
*p<.05
Model 1
< .001
48
A fourth model, Model 4, was constructed to test invariance of the measurement
indicators of the latent endogenous variables across time as detailed below.
Measurement Invariance
A concern that relates to longitudinal studies is whether constructs
measured across waves are equivalent (Byrne, Shavelson, & Muthen, 1989;
Pentz & Chou, 1994). Testing for measurement invariance addresses whether a
particular measurement instrument operates equivalently across different
populations (e.g., gender, age, race) and/or on the same population
longitudinally. It concerns primarily the invariance of regression intercepts, factor
loadings (regression slopes), and error/uniqueness variances. Joreskog (1971)
developed a procedure that enabled researchers to test for invariance
simultaneously across groups and waves.
Conventionally, tests of invariance begin with a baseline model. This
model represents the most parsimonious as well as substantively meaningful
model specification (Byrne, 1994). In the present study, Model 3 was chosen for
this purpose as the most parsimonious and meaningful model.
In testing for invariance, each measurement indicator for the latent
construct was tested for equivalence across waves. A fourth model (Model 4)
tested the loadings of the observed indicators on the endogenous latent
variables. Equality constraints were imposed on loadings in Model 4 so that each
49
observed indicator at Wave 1 was equal to the corresponding indicator at Wave
2. For example, the loadings of IADL at Wave 1 and IADL at Wave 2 were
forced to be equal. Notice that loadings to IADL and intimacy were set to one for
scaling purposes in Waves 1 and 2.
As shown in Table 5a, although the overall chi-square for Model 4 was
significant x2(88, N = 140) = 118.76, p < .05, practical fit indices suggested an
adequate fit to the data; NNFI = .96; CFI = .97; IFI = .97; RMSEA = .05.
Comparing Model 4 to Model 3, the chi-square difference test, x2d,ff = 4.65, (2 =
n.s.), was not statistically significant, indicating that constraints imposed on
Model 4 were reasonable (See Table 5b). On the basis of this result, it was
concluded that the indicators measuring the latent variables were invariant over
time. Parameter estimates from Model 4 are used to present the results below.
Loadings on the latent variables.
Table 6 shows the loadings of the indicators on the four latent variables of
Model 4. Inspection of the loadings indicated that all loadings were statistically
significant (2 < .001) and substantial, ranging from .62 to .91 (completely
standardized estimates). The completely standardized loadings for IADL and
ADL were .87 and .80 in Wave 1 and .91 and .79 in Wave 2. The completely
standardized loadings for indicators of the latent construct, relationship quality,
ranged from .62 to .82 in Wave 1 and .73 to .91 in Wave 2. The most prominent
50
Table 6: Loadings for Structural Model 4
Parameter
estimates
Unstandardized
estimates
Completely
standardized
estimates
t-ratio
IADL, Wave 1<== Amount of care given,
Wave 1
1.00®
.87
ADL, Wave 1<== Amount of care given,
Wave 1
.99
.80
8.92
Closeness, Wave 1<== Relationship
quality, Wave 1
.72
.62
11.25
Good relationship quality, Wave 1
<== Relationship quality, Wave 1
.84
.73
12.82
Attachment, Wave 1<==
Relationship quality, Wave 1
.84
.64
12.04
Intimacy, Wave 1<== Relationship
Quality, Wave 1
1.00ED
.82
IADL, Wave 2<== Amout of care given,
Wave 2
1.00®
.91
ADL, Wave 2<== Amout of care given,
Wave 2
.99
.79
8.92
Closeness, Wave 2<== Relationship
quality, Wave 2
.72
.75
11.25
Good relationship quality, Wave 2
<== Relationship quality, Wave 2
.84
.73
12.82
Attachment, Wave 2<== Relationship
quality, Wave 2
.84
.74
12.04
1.00®
.91
Intimacy, Wave 2,<== Relationship
quality, Wave 2
Note. IADL - Instrumental Activities of Daily Living; ADL - Activities of Daily
Living.
t-ratios above 3.21 are significant at p. < .001
ED Parameters set to 1.0
51
indicator for measuring relationship quality was intimacy (.82 Wave 1 and .91 at
Wave 2) followed by good relationship (.73 Wave 1, .73 Wave 2), attachment
(.64, Wave 1, .74, Wave 2), and closeness (.62 Wave 1, .75 Wave 2).
Path coefficients.
The maximum-likelihood estimates and associated t-values for the
structural parameter from Model 4 are shown in Table 7. Examination of the
hypothesized completely standardized paths in the model indicated that
hypothesis 1 was supported, showing that the latent construct, amount of care
given, is stable over time, (completely standardized estimate,
= .81, t = 11.48,
< .001). Results also showed that perceived relationship quality was stable
over time (completely standardized estimate, 0 = .93,1 = 13.45, 2 < .001), thus
supporting hypothesis 2.
The next two hypotheses, H3 and H4, dealt with the relationship between
amount of care given and relationship quality at each wave. H3 was not
supported (completely standardized estimate,
= -.01, t = -.08, p = n.s.) in that
amount of care given was not associated with relationship quality at Wave 1. H4,
however, was supported. As predicted, as the amount of care given increased,
perceived relationship quality decreased (completely standardized estimate, 13 = .16,
= -2.67,
< .01).
