Weekly topic on: lifecycle perspectives
Reading:
What Makes Us Happy” article from the Atlantic Monthly (June, 2009)
An, J.S., & Cooney, T.M. (2006). Psychological well-being in mid to late life: The role of generativity development and parent-child relationships across the lifespan.International Journal of Behavioral Development, 30(5), 410-421.
Film:
“Beginners” (d. Mike Mills, 2010, USA)
To get full creditfor WEEKLY response papers, follow these rules:
1-page ( no more or less)
Single-spaced
International Journal of Behavioral Development
2006, 30 (5), 410–421
© 2006 The International Society for the
Study of Behavioural Development
http://www.sagepublications.com
DOI: 10.1177/0165025406071489
Psychological well-being in mid to late life: The role of generativity
development and parent–child relationships across the lifespan
Jeong Shin An
Teresa M. Cooney
Seoul Cyber University, South Korea
University of Missouri–Columbia, USA
This study examined the association between generativity and psychological well-being for a subsample of 1882 mid- to late-life parents using the MIDUS data set. Guided by Erikson’s theory of
psychosocial development, we tested a structural model of psychological well-being that also included
direct and indirect effects (via generativity) of remembered pre-adult relationships with parents and
current parental experiences with offspring on well-being. Respondents who recalled positive,
trusting relationships with parents in childhood reported more positive parental experiences with
their adult offspring and better psychological well-being. Current parental experiences had both
indirect and direct effects on well-being too, but generativity had the strongest direct effects. Thus,
it appears that the achievement of generativity plays a substantial role in well-being in mid- and late
life. Findings also reveal that the impact of generativity on well-being is stronger for females than
males. Implications for intervention with older adults, such as promoting volunteer work, are
discussed.
Keywords: adult well-being; generativity; parent–child relations
being, as they operate both directly and via generativity. Childhood relationships with parents are examined because Erikson
(1963) believed that healthy development resulted from
dealing successfully with a series of stage-specific developmental challenges across the lifespan. Plus, he specifically
acknowledged the role of early family experiences in his
discussions of generativity. Current parenting experiences with
offspring are included in the model because Erikson
considered parenting to be one way in which adults could
confront issues of generativity. Moreover, empirical evidence
reveals a significant link between midlife and elderly parents’
relationships with offspring, and their well-being and adjustment. An important theoretical contribution of this study is
that it empirically tests whether generativity is the mediating
mechanism between early family experience and adult wellbeing, and between current parenting experiences and wellbeing in mid and late life.
With more adults living into old age (US Census Bureau,
2004), there is heightened concern about making the later
years of life quality ones – characterized by good health and
positive well-being. The United States, for example, has
launched a major initiative, Healthy People 2010, aimed at
helping Americans live longer and healthier lives (US Department of Health & Human Services, 2005). A key aspect of
positive health is psychological well-being, which involves such
components as quality social relations, positive self-regard,
purpose in life, and mastery over one’s environment. These are
considered core elements to a “good life” for adults (Ryff &
Singer, 1998).
Healthy adjustment in mid to later adulthood, specifically
psychological well-being, is the focus of this study. In accord
with the view that social relationships play a significant role in
adults’ well-being and positive evaluations of life (Diener, Suh,
Lucas, & Smith, 1999; Markus, Ryff, Curhan, & Palmersheim,
2004; Shaw, Krause, Chatters, Connell, & Ingersoll-Dayton,
2004), we test a model that explains well-being in relation to
both childhood and concurrent relational experiences.
Erikson’s (1963) theory of psychosocial development frames
the study. This theory emphasizes the critical role that caring
relationships with others play in development, beginning early
in life as young children build trust in their caregivers and the
world, later in young adulthood as intimacy issues take center
stage, and again in mid to late life when generativity – showing
care and concern for the next generation – becomes important
(Erikson, 1963). Given our interest in mid- to late-life development, Erikson’s theory leads us to test the role of generativity
as a central mechanism influencing well-being, as illustrated in
Figure 1. Additionally, we explore the influence of both
current parenting experiences with adult offspring and early
relationships with one’s own parents on mid- to late-life well-
Erikson’s (1963) theory of psychosocial development
includes progression through eight stages over the lifespan.
