University of Southern California Dissertations and Theses

Thinking generatively versus acting generatively: exploring the associations of generative self-concept and generative activity with cognitive function among older adults

(USC Thesis Other)

Thinking generatively versus acting generatively: exploring the associations of generative self-concept and generative activity with cognitive function among older adults

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THINKING GENERATIVELY VERSUS ACTING GENERATIVELY:
EXPLORING THE ASSOCIATIONS OF
GENERATIVE SELF-CONCEPT AND GENERATIVE ACTIVITY
WITH COGNITIVE FUNCTION AMONG OLDER ADULTS

A dissertation presented to the Graduate School of the University of Southern California
in partial fulfillment of the requirements for the degree of
Doctor of Philosophy in Gerontology

by

Elizabeth Hagood Prickett
Bachelor of Arts, Wake Forest University, 2006
Master of Science, Virginia Commonwealth University, 2009

University of Southern California

Los Angeles, California

August 8, 2017

ii

DEDICATION

I dedicate this dissertation to my grandfather, James White,

who first inspired me to study gerontology.

iii

ACKNOWLEDGEMENTS

I owe a tremendous debt of gratitude to those whose assistance and support made this
work possible. First, I wish to extend my heartfelt thanks to my advisor, Dr. Tara Gruenewald,
for her mentorship over the past five years. Her dedicated and caring guidance enriched my
doctoral training greatly, and I am deeply grateful for the privilege of working with her. Tara, I
can never thank you enough for all that you have done for me, but please know how very much I
appreciate your support, your encouragement, and your incredible example of generativity.
I wish to also thank my co-advisor, Dr. Susan Enguidanos, for all of her kindness and
support. Susan, thank you for welcoming me so warmly into your lab, for your thoughtful
review of my dissertation research, and for being such a positive and encouraging presence when
I needed that most. I feel very lucky to have had the opportunity to know you and to work with
you.
I wish to extend my utmost gratitude to two invaluable members of my dissertation
committee, Dr. Cleopatra Abdou-Kamperveen and Dr. Kate Wilber. Cleopatra and Kate, thank
you so much for your insights on my dissertation research, as well as for your mentorship
throughout my time in the doctoral program. I am extremely grateful for your kindness and for
the privilege of learning from you.
Thank you also to the faculty, administrators, and staff at the Leonard Davis School of
Gerontology. I deeply appreciate the opportunity to study at USC and the support which made
that possible.
I also wish to thank my wonderful friends and labmates, Elena Gonzalez, Molli
Grossman, Shivanti Kariyawasam, Nicole Marcione, Joohong Min, Laura Pomatto, and Diana
iv

Wang, for their steadfast support. I could not have made it through this journey without you, and
I treasure your friendship.
Finally, I wish to thank my incredible family for all of their love and encouragement. I
am grateful to my parents, Dr. E. Scott Hagood and Mrs. Barbara White Hagood, for giving me
the strength and courage to pursue my dreams, even when that meant moving across the country
to begin the doctoral program at USC. Dad and Mom, I love you so much, and your support has
made my dreams become realities. I am thankful to my sisters, Sarah Milton and Virginia Anne
Adams, for being my best friends and bright spots in life. Thank you to my grandmother, Mrs.
Susie White, for being an extraordinary example of successful aging and a constant source of
inspiration. I owe a very special thanks to my uncle, Mr. Jerry Boxman, for generously and
uncomplainingly helping me to move to California so that I could begin the Ph.D. program.
Lastly, I wish to thank my husband, Mr. Drew Prickett, for his unfaltering love and
support. Drew, you inspire me every single day, and I can’t imagine life without you. Thank
you for being there for me through anything and everything. I love you, and I can’t wait for our
journey ahead.

v

TABLE OF CONTENTS
Dedication…………………………………………………………………………………………ii
Acknowledgements….………………………………………………………………………...…iii
Abstract…………………………………………………………………………………………...xi
Chapter 1: Background……………………………………………………………………………1
1A. Background…………………………………………………………...………………1
1B. The Origins of Generativity……………………………….………………………….4
1C. Generativity and Health……………………………………………………………....7
1D. Generativity and Cognition…………………………………………….……………..9
1E. Dissertation Objectives………….…………………………………………...………23
Chapter 2: Study 1…………………………………………………………………………….…25
2A. Background………………………………………….…………………………..…..25
2B. Significance……………………………………………………………….……..…..27
2C. Methods…………………………………………….…….……………….……..…..28
2D. Aims & Hypotheses…………………………….…….……………………………..37
2E. Results……………………………….………………………………………..……..39
2F. Discussion……………………….…………………………………………………...49
Chapter 3: Study 2…………………………………………………………………………….…53
3A. Background……………………………….……………………………………..…..53
3B. Significance………………………….…………………………………………..…..55
3C. Methods……………….……………………….………………………….……..…..56
3D. Aims & Hypotheses…………………….…………….……………………………..64
3E. Results…………………………….…………………………………………..……..66
vi

3F. Discussion…………………………………………………………………………....83
Chapter 4: Study 3…………………………………………………………………………….…90
4A. Background………………………….…………………………………………..…..90
4B. Significance…………………………………………………………….………..…..91
4C. Methods……………………………………………..…………………….……..…..92
4D. Aims & Hypotheses……………………………………..…………………………103
4E. Results…………………………………………………………………...…..……..106
4F. Discussion………………………………….……………………………………….127
Chapter 5: Summary & General Discussion….………………………………………………...133
5A. Study 1……………………………………………………………………...……...133
5B. Study 2………………………….……………………………………………...…..134
5C. Study 3.……………………………… ……….……………………………..…….135
5D. General Discussion……………………………………….……………….……….137
5E. Conclusion..…………………………………………………….…………………..147
References………………………………………………………………………………………148

vii

LIST OF FIGURES
Figure 1-1: Overview of prospective mediating pathways between generativity and cognitive
function….…………………………………………………………………………….…15
Figure 2-1: Outline of path modeling strategy utilized in Study 1 to assess generative self-
concept—cognition associations...…………………………………………………….…37
Figure 2-2: Path model of associations between change in self-perceived generative contributions
and episodic memory performance………………………………………………….…...44
Figure 2-3: Path model of associations between change in self-perceived generative contributions
and executive function………...…………………………………………...………….…45
Figure 2-4: Path model of associations between change in self-perceived generative
characteristics and episodic memory performance…………………………...……….…46
Figure 2-5: Path model of associations between change in self-perceived generative
characteristics and executive function………...…………………………………….…...48
Figure 3-1: Outline of path modeling strategy utilized in Study 2 to assess generative activity—
cognition associations…………………………………………..….………………….…64
Figure 3-2: Path model of associations between change in frequency of volunteering with
children/young people and change in verbal memory performance.………..………...…73
Figure 3-3: Path model of associations between change in frequency of volunteering with
children/young people and change in working memory performance.…………….….…75
Figure 3-4: Path model of associations between change in frequency of other forms of
volunteering/charity work and change in verbal memory performance.……………...…76
Figure 3-5: Path model of associations between change in frequency of other forms of
volunteering/charity work and change in working memory performance.…………...….78
viii

Figure 3-6: Path model of associations between change in caregiving frequency and change in
verbal memory performance.…………..………………….………………………......…79
Figure 3-7: Path model of associations between change in caregiving frequency and change in
working memory performance.…………………………………….………………….…81
Figure 4-1: Outline of path modeling strategy utilized in Study 3 to assess generative self-
concept—cognition associations and generative activity—cognition associations..…...102
Figure 4-2: Path model of associations between change in self-perceived generative contributions
and change in episodic memory performance………………………………………….112
Figure 4-3: Path model of associations between change in self-perceived generative contributions
and change in executive function………...…………………………………….…….…113
Figure 4-4: Path model of associations between change in self-perceived generative
characteristics and change in episodic memory performance…………………….….…115
Figure 4-5: Path model of associations between change in self-perceived generative
characteristics and change in executive function………...………………………….…117
Figure 4-6: Path model of associations between change in frequency of volunteering and change
in episodic memory performance……………………………………..…………..….…118
Figure 4-7: Path model of associations between change in frequency of volunteering and change
in executive function……………………………………...……………………....….…120
Figure 4-8: Path model of associations between change in frequency of emotional support
provision and change in episodic memory performance………………………..…..….121
Figure 4-9: Path model of associations between change in frequency of emotional support
provision and change in executive function…………………………………..…..….…123
ix

Figure 4-10: Path model of associations between change in frequency of instrumental support
provision and change in episodic memory performance…………………………….…124
Figure 4-11: Path model of associations between change in frequency of instrumental support
provision and change in executive function………………………………………….…126

x

LIST OF TABLES
Table 2-1: Study 1 Descriptive Statistics……………………………………………………..….41
Table 2-2: Direct Effects of Generative Self-Concept Predictors on Cognitive Function………48
Table 3-1: Study 2 Descriptive Statistics………………………………………………………...69
Table 3-2: Direct Effects of Generative Activity Predictors on Cognitive Function……………81
Table 4-1: Study 3 Descriptive Statistics……………………………………………………….108
Table 4-2: Direct Effects of Generative Self-Concept and Generative Activity Predictors on
Cognitive Function……………………………………………………………………...126

xi

ABSTRACT
The developmental phenomenon known as generativity collectively refers to those
behaviors and affective states which originate from the desire to contribute in a positive and
productive manner to the welfare of others. Though first identified as a psychosocial impulse of
midlife, generativity has since been recognized for its developmental significance in older
adulthood. Generativity has been shown to predict a number of desirable outcomes in later life,
including those related to psychological well-being (e.g., greater autonomy, mastery, and
purpose in life, as well as fewer depressive symptoms) and physical health (e.g., lesser risk of
disability and mortality). Preliminary evidence suggests that generativity may also hold promise
as a tool to enhance cognitive function in older adulthood, though only a small handful of studies
have investigated this possibility in short-term, experimental contexts. In order to more robustly
evaluate generativity’s potential to facilitate cognitive vitality across the second half of life,
larger-scale longitudinal investigations of generativity—cognition associations are warranted.
The current set of dissertation studies sought to examine longitudinal associations
between both (1) generative contributory activity (encompassing volunteering, caregiving, and
social support provision) and cognition; and (2) generative self-concept (encompassing self-
perceptions of generative contributory value and self-perceptions of generative characteristics)
and cognition. To this end, a series of three studies was conducted using data from two national
panel studies of health and development among American elders, including the National Survey
of Midlife Development (MIDUS) and the Health and Retirement Study (HRS). Study 1 used
data from the first (1995—1996) and second (2004—2006) waves of MIDUS to investigate
associations between generative self-concept and cognitive function, while Study 2 used data
from the 2008—2014 waves of HRS to assess associations between generative activity and
xii

cognitive function. Study 3 used data from the second and third (2013—2014) waves of MIDUS
to comparatively investigate associations between the respective generativity predictors and
cognition. For each study, path analyses with maximum likelihood estimation were used to
model change in generativity as a predictor of change in cognitive function across time. Tests of
multiple mediation were incorporated in each analysis in order to explore hypothesized
mechanistic pathways through which generativity may influence cognition.
The results of these investigations ultimately showed modest positive associations
between both generative activity and cognition and generative self-concept and cognition,
respectively. The magnitude of these associations was approximately equivalent, suggesting that
elders’ generative activity engagement and their generative self-concept dually contribute to
enhancements in cognitive functioning in later life. The two parameters also appeared to
mutually reinforce one another such that increases in generative activity were associated with
enhancements in generative self-concept, and vice versa. In addition, the mediation analyses
showed that both generativity predictors were independently associated with a number of
favorable health and psychosocial outcomes, though none of these prospective mediators
accounted for a substantial proportion of the observed generativity—cognition associations.
Taken together, the body of work detailed here provides the first evidence from large-
scale, longitudinal studies that generativity may serve to bolster cognitive function in later life.
This research supports prior empirical work which indicates that generativity may promote
successful aging among older adults. Though further investigation is necessary to corroborate
the present findings, they suggest the compelling possibility that the growing population of
American elders may be able to minimize the effects of age-related cognitive decline through the
contributory activities and feeling states of generativity.
1