52
Table 7: Parameter Estimates of the Hypothesized Model (Beta)
Parameter
estimates
Amount of care given, Wave 1
==> Amount of care given, Wave 2
Unstandardized
estimates
Completely
standardized
estimates
t-ratio
.99
.81
11.48
Amount of care given, Wave 1
==>Relationship quality, Wave 1
-.01
-.01
-.08
Relationship quality, Wave 1
==> Relationship quality, Wave 2
1.11
.93
13.45
Relationship quality, Wave 2
==> Amount of care given, Wave 2
-.14
-.16
-2.67
Note. t-ratios above 2.58 and 3.21 are significant at p < .01 and p < .001
respectively.
53
Effects of the exogenous variables.
The last four hypotheses dealt with the relationships between the
exogenous variables (i.e., mothers' age daughters' perception of mothers' health,
coresidence, and past relationship quality) and the latent endogenous variables,
amount of care given and relationship quality at Wave 1. Table 8 presents the
maximum-likelihood estimates (both unstandardized and completely
standardized) and associated t-values. Results indicate that H5was supported.
Mothers' age was positively associated with the initial amount of care given at
Wave 1. Mothers who were older reported needing more care than their younger
counterpart (completely standardized estimate, y = .35, t = 4.26, 2 < .001). H6
was also supported. Daughters who perceived their mothers to be in good health
provided less care in the form of IADLs and ADLs (completely standardized
estimate, y = -.42, t = -5.04, 2 < .001). Additionally, hypothesis H7 was also
supported. Daughters who lived with their mothers provided more care than
those who did not (completely standardized estimate, y =.30., t = 3.70, p < .001).
Finally, support was also observed for H8, which showed that perceived past
relationship quality is positively related to perceived current relationship quality
(completely standardized estimate, y = .51, t = 6.10, 2 < .001). Taken together,
results from these analyses offered empirical support for the hypothesized
structural relationships specified in Model 4.
54
Table 8: Parameter Estimates of the Exogenous Variables to Latent Variables
Parameter
estimates
Unstandardized
estimates
Completely
standardized
estimates
t-ratio
Mothers', age, Wave 1==> Amount of care
given, Wave 1
.02
.35
4.26
Daughters' perception of mothers' health,
Wave 1==> Amount of care given, Wave 1
-.29
-.42
-5.04
Coresidence, Wave 1==> Amount of care
given, Wave 1
.44
.30
3.70
Past relationship quality, Wave 1==>
Relationship quality, Wave 1
.34
.51
6.10
Note. t-ratios above 3.21 are significant at p < .001.
55
Chapter Five
Discussion and Conclusion
The purpose of this study was to examine longitudinally the effect
of the amount of care given on daughters' perceptions of their relationship
with their care-receiving mothers. Specifically, this study used data from
140 caregiving daughters to test the elaborated wear-and-tear hypothesis,
which proposes that daughters' perception of relationship quality with their
mothers will deteriorate as the amount of care given increases.
Caregiving daughters in the present study reported helping their
mothers with IADLs such as shopping and errands, indoor maintenance,
and bureaucratic mediation. Nearly one in five gave help with ADLs such
as bathing and dressing. The need for help with these activities is
characteristic among the elderly population with similar needs (Stone et
al., 1987). Additionally, the rate of care provided for ADLs is comparable
to the rates of care for women over age 75 (Spitze & Logan, 1989).
Caregiving daughters, on average, reported high attachment and intimacy
scores and good and close relationships with their care-receiving mothers.
56
Primary Findings
Stability of the endogenous latent variables.
The first two hypotheses dealt with the stability of the two
endogenous latent constructs, amount of care given and relationship
quality. H1 and H2 predicted that amount of care given and relationship
quality would remain stable over time. The results supported both
hypotheses. The amount of care given by caregiving daughters to their
physically dependent elderly mothers ramained stable from Wave 1 to
Wave 2. This is not surprising given that these mothers had chronic rather
than acute health conditions, such as arthritis and sensory impairments,
which are common among elderly persons. Such conditions tend to
change only gradually over time. Although the needs of care-receiving
elderly mothers may gradually increase, the finding that amount of care
given is stable over two years suggests that caregiving daughters have
time to adapt and adjust to the stresses of giving care (Stephens & Zarit,
1989).
That relationship quality is stable over time also is not surprising.
Others have found that adult children who perceived their relationships
with their parents as being warm, intimate, and close in early life, tend to
report a feeling of intimacy and closeness in mid- and later life as well,
57
suggesting that such relationships are stable (Rossi & Rossi, 1990;
Whitbeck et al., 1994). From a life span perspective, (Kahn & Antonucci,
1980), mother-daughter relationships in mid- and later life are a
continuation of early relationships, and they reflect past intimate
experiences and histories. Perhaps relationship stability is due also to the
socialization process that encourages parents and children to keep
connected to each other throughout the life course (Atkinson, 1989). For
example, women, especially mothers and daughters, have been
socialized to be active participants in maintaining kin relationships.
Another reason may be due to the high attachment levels between
mothers and daughters. Research has shown that mothers and daughters
in mid- and later life show high levels of attachment by living in close
proximity to each other, communicating with each other by telephone or
letters, exchanging periodic visits, and sharing common experiences that
lead to increased emotional intimacy (Cicirelli, 1983; Lawton et al., 1994;
Umberson, 1992). Finally, most daughters (n = 127, 91%) were mothers.
Hess and Waring (1978) argued that stability in relationship quality may
be due to the common role experience daughters shared with their
mothers by becoming parents themselves (Walker, Thompson, & Morgan,
1987).
58
Relationship between amount of care given and relationship quality.
Hypotheses 3 and 4 dealt with the relationship between amount of
care given and relationship quality at each wave. The literature has shown
that high amounts of care given may result in detrimental effects for
caregiving daughters, (i.e., Lang & Brody; 1983; Townsend et al., 1989).