Each stage is marked by a unique developmental challenge
that must be appropriately resolved so that optimal development can be achieved at that and later life stages. This
research considers the first and seventh stages. According to
Erikson (1968), during the first developmental stage the
young child engages in “receiving and accepting what is
given” (p. 99) and needed from the caregiver and, as a result,
develops either trust or mistrust in others and the world. This
trust–mistrust crisis focuses heavily on caring and interpersonal relations with others as the infant comes to develop a
Correspondence should be sent to Teresa M. Cooney, Human
Development and Family Studies, 314 Gentry Hall, University
of Missouri–Columbia, Columbia, MO 65211, USA; e-mail:
CooneyT@missouri.edu
Erikson’s theory of psychosocial development
INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2006, 30 (5), 410–421
sense of whether parental caregivers will reliably attend to
his/her needs and demonstrate consistent care and understanding (Sorell & Montgomery, 2001). Parental affection is
surely linked to caregiver concern and sensitivity, and thus
also contributes to the development of concern. As Rossi
(2001) argues “concern for others springs from the seedbed
of family affection laid down during the years of dependency
in infancy and childhood” (p. 228). Therefore, early childhood relationships and resolution of this initial developmental crisis surrounding trust are central to development
across the lifespan.
Erikson’s seventh stage occurs in middle age and involves
resolution of the generativity versus stagnation crisis, which
centers on the “need to be needed” (Erikson, 1968, p. 138),
caring for the next generation, and contributing to future
generations through productive and creative endeavors. Again,
as in childhood, caring relations are central to this crisis, thus
it is logical that the achievement of generativity would build on
trust developed early in life. Erikson (1968) claimed that trust
established in early childhood contributes to later development
because “in learning to get somebody to do for him what he
wishes to have done” (and needs to have done) a child
“develops the necessary groundwork . . . ‘to get to be’ the giver
. . . and eventually become a giving person” (p. 99). Connection between one’s relationship with parent caregivers in childhood and generativity in adulthood is therefore noted in
Erikson’s theory.
Generativity can be achieved in multiple ways. Parenting
offers one opportunity for individuals to guide and care for the
next generation, yet Erikson considered it neither necessary
nor sufficient for achieving generativity. Indeed, he claimed
that some parents failed in “the ability to develop true care”
(1968, p. 138), a problem he believed originated in early childhood. Generativity also can be derived through altruistic
activities such as volunteering and mentoring, or through other
forms of community involvement and productive contributions. Measures and recent study of generativity reflect these
various components (Keyes & Ryff, 1998). Moreover, some
developmentalists posit that in resolving the generativity crisis,
men and women emphasize different components of generativity, with public (i.e. work, civic activity) contributions taking
center stage for men, and family investments dominating
women’s lives (Sorell & Montgomery, 2001).
Though Erikson stressed that positive development
depended on resolution of distinct tasks at successive life
stages, he did not view development as unidirectional. Rather,
he argued that the initial resolution of particular developmental tasks does not absolutely determine subsequent
developmental because, across life, crises of the earlier stages
can be reworked and resolved (Sorell & Montgomery, 2001).
Hence, in considering well-being and development in adulthood, it is critical to look beyond early childhood influences to
subsequent experiences, such as parenting one’s own children,
which may promote the reworking of previously encountered
developmental crises.
In sum, Erikson’s theory provides a foundation for our
study of well-being in mid and late life. The central question
we raise is what role does generativity play in adult psychological well-being? Is generativity a mechanism through which
past and current family relationships influence well-being, or
do these experiences have direct effects on well-being? Lastly,
does the link between well-being and these predictors differ
by gender?
411
Empirical evidence of the structure of
well-being in adulthood
In line with Erikson’s theory, we hypothesize an influence of
both childhood relationships with parents (pathway A – Figure
1), and generativity (pathway E) on adult well-being.We briefly
review empirical support for these two direct paths below.
Pathway C – denoting an influence of early experiences on
generativity, however, has not been empirically demonstrated.
Testing of this pathway is one contribution of this study.