CHAPTER 1: BACKGROUND

1A. Background
Over the course of the next forty years, the population of older adults in the United States
and in countries around the world will grow at a pace which is unprecedented in the modern era.
In the U.S. alone, the population of individuals age 65 and older will reach 88 million by 2050,
an increase from 46 million in 2014 (U.S. Census Bureau, 2015). The expansion of this
population will be accompanied by a concurrent increase in the generative, or contributory,
potential of this group. Generativity on the part of older adults has the capacity to improve life at
both the societal and individual levels. For society, the generative contributions of elders can
elicit increases in intergenerational exchange, eliminate ageist beliefs, and create more vibrant
and cohesive communities in general. The non-profit organization Encore, for instance,
capitalizes on the generative potential of American retirees in order to address critical social
issues such as poverty, civil rights, health inequalities, and access to educational opportunities
(Encore, 2014). For individuals, both generative activity and one’s generative self-concept have
the distinct capacity to shape health outcomes among older adults. Prior evidence from the
generativity-based Experience Corps program demonstrates that a variety of health benefits are
to be had from generative activity, including lower risk of impairment in activities of daily
living, higher levels of physical activity, and fewer depressive symptoms (Fried et al., 2004;
Hong & Morrow-Howell, 2010). In addition, older adults who perceive that they are socially
and generatively valuable show lesser risk of disability and death, as well as lesser risk of
placement in institutional care facilities, than those who do not (Grand, Grosclaude, Bocquet,
Pous, & Albarede, 1988, 1990; Gruenewald, Karlamangla, Greendale, Singer, & Seeman, 2007,
2009; Gruenewald, Liao, & Seeman, 2012; Okamoto & Tanaka, 2004; Pitkala, Laakkonen,
2

Strandberg, & Tilvis, 2004). Preliminary evidence likewise suggests that one’s generative
activities and conceptualizations of generative value may have beneficial effects for cognition
among older adults (Carlson et al., 2008, 2009, 2015; Hagood & Gruenewald, 2016). For a vital
health outcome such as cognitive function, generativity may thus represent a unique means of
enhancing functioning during a period of the life course that is traditionally characterized by
declines in fluid abilities (Baltes, 1993; Craik & Bialystok, 2006).
In the present era of rapid population aging, cognitive function is frequently cited as a
barometer of older adults’ health and well-being, as well as their capacity to engage productively
with the world around them. Changes in cognitive function across the life course are well-
documented and follow the “classic” pattern of cognitive aging, in which measures of fluid
intelligence (e.g., working memory, verbal memory, processing speed, reasoning ability) show
marked decline beginning in early adulthood, while measures of crystallized intelligence (e.g.,
vocabulary and cultural knowledge) are generally upheld across the life course (Baltes, 1993;
Park & Schwartz, 2000; Salthouse, 1991). Impairments in fluid abilities due to cognitive aging
are associated with functional decline in prefrontal cortex (Peters, 2006), anterior cingulate
cortex (Pardo et al., 2007), and hippocampus (Driscoll et al., 2003), including decreased
neuronal metabolism, subsequent neuronal atrophy, and reductions in synaptic density and
complexity. Though normative in cognitive aging, these changes can and do lead to deficits in
global cognitive functioning. Such deficits accrue over time until they reach a point of
manifestation as cognitive impairment, or the inability to perform tasks which require skills
involving memory, reasoning, inhibition, and processing speed. Recent estimates reveal that
between 16—20% of the elderly population suffer from mild cognitive impairment (Roberts &
Knopman, 2013), 11—14% suffer from dementia (Kasper, Freedman, & Spillman, 2014;
3

Plassman et al., 2007) and 10—12% suffer from Alzheimer’s disease (Plassman et al., 2007). As
the population of older adults grows in the coming years, these numbers are expected to increase
profoundly. The Centers for Disease Control and Prevention (2011) report that the number of
older Americans with Alzheimer’s disease – the least prevalent type of cognitive impairment
among the subtypes listed above – will increase from 5 million to 13 million by 2050. This shift,
along with increases in the number of American elders afflicted by mild cognitive impairment
and dementia, will lead to massive increases in caregiving costs for cognitively impaired
individuals, not to mention losses in the capacity to live independently and to make productive
contributions to society.
In order to sustain a society of productive elders in the face of current population aging
and health trends, interventions which can effectively curb the effects of cognitive aging must be
explored and evaluated. As a natural developmental impulse of older adulthood, generativity
may represent an apt and readily available means of promoting engagement with life and
sustaining high levels of cognitive functioning during this phase of the life course. Though
short-term experimental interventions suggest that generativity may enhance cognitive vitality
among older adults (Carlson et al., 2008, 2009, 2015), its associations with cognition have yet to
be assessed in longitudinal contexts, thereby limiting the utility of these investigations in terms
of their ability to predict whether generativity is associated with improved cognitive functioning
over longer periods of time. In order to thoroughly evaluate these associations, longitudinal
analyses using large-scale and nationally representative generativity and cognition data are an
important step forward. The results of such longitudinal investigations will be presented in the
coming chapters. However, prior to examining these results, it is first necessary to explore the
4

conceptual origins of generativity, its developmental significance in older adulthood, and its role
as a potential determinant of health and cognitive vitality in later life.
1B. The Origins of Generativity
Generativity is a relatively new paradigm in human developmental psychology. The term
“generativity” was first coined by psychologist Erik Erikson in his 1950 model of lifespan
development. In Erikson’s conceptualization, generativity represented the seventh of eight
developmental “tasks” of life, and he defined it as “the interest in establishing and guiding the
next generation” (Erikson, 1950, p. 231). Erikson paired generativity with its theoretical
antithesis, stagnation, which represented the “interpersonal impoverishment” (Erikson, 1950, p.
231) that results from a preoccupation with the self and a lack of concern for the needs of others.
Within Erikson’s model, the generativity—stagnation crisis was positioned in the years of
middle adulthood, just after the intimacy—isolation crisis of early adulthood and just prior to the
integrity—despair crisis of older adulthood.
After the publication of Erikson’s original commentaries on generativity, further
conceptual development of the construct occurred slowly. Only about 25 years ago did the first
full-fledged conceptual model of generativity emerge in the literature. This model, offered by
the team of McAdams and de St. Aubin (1992), posited that generativity was made up of seven
constituent psychosocial features, including (1) cultural demand for generativity; (2) one’s inner
desire for generativity (subdivided into a desire for symbolic immortality and a desire to feel
needed by others); (3) concern for future generations; (4) a “belief in the species” (Erikson,
1963, p. 267), which reflects individuals’ belief in the inherent goodness of humanity and which
prompts them to nurture and further it through generativity; (5) generative commitments,
including generative goals and decisions; (6) generative actions; and finally (7) narration of
5

generativity within the broader life course. This model, though illuminating in its identification
of distinct generativity parameters, did little to encourage conceptual parsimony in the study of
generativity. However, in examining the seven features set forth here, it is useful to classify each
item according to whether it represents a psychological impulse encompassing one’s intrinsic
needs, feelings, and beliefs (e.g., #2, 3, 4, and 7) or whether it involves social impulses and
dynamics which potentially drive generative behavior (e.g., #1, 5, and 6). Though it is
impossible to completely disentangle psychological and social influences in the study of human
development, this classification scheme of psychological generativity (involving an individual’s
generative self-concept) versus social generativity (involving one’s generative activities and
relationships) is embraced in the current manuscript and may represent a useful tool for future
studies of generativity. In fact, some scholars have already made use of this dichotomization in
their research on generativity. Stewart and Vandewater (1998), for instance, hypothesized that
generativity can be broken down into the psychological components of desire and
accomplishment and the social component of one’s involvement in actual generative activity.
Despite the conceptual development noted above, generativity’s initial designation as a
midlife phenomenon led many theorists to overlook its developmental salience in other parts of
the life course, particularly older adulthood. To some extent, this pattern of theoretical neglect
endures today, although a number of scholars have explored later life generativity in greater
depth. Work by Kotre (1984), for instance, sought to expand Erikson’s vision, moving beyond
stage-based conceptualizations of generativity towards a more expansive temporal understanding
of the construct in the context of the adult life course. Kotre (1984) argued that generativity
embodies “a desire to invest one’s substance in forms of life and work that will outlive the self”
(p. 10). From this perspective, generativity may serve a particularly critical function in older
6

adulthood given that it reinforces one’s sense of a personal contributory legacy, which in turn
confers a sense of symbolic immortality in the face of death (in keeping with terror management
theory, per Becker, 1973). As noted above, McAdams and de St. Aubin (1992) expounded on
this notion in their conceptual model of generativity, which posited that one’s “inner desire” for
generativity is prompted by both a desire for symbolic immortality and a desire to feel needed by
others. Subsequent empirical investigations have corroborated generativity’s role as an
important developmental aspiration of later life, especially those which have evaluated the effect
of mortality awareness on generativity. For example, reminders of death have been shown to
evoke generative strivings among older adults (Maxfield et al., 2014), as have constraints on
future time perspective among elders in studies of socioemotional selectivity theory (Lang &
Carstensen, 2002). Notably, such effects were not observed among younger adults in these
investigations, suggesting that generativity may occupy a unique developmental function in the
years of later life.
Intriguingly, Erikson himself, along with Erikson and Kivnick, suggested in 1986 that
generative strivings do persist into older adulthood and that one’s generative pursuits during this
part of the life course may be less prescriptive and more personally fulfilling than those available
in earlier life. Accordingly, contemporary definitions of generativity are more flexible than
Erikson’s original 1950 conceptualization. Today, generativity is most productively defined as a
set of feeling states (i.e., one’s generative self-concept) and behaviors (i.e., one’s generative
activities) which derives from a desire to positively contribute to the welfare of others. It is
important to acknowledge that generativity is not driven by purely selfless motives, but has
instead been described as occurring through a combination of both selfless and self-concerned
impulses (McAdams & Logan, 2003). Those who engage in generativity make productive
7

contributions to the world around them in a manner that is both selfless (through altruistic, other-
oriented contributory behavior) and self-preserving (through the establishment of a personal
contributory legacy that will effectively “outlive the self” [Kotre, 1984, p. 10] and reinforce a
sense of purpose and meaning in the latter years of life). The recipient of generative concern or
actions may be a younger individual, but this is not an absolute prerequisite for generativity.
Though often associated with the activity of parenting, generativity does not necessarily entail
conceiving and raising children, as both parents and childless individuals can engage in
generativity and have been shown to reap equivalent psychological rewards from the same
(Rothrauff & Cooney, 2008). Moreover, a generative actor is not necessarily an individual in
midlife, as both young adults (McGuire, 2005), those in the very final stages of life (Hauser,
2013), and those in between (Keyes & Ryff, 1998; McAdams, de St. Aubin, & Logan, 1993;
Stewart & Vandewater, 1998) have been shown to engage in generativity.
1C. Generativity and Health
Recognition of the developmental significance of one’s generative (i.e., social
contributory) identity in older adulthood has led scholars to hypothesize that generativity may
facilitate successful aging (Fisher, 1995; Villar, 2012). In support of this premise, a number of
empirical investigations, both observational and experimental in nature, have been initiated to
explore associations between generativity and health. Observational studies, which have tended
to focus on aspects of elders’ generative self-concept, indicate that a greater sense of generative
contributory worth in later life is associated with better psychological and physical health. More
positive self-perceptions of one’s generativity, for instance, are associated with a number of
indicators of psychological well-being among elders, including greater feelings of self-efficacy, a
greater sense of mastery, greater social connectedness, greater overall social integration, and
8