Daughters had been providing care to their dependent elderly mothers
prior to the onset of the study. H3 predicted that care given in Wave 1
would be negatively related to perceived relationship quality in Wave 1.
This hypothesis was not supported. No association between amount of
care given at Wave 1 and relationship quality at Wave 1 was found. The
amount of care provided by caregiving daughters was minimal in Wave 1.
For example, approximately two-thirds of the caregiving daughters
reported minimal or no care for the following IADLs in Wave 1: financial
management, food preparation, outdoor maintenance, and financial
support. The need for personal care was also reported to be minimal or
nonexistent for the majority of daughters. Perhaps the type of care
provided by daughters (i.e., shopping, indoor maintenance, bureaucratic
mediation) is not particularly stressful.
H4, however, was supported. H4 predicted that over time, amount
of care given would be negatively related to relationship quality.
Daughters who provided increased care in the form of IADLs and ADLs to
59
their dependent mothers from Wave 1 to Wave 2 showed a decline in
relationship quality in Wave 2. In other words, relationship quality declined
over time in response to an increased amount of care given. This finding
is consistent with the elaborated wear-and-tear hypothesis in that
providing more care erodes relationship quality over time even though it
does not do so initially (Pear lin et al., 1990; Scharlach, 1987; Sheehan &
Nuttall, 1988; Townsend et al., 1989; Walker et al., 1996). As mothers
need more care, perhaps they are less able to contribute to the
relationship, thus leading to declines in relationship quality (Dowd, 1975).
It is important to note that although relationship quality decreased
slightly, the means of the observed relationship quality indicators showed
that caregiving daughters, on average, reported having good and close
relationships with their care-receiving mothers, relationships that were
characterized by attachment and intimacy. This is not surprising given
other research indicating close relationships throughout adulthood
between mothers and daughters (Baruch & Barnett, 1983). Furthermore,
Abel (1990) found that the mother-daughter bond strengthened as a result
of caregiving; that is, daughters reported feeling very attached and had
strong feelings of affection for their dependent parent.
The high relationship quality scores found in this study may be due
also, in part, to the nature of the sample. As stated previously,
respondents volunteered for this study, probably overrepresenting positive
situations. Additionally, only a small percentage of mothers in this study
60
required assistance with personal care, which may be more stressful to
relationships. Finally, that this study examined daughters whose mothers
had acute or chronic health problems and not any form of dementia may
account for the daughter's high perception of relationship quality. Studies
have shown that caring for someone with Alzheimer's Disease or other
forms of dementia is stressful and burdensome for the caregiver (Birkel,
1987; Schulz & Williamson, 1991; Zarit, Todd, & Zarit, 1986).
Relationship between covariates and endogenous latent variables.
Hypotheses H5 through H8 dealt with the relationships between the
covariates and the endogenous latent variables. As predicted, mothers'
age was positively associated with the initial amount of care given (H5).
This substantiates earlier findings that increasing age is related to poorer
health (Cicirelli, 1981; Hareven, 1995; Kinsella, 1995). The results also
showed that daughters who perceived their mothers to be in better health
provided less care in the form of IADLs and ADLs (H6), thus supporting
previous findings that poorer health is associated with an increase in
helping behaviors by caregiving daughters (Abel, 1986; Archbold, 1983;
Lang & Brody, 1983; Markides & Cooper, 1989; Walker et al., 1992).
61
As stated previously, 20% of the caregiving daughters in this study
lived with their dependent elderly mothers. As H7 predicted, the results
showed that caregiving daughters who lived with their mothers provided
more care than caregiving daughters who lived separately. A coresidential
caregiving daughter may do all of her mother's cooking and laundry, for
example, even if the mother is not dependent on that help, because the
daughter does these activities in her homemaking role. Others have
supported this finding as well (Chappell, 1991; Lang & Brody, 1983;
Walker & Pratt, 1991; Walker et al., 1996).
H8 predicted that daughters' perceived past relationship quality
with their mothers would be positively related to their perceptions of
current relationship quality. This hypothesis was supported. From a social
learning perspective (see Patterson, 1982, 1986), Whitbeck et al., (1991)
found that learned patterns of interactions from the early parent-child
relationship did shape the current elderly parent-adult child relationship
Rossi and Rossi (1990) also found that parents and children mutually
reinforce each other throughout life in the way they communicate,
negotiate and renegotiate rules and roles, and deal with family conflicts.
The continuous reinforcement of interactional patterns between parents
and children caries over to current relationship quality.
62
Limitations and Strengths
Limitations.
As with all research, there are limitations to the present study. First,
results of this study cannot be generalized to the family caregiving
population because it was selective, excluding caregiving spouses, sons,
and care-receiving fathers. Respondents were recruited both from urban
and rural areas, and having a wide range of caregiving daughters were
included who were similar to representative samples of White daughters
providing care to an impaired elderly parent(s). Second, this study
focused on caregiving daughters whose mothers did not have cognitive
deficiencies. As noted in the intergenerational and caregiving literatures,
giving care to a person with some form of dementia may cause more
wear-and-tear on the caregiver than the type of caregiving represented in
this study (e.g., Blieszner & Shifflift, 1990). A third limitation is that 98% of
the sample in this study is White so the results cannot be generalized to
other ethnic/racial populations. A final limitation is that the geographical
characteristics of the Pacific Northwest may be different from other
locations.
63
Strengths.