Developmental antecedents of well-being
Both clinical and non-clinical research documents a connection between childhood relationships with parents and adjustment in adulthood. Specifically, clinical evidence links negative
interactions with parents in childhood (e.g., lack of attention
and affection) to severe interpersonal hostility (Luecken,
2000), depression (Parker, 1983; Richman & Flaherty, 1986)
and other psychological disorders in adulthood (Kessler,
Davis, & Kendler, 1997). Retrospective reports of family
experiences during childhood, drawn from non-clinical
samples, reveal similar associations. A consistent finding is that
memories of less affectionate and less supportive parenting in
childhood predict adult depression (Shaw et al., 2004;
Whitbeck et al., 1992). Finally, even prospective data indicate
a connection between received parenting in childhood and
adult well-being. Franz, McClelland, and Weinberger (1991)
found significant correlations between mothers’ reports of
warm, affectionate parenting behavior when their children
were five and positive well-being reported by the children at
age 36. A limitation of this latter finding, however, is that wellbeing was narrowly conceptualized as adults’ ability to
maintain good family and friend relationships, and a happy
marriage. It would thus be useful to examine this issue with a
broader, validated measure of adult well-being.
The connection between adult development – specifically
generativity – and well-being has received recent widespread
attention. Generativity is considered a major socio-personal
resource contributing to personal and social worth, and ultimately to quality of life (Keyes & Ryff, 1998), and empirical
evidence links generativity to life satisfaction and psychological well-being (Ackerman, Zuroff, & Moskowitz, 2000;
Grossbaum & Bates, 2002; Kinnevy & Morrow-Howell, 1999;
McAdams, de St. Aubin, & Logan, 1993; Stewart, Ostrove, &
Helson, 2001). Yet, significant associations between generativity and well-being are documented primarily for what are
labeled generative concerns (motivation to help the next
generation), rather than for generative behaviors (de St. Aubin
& McAdams, 1995; Grossbaum & Bates, 2002; McAdams et
al., 1993). Thus, in testing the effect of generativity on wellbeing and its mediating role between early received parenting
and well-being, multiple dimensions of generativity should be
examined.
Current parenting experiences and well-being
In Figure 1, we examine the influence of adults’ current
parenting experiences on well-being both directly (pathway D)
and via generativity (pathway F). Further, given research
evidence of continuity in parenting across generations
(pathway B), the combined B and D pathways in Figure 1 may
provide partial explanation for how early received parenting
412
AN AND COONEY / PSYCHOLOGICAL WELL-BEING
Generativity
in adulthood
Generative Qualities
Societal Contributions
Civic Obligations
Obligations for
Volunteer Work
E
C
Childhood
Relationships with Parents
F
Psychological Well-being
in adulthood
A
Mother Affection
Father Affection
Relationship with Mother
Relationship with Father
Autonomy
Environmental Mastery
Personal Growth
Purpose in Life
Self-Acceptance
B
D
Current Parental
Experiences
in adulthood
Parental Satisfaction
Parental Dissatisfaction
Parental Obligations
Relationships with Child
Figure 1.
Conceptual model of mediating mechanisms.
affects adult well-being (a finding in the literature summarized
above).
Evidence for pathway B in Figure 1 comes from studies
finding that adults who recall rejecting, harsh parenting in
childhood tend to use similar behaviors with their own
offspring (Simons, Whitbeck, Conger, & Wu, 1991; Whitbeck
et al., 1992). Prospective parent report data (Thornberry,
Freeman-Gallant, Lizotte, Krohn, & Smith, 2003) and observational data on parent behavior (Belsky, Jaffee, Sligo,
Woodward, & Silva, 2005) also reveal significant connections
between parenting received in childhood and similar parenting
of one’s own children in adulthood, as proposed in pathway B.
Evidence for pathway F is interesting to note because most
is derived from research on the parenting of relatively young
children by young parents (mostly fathers), rather than older
parents for whom generativity issues should be more salient.
Studies show that generative behaviors and feelings are reportedly stronger for fathers who are more involved with child care
(Snarey, 1993), have greater contact with their children
(Eggebeen & Knoester, 2001), place greater importance on the
parenting role (Christiansen & Palkovitz, 1998), and express
greater parenting satisfaction (Heath & Heath, 1991). But the
average age of parents in most studies ranges from 35 to 40
(Christiansen & Palkovitz, 1998; McKeering & Pakenham,
2000). The current study therefore provides a needed test of
the connection between parenting and generativity for middleaged and older parents.