fewer depressive symptoms (Gruenewald et al., 2007, 2009), as well as autonomy, self-
acceptance, purpose in life, positivity toward others, and personal growth (Rothrauff & Cooney,
2008). In terms of physical well-being, older adults who report greater feelings of usefulness to
others experience more favorable trajectories of physical functioning and health with advancing
age, including lower odds of disability and impairment in activities of daily living, lower risk of
placement in institutional care facilities, and lower overall mortality risk (Grand et al., 1988,
1990; Gruenewald et al., 2007, 2009, 2012; Okamoto & Tanaka, 2004; Pitkala et al., 2004).
Evidence from the experimental realm likewise supports the idea that generativity may promote
successful aging. For example, volunteers in the Experience Corps program, a generativity-
based randomized controlled intervention in which older adults engage as mentors and tutors to
underprivileged schoolchildren for no less than 15 hours per week, have shown a number of
improvements in their overall health and well-being. Relative to wait-listed and matched
controls, Experience Corps volunteers display lesser impairment in activities of daily living,
higher levels of physical activity, fewer depressive symptoms, and higher levels of perceived
social support (Fried et al., 2004; Hong & Morrow-Howell, 2010), as well as a qualitatively
greater sense of personal achievement and enjoyment of life (Varma et al., 2014). In addition,
engagement in the Experience Corps program led to overall enhancements in self-perceptions of
generativity among volunteers as opposed to the wait-listed control group (Gruenewald et al.,
2016).
The health benefits of generativity in older adulthood do not appear to be limited to the
spheres of physical and psychological well-being, but also extend into the domain of cognitive
performance. Empirical evidence regarding generativity—cognition associations will be
discussed in detail in the following section, as well as theoretical models which may account for
9

these associations and mechanistic pathways linking generativity and cognitive function in later
life.
1D. Generativity and Cognition
1D(a). Empirical Evidence Linking Generativity with Cognition
Though analyses assessing generativity—cognition associations are few in number and
have been exclusively experimental in design, they nevertheless provide important insight into
the role that generativity may play in shaping cognitive function in older adulthood. These
studies will be explored below.
First among these investigations are Experience Corps studies of the impact of generative
activity engagement on cognitive performance among older adult program participants. In
comparison with wait-listed controls, Experience Corps volunteers exhibited better overall
executive function and memory performance (Carlson et al., 2008), increased cortical and
hippocampal volume (Carlson et al., 2015), and enhanced cortical plasticity and cognitive
reserve as assessed with functional magnetic resonance imaging (Carlson et al., 2009).
Importantly, the favorable neuroplastic changes that were observed among Experience Corps
volunteers occurred in brain regions that are subject to both anatomical and functional
deterioration with aging (e.g., prefrontal cortex, anterior cingulate cortex, hippocampus).
Together, these findings indicate that generative activity may stimulate beneficial changes in the
aging brain, effectively diminishing age-related cognitive decline. Though mediators of
generative activity—cognition associations have yet to be conclusively evaluated, Experience
Corps investigators hypothesize that the intensive generative activity fundamental to the
Experience Corps volunteering role may promote enhanced cognition through several pathways
(Fried et al., 2004), including (1) physical activity (through the daily commute to the
10

volunteering site, as well as through other volunteer activities that require walking, stair
climbing, etc.); (2) social engagement (through interaction with students, teachers, fellow
volunteers, etc.); and (3) cognitive stimulation (through novel and potentially cognitively
challenging activity in the volunteer role). More information regarding potential mechanisms
through which generativity may influence cognitive function will be discussed in section 1D(c)
below.
Secondly, an experimental priming investigation conducted in our laboratory showed that
elders’ conceptualizations of their generative contributory value may have robust effects on
cognitive performance (Hagood & Gruenewald, 2016). Older adult participants in this study
were randomly assigned to read either a positively-themed news article which highlighted the
generative contributions of elders or a negatively-themed article which described elders as a
societal burden with little contributory potential. Participants who read the negative priming
article recalled three fewer words on average in the post-priming verbal memory assessment in
comparison with those who read the positive article (d = 0.75), suggesting that sufficient
internalization of the primes occurred to affect cognitive processing. This finding indicates that
older adults’ perceptions of their generative contributory value may strongly influence cognitive
functioning in later life.
While compelling, the evidence presented above begs additional questions regarding the
relationship between generativity and cognition. First, why exactly is generativity expected to
influence cognition? And second, how may generativity actually be linked with cognitive
functioning? These questions will be respectively examined in sections 1D(b) and 1D(c) below.

11

1D(b). Theoretical Perspectives on Generativity—Cognition Associations
Several conceptual models from the gerontological literature are helpful in providing a
theoretical framework through which to understand potential associations between generativity
and cognition in older adulthood. These include: (1) the socioemotional selectivity theory of
Carstensen and colleagues (Carstensen, 1995; Carstensen, Isaacowitz, & Charles, 1999); (2) the
broaden-and-build theory of positive emotions (Fredrickson, 2004); and (3) the mental exercise
hypothesis of cognitive aging, known more colloquially as the “Use It or Lose It” hypothesis
(Salthouse, 2006; Schooler, 2007).
1D(b-1). Socioemotional selectivity theory. Socioemotional selectivity theory is a
developmental theory of social motivation which posits that one’s awareness of time boundaries,
or, put differently, one’s “future time perspective” (Lang & Carstensen, 2002, p. 125), dictates
individuals’ selection of social goals. That is, when individuals perceive their time boundaries to
be unlimited, as in youth, they are motivated to focus on the cultivation of novel relationships
and information which could be of use in the future. Conversely, when individuals perceive their
time boundaries to be limited, as in older adulthood, their social goals shift such that they tend to
focus on familiar and emotionally fulfilling relationships which promote emotional stability near
the end of life. As noted above, studies of socioemotional selectivity theory have shown that
constraints on future time perspective among older adults are associated with the prioritization of
goals related to generativity (Lang & Carstensen, 2002). This evidence suggests that
generativity may represent a prominent socioemotional goal of later life. A corollary of
socioemotional selectivity theory, the positivity effect (Carstensen & Mikels, 2005; Mather &
Carstensen, 2005; Reed & Carstensen, 2012), may serve as an important theoretical linkage
between generativity and cognitive functioning among older adults. The positivity effect refers
12

to an observable bias in elders’ cognitive processing for positive as opposed to negative
information. This tendency has been documented in tasks involving both attention and memory
(Mather & Carstensen, 2005), with older adults showing high levels of cognitive performance in
tasks involving the selection of and memory for positive information. Notably, these effects are
apparent even as abilities in other cognitive domains are observed to decline (Mather &
Carstensen, 2005). Though positivity effects in activities related to generativity have yet to be
investigated, it follows that elders’ prioritization of emotionally fulfilling goals such as
generativity would be accompanied by positivity effects in associated cognitive tasks among
generatively-oriented older adults. For instance, socioemotional selectivity theory would predict
that elders would be highly motivated to develop a sense of their generative contributory legacy
in later life as a means of promoting emotional well-being. In so doing, positivity-based memory
processes may be activated in order to allow elders to selectively recall past generative
achievements.
1D(b-2). Broaden-and-build theory. The broaden-and-build theory of positive
emotions was formally articulated in 2004 by psychologist Barbara Fredrickson. The broaden-
and-build theory posits that the experience of positive emotions effectively expands an
individual’s mindset to encompass new and potentially stimulating “thought—action”
possibilities, which thus broaden the sense of what is possible in one’s life. This broadened
mindset then hypothetically results in engagement with new ideas and with new activities and
relationships, which then build the individual’s portfolio of psychological resources and skills.
For older adults who have perhaps reached a point of transition in their lives due to retirement
from work or another type of role shift, generativity may represent a very real means of
broadening one’s sense of what the latter part of life can become. In broadening one’s mindset
13

to reconceptualize later life as a period of generative opportunity, elders can then identify
purposeful and meaningful roles that they wish to fulfill in their social worlds. In fulfilling these
roles, older adults then build psychological and social competency across multiple domains.
Indeed, the generativity research reveals that both generative activity and self-perceptions of
generative contributory value are associated with a greater sense of social connectedness (Fried
et al., 2004), fewer depressive symptoms, greater feelings of mastery and self-efficacy
(Gruenewald et al., 2007, 2009), as well as autonomy, self-acceptance, purpose in life, positivity
toward others, and personal growth (Rothrauff & Cooney, 2008). The experience of such
positive emotions through generativity, in turn, may promote better cognitive functioning.
Higher levels of positive affect, for instance, are known to predict better working memory and
decision-making abilities among elders (Carpenter, Peters, Västfjäll, & Isen, 2013). Higher
levels of positive emotion may also act to improve cognition by downregulating the production
of stress hormones which are known to cause cognitive performance deficits. Stress pathways
such as the hypothalamic-pituitary-adrenal (HPA) axis are activated in response to negative
feeling states such as fear, sadness, and stress, producing the stress hormone cortisol.
Chronically elevated cortisol levels are associated with cognitive impairments in the domains of
processing speed, verbal memory, verbal learning, and hand-eye coordination (Lee et al., 2007).
However, according to the broaden-and-build theory, engagement in psychologically gratifying
and enriching activities such as generativity would effectively suppress HPA activity through
positive emotions surges, thereby lowering systemic cortisol levels and promoting better
cognitive function.
1D(b-3). Mental exercise hypothesis. The mental exercise hypothesis refers to the
popular notion that one’s cognitive faculties are maintained only through use, or “mental
14

exercise.” This hypothesis is known among lay people as the “Use It or Lose It” maxim.
Though the validity of the mental exercise hypothesis has been brought into question over the
past 15 years (see Salthouse, 2006), other researchers argue that ongoing engagement in
cognitively stimulating activities may decrease the effects of age-related cognitive decline
(Hultsch, Hertzog, Small, & Dixon, 1999; Schooler, 2007). In support of the latter viewpoint,
some studies suggest that the repetitive activation of cortical neurons may actually have
protective effects on global cognitive function, whereas the repeated use of other cell types in the
human body may actually lead to their degeneration (Swaab, 1991). Despite the disputed nature
of the mental exercise hypothesis, generative activity affords an intriguing opportunity to test this
premise. It is possible that engagement in generative activities may promote the maintenance of
cognitive abilities over time among older adults, though the degree of impact may be highly
dependent upon the nature of the generative activity at hand. Caregiving, for instance, represents
a highly valuable, though highly stressful, social contributory role. Engagement in caregiving
may not foster the same benefits for cognition as would participation in a more novel,
cognitively stimulating, and gratifying activity such as serving as an academic mentor to
underprivileged children.
1D(c). Mechanisms Linking Generativity and Cognition
The mechanisms through which generativity – both in the form of generative activity and
in the form of one’s generative self-concept – may impact cognitive performance in older
adulthood are as of yet unknown. A number of potential mediating pathways exist, including
those of health status, physical activity, affective well-being, social engagement, and cognitive
stimulation. Each of these mediators will be discussed below, as well as evidence which
supports their associations with cognitive function. Figure 1 presents an overview of these
15

potential mediating pathways, as well as those which will be investigated subsequently in the
studies contained within this manuscript.

Figure 1-1. Overview of prospective mediating pathways between generativity and cognitive
function.