A major strength of this study is that it investigated longitudinally
the effect of the amount of care given on perceived relationship quality
between mothers and daughters in middle and later life, thus filling a gap
in both the gerontological and intergenerational literatures. Another
strength of the study is that it used structural equation modeling
methodology to compensate for previous methodological problems. A
third strength was controlling for the care recipients' impairment; that is,
this study only focused on care recipients who had acute or chronic health
conditions. A fourth strength was that the research examined only the
mother-daughter relationship, thus, controlling for the gender of the
caregiver and the care-recipient as well as excluding daughters with
married mothers as to reduce confounding effects in the data (Boyd,
1989; Horowitz, 1985; Stafford, 1988). Finally, this research addressed
directly the problem of missing data that often plagues longitudinal studies
by using EM Algorithm to account for missing data patterns (Dempster et
al., 1977; Little & Rubin, 1987; Malone Beach & Zarit, 1991; Mancini &
Blieszner, 1989; Young & Kahana, 1989). These strengths offset many
weaknesses in the caregiving literature and help to minimize some of the
study's limitations.
64
Implications
Although the finding that an increase in the amount of care given
over time leads to a decrease in relationship quality, caregiving daughters
generally reported good and close relationships with their care-receiving
mothers. Service providers should target those daughters who are giving
increasing amounts of care by offering respite services that allow them to
take a break from the demands of caregiving. Service providers should
also make classes available in stress management.
The finding that coresidential daughters provide more care is
important in that such individuals may need to learn coping skills to
reduce potential stress and burden as the amount of care provided
increases. Additionally, service providers should target daughters who
report having a poor relationship with their mothers. Interventions such as
individual or family counseling may help them focus on positive
relationship aspects rather than the negative aspects of the relationship.
Directions for Future Research
The findings reported here are based on the average caregiving
daughter and does not address the issue of individual variability within
and between caregivers. As Walker et al.(1996) and Townsend et al.
65
(1989) found, there is mounting evidence of variability among caregivers.
Techniques such as latent growth cure methodology will help in assessing
the variability among caregiving daughters in the caregiving literature.
Further, more research is needed to examine the association between
relationship quality and amount of care given. For example, what other
factors contribute to decline in relationship quality? Do mothers and
daughters argue in their perceptions of the amount of care given and
relationship quality? If not, how do they differ? Finally, when the amount of
care given remains stable for many years, does relationship quality also
remain stable? These are just a few of the questions that should be
addressed in the family caregiving literature.
66
Bibliography
Abel, E. K. (1989). The ambiguities of social support: Adult daughters
caring for frail elderly parents. Journal of Aging Studies, 3, 211-230.
Abel, E. K. (1986). Adult daughters and care of the elderly. Feminist
Studies, 12, 479-493.
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in
practice: A review and recommended two-step approach. Psychological Bulletin,
103.411 -423.
Aquilino, W. S. (1990). The likelihood of parent-adult child coresidence:
Effects of family structure and parental characteristics. Journal of Marriage and
the Family, 52, 405-419.
Aquilino, W. S., & Supple, K. R. (1991). Parent-child relations and parent's
satisfaction with living arrangements when adult children live at home. Journal of
Marriage and the Family, 53, 13-27.
Archbold, P. G. (1983). Impact of parent-caring on women. Family
Relations, 32, 39-45.
Aronson, J. (1992). Women's sense of responsibility for the care of old
people: But who else is going to do it? Gender and Society, 6, 8-29.
Atkinson, M. P. (1989). Conceptualizations of the parent-child relationship:
Attachment, crescive bonds, and identity salience. In J. A. Mancini (Ed.), Aging
parents and adult children (pp. 81-98). Lexington, MA: D. C. Heath.
Atkinson, J., & Huston, T. L. (1984). Sex role orientation and division of
labor in early marriage. Journal of Personality and Social Psychology, 46, 330345.
67
Barer, B. M., & Johnson, C. L. (1990). A critique of the caregiving
literature. The Gerontologist, 30, 26-29.
Barnett, R. C., Kibria, N., Baruch, G. K., & Pleck, J. H. (1991). Adult
daughter-parent relationships and their associations with daughters' subjective
well-being and psychological distress. Journal of Marriage and the Family, 53,
29-42.
Baruch, G., & Barnett, R. C. (1983). Adult daughters' relationships with
their mothers. Journal of Marriage and the Family, 45, 601-606.
Barusch, A. S. (1988). Problems and coping strategies of elderly spouse
caregivers. The Gerontologist, 28, 677-685.
Barusch, A. S., & Spaid, W. M. (1989). Gender differences in caregiving:
Why do wives report greater burden? The Gerontologist, 29, 667-676.
Bengtson, V. L., Cutler, N. E., Mangen, D. J., & Marshall, V. W. (1985).
Generations, cohorts, and relations between age groups. In R. H. Binstock & E.
Shanas (Eds.), Handbook of aging and the social sciences (2nd ed., pp. 304338). New York, New York: Van Nostrand.
Bengtson, V. L., & Roberts, R. L. (1991). Intergenerational solidarity in
aging families: An example of formal theory construction. Journal of Marriage
and the Family, 53, 856-870.
Bent ler, P. M. (1988). Causal modeling via structural equation systems. In
J. R. Nesselroade & R. B. Cattell (Eds)., Handbook of multivariate experimental
psychology (2nd ed., pp. 317-335). New York: Plenum.
Bent ler, P. M. (1980). Multivariate analysis with latent variables: Causal
modeling. Annual Review of Psychology, 31, 419-456.
Blieszner, R., & Mancini, J. (1987). Enduring ties: Older adults' parental
role and responsibilities. Family Relations, 36, 176-180.