This study also extends research on how parenting impacts
women’s generativity development and well-being. Though
women express more generative concerns (Wilson & Musick,
1997) and obligation to help others than men (Keyes & Ryff,
1998), and engage in more generative behavior (Hart,
McAdams, Hirsch, & Bauer, 2001; Wilson & Musick, 1997),
it is suggested that parenting is less important to their
expression of generativity than it is for men. This assertion is
based on a study of fairly young mothers and fathers
(McKeering & Pakenham, 2000). Parenting may contribute
more to the achievement of generativity among older cohorts
because women’s family roles are more central in the lives of
these cohorts than for younger cohorts coming of age today.
Considering gender differences within a middle-aged and older
sample is thus critical.
Substantial evidence that adult child–parent relationships
affect parental outcomes leads to consideration of possible
direct effects of current parenting experiences on adult wellbeing as well (pathway D). Evaluations of adult offspring and
their accomplishments (e.g. financial independence, marital
stability) have been associated with parental well-being
(Aquilino, 1996; Pillemer & Suitor, 1991), and studies consistently reveal a link between adult child–parent relationship
quality and parental well-being (Koropeckyj-Cox, 2002; Ryff
& Heidrich, 1997; Umberson, 1996). Specifically, strained
relationships and problematic experiences with adult offspring
are associated with heightened depression and drinking among
parents, as well as reduced life satisfaction and well-being.
Hagestad (1986) contends that parents gain a sense of
accomplishment when their children manage “life tasks on
time and in the normal sequence” (p. 685), such as graduating from school, starting work, and building their own families.
Presumably, generativity results partly from the “reflected
glory” parents derive from their offspring’s successful mastery
of adult developmental tasks (Hagestad, 1986). Yet, empirical
INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2006, 30 (5), 410–421
evidence is needed to determine whether current experiences
with adult offspring impact parents’ well-being directly, or
whether parenting experiences operate only through their
influence on parents’ achievement of generativity. The current
study addresses this issue.
Methods
Data
This analysis uses nationally representative data from the 1995
National Survey of Midlife Development in the United States
(MIDUS) (Brim et al., 1996). MIDUS respondents were
drawn through random-digit-dialing and included noninstitutionalized, English-speaking adults living in the contiguous 48 states. The sample of 4242 adults, aged 25 to 74,
responded to an initial telephone interview and completed a
follow-up mail questionnaire. Response rates for the telephone
interview and mail questionnaire (of telephone respondents
only) were 70 and 87%, respectively. The sample was stratified
by age and sex, with an over-sampling of men ages 65 to 74.
Sample weights were developed to adjust for sample design,
yet unweighted data are used here because results of previous
multivariate analyses (Keyes & Ryff, 1998) were not affected
by sample weighting and weights are problematic for Structural
Equation Modeling procedures used in this study.
Only respondents with at least one biological, adopted, or
stepchild over age 18 are included in the analysis so that
parental experiences with adult offspring can be examined in
relation to well-being. There were 1,980 sample respondents
(ages 35–74) who had an adult child. Fifteen of them were
under age 35 and were excluded from the analysis due to being
outliers on the age variable. (Studies of midlife generally limit
their samples to persons over age 35.) Cases with missing data
that were excluded from the analyses differed significantly on
one variable; respondents with missing data were less educated
than the analytic sample. After checking and deleting multivariate outliers to meet normality and linearity assumptions,
the analyses include 1,882 parents.
Measures
MIDUS is a unique multidisciplinary study in that it provided
assessment of numerous psychological constructs within a
large, nationally representative sample. To administer a survey
with such an expansive scope and large sample, MIDUS
investigators conducted six pilot studies (some using national
samples) to create shortened measures of key social and
psychological constructs (Brim, Ryff, & Kessler, 2004). Below
we describe the items constituting the measures for this study.
Generally, the source of these items is not provided in MIDUS
documentation. Where possible we note sources referred to in
other publications using MIDUS data.
Given the extensive pilot testing of the MIDUS survey
instrument, we have confidence in the validity of the measures
that we use. Furthermore, we subjected all of our key predictor and outcome variable items to factor analysis to determine
whether key indicators were unidimensional or multidimensional in nature. This analysis suggested that a multidimensional approach to these concepts was warranted. Below
we describe the constructed subscales for each latent indicator
in our model.
413
Childhood relationships with parents are assessed with
questions asking respondents to recall their relationships with
their parents when they were growing up. Because these questions are retrospective and did not specify a period in childhood to target, they are best interpreted as pre-adult memories.