1D(c-1). Health status. Generativity among older adults may have important
downstream effects on global measures of health, such as level of impairment in activities of
daily living (ADLs), chronic disease burden, and body mass index. In theory, those who are
generative interact in a positive and productive way with the world around them, triggering
enhancements in affect and activity which can lead to such broader improvements in health and
functioning (Villar, 2012). While health status may ultimately represent a proxy for parameters
such as physical activity (see section 1D(c-2) below), affective well-being (see section 1D(c-3)
16

below), and sociodemographic characteristics such as sex, race, and income, it is nevertheless
important to independently investigate the potential mediating role of health status variables as
they relative to generativity in later life. Those who engage in generativity and perceive
themselves to be generatively valuable to others demonstrate better profiles of both physical and
psychological health, as documented in section 1C above. Related research from the generativity
domain of volunteering indicates that engagement in volunteering is associated with better self-
rated health, lesser overall functional impairment, lesser hypertension risk, and lesser risk of
death (Burr, Tavares, & Mutchler, 2011; Oman, Thoresen, & McMahon, 1999; Tang, 2009).
Improved health and well-being resulting from generative engagement may consequently operate
to promote better functioning in the sphere of cognition. Indeed, the literature reveals that
individuals’ health status is a robust predictor of their cognitive functioning. Level of ADL
impairment, for instance, displays a strong negative correlation with cognitive function among
older adults (Gill, Richardson, & Tinetti, 1995; Tuokko, Morris, & Ebert, 2005). Hypertension,
a prominent public health concern in the United States impacting more than 60 percent of those
65 and older (Centers for Disease Control and Prevention, 2016), is likewise associated with
cognitive decline (Kececi-Savan et al., 2016; Obisesan, 2009; Obisesan et al., 2008). Diabetes,
another chronic disease which disproportionately impacts older adults, has nearly quadrupled in
prevalence over the past 30 years (Centers for Disease Control and Prevention, 2014, 2015a) and
is associated with both declines in processing speed and the development of Alzheimer’s disease
(Arvanitakis, Wilson, Bienias, Evans, & Bennett, 2004). In addition, higher body mass index
across adulthood predicts deficits in executive function (Gunstad et al., 2007) and is also
associated with decreases in gray matter volume in prefrontal cortex (Walther, Birdsill, Glisky,
& Ryan, 2010).
17

1D(c-2) Physical activity. Generativity may operate to increase physical activity among
older adults through several means. First, it may prompt individuals to engage in contributory
activity which involves physical exertion (e.g., volunteering to build houses for Habitat for
Humanity). Second, a more positive generative self-concept and sense of one’s contributory
value may prompt older adults to be more physically active on principle as a means of
maintaining such value. Regardless, engagement in physical activity is known predictor of
cognitive function among older adults. Nearly all empirical investigations of physical activity—
cognition associations, including both observational and intervention-based studies, have shown
that physical exercise, particularly that which is aerobic in nature, is associated with improved
cognitive performance and reduced risk of neurodegenerative disease among older individuals
(for review, see Bherer, Erickson, & Liu-Ambrose, 2013). Physical activity, then, may prevent
age-related cognitive decline which is known to occur in biologically-based, fluid abilities, such
as processing speed, memory, and reasoning. Physical exercise is believed to act on the
biological substrates which facilitate these aspects of cognitive function, promoting favorable
neurotrophic and neuroplastic changes which can give rise to global enhancements in
performance. Research has shown that as little as six months of participation in aerobic exercise
can induce compensatory neurophysiological changes in the aging brain (Erickson & Kramer,
2009). These changes may include angiogenesis, synaptogenesis, and neurogenesis, all of which
are induced by physical exercise in aging animals (Eadie, Redila, & Christie, 2005; Isaacs,
Anderson, Alcantara, Black, & Greenough, 1992; van Praag, Shubert, Zhao, & Gage, 2005) and
which are thought to subserve cognitive function. In humans, aerobic exercise has been shown
to increase hippocampal volume and improve memory (Erickson et al., 2010). However, the
extent to which generative activity may impact cognitive function will depend upon whether the
18

generative activity in question involves legitimate physical exertion. Without such exertion,
physical activity may fail to mediate the generative activity—cognition association, or it may
mediate the association to a lesser degree.
1D(c-3). Affective well-being. Affective well-being may serve as another potential
mediator of the association between generativity and cognition among older individuals, as the
act of contributing meaningfully to others may have profound emotional and affective
significance during this part of the life course (see sections 1B and 1D(b-1) above). Generative
attitudes have shown positive associations with a number of indicators of emotional well-being
among older adults, including autonomy, mastery, personal growth, positivity towards others,
purpose in life, and self-acceptance (Rothrauff & Cooney, 2008; for review of well-being in
older adulthood, see Ryff, 1995). Similarly, it is logical to hypothesize that generative actions
would induce these favorable affective states, though the correlation between generative actions
and affect has not been thoroughly tested. Preliminary qualitative research from the Experience
Corps model suggests that older adult volunteers reap unique psychosocial rewards from the act
of making generative contributions (Varma et al., 2014), which may in turn promote affective
well-being. Such rewards include the opportunity to help children who may be in need, the
development of bonds with children, personal fulfillment derived from the volunteer role,
feelings of personal growth and achievement, a sense of heightened activity, and a sense of
increased social connectedness. These psychosocial benefits may then confer increases in the
affective indicators listed above (e.g., mastery, purpose in life, positivity, etc.). Quantitative
work from the Experience Corps program corroborates this idea, as generatively engaged
volunteers displayed fewer depressive symptoms than did wait-listed controls (Hong & Morrow-
Howell, 2010). Related work from the volunteering literature has also shown positive
19

associations between contributory activity and affective well-being. For instance, the number of
hours of annual volunteer work predicts greater overall happiness, higher life satisfaction, greater
self-esteem, higher levels of personal control, and fewer depressive symptoms among older
adults (Lum & Lightfoot, 2005; Morrow-Howell, Hinterlong, Rozario, & Tang, 2003; Thoits &
Hewitt, 2001; Van Willigen, 2000). Altruistic behavior has also been associated with a number
of favorable psychological outcomes, including higher levels of positive affect and better overall
mental health (Dulin & Hill, 2003; Post, 2005; Schwartz, Meisenhelder, Ma, & Reed, 2003).
If contributory and altruistic (e.g., generative) feeling states and behaviors presumably
lead to greater affective well-being, then it is reasonable to assume that generativity will promote
the same benefits for cognitive function that affective well-being is known to confer. More
positive affect, for instance, is associated with enhanced working memory performance and
better decision-making abilities among older adults (Carpenter, Peters, Västfjäll, & Isen, 2013),
as noted in section 1D(b-2) above. These effects may occur through executive control
mechanisms which detect the level of threat present in a particular piece of information and then
allocate cognitive resources appropriately. When that information is positive and emotionally
salient (as in the case of the positive feeling states that may result from generative engagement),
it appears to consume fewer cognitive resources at the cortical level than would negative and
potentially threatening information, which facilitates enhancements in executive function,
including working memory, planning, reasoning, task alternation, and inhibitory processes
(Pessoa, 2009). The positive emotions attained through generative activity may also lead to
improved cognition by downregulating the production of stress hormones that are known to
hamper performance. As discussed previously, the body’s physiological stress pathways (e.g.,
the HPA axis) are activated in response to negative feeling states such as fear, sadness, and
20

stress. The chief downstream product of the HPA axis is the stress hormone cortisol, which has
known detrimental effects on cognitive function (for review, see McEwen & Sapolsky, 1995).
Elevated levels of salivary cortisol are associated with impaired performance across multiple
spheres of performance, including processing speed, verbal memory, verbal learning, and hand-
eye coordination (Lee et al., 2007). Cortisol’s effects on cognitive function are particularly
robust for memory. High levels of cortisol can induce neuronal dysfunction and death in the
aging hippocampus, leading to performance deficits in both learning and memory. Lastly, it is
possible that the positive feelings which result from older adults’ generativity can lead to
enhancements in other domains of psychological well-being that are associated with cognition,
including self-perceptions of aging (e.g., satisfaction with aging, subjective age). More positive
self-perceptions of aging have been empirically associated with both better memory and better
executive function (Stephan, Caudroit, Jaconelli, & Terracciano, 2013).
1D(c-4). Social activity. Social engagement may function as an additional mediator of
the association between generativity and cognition among older individuals. Generative
activities inherently involve some degree of social interaction, and older adults who are
generatively engaged may thus reap many of the known benefits of social engagement for
cognitive function. Higher levels of social activity, for instance, are associated with significantly
reduced risk of impairment in episodic memory, semantic memory, working memory, processing
speed, and visuospatial abilities (James, Wilson, Barnes, & Bennett, 2011). Social connectivity
and support from others, which may be gained from social activity, are associated with improved
cognitive performance in the domains of executive function and memory, lesser overall cognitive
decline, and lesser risk of dementia (Barnes, Mendes de Leon, Wilson, Bienias, & Evans, 2004;
Crooks, Lubben, Petitti, Little, & Chiu, 2008; Holtzman et al., 2004; Seeman, Lusignolo, Albert,
21

& Berkman, 2001; Seeman et al., 2011). Conversely, low feelings of social connectedness and
low levels of social engagement are predictive of cognitive decline among older adults
(Zunzunegui, Alvarado, Del Ser, and Otero, 2003). Together, these findings suggest that the
social connectivity which is to be had from generative engagement may bolster cognitive
function in later life. Mechanistically, this may occur by decreasing stress responsiveness and
downregulating stress pathways. As explained previously, such downregulation effectively
decreases the amount of circulating cortisol in the body, which might otherwise impair cognitive
function. Social interconnectedness may also act to decrease blood pressure and increase heart
rate variability, both of which are associated with better profiles of cardiovascular health. These
more favorable cardiovascular profiles are in turn associated with better cognitive performance
and with lesser risk of cognitive impairment with aging (Albinet, Boucard, Bouquet, &
Audiffren, 2010; Elias, Wolf, D’Agostino, Cobb, and White, 1993; Qiu, Winblad, & Fratiglioni,
2005).
1D(c-5). Cognitive stimulation. Lastly, it is important to recognize that cognitive
stimulation may also mediate the association between generativity and cognitive performance.
Generative activities which involve novelty or complexity (i.e., which are cognitively
stimulating) may exert strong effects on downstream cognitive performance, whereas those
which do not may have negligible effects on cognition. Thus the cognitive complexity of the
generative activity at hand will likely dictate the extent to which cognitive stimulation mediates
generative activity—cognition associations. This is an important point of differentiation
amongst generative activities – that not all generative contributions will necessarily entail
cognitive stimulation. In particular, those which are rote or obligatory in nature (e.g., long-term
caregiving for a chronically ill spouse or relative) may involve very little, if any, cognitive
22

stimulation. However, those which are new and challenging may involve substantial levels of
stimulation. Prior research on the Senior Odyssey program, a lifestyle intervention in which
teams of older adults are routinely exposed to novel and substantively complex problems in a
competitive context, has shown that involvement in the intervention is associated with improved
processing speed, reasoning, visuospatial ability, verbal fluency, and working memory (Stine-
Morrow, Parisi, Morrow, & Park, 2008). Findings from the Experience Corps program also
suggest that cognitive stimulation may bolster cognitive performance, as older adult volunteers
involved in the novel activities of the Experience Corps intervention (e.g., navigating to and
within the volunteering site, teaching and mentoring children across multiple academic areas,
interacting with teachers and other volunteers, etc.) showed improved executive function and
memory, as well as favorable neuroplastic changes in cortex (Carlson et al., 2008, 2009, 2015).
These findings suggest that generative activities may have the capacity to strongly impact
cognitive function, but that the degree of impact depends upon the frequency and intensity of
cognitive stimulation involved in the generative activity at hand.
Taken together, the mediators discussed above provide additional impetus to investigate
associations between generativity and cognition. Above and beyond their role as mediators of
generativity—cognition associations, they are also of interest insofar as they represent distinct
outcomes which may be associated with generativity independent of cognition. The current set
of dissertation studies affords an excellent opportunity to assess such associations using large-
scale, nationally representative survey data, not to mention the more focal association between
generativity and cognition function in older adulthood.

23

1E. Dissertation Objectives
The proposed set of dissertation studies will address a number of important gaps in our
knowledge regarding how generativity might be linked to cognitive vitality over time in older
adulthood. At a broad level, this body of work will seek to evaluate in a comparative context
whether it is one’s generative self-concept (i.e., one’s generative characteristics and feelings of
generative contributory value) or one’s generative actions per se which are more strongly
associated with cognition over time. This goal was achieved through a series of three
investigations with the following specific aims. First, Study 1 examined longitudinal
associations between elders’ generative self-concept and cognition using data from the National
Survey of Midlife Development (MIDUS). This investigation capitalized on the availability of
assessments of self-perceived generativity (both self-perceived generative contributions to others
and self-perceived generative characteristics) among older adult participants in the MIDUS
study. MIDUS also offers performance-based measures of executive function and episodic
memory, which were explored as composite representations of cognitive ability in the present
analyses. These measures permitted exploration of how longitudinal change in generativity is
linked to levels of cognitive function across older adulthood. Study 2 investigated longitudinal
associations between generative activity and cognition among older adults in the Health and
Retirement Study (HRS). HRS features a longitudinal, repeated-measurement design which
facilitates the prospective prediction of cognitive function in relation to levels of generative
engagement over time, as well as the examination of potential converse associations between the
two (i.e., prospective prediction of generative engagement as a function of levels of cognitive
performance over time). Finally, Study 3 again took advantage of data from the MIDUS study to
investigate longitudinal associations between both generative self-concept and cognition and
24

generative activity and cognition. Unlike either of the previous studies, this investigation
facilitated robust empirical comparisons of the independent and interactive influences of
generative activity engagement and generative self-concept on cognitive performance over time.
Studies 1—3 also concomitantly evaluated mechanistic pathways through which generativity
may be linked to cognitive performance in order to inform our understanding of how one’s
generative activity and generative self-concept might impact cognition. Altogether, these
investigations aimed not only to enhance our understanding of the associations between
generativity and cognitive function in older adulthood, but to advance our knowledge of the
relative and interacting influences of generative activity and generative self-concept on
functional outcomes in later life.