68
Blieszner, R., & Shifflift, P. A. (1990). The effects of Alzheimer's disease
on close relationships between patients and caregivers. Family Relations, 39,
57-62.
Bollen, K. E. (1989). Structural equations with latent variables. New York:
John Wiley.
Boyd, C. J. (1989). Mothers and daughters: A discussion of theory and
research. Journal of Marriage and the Family, 51, 291-301.
Bretherton, I., Biringen, Z., & Ridgeway, D. (1991). The parental side of
caregiving. In K. Pillemer & K. McCartney (Eds.), Parent-child relations in later
life (pp. 1-24). Hillsdale, NJ: Lawrence Erlbaum.
Brody, E. M. (1986). Filial care of the elderly and changing roles of women
(and men). Journal of Geriatric Psychiatry, 14, 175-200.
Brody, E. M. (1985). Parent care as a normative family process. The
Gerontologist, 25, 19-29.
Brody, E. M., Johnsen, P. T., & Fulcomer, M. C. (1994). What should adult
children do for elderly parents: Opinions and preferences of three generations of
women. Journal of Gerontology, 39, 736-746.
Brody, E. M., Litvin, S. J., Albert, S. M., & Hoffman, C. J. (1994). Marital
status of daughters and patterns of parent care. Journal of Gerontology: Social
Sciences, 49, S95-S103.
Brody, E. M., & Schoonover, C. B. (1986). Patterns of parent care when
adult daughters work and when they do not. The Gerontologist, 26, 372-381.
Bromberg, E. M. (1982-1983). Mother-daughter relationships in later life:
Negating the myth. Aging, 340, 15-20.
69
Browne, K. D., & Cudeck, R. (1993). Alternative ways of assessing model
fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation models (pp.
136-162). Newbury Park, CA : Sage.
Byrne, B. M. (1994). Structural equation modeling with EQS and
EQSndows: Basic concepts, applications, and programming. Thousand Oaks,
CA: Sage Publications, Inc.
Byrne, B. M., Shavelson, R. J., & Muthen, B. (1989). Testing for
invariance of factor covariance and mean structures: The issue of partial
measurement invariance. Psychological Bulletin, 105, 456-466.
Caspi, A., & Elder, G. H. (1988). Emergent family patterns: The
intergenerational construction of problem behavior and relationships. In R. Hinde
& J. Stevenson-Hinde (Eds.), Relationships within families. Oxford: Clarendon
Press.
Chappell, N. L. (1991). Living arrangements and sources of caregiving.
The Gerontologist, 46, S1-S8.
Cicirelli, V. G. (1991). Attachment theory in old age: Protection of the
attached figure. In K. Pillemer & K. McCartney (Eds.), Parent-child relations in
later life (pp. 25-41). Hillsdale, NJ: Lawrence Erlbaum.
Cicirelli, V. G. (1983). Adult children's attachment and helping behavior to
elderly parents: A path model. Journal of Marriage and the Family, 45, 815-825.
Cicirelli, V. G. (1981). Helping elderly parents: The role of adult children.
Boston, MA: Auburn House Publishing Company.
Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood
estimation with incomplete data via the EM algorithm. Journal of Research and
Statistical Society, B39, 1-38.
70
DeWitt, D. J., blister, A. V., & Burch, T. K. (1988). Physical distance and
social contact between elders and their adult children. Research on Aging, 10,
56-80.
Dowd, J. J. (1975). Aging as exchange: A preface to theory. Journal of
Gerontology, 30, 584-594.
Dwyer, J. W., & Coward, R. T. (1991). A multivariate comparison of the
involvement of adult sons versus daughters in the care of impaired parents.
Journal of Gerontology: Social Sciences, 46, S259-S269.
Dwyer, J. W., & Seccombe, K. (1991). Elder care as family labor: The
influence of gender and family position. Journal of Family Issues, 12, 229-247.
Frankel, B. G., & DeWitt, D. J. (1988). Geographic distance and
intergenerational contact: Methodological issues. Journal of Aging Studies, 34,
238-250.
Gerstel, N., & Gallagher, S. K. (1993). Kinkeeping and distress: Gender,
recipients of care, and work-family conflict. Journal of Marriage and the Family,
55 598-607.
Haley, W. E., & Pardo, K. M. (1989). Relationship of severity of dementia
to caregiving stressors. Psychology and Aging, 4, 389-392.
Hareven, T. K. (1995). Historical perspectives on the family and aging. In
R. Blieszner & V. H. Bedford (Eds.), Handbook of aging and the family (pp. 1331). Westport, CT: Greenwood Press.
Harootyan, R. A., & Bengtson, V. L. (1994). Intergenerational linkages:
The context of the study. In V. L. Bengtson & R. A. Harootyan (Eds.),
Intergenerational linkages: Hidden connections in American society. (pp. 1-18).
New York, NY: Springer Publishing Co.
Harris, D. (1990). Sociology of aging. New York, NY: Harper & Row.
71
Hayuduk, L. A. (1987). Structural equation modeling with LISREL.
Baltimore, MD: Johns Hopkins.
Hess, B., & Waring, J. M. (1978). Parent and child in later life: Rethinking
the relationship. In R. M. Lerner & G. B. Spanier (Eds.), Child influences on
marital and family interaction (pp. 241-273). New York: Academic.
Himes, C. L. (1994). Parental caregiving by adult women: A demographic
perspective. Research on Aging, 16, 191-211.