Thus, we label this dimension Remembered Pre-Adult
Relationships with Parents. One item asked respondents to rate
their overall relationship with each parent (separately). Six
additional items (each asked in relation to each parent)
addressed parenting behaviors received in childhood that
display affection (Rossi, 2001) and support (Shaw et al.,
2004), and that we assert contribute to interpersonal trust. All
items on this factor are reverse coded so that higher summed
scores reflect more positive parenting in childhood. The scale
for mothers obtained a Cronbach’s alpha of .90, and that for
fathers a .92. Appendix A lists the items used for this domain.
Current parental experiences are measured with 14 questions that asked respondents to evaluate various aspects of
their relationships with their adult offspring. These items (see
Appendix B) were analyzed by a principle component factor
analysis and clustered as four factors: satisfaction with children
(α = .79), dissatisfaction with children (α = .42), relationship
quality with children (α = .71), and feelings of obligation to
children (α = .71). High scores represent more parental satisfaction and dissatisfaction, better quality of relationship, and
greater obligation.
Recent conceptualizations of generativity involve multiple
dimensions. MIDUS included 19 items that at face value
address various dimensions of generativity. A principle
component factor analysis of these items resulted in four
factors that we labeled to reflect as closely as possible similar
components of generativity used in the literature. They
include:
1 Generative qualities (α = .85) – six items from the Loyola
Generativity Scale (McAdams & de St. Aubin, 1992), which
reflect a view of self as having generative characteristics,
based on behavioral interactions with others. These items do
not specifically name one’s own children among this group
of others;
2 Civic obligations (α = .78) – six items used by Keyes and
Ryff (1998) that reflect obligation to be an involved citizen;
3 Obligations to volunteer work (α = .75) – three items
concerning feelings of duty to help others in need
(McAdams et al., 1993); and
4 Societal contributions (α = .87) – five items focused on the
thought and effort involved in making contributions to the
welfare of others. Three items from this dimension are
similar to ones used by Keyes and Ryff (1998) to measure
“generative concern,” but we label the dimension “societal
contributions” because two of the items ask for a self-rating
of one’s actual contributions to the well-being of others,
which involves more than feelings of concern.
All items in the generativity domain are reported in
Appendix C. Finally, note that these four components are
similar to what Snarey (1993) referred to as societal generativity. We omitted items regarding actions and feelings toward
one’s own children, which Snarey calls parental generativity, to
avoid confounding our measures of current parental experiences and generativity.
Psychological well-being is measured with a shortened
version of Ryff’s (1989) psychological well-being scale. This
version has three items for each of Ryff’s six dimensions of
414
AN AND COONEY / PSYCHOLOGICAL WELL-BEING
well-being and has demonstrated adequate validity (Ryff &
Keyes, 1995). We only use five of the six distinct dimensions of
well-being in the analysis: autonomy, environmental mastery,
personal growth, purpose in life, and self acceptance. The
subscale addressing positive relations with others is omitted
from the outcome measure to avoid confounding aspects of the
dependent variable with predictors in the model, specifically
current parental experiences. Scale scores are recoded so that
higher total scores reflect greater psychological well-being. The
Cronbach’s alphas for these five subscales range from .38 to .59,
which are adequate given the small number of items on each
subscale. These reliabilities are similar to those obtained when
using the entire MIDUS sample.
Data analyses
After comparing descriptive statistics for men and women on
all study variables, we conducted structural equation modeling
(SEM) procedures using the AMOS (Analysis of Moment
Structure) program (Arbuckle & Wothke, 1999) to estimate
the conceptual model in Figure 1. Direct effects of remembered pre-adult relationships with parents, current parental
experiences, and generativity on psychological well-being were
estimated. Mediating relationships and possible gender
differences in the models were also examined.
SEM was performed in several steps. First, a measurement
model was developed using Confirmatory Factor Analysis to
determine how constructs in the hypothesized model correspond to each other. Second, a structural model, based on the
measurement model, was developed to evaluate the magnitude
of associations among the latent variables. Next, mediation
effects were examined by using constrained paths between
predictor and outcome variables. Finally, using a multi-sample
model, differences in the fit of the structural models for men
and women were examined.
Results
Preliminary Analyses
The sub-samples of 876 males and 1,006 females are fairly
similar on background characteristics. The average age of men
(55.74) is somewhat older than that of women (54.64, t =
2.11, p
Order an Essay Now & Get These Features For Free:
Turnitin Report
Formatting
Title Page
Citation
Outline