25

CHAPTER 2: STUDY 1

2A. Background
When scholars of human development attempt to convey what it means to be generative,
they very often rely on definitions of generativity which involve contributory activity per se.
Often cited as examples of generativity are the activities of parenting, volunteering, and
caregiving. Each of these activities represents a dynamic form of generative exchange in which
a generative giver seeks to provide for the well-being of the generative care recipient. Though
these and other forms of generative activity are fundamental to the development of one’s
generative identity, they do not encapsulate it. In a very real sense, individuals may also live out
their generativity through their own mental perceptions of their generative character and
generative contributory worth. These perceptions of generativity will be collectively referred to
here as one’s generative self-concept.
For older adults, the attainment of a positive generative self-concept may represent a vital
aspect of healthy development, especially as individuals approach the end of life and begin to
grapple with the broader meaning, purpose, and impact of their lives. One’s generative self-
concept is believed to have strong emotional and affective significance in later life, particularly
given that the establishment of a sense of generative worth equates to the extension of the self
beyond one’s own death. As noted previously, generativity has been defined as “a desire to
invest one’s substance in forms of life and work that will outlive the self” (Kotre, 1984, p. 10).
Generativity is also believed to be motivated by what scholars McAdams & de St. Aubin (1992)
have called “a desire for symbolic immortality” (p. 1005). This sense of immortality, in turn,
may provide emotional solace in the latter years of life as individuals approach death.
Intriguingly, prior empirical work has shown that the relationship between generativity and
26

psychological well-being is mediated by one’s desire for symbolic immortality (Huta & Zuroff,
2007). Those who attain higher levels of psychological well-being through the development of
a robust generative self-concept may then be better positioned to age successfully. As evidence
of this, noted gerontologists Baltes and Baltes (1990) have cited generativity as a manifestation
of optimal later life development in their writings on the psychological aspects of successful
aging.
As previously reviewed in sections 1C and 1D, generative self-concept demonstrates
significant associations with indicators of health and well-being in older adulthood. For
instance, higher self-perceptions of generative contributory value among elders predict lower risk
of functional disability, lower risk of placement in institutional care facilities, and lower risk of
death (Grand et al., 1988, 1990; Gruenewald et al., 2007, 2009, 2012; Okamoto & Tanaka, 2004;
Pitkala et al., 2004). A higher sense of one’s self-perceived generative character and
contributions also predicts better emotional well-being, including greater feelings of self-
efficacy, a greater sense of mastery, greater social connectedness, greater overall social
integration, and fewer depressive symptoms (Gruenewald et al., 2007, 2009), as well as
autonomy, self-acceptance, purpose in life, positivity toward others, and personal growth
(Rothrauff & Cooney, 2008). Though generative self-concept—cognition associations have yet
to be empirically evaluated in observational and longitudinal research, it is hypothesized that
such associations would mirror those of the physical and psychological health outcomes cited
above. That is, a more positive generative self-concept would theoretically promote better
cognitive functioning among older individuals over time.

27

2B. Significance
Though individuals possess a considerable degree of autonomy in shaping their
generative character and sense of social contributory value, it is important to note that one’s
generative self-concept does not develop in a social vacuum. Societal perceptions of elders’
collective contributory characteristics and contributory value invariably affect individual
perceptions of the same. While the aging of the world’s population has brought discourse
regarding older adults’ generative contributory value to a global stage, much of that discourse is
patently negative, focusing on older individuals’ lack of contributory worth instead of their
exemplification of it. Today, crucial policy decisions regarding entitlement spending, the
distribution of health care resources, and the funding of community programs are made on the
basis of the assumption that older individuals fundamentally lack generative characteristics and
generative worth. In an article that appeared in The New Republic in 1988, journalist Henry
Fairlie remarked, “Something is wrong with a society that is willing to drain itself to foster such
an unproductive section of its population, one that does not even promise (as children do) one
day to be productive” (p. 19). Such sentiments reflect the unfounded, yet widespread belief that
being aged means being useless. In the United States, this perception is longstanding and has
been described as an unseen and culturally acceptable form of bigotry (Butler, 1969).
However unknowingly, the acceptance and dissemination of negative beliefs regarding
older adults’ generative contributory value may be doing significant harm to elders’ cognitive
well-being. Prior investigations of exposure to more generalized types of negative aging
stereotypes show that both short- and long-term contact with such ideologies elicits cognitive
deficits (Hess, Auman, Colcombe, & Rahhal, 2003; Hess, Emery, & Queen, 2009; Hess &
Hinson, 2006; Hess, Hinson, & Statham, 2004; Levy, 1996; Levy & Langer, 1994). In an
28

analogous manner, older adults’ generative self-concept may be significantly altered by exposure
to negative societal conceptualizations of elders’ contributory value. Over time, chronic
exposure to such negative conceptualizations may lead to appreciable levels of cognitive decline
– decline which is largely avoidable. To date, only one known empirical investigation has
sought to examine the impact of exposure to positive versus negative stereotypes regarding
elders’ contributory value on cognitive performance, and this study showed that older adults who
were exposed to such negative stereotypes experienced significant and immediate impairments in
their memory performance relative those exposed to positive stereotypes of elders’ contributory
value (Hagood & Gruenewald, 2016). This result suggests the alarming possibility that negative
societal beliefs about older adults’ social contributory value may be contributing to cognitive
decline over time among both present and future generations of elders.
As a preliminary step towards understanding the nature of associations between elders’
generative self-concept and cognitive functioning over time, the current study sought to examine
longitudinal associations between older Americans’ generative self-concept (specifically their
self-perceived generative contributions and self-perceived generative characteristics) and their
cognitive performance using data from the National Survey of Midlife Development, or MIDUS.
Additional detail regarding this investigation, including its methodology, aims, and measures,
will be provided subsequently in section 2C.
2C. Methods
2C(a). Dataset
The MIDUS study is a longitudinal survey of health and developmental trajectories
across the adult life course (Brim et al., 2016; Ryff et al., 2012; Ryff & Lachman, 2013). At the
baseline wave in 1995, a sample of adults ranging in age from 25 to 74 (n = 7,108) were enrolled
29

via random digit dialing (RDD) and targeted sampling of twins and siblings of RDD
respondents. A follow-up wave, MIDUS II (n = 4,963), was initiated in 2004 and included a
number of focused substudies in addition to the replication of the original measures surveyed in
MIDUS I. One such substudy, the MIDUS II Cognitive Project, assessed cognitive function via
the Brief Test of Adult Cognition by Telephone, or BTACT (Lachman & Tun, 2008; Tun &
Lachman, 2005; Tun & Lachman, 2006), and is the source of the cognitive performance
measures used in the present study. As cognitive function measures were first collected at
MIDUS II with the initiation of the Cognitive Project, such data is not available within MIDUS I
and thus could not be incorporated into the present analyses.
2C(b). Participants
As the primary goal of the current investigation was to determine whether one’s
generative self-concept is related to cognitive function in the second half of life, participants age
50 and older at the baseline wave of assessment (MIDUS I) were selected for inclusion in the
present analyses. For this age cohort, generativity is hypothesized to represent a central and
growing developmental concern, while declines in fluid intelligence are also becoming more
precipitous in nature given typical patterns of age-related decline across the life course (Hedden
& Garieli, 2004). The analytic sample was also limited to those participants who completed both
the MIDUS general survey at waves I and II, as well as the Cognitive Project at MIDUS II,
giving a total analytic sample of 1,552 individuals.
2C(c). Measures
2C(c-1). Generative self-concept. Individuals’ generative self-concept was dually
conceptualized as self-ratings of contributory behavior toward others (e.g., self-perceived
generative contributions) and self-ratings of generative character (e.g., self-perceived generative
30

characteristics). The operationalization of these constructs within the MIDUS survey is
discussed below.
Self-perceived generative contributions. The variable of self-rated generative
contributions reflects respondents’ perceptions of their current contributions to the well-being of
others in terms of emotional, instrumental, financial, and material support given. Respondents
were asked to rate their present level of contributions to others on an 11-point scale, with a score
of 0 representing one’s “worst possible contribution” and a score of 10 representing one’s “best
possible contribution.” This measure was surveyed at both the baseline (MIDUS I) and 10-year
follow-up (MIDUS II) waves. Change in self-reported generative contributions was computed
by subtracting MIDUS I scores from MIDUS II scores, yielding a continuous change score
ranging from -10 to +10.
Self-perceived generative characteristics. The variable of self-rated generative
characteristics reflects respondents’ perceptions of their own generative character (e.g., their
generative traits and achievements). This measure is an abbreviated 6-item version of the
original 20-item Loyola Generativity Scale, which was developed in 1992 by McAdams and de
St. Aubin. Respondents were asked to indicate their agreement (1 = “A lot”; 4 = “Not at all”)
with each of the following 6 statements regarding their generative character: (1) “Others would
say that you have made unique contributions to society”; (2) “You have important skills you can
pass along to others”; (3) “Many people come to you for advice”; (4) “You feel that other people
need you”; (5) “You have had a good influence on the lives of many people”; and (6) “You like
to teach things to people.” These six items were reverse coded and summed to produce a scale
measure of self-perceived generative character ranging from 6 (lowest self-perception of
generative character) to 24 (highest self-perception of generative character). The 6-item Loyola
31

measure has demonstrated high internal consistency amongst MIDUS I and II participants, with
reliability coefficients of 0.84 and 0.85, respectively (Brim et al., 2009a; Brim et al., 2009b). As
with the generative contributions variable, change in self-perceived generative characteristics
was calculated by subtracting MIDUS I scores from MIDUS II scores.
2C(c-2). Cognition. Cognitive function was assessed at MIDUS II with the Brief Test of
Adult Cognition by Telephone (BTACT). This short cognitive battery includes six subtests
which assess working memory, episodic verbal memory, verbal fluency, reasoning, processing
speed, and task switching ability via telephone (Lachman & Tun, 2008; Tun & Lachman, 2005;
Tun & Lachman, 2006). In comparison with in-person cognitive assessments, the BTACT
demonstrates criterion validity, and it has also shown good test-test reliability (Lachman,
Agrigoroaei, Tun, & Weaver, 2014). Factor analyses with the BTACT subtests revealed that a
two-factor solution fit the observed measurement loadings best, producing two distinct factors:
(1) episodic verbal memory; and (2) executive function (Lachman et al., 2014). Consistent with
this factor structure and with previous investigations utilizing the BTACT (Agrigoroaei &
Lachman, 2011; Seeman et al., 2011; Stephan et al., 2013), the present analyses combine
participant scores from the six BTACT subtests into composite measures of executive function
and episodic memory. Detailed descriptions of the subtests which comprise these measures are
provided in the BTACT documentation (Tun & Lachman, 2005; Lachman et al., 2014); brief
summaries of each are provided below.
Episodic memory composite score. The episodic memory composite score incorporates
standardized scores from the immediate and delayed iterations of the Rey Auditory-Verbal
Learning Test (RAVLT) as administered during the BTACT. During each administration of the
RAVLT, 15 words are read aloud to the participant at a pace of one word per second, and then
32