Hogan, D. P., Eggebeen, D. J., & Clogg, C. C. (1993). The structure of
intergenerational exchange in American families. American Journal of Sociology,
98 1428-1458.
Horowitz, A. (1985). Sons and daughters of caregivers to older parents:
Differences in role performance and consequences. The Gerontologist, 25, 612617.
Hoyle, R. H. (1995). The structural equation modeling approach: Basic
concepts and fundamental issues. In R. H. (Ed.), Structural equation modeling:
Concepts, issues, and applications (pp. 1-15). Westport, CT: Greenwood Press.
Jarett, W. H. (1985). Caregiving within kinship systems: Is affection really
necessary? The Gerontologist, 25, 5-10.
Johnson, C. L., & Catalano, D. J. (1983). A longitudinal study of family
supports to impaired elderly. The Gerontologist, 23, 612-618.
Joreskog, K. G. (1971). Statistical analyses of sets of congeneric tests.
Psychometrika, 36, 109-134.
Joreskog, K. G., & Sorbom, D. (1993). LISREL 8: Structural equation
modeling with the SIMPLIS command language. Chicago, IL: Scientific Software
International.
72
Inc.
Kart, C. S. (1985). The realities of aging. Boston, MA: Allyn and Bacon,
Kinsella, K. (1995). Aging and the family: Present and future demographic
issues. In R. Blieszner & V. H. Bedford (Eds.), Handbook of aging and the family
(pp. 32-56). Westport, CT: Greenwood Press.
Kunkel, S. R., & Applebaum, R. A. (1992). Estimating the prevalence of
long-term care for disability for an aging society. Journal of Gerontology: Social
Sciences, 47, S253-S260.
Lang, A. M., & Brody, E. M. (1983). Characteristics of middle-aged
daughters and help to their elderly dependent mothers. Journal of Marriage and
the Family, 45, 193-202.
Lawton, L., Silverstein, M., & Bengtson, V. L. (1994). Affection, social
contact, and geographic distance between adult children and their parents.
Journal of Marriage and the Family, 56, 57-68.
Lawton, L., Silverstein, M., & Bengtson, V., L. (1994). Solidarity between
generations in families. In V. L. Bengtson & R. A. Harootyan (Eds.),
Intergenerational linkages: Hidden connections in American society. (pp. 19-42).
New York, NY: Springer Publishing Co.
Lee, G. R. (1987). Aging and intergenerational relations. Journal of Family
Issues. 8. 448-450.
Lee, G. R., & Ellithorpe, E. (1982). Intergenerational exchange and
subjective well-being among the elderly. Journal of Marriage and the Family, 44,
217-224.
Levitt, M. J., Guacci, N., & Weber, R. A. (1992). Intergenerational support,
relationship quality, and well-being. Journal of Family Issues, 13, 465-481.
73
Little, R. J. A., & Rubin, D. B. (1987). Statistical analysis with missing
data. New York: John Wiley & Sons.
Malone Beach, E. E., & Zarit, S. H. (1991). Current research issues in
caregiving to the elderly. International Journal of Aging and Human
Development, 32, 103-114.
Mancini, J. A., & Blieszner, R. (1989). Aging parents and adult children:
Research themes in intergenerational relations. Journal of Marriage and the
Family, 51, 275-290.
Markides, K. S., & Cooper, C. L. (1989). Aging, stress, social support, and
health: An overview. In K. S. Markides & C. L. Cooper (Eds.), Aging, stress, and
health (pp.1-12). NY: John Wiley & Sons.
Mathews, S. H. (1988). The burdens of parent care: A critical evaluation of
recent findings. Journal of Aging Studies, 2, 157-165.
McDonald, G. W. (1981). Structural exchange and marital interaction.
Journal of Marriage and the Family, 43, 825-839.
Miller, B. (1989). Adult children's perceptions of caregiver stress and
satisfaction. Journal of Applied Gerontology, 8, 275-293.
Miller, B., McFall, S., & Montgomery, A. (1991). The impact of elder
health, caregiver involvement, and global stress on two dimensions of caregiver
burden. Journal of Gerontology: Social Sciences, 46. S9-S19.
Montgomery, R. J. V., Gonyea, J. G., & Hooyman, N. R. (1985).
Caregiving and the experience of subjective and objective burden. Family.
Relations, 34, 19-25.
Montgomery, R. V. J., & Kamo, Y. (1989). Parent care by sons and
daughters. In J. Mancini (Ed.), Aging parents and adult children (pp. 213-230).
Lexington, MA: Lexington Books.
74
Motenko, A. K. (1989). The frustrations, gratifications, and well-being of
dementia caregivers. The Gerontologist, 29, 166-172.
Muthen, B., Kaplan, D., & Hollis, M. (1987). Structural equation modeling
for data that are not missing completely at random. Psychometrika, 51, 431-462.
Nunnally, F. C. (1978). Psychometric theory. New York: McGraw-Hill Book
Co.
Patterson, G. R. (1986). Performance models for antisocial boys.
American Psychologist, 41, 432-444.
Patterson, G. R. (1982). Coercive Family Process. Eugene, OR: Castillia.
Pear lin, L. I., Mullan, J. T., Semple, S. J., & Skaff, M. M. (1990).
Caregiving and the stress process: An overview of concepts and their measures.
The Gerontologist, 30, 583-591.
Pentz, M. A., & Chou, C. P. (1994). Measurement invariance in
longitudinal clinical research assuming change from development and
intervention. Special Section: Structural equation modeling in clinical research.
Journal of Consulting and Clinical Psychology, 62, 450-462.