the participant is given one minute to recall as many words as possible. This assessment has a
possible range of 0—15 correctly recalled words. Scores from the immediate and delayed
administrations were standardized and averaged to produce a z-scored episodic memory
composite measure.
Executive function composite score. The executive function composite score was
calculated as the average of standardized scores from working memory, verbal fluency,
reasoning, processing speed, and task alternation assessments, giving a z-scored measure of
cognitive performance in this domain. Working memory was assessed with a backwards digits
test in which the participant listens to a list of numbers and is then asked to recall the list in
reverse order. The number of digits in the list gradually increases with each iteration of the test
(from two to eight digits), and the longest span of digits that is correctly recalled represents the
final score (range: 0—8). Verbal fluency was assessed with a listing assessment in which the
respondent is asked to list in one minute as many examples as possible from the category of
animals. The range of responses for this measure was restricted only by the time limit.
Reasoning was assessed using the number of correctly completed problem sets in a number
series completion test. In this test, participants are presented with a list of numbers and are asked
to provide the subsequent number in the series. A total of five lists of varying difficulty are
presented to the participant, and one point is awarded for each correctly completed number series
(range: 0—5). Processing speed was assessed using scores on a backwards counting test in
which the participant is asked to count backwards from 100. The final score for this measure
represents the number of correct responses reported. Task alternation was measured using a
stop/go test in which the participant is asked to respond with the word “stop” upon hearing the
word “red” and to respond with the word “go” upon hearing the word “green.” After this
33

exercise, participants were then prompted to give the reverse response (i.e., to say “go” when
presented with the word “red”). Finally, participants were tested on a composite task alternation
exercise via a prompt to switch back and forth between these congruent and incongruent
response modes. Task alternation was ultimately assessed by calculating the mean of response
times for the switching and non-switching trials.
2C(c-3). Covariates. The following variables were included as covariates in models of
associations between generative self-concept and the cognitive function measures described
above. Variables which were examined as prospective mediators of generativity—cognition
associations are noted in the summaries below. The latter set of mediating variables
encompassed all conceptual domains outlined in section 1D(c) above, with the exception of
cognitive stimulation, which was not queried within MIDUS. In addition, the generative activity
of volunteering was investigated as a potential mediator of generative self-concept—cognition
associations as noted below.
Sociodemographic factors. Age, education, race, and sex were included in the present
models as sociodemographic covariates. Age was represented as a continuous variable, while
sex, race, and level of educational attainment were represented categorically (i.e., male/female;
white/non-white; high school degree or less/some college or more).
Health status. Body mass index (BMI), the number of major health conditions, and the
number of impairments in activities of daily living (ADLs) were included as health status
covariates and as potential mediators of generativity—cognition associations. BMI was
calculated as the respondent’s weight in kilograms divided by height in meters squared. The
number of major health conditions was calculated as in previous MIDUS investigations (e.g.,
Gruenewald et al., 2012) as the sum of nine health conditions, including AIDS, autoimmune
34

disorders, cancer, diabetes, heart disease, hypertension, lung problems, neurological disorders,
and stroke. The number of major conditions was topcoded at 5 conditions. The number of ADL
impairments was calculated as the sum of limitations in the following activities on a reverse-
coded scale of 1 (“Not at all”) to 4 (“A lot”): Lifting or carrying groceries; bathing or dressing;
climbing several flights of stairs; bending, kneeling, or stooping; walking several blocks; and
moderate-intensity activities such as household chores. Change in these variables was calculated
as the difference in scores from the MIDUS I assessment to the MIDUS II assessment ten years
later.
Health behavior. Level of physical activity was included in the current analyses as a
health behavior mediator, while smoking status was included as a prospective covariate.
Physical activity at both MIDUS I and MIDUS II was calculated as the sum of moderate and
vigorous activity, both of which were measured on a reverse-coded scale of 1 (“Never”) to 6
(“Several times a week or more”). These measures were summed to give composite measures of
overall physical activity, with vigorous physical activity weighted by a factor of 1.5 (as in
Gruenewald et al., 2012). Ten-year change in physical activity was established by subtracting
MIDUS I physical activity level from that at MIDUS II. Change in smoking status from MIDUS
I to MIDUS II was determined by creating status categories (e.g., current smoker, former
smoker, never smoked) and generating dummy variables from these categorizations reflecting
those who identified themselves as consistent non-smokers at both MIDUS I and II, those who
identified as former smokers by MIDUS II, and those who identified as consistent smokers at
both MIDUS I and II.
Social contact. Several potential generativity—cognition mediators were investigated in
the domain of social contact, including frequency of contact with (1) family, (2) friends, and (3)
35

neighbors. Contact with family members and friends was measured on a reverse-coded 8-point
scale (1 = “Never or hardly ever”; 8 = “Several times a day”), while contact with neighbors was
measured on reverse-coded 6-point scale (1 = “Never or hardly ever”; 6 = “Almost every day”).
For each of these dimensions, ten-year change in the level of contact was assessed by subtracting
MIDUS I contact frequencies from MIDUS II contact frequencies.
Social support. Potential social support mediators included the respective levels of
emotional and instrumental support given to others. Emotional support was assessed as the total
number of hours spent giving such support (e.g., comforting, listening, giving advice) each
month to a spouse/partner, parent, in-laws, children/grandchildren, and other family
members/friends. Instrumental support was assessed as the total number of hours each month
spent giving such support (e.g., giving unpaid assistance to others with needs such as household
chores and transportation) to a parent, in-laws, children/grandchildren, and other family
members/friends.
Affective well-being. Affective well-being covariates included positive and negative
affect, both of which were represented as six-item scale measures of the respective affective
dimensions. Positive affect was assessed as how often participants indicated feeling “cheerful,”
“in good spirits,” “extremely happy,” “calm and peaceful,” “satisfied,” and “full of life” over the
past 30 days, while negative affect was assessed as how often participants indicated feeling “so
sad nothing could cheer you up,” “nervous,” “restless or fidgety,” “hopeless,” “that everything
was an effort,” and “worthless” over that same period. Scores for each of the six items were
reverse coded and summed, producing a 5-point scale score (1 = “None of the time”; 5 = “All of
the time”). Respective change in positive and negative affect from baseline to follow-up was
calculated by subtracting affect scores at MIDUS I from affect scores at MIDUS II.
36

Productive engagement. Participants’ total engagement in volunteering was investigated
as a generative activity mediator in the current analyses, while one’s work status was included as
a covariate. Total hours of volunteering at both MIDUS I and MIDUS II was calculated by
summing participants’ total engagement in volunteering in hospitals or health care settings, in
schools, for political causes, and for other organizations or charities. Change in total
volunteering was calculated by subtracting the total hours of engagement at MIDUS I from that
at MIDUS II. Change in work status (i.e., paid employment status) from MIDUS I to MIDUS II
was determined by creating status categories reflecting ongoing employment from baseline to
follow-up, no employment during this period, dropping employment during this period, and
adding employment during this period. Dummy variables were then created to reflect changes in
work status from baseline to follow-up.
2C(d). Analyses
Study 1 utilized full path analyses with maximum likelihood estimation to model
associations between change in the following generative self-concept predictors from MIDUS I
to MIDUS II and cognitive performance in the following domains at MIDUS II:
(1) Self-perceived generative contributions and episodic memory
(2) Self-perceived generative contributions and executive function
(3) Self-perceived generative characteristics and episodic memory
(4) Self-perceived generative characteristics and executive function
For each of the four sets of analyses, regression models of generativity—cognition
associations were produced using the statistical tools available in MPlus (version 7.4), with tests
of multiple mediation incorporated for each analysis as shown in Figure 2-1. Note that each
analysis controlled for baseline generativity, as well as age, sex, race, education, work status,
37

partner status, and smoking behavior. Maximum likelihood estimation with robust standard
errors was used to address the small degree of incomplete data (0.19%—9.15%) for the study
variables.

Figure 2-1. Outline of path modeling strategy utilized in Study 1 to assess generative self-
concept—cognition associations.

2D. Aims & Hypotheses
Using the path modeling strategy described above, Study 1 sought to achieve the
following five aims and tested the following four hypotheses:
(1) Longitudinal associations between change in self-perceptions of generative
contributions to others from MIDUS I to II were investigated as a predictor of
episodic memory at MIDUS II (Aim 1), with the expectation that generativity—
38

cognition associations would be significant and positive, such that more positive
change in elders’ self-perceptions of their generative contributory value would
predict higher levels of memory performance at MIDUS II (Hypothesis 1).
(2) Longitudinal associations between change in self-perceptions of generative
contributions to others from MIDUS I to II were investigated as a predictor of
executive function at MIDUS II (Aim 2), with the expectation that generativity—
cognition associations would be significant and positive, such that more positive
change in elders’ self-perceptions of their generative contributory value would
predict higher levels of executive function at MIDUS II (Hypothesis 2).
(3) Longitudinal associations between change in self-perceptions of generative
characteristics from MIDUS I to II were investigated as a predictor of episodic
memory at MIDUS II (Aim 3), with the expectation that generativity—cognition
associations would be significant and positive, such that more positive change in
elders’ self-perceptions of their generative characteristics would predict higher
levels of memory performance at MIDUS II (Hypothesis 3).
(4) Longitudinal associations between change in self-perceptions of generative
characteristics from MIDUS I to II were investigated as a predictor of executive
function at MIDUS II (Aim 4), with the expectation that generativity—cognition
associations would be significant and positive, such that more positive change in
elders’ self-perceptions of their generative characteristics would predict higher
levels of executive function at MIDUS II (Hypothesis 4).
(5) Indicators from the five hypothesized mediating pathways identified in Figure 2-1
(health status, physical activity, affective well-being, social activity, and the
39

generative activity of volunteering) were explored as mediators of generative self-
concept—cognition associations (Aim 5). No a priori hypotheses were specified with
regard to the relative strength of these mediators in terms of their associations with
elders’ generativity and cognitive function. Instead, differences in the extent to which
they individually mediate these associations were investigated in an exploratory
manner.
2E. Results
2E(a). Descriptive Results
Descriptive statistics for all variables incorporated in the current analyses are presented in
Table 2-1. The average age of respondents in the sample was approximately 60 years at the
baseline MIDUS I assessment. The sample was predominantly white, with a slightly greater
proportion of females than males. About 60 percent of the sample attained a high school
diploma or less. In terms of health status at baseline, the mean body mass index of the sample
was 27, and both the number of major health conditions and the number of ADL impairments
were approximately 1. Respondents tended to show slight increases in these health status
measures over the 10-year follow-up period. The sample demonstrated mid-range levels of
average physical activity and tended to decrease in total physical activity over follow-up. In
addition, the respondents were evenly split between those who were consistent non-smokers over
the follow-up period (45%) and those who were former smokers by follow-up (45%), with
consistent smokers comprising a lesser proportion (10%) of the sample.
In terms of affective well-being, most respondents reported relatively low levels of
negative affect and relatively high levels of positive affect at baseline, with slight decreases in
negative affect and slight increases in positive affect observed over the follow-up period. The
40

majority of the sample maintained a marital or cohabitating partner over the follow-up period
(67%). The sample also displayed moderate to high levels of average contact with family,
friends, and neighbors at baseline, values which remained relatively stable over the 10-year
follow-up period. Respondents provided low to mid-range levels of emotional and instrumental
support to others at baseline, and these levels of support provision tended to increase slightly
over the follow-up period, particularly in the domain of instrumental support.
Most of the sample was not working for pay at either baseline or follow-up (46%), while
21 percent reported working at both baseline and follow-up, 28 percent had stopped working by
follow-up, and 5 percent had started working by follow-up. In terms of engagement in
volunteering, the sample averaged 7 hours of volunteering per month and showed slight
increases in total monthly volunteering over the follow-up period.
Self-reported generative contributions to others at baseline were above the midpoint of
the 0—10 scale (M = 6.81; SD = 2.18). On average, respondents tended to decrease slightly in
their perceived generative contributions over the 10-year follow-up period from MIDUS I to
MIDUS II (M = -0.33; SD = 2.44). In terms of self-perceived generative characteristics,
respondents displayed moderate to high self-ratings, averaging 17.13 (SD = 3.81) across the
possible range of 6 to 24 points. Over the follow-up period, respondents declined very slightly in
their self-ratings of generative characteristics (M = -0.23; SD = 3.17).
Finally, respondents demonstrated mid-range scores for both the z-scored episodic
memory composite measure (M = -0.29; SD = 0.99) and the executive function composite
measure (M = -0.35; SD = 0.91) at the 10-year MIDUS II follow-up assessment.