Pratt, C. C., Schmall, V., Wright, S., & Cleland, M. (1985). Burden and
coping strategies of caregivers to Alzheimer's patients. Family Relations, 34, 2733.
Pruchno, R. A., & Resch, N. L. (1989). Husbands and wives as
caregivers: Antecedents of depression and burden. The Gerontologist, 29, 159165.
Rakowski, W., & Clark, N. M. (1985). Future outlook, caregiving, and carereceiving in the family context. The Gerontologist, 25, 618-623.
75
Reed, B. R., Stone, A. A., & Neale, J. M. (1990). Effects of caring for a
demented relative on elders' life events and appraisals. The Gerontologist, 30,
200-205.
Roberts, R. L., Richards, L. N., & Bengtson, V. L. (1991).
Intergenerational solidarity in families that bind. Marriage and Family Review, 13,
11-43.
Robinson, B., & Thurnher, M. (1979). Taking care of aged parents: A
family cycle transition. The Gerontologist, 19, 586-593.
Rossi, A. S., & Rossi, P. H. (1990). Of human bonding: Parent-child
relations across the life course. New York: Aldine de Gruyter.
Scharlach, A. E. (1994). Caregiving and employment: Competing or
complementary roles? The Gerontologist, 34, 378-385.
Scharlach, A. E. (1987). Role strain in mother-daughter relationships in
later life. The Gerontologist, 27, 627-631.
Schulz, R., & Williamson, G. M. (1991). A 2-year longitudinal study of
depression among Alzheimer's caregivers. Psychology and Aging, 6, 569-578.
Shanas, E. (1979). The family as a social support system in old age. The
Gerontologist, 19, 169-174.
Sheehan, N. W., & Nuttall, P. (1988). Conflict, emotion, and personal
strain among family caregivers. Family Relations, 37, 305-309.
Silverstein, M., Lawton, L., & Bengtson, V. L. (1994). Types of relations
between parents and adult children. In V. L. Bengtson & R. A. Harootyan (Eds.),
Intergenerational linkages: Hidden connections in American society. (pp. 43-76).
New York, NY: Springer Publishing Co.
76
Spitze, G., & Logan, J. (1990). Sons, daughters, and intergenerational
support. Journal of Marriage and the Family, 52, 420-430.
Stephens, S. A., & Christianson, J. B. (1986). Informal care of the elderly.
Lexington, MA: D.C. Heath.
Stephens, M. A., & Zarit, S. H. (1989). Symposium: Family caregiving to
dependent older adults: Stress, appraisal, & coping. Psychology and Aging, 4,
387-388.
Stoller, E. P. (1985). Exchange patterns in the informal support networks
of the elderly: The impact of reciprocity on morale. Journal of Marriage and the
Family, 47, 335-341.
Stoller, E. P. (1983). Parental caregiving by adult children. Journal of
Marriage and the Family, 45, 851-858.
Stoller, E. P., & Pugliesi, K. L. (1989). Other roles of caregivers:
Competing responsibilities or supportive resources. Journal of Gerontology:
Social Sciences, 44, S231-S238.
Stone, R., Cafferata, G. L., & Sangl, J. (1987). Caregivers of the frail
elderly: A national profile. The Gerontologist, 27, 616-626.
Stueve, A., & O'Donnell, L. (1984). The daughter of aging parents. In G.
B. Barach & J. Brooks-Gunn (Eds). Women in midlife. (pp. 203-225). New York:
Plenum.
Stull, D. E., Bowman, K., & Smerglia, V. (1994). Women in the middle: A
myth in the making? Family Relations, 43, 319-324.
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups.
New York: John Wiley & Sons.
77
Thompson, L., & Walker, A. J. (1984). Mothers and daughters: Aid
patterns and attachment. Journal of Marriage and the Family, 46, 313-322.
Townsend, A., Noelker, L., Deimling, G., & Bass, D. (1989). Longitudinal
impact of interhousehold caregiving on adult children's mental health.
Psychology and Aging, 4, 393-401.
Treas, J. (1977). Family support systems for the aged: Some social and
demographic considerations. The Gerontologist, 17, 486-491.
Umberson, D. (1992). Relationships between adult children and their
parents: Psychological Consequences for both generations.
Journal of Marriage and the Family, 54, 664-674.
Umberson, D. (1989a). Parenting and well-being: The importance of
context. Journal of Family Issues, 10, 427-439.
Umberson, D. (1989b). Relationships with children: Explaining parents'
psychological well-being. Journal of Marriage and the Family, 51, 999-1012.
U.S. Bureau of the Census. (1994). Current Population Reports, P251095. Washington, DC: U.S. Government Printing Press.
Walker, A. J., & Allen, K. A. (1991) Relationships between caregiving
daughters and their elderly mothers. The Gerontologist, 31, 389-396.
Walker, A. J., & Pratt, C. C. (1991). Daughters' help to mothers:
Intergenerational aid versus caregiving. Journal of Marriage and the Family, 53,
3-12.
Walker, A. J., & Thompson, L. (1983). Intimacy and intergenerational aid
and contact among mothers and daughters. Journal of Marriage and the Family,
45. 841-849.
78
Walker, A. J., Acock, A. C., Bowman, S. R., & Li, F. (1996). Amount of
care given and caregiving satisfaction: A latent growth curve analysis. Journal of
Gerontology: Psychological Sciences, 51, P130-P142.
Walker, A. J., Martin, S. K., & Jones, L. L. (1992). The benefits and costs
of care receiving for daughters and mothers. Journal of Gerontology: Social
Sciences, 47, S130-S139.
Walker, A. J., Pratt, C. C., & Eddy, L. (1995) Informal care to aging family
members: A critical review. Family Relations, 44, 402-411.
Walker, A. J., Pratt, C. C., & Oppy, N. C. (1992). Perceived reciprocity in
family caregiving. Family Relations, 41, 82-85.
Walker, A. J., Shin, H. Y., & Bird, D. N. (1990). Perceptions of relationship
change and caregiver satisfaction. Family Relations, 39, 147-152.
Walker, A. J., Thompson, L., & Morgan, C. S. (1987). Two generations of
mothers and daughters: Role position and interdependence. Psychology of
Women Quarterly, 11, 195-208.
Walker, A. J., Pratt, C. C., Shin, H. Y., & Jones, L. L. (1990). Motives for
parental caregiving and relationship quality. Family Relations, 39, 51-56.
Ward, R. A., & Spitze, G. (1992). Consequences of parent-adult child
coresidence. Journal of Family Issues, 13, 553-572.
Wellman, B., & Wortley, S. (1990). Different strokes from different folks:
Community ties and social support. American Journal of Sociology, 96, 558-588.
Whitbeck, L. B., Hoyt, D. R., & Huck, S. M. (1994). Early family
relationships, intergenerational solidarity, and support provided to parents by
their adult children. Journal of Gerontology: Social Sciences, 49, S85-S94.
79
Whitbeck, L. B., Simons, R. L., & Conger, R. D. (1991). The effects of
early family relationships on contemporary relationships and assistance patterns
between adult children and their parents. Journal of Gerontology: Social
Sciences, 46, S330-S337.
Young, R. F., & Kahana, E. (1989). Specifying caregiver outcomes:
Gender and relationship aspects of caregiving strain. The Gerontologist, 29, 660666.
Zarit, S. H., Reeves, K. E., & Bach-Peterson, J. (1990). Relatives of the
impaired elderly: Correlates of feelings of burden. The Gerontologist, 20, 649655.
Zarit, S. H., Todd, P. A., & Zarit, J. M. (1986). Subjective burden of
husbands and wives as caregivers: A longitudinal study. The Gerontologist, 20,
260-266.
80
Appendices
81
Appendix A
Caregiving Tasks and Activities
In the past 24 hours, which of the following did you do or arrange to have done
for your mother? (We will also ask you about how much time or money was
involved in each of these activities.)
A. Shopping/Errands
B. Indoor Maintenance
Went grocery shopping
Made the bed
Ran errands (to bank, cleaners,
gas station, etc.)
Changed the bed
Drove mother somewhere
Bought household supplies
Bought clothes
Made an expensive purchase
C. Financial Tasks
Paid bills
Wrote checks
Budgeted, planned, or
reviewed expenses
D. Food Preparation/Clean-Up
Made breakfast
Made lunch
Made dinner
Did the dishes
Straightened the house,
put things away
Cleaned the house (mopped,
waxed, dusted, vacuumed)
Took out the garbage
Did the laundry
Ironed clothes
Decorated the house, arranged
furnishings
Did repairs around the house
E. Outdoor Maintenance
Gardened (pulled weeds, planted
vegetables, cut shrubs)
Repaired car or other vehicle
Washed or waxed car or other
vehicle
82
F. Personal Care
G. Financial Contributions
Fed your mother
Paid for groceries
Changed or dressed your mother
Paid for utilities (gas, water,
electricity)
Bathed your mother
Put your mother to bed
Helped use toilet or bed pan
Helped your mother exercise
Paid for household supplies
Paid for clothes
Paid rent/mortgage
H. Psychological Support
Combed your mother's hair
Listened to mother
Helped with medication
Discussed problems with mother
Helped in and out of bed
Gave understanding
Helped to walk
Gave advice
Put on make-up
Cheered up
Helped with stairs
Improved mother's morale
Changed diapers
Gave warmth
Helped in and out of chairs
Gave affection
I. Leisure Activities
Read to your mother
Gave respect
J. Bureaucratic Mediation
Took mother to church
Got information for mother
Took mother to a meeting
Made application for mother
Took mother to visit friends
Arranged for service for mother
Took mother to restaurant
Filled out forms
Went for a walk with mother
Talked with doctors for mother
83
I. Leisure Activities (continued)
K. General Activities
Played a table game (cards)
Talked with mother on telephone
Took to park or picnic
Talked/chatted with mother in
person
Took shopping or browsing
Took for a drive
Adapted from Atkinson and Huston (1984).
84
Appendix B
Intimacy Scale
1.
We want to spend time together.
2.
She shows that she loves me.
3.
We're honest with each other.
4.
We can accept each other's criticisms of our faults and mistakes.
5.
We like each other.
6.
We respect each other.
7.
Our lives are better because of each other.
8.
We enjoy the relationship.
9.
She cares about the way I feel.
10.
We feel like we're a unit.
11.
There's a great amount of unselfishness in our relationship.
12.
She always thinks of my best interest.
13.
I'm lucky to have her in my life.
14.
She always makes me feel better.
15.
She is important to me.
16.
We love each other.
17.
I'm sure of this relationship.
85
Appendix C
Attachment Scale
1.
We're dependent on each other.
2.
We anticipate each other's mood.
3.
We nurture each other.
4.
I feel like I want to support her.
5.
She is closer to me than others are.
6.
We're emotionally dependent on each other.
7.
When we anticipate being apart, our relationship intensifies.
8.
We anticipate each other's needs.
9.
Our best times are with each other.