41

Table 2-1. Study 1 Descriptive Statistics; MIDUS I (1995—1996) & MIDUS II (2004—
2006); n = 1,552

Generative self-concept variables n Mean (SD) Range
Baseline self-perceived generative
contributions 1,513 6.81 (2.18) 0—10
Change in self-perceived generative
contributions 1,434 -0.33 (2.44) -10—10

Baseline self-perceived generative
characteristics 1,531 17.13 (3.81) 6—24
Change in self-perceived generative
characteristics 1,496 -0.23 (3.17) -11—18

Sociodemographic variables n Mean (SD) Range
Age (at baseline) 1,552 59.64 (6.75) 50—75
Gender: 1,552

Female 863 (55.6%)

Male 689 (44.4%)

Race: 1,531

White 1,448 (94.6%)

Non-White 83 (5.4%)

Education: 1,548

High school or less 915 (59.1%)

Some college or more 633 (40.9%)

Health status variables n Mean (SD) Range
# Major health conditions at baseline 1,552 0.79 (0.91) 0—5
Change in # major health conditions
1,552
0.53 (0.98) -3—5
Body mass index at baseline
1,491
27.17 (4.83) 15.45—48.82
Change in body mass index
1,410
0.59 (2.80) -23.20—14.41
# ADL impairments at baseline
1,546
1.47 (0.66) 1—4
Change in # ADL impairments
1,538
0.33 (0.68) -2.67—3.00

Health behavior variables n Mean (SD) Range
Physical activity level at baseline 1,544 10.95 (3.34) 2.50—15.00
Change in physical activity level 1,492 -4.33 (3.74) -12.50—10.00
Change in smoking status: 1,541
Consistent non-smoker 687 (44.6%)
Former smoker 696 (45.2%)
Consistent smoker 158 (10.3%)

42

Social contact variables n Mean (SD) Range
Change in partner status: 1,552
Partner at both 1034 (66.6%)
No partner at either 342 (22.0%)
Gained partner 45 (2.9%)
Lost partner 131 (8.4%)
Freq. of contact with family at baseline 1,519 5.98 (1.49) 1—8
Change in freq. of contact with family
1,504 0.15 (1.60) -7.00—7.00
Freq. of contact with friends at baseline
1,536 5.57 (1.59) 1—8
Change in freq. of contact with friends
1,512 0.20 (1.72) -7.00—6.00
Freq. of contact with neighbors at
baseline 1,534 5.16 (1.14) 1—6
Change in freq. of contact with neighbors 1,512 0.01 (1.24) -5.00—5.00

Social support variables n Mean (SD) Range
Emotional support given at baseline 1,522 45.36 (57.64) 0—250
Change in emotional support given 1,469 1.97 (66.52) -250—247
Instrumental support given at baseline 1,529 5.77 (12.53) 0—100
Change in instrumental support given 1,479 12.20 (26.58) -100—100

Productive engagement variables n Mean (SD) Range
Volunteering hours at baseline 1,522 7.13 (16.08) 0—138
Change in volunteering hours 1,456 1.83 (20.39) -138—197
Change in work status: 1,543
No work 712 (46.1%)
Added work 72 (4.7%)
Dropped work 434 (28.1%)
Working at both 325 (21.1%)

Affective well-being variables n Mean (SD) Range

Baseline negative affect 1,545 1.43 (0.54) 1—5
Change in negative affect 1,523 -0.01 (0.51) -2.83—3.00
Baseline positive affect 1,546 3.50 (0.69) 1—5
Change in positive affect 1,538 0.06 (0.65) -3.00—4.00

Cognitive performance variables n Mean (SD) Range

Executive function 1,549 -0.35 (0.91) -4.64—2.49
Episodic memory 1,546 -0.29 (0.99) -2.59—3.60

43

2E(b). Results of Path Analyses
2E(b-1). Hypothesis 1: Change in self-perceptions of generative contributions will
positively predict episodic memory at follow-up. As illustrated in Figure 2-2, 10-year change
in self-perceived generative contributions to others was significantly predictive of episodic
memory performance at the MIDUS II assessment ( β = 0.104, B = 0.042; p < 0.001), thus
corroborating Hypothesis 1. Though not shown in Figure 2-2, baseline levels of self-perceived
contributions likewise predicted memory performance ( β = 0.066, B = 0.029; p = 0.031).
The direct, unmediated effect of self-perceived generative contributions on memory
performance occupied a substantial proportion (91%) of the total effect. While the total indirect
effect of the prospective mediators did not reach significance at the level of p = 0.005, change in
self-perceived generative contributions was significantly associated with several of these
variables. Notable among these are change in positive affect ( β = 0.126, B = 0.034; p < 0.001),
change in negative affect ( β = -0.110, B = -0.023; p < 0.001), change in contact with friends ( β =
0.111, B = 0.078; p < 0.001), change in emotional support provided to others ( β = 0.122, B =
3.33; p < 0.001), change in instrumental support provided to others ( β = 0.177, B = 1.92; p <
0.001), and change in frequency of volunteering ( β = 0.208, B = 1.73; p < 0.001).

44

Figure 2-2. Path model of associations between change in self-perceived generative
contributions and episodic memory performance. Model adjusted for age, sex, race, education,
work status, partner status, smoking behavior, and baseline self-perceptions of generativity.
Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

2E(b-2). Hypothesis 2: Change in self-perceptions of generative contributions will
positively predict executive function at follow-up. Figure 2-3 demonstrates the results of
generative contributions—executive function path analyses. As shown, change in self-
perceptions of one’s generative contributions to others from MIDUS I to MIDUS II significantly
predicted executive function at MIDUS II ( β = 0.077, B = 0.028; p = 0.006), providing support
for Hypothesis 2. Though not shown in Figure 2-3, baseline levels of self-perceived
contributions were not associated with executive function at the 10-year follow-up.
The direct, unmediated effect of self-perceived generative contributions on executive
function accounted for 88% of the total effect, with the remaining 11% attributable to the indirect
effects. As with the episodic memory outcome, the aggregate indirect effects were not
45

significant. In addition, self-perceived generative contributions were again predictive of several
other indicators of health and well-being among MIDUS respondents, including change in
positive affect ( β = 0.126, B = 0.034; p < 0.001), change in negative affect ( β = -0.111, B = -
0.023; p < 0.001), change in contact with friends ( β = 0.111, B = 0.078; p < 0.001), change in
emotional support provided to others ( β = 0.122, B = 3.33; p < 0.001), change in instrumental
support provided to others ( β = 0.178, B = 1.93; p < 0.001), and change in volunteering ( β =
0.207, B = 1.73; p < 0.001).

Figure 2-3. Path model of associations between change in self-perceived generative
contributions and executive function. Model adjusted for age, sex, race, education, work status,
partner status, smoking behavior, and baseline self-perceptions of generativity. Standardized
coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

46

2E(b-3). Hypothesis 3: Change in self-perceptions of generative characteristics will
positively predict episodic memory at follow-up. As shown in Figure 2-4, change in self-
perceived generative characteristics from MIDUS I to MIDUS II did not significantly predict
episodic memory at MIDUS II, leaving Hypothesis 3 unsupported. Baseline self-perceived
generative characteristics did not predict episodic memory performance, either.
Despite the lack of significant associations between self-perceived generative
characteristics and memory performance, change in generative characteristics was significantly
predictive of several other variables, including change in positive affect ( β = 0.160, B = 0.033; p
< 0.001), change in negative affect ( β = -0.118, B = -0.019; p < 0.001), change in contact with
friends ( β = 0.139, B = 0.076; p < 0.001), change in instrumental support provided to others ( β =
0.120, B = 1.00; p < 0.001), and change in volunteering ( β = 0.162, B = 1.04; p < 0.001).

Figure 2-4. Path model of associations between change in self-perceived generative
characteristics and episodic memory performance. Model adjusted for age, sex, race, education,
47

work status, partner status, smoking behavior, and baseline self-perceptions of generativity.
Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

2E(b-4). Hypothesis 4: Change in self-perceptions of generative characteristics will
positively predict executive function at follow-up. As with the outcome of episodic memory
performance, change in self-perceived generative characteristics did not predict executive
function at the 10-year follow-up. This result (illustrated in Figure 2-5) fails to support
Hypothesis 4. In addition, baseline self-perceived generative characteristics at MIDUS I were
not associated with episodic memory.
Again, despite the lack of significant associations between self-perceived generative
characteristics and executive function, change in generative characteristics was significantly
predictive of change in positive affect ( β = 0.160, B = 0.033; p < 0.001), change in negative
affect ( β = -0.118, B = -0.019; p < 0.001), change in contact with friends ( β = 0.139, B = 0.076;
p < 0.001), change in instrumental support provided to others ( β = 0.120, B = 1.00; p < 0.001),
and change in volunteering ( β = 0.162, B = 1.04; p < 0.001).

48

Figure 2-5. Path model of associations between change in self-perceived generative
characteristics and executive function. Model adjusted for age, sex, race, education, work status,
partner status, smoking behavior, and baseline self-perceptions of generativity. Standardized
coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

The observed direct effects of the generative self-concept predictors investigated in
sections 2E(b-1) through 2E(b-4) are summarized in Table 2-2 below:
Table 2-2. Direct Effects of Generative Self-Concept Predictors on Cognitive Function;
MIDUS I (1995—1996) & MIDUS II (2004—2006); n = 1,552
Generative Self-Concept Predictors

Episodic
Memory
Performance
Executive
Function

Change in self-perceived generative contributions 0.104 (0.042)** 0.077 (0.028)**
Baseline self-perceived generative contributions 0.066 (0.029)* 0.046 (0.019)

Change in self-perceived generative characteristics 0.018 (0.005) 0.018 (0.005)
Baseline self-perceived generative characteristics 0.043 (0.011) 0.021 (0.005)

Note. Unstandardized coefficients in parentheses; * p < 0.05, * p < 0.01
49

2F. Discussion
The present analyses are the first of their kind to examine longitudinal associations
between self-perceptions of generativity and cognitive performance among older adults. The
analyses examined two distinct indicators of generative self-concept (1—self-perceived
generative contributions to others; and 2—self-perceived generative characteristics), as well as
two indicators of cognitive performance (1—episodic verbal memory; and 2—executive
function), and the associations between the respective generativity predictors and cognitive
functioning outcomes through path modeling with accompanying tests of multiple mediation.
The results demonstrate that positive change in older adults’ self-perceptions of their
generative contributions to others over a period of 10 years is associated with small
enhancements in both memory performance and executive function at the end of that time
period. Associations between change in self-perceived generative contributions and memory
were slightly larger than were those between self-perceived generative contributions and
executive function, suggesting that one’s perceptions of his or her contributory value may have
relatively stronger effects on memory function than on executive processes. In addition, baseline
self-perceptions of one’s generative contributions to others also predicted memory performance
at follow-up, suggesting that one’s prior perceptions of generative contributory value may also
contribute to memory at subsequent points in later life. The latter finding mirrors that of
Gruenewald and colleagues (2012), who found that generativity levels in the early stages of older
adulthood are important predictors of health and functioning at later points in older adulthood.
In sum, the current results suggest that both previously established self-perceptions of generative
contributory worth and changes in those perceptions appear to predict cognitive function as one
moves into the second half of life.
50

Intriguingly, the present investigation showed no significant associations between 10-
year change in older adults’ self-perceived generative characteristics and either episodic memory
or executive function. Baseline self-perceptions of generative characteristics were likewise not
associated with cognition in either performance domain. These findings collectively imply that
one’s sense of his or her generative character – both at early points in older adulthood and as
individuals move further into later life – may not predict cognitive function. The discrepancy in
results for the generative characteristics measure as compared with the generative contributions
measure may point to the idea that the latter is a more appropriate proxy for individuals’ present
conceptualizations of their social contributory value, particularly given that self-perceptions of
generative contributions were more strongly associated with actual generative activity (e.g.,
frequency of volunteering, frequency of providing emotional support, and frequency of
providing instrumental support) than were self-perceptions of generative characteristics. The
generative characteristics measure is also somewhat of an amalgamation of both present and past
generativity, as seen in the combination of present tense statements (e.g., “You like to teach
things to people”) along with past tense statements (e.g., “You have had a good influence on the
lives of many people”). As such, it may not provide the most robust representation of
individuals’ generative self-concept at the present time, whereas the contributions measure
directly queries individuals’ current contributions to others and may more accurately reflect
recent fluctuations in participants’ contributory activity and associated cognitive—affective
states. This, in turn, may account for the discrepancy in significance between the two
generativity predictors and cognitive function.
The significant association between self-perceived generative contributions to others and
cognition warrants further discussion here. First, it is important to note that there were no
51

significant indirect mediation effects observed in either the episodic memory analysis (see
section 2E(b-1)) or the executive function analysis (see section 2E(b-2)). This suggests that
change in elders’ sense of their generative contributory value may have independent and unique
effects on cognitive function. At a general level, the associations between one’s self-perceived
generative contributions and cognitive function provide support for the idea that generativity
may promote successful aging in later life (per Villar, 2012). Of note in this regard is that
positive change in self-perceived generative contributions to others among MIDUS respondents
not only predicted enhanced cognitive functioning in the domains of both memory and executive
function, but also predicted a number of other indicators of successful aging as conceptualized in
the original model of the same (i.e., maintenance of cognitive and physical ability, prevention of
disability and disease, and sustained engagement with life; Rowe & Kahn, 1997). These
indicators include lower risk of disability (as assessed with change in ADL impairment) and
greater engagement with life (as assessed with change in the frequency of volunteering,
emotional support provision, and instrumental support provision). These findings point to the
idea that enhancements in elders’ generative self-concept over time may positively influence
trajectories of cognition, health, and overall capacity for engagement.
Several limitations must be addressed regarding the current analysis. First, it is critical to
note that the findings presented here may not be generalizable to all racial groups given the low
proportion of non-white respondents surveyed in MIDUS. A second limitation of the present
study is that cognitive performance data were not collected at MIDUS I, thus preventing baseline
levels of cognitive function from being included as covariates. The latter constraint, however,
will be removed in the investigation presented in Study 3, which incorporated both baseline and
follow-up assessments of cognition from subsequent waves of MIDUS. Lastly, some scholars
52

have pointed out the lack of suitability of the Loyola Generativity Scale for assessing
generativity among older adults (Schoklitsch & Baumann, 2012), which could account for its
lack of association with cognition in the present study. In light of this perspective, new
generativity scales which specifically query individuals’ generative self-concept in later life
should be created for use in future research.
In aggregate, the analyses presented here represent an important addition to the growing
body of work on the relationship between generativity and health among older adults. While
generativity may positively influence health outcomes in a number of domains, associations
between generativity and cognitive performance may be especially important. As the population
of older individuals increases in future years, opportunities for maintaining and improving
cognitive performance will become increasingly meaningful at both the individual and societal
levels. For older individuals, a pathway to cognitive fitness may exist in one’s self-perceptions
of generative contributory value, and those perceptions should be nourished accordingly. For
society, it will more important than ever to recognize the powerful resource that exists in older
adults’ generativity.

53

CHAPTER 3: STUDY 2

3A. Background
As the population of older adults expands in coming decades, the generative potential of
this group will increase in tandem, reaching a threshold which is historically unparalleled.
Accordingly, older Americans have been described as our nation’s “only increasing natural
resource” (Freedman, 2002, p. 28). While elders’ increasing contributory potential represents an
indisputable societal windfall, it is also affords an important opportunity to assess the value of
generative activity as a potential tool for enhancing cognition across older adulthood. As
indicated in section 1D(a) above, generative activity interventions such as the Experience Corps
model suggest that generative engagement has significant potential to promote cognitive vitality
in later life (Carlson et al., 2008, 2009, 2015). While cognitive benefits of generative activity
engagement have been demonstrated in the context of the Experience Corps intervention, much
less is known about the capacity of everyday generative activity (e.g., less intensive forms of
volunteering, caregiving, and other contributory activity) to promote cognitive functioning
among older adults. Given the profound anticipated aging of the population over the course of
the next 30 years and typical patterns of age-related cognitive decline in the second half of life
(Baltes, 1993; Craik & Bialystok, 2006; Hedden & Garieli, 2004), it is more important than ever
to ascertain how generative activity may relate to, and perhaps promote, cognition across older
adulthood.
Generative activity can take on many forms in older adulthood, and older individuals are
arguably the most generative of all age groups given the overall volume of their contributions
and the number of spheres in which they make contributions. Older adults display an innate
inclination towards pro-social contributory roles (Dávila & Díaz-Morales, 2009), and they are
54

active as volunteers, as custodial grandparents, and as caretakers for their peers, among other
generative pursuits. The Bureau of Labor Statistics (2014) reported that Americans age 65 and
older volunteered 86 hours per year on average in 2013, a figure which exceeded all other age
groups in the country in terms of both hourly activity and the total economic value of that
activity. In addition, this figure is projected to increase significantly from 9 million to 13
million, or by approximately 50 percent, by the year 2020 (Corporation for National and
Community Service, 2007). The Corporation for National and Community Service (2007) also
highlights a growing trend among older adults to volunteer with or on behalf of youth, a
tendency which may signal a heightening of generative impulses. In terms of grandparent
caregiving, 34 percent of the nation’s 2.7 million grandparent caregivers with primary
responsibility for a child are older adults (Livingston, 2013). Older individuals also engage in
high-frequency and high-intensity caregiving to other family members more often than any other
age group, with the number of hours spent in caregiving per week increasing with the age of the
caregiver (National Alliance for Caregiving & AARP, 2004, 2015; Partnership for Solutions,
2004). The AARP Public Policy Institute reports that unpaid family caregiving, the total
economic value of which is estimated to be $470 billion and which equals that of Apple, Hewlett
Packard, IBM, and Microsoft combined, will persist as the top source of long-term care services
in the United States as the population ages in the coming years (Reinhard, Feinberg, Choula, &
Houser, 2015).
In investigating older adults’ engagement in generative activity, it is important to dually
consider the motivations that either underlie or accompany that activity. While contributory
activities are inherently aimed at bettering the welfare of others, it is not necessarily the case that
such engagement is volitional. For instance, older adults who serve as caregivers to a
55

chronically ill or disabled spouse may do so not out of a sense of personal gain or enjoyment, but
out of a sense of obligation. Those who engage in compulsory forms of caregiving often face
situations which are fundamentally different than those encountered within volitional activities,
especially in terms of the emotional duress that they involve (per the stress process model of
caregiver burden; Aneshensel, Pearlin, Mullan, Zarit, & Whitlatch, 1995), and they may likewise
involve different motives for participation. The prevalence of these types of caregiving has also
increased among older adults over the past decade. Caregiving for a chronically ill relative or
spouse, for example, has increased from 33 to 53 percent among those 50 years of age and older
(National Alliance for Caregiving and AARP, 2004, 2015). However, it is not necessarily the
case that those who engage in compulsory caregiving are wholly motivated by obligation. On
the contrary, one’s desire to contribute in a generative manner (that is, to contribute meaningfully
to the well-being of a care recipient) may strongly motivate care provision and may positively
influence one’s appraisal of the caretaking role and of contributory roles more generally. More
positive appraisals of contributory roles, in turn, may bolster cognitive functioning. Conversely,
more negative appraisals of contributory roles may have detrimental effects on cognitive
function.
3B. Significance
In reviewing the varied forms of contributory behavior cited above and the apparent
frequency of engagement in such behavior, it can be concluded that the generative activities of
older adults are extensive. However, even among those who are not presently engaged in
volitional generative activity, such behavior is modifiable. The Experience Corps program has
shown that it is possible to experimentally alter elders’ levels of generative activity and to
achieve impressive subsequent health benefits from that intervention, including improved
56

cognitive performance (Carlson et al., 2008, 2009). This evidence collectively suggests that
older adults possess high motivation for generative activity, high prevalence of current
generative engagement, and proven modifiability of generative behavior, all of which are
important reasons to investigate associations between generative activity and cognition from the
standpoint of public health.
At present, however, no known empirical studies have attempted to track longitudinal
associations between generative activity and cognitive function. The current study sought to fill
this empirical void by investigating associations between frequency of engagement in generative
activity and cognitive performance over time among American elders in the Health and
Retirement Study.
3C. Methods
3C(a). Dataset
Study 2 utilized data from the Health and Retirement Study (HRS). Funded by the
National Institute on Aging, HRS was designed to assess the health and well-being of America’s
older adults in the face of the rapid impending aging of the national population (Institute for
Social Research, 2015). The baseline assessment of HRS was initiated in 1992 and has been
followed by 10 subsequent biennial assessment waves, generating a longitudinal survey series
spanning 11 waves. While cognitive function has been surveyed since the baseline wave of HRS
in 1992, HRS did not survey psychosocial factors until the 2004 wave, and specific survey items
querying generative activity were not incorporated until 2008. However, the 2008/2010 and
2012/2014 psychosocial waves each surveyed this information, generating two waves of data
(baseline and follow-up, respectively) for use in the present longitudinal analysis. Importantly,
the psychosocial survey component of HRS, also known as the “leave-behind questionnaire,”
57

was not administered to all HRS participants at each biennial survey wave, but was instead
administered to alternating subsamples every four years (see Smith et al., 2013, for complete
sampling information). In 2008, a random 50 percent of the total HRS sample completed the
leave-behind questionnaire (n = 7,073), with follow-up and additional new sampling for this
cohort occurring four years later in 2012 (n = 7,412). In 2010, the remaining 50 percent of the
HRS sample completed the leave-behind questionnaire (n = 8,332); follow-up for this group took
place in 2014 (n = 7,244).
3C(b). Participants
The current analysis focused upon the respective 2008/2012 and 2010/2014 leave-behind
subsamples, as participants in these samples completed the generative activity measures which
are central to the research conducted here. The analytic sample was comprised of adults age 50
and older who completed both the HRS leave-behind questionnaire in 2008/2012 and the
cognitive substudy at those timepoints (n = 3,037), as well as the sample of adults age 50 and
older who completed both the leave-behind questionnaire in 2010/2014 and the cognitive
substudy at those waves (n = 3,152). In sum, this produced a composite sample of 6,189 unique
participants.
3C(c). Measures
3C(c-1). Generative activity predictors. The following generative activity measures
were examined as predictors of cognitive performance: (1) Caring for a sick or disabled adult;
(2) Performing volunteer work with children or young people; and (3) Performing other types of
volunteer/charity work. For each of the respective activity types, frequency of engagement was
measured on a 6-point ordinal scale (1= “Daily”; 2 = “Several times a week”; 3 = “Once a
week”; 4 = “Several times a month”; 5 = “At least once a month”; 6 = “Not in the last month”).
58

At the 2010 wave, an additional category (7 = “Never/Not relevant”) was added, which was
carried forward into the 2012 and 2014 waves. These scales were reverse coded such that lower
values are indicative of lesser engagement across all activity types and all waves (Values 6 and 7
= 0; 5 = 1; 4 = 2; 3 = 3; 2 = 4; 1 = 5). Finally, change in generative activity from baseline to
follow-up was calculated by subtracting participants’ generative activity frequency values at the
baseline assessment (2008/2010) from their frequency values at the follow-up assessment
(2012/2014).
3C(c-2). Cognitive performance outcomes. Two cognitive function outcomes were
examined in the present study, including verbal memory and working memory. Detailed
explanations of these measures are available in Ofstedal, Fisher, and Herzog (2005) and in
Fisher, McArdle, McCammon, Sonnega, and Weir (2014); brief summaries are provided here.
Verbal memory. Verbal memory was measured in HRS using a 10-word memory
assessment in which a list of 10 words (e.g., “sky,” “book,” “doctor”) was read aloud to the
respondent. Immediately afterward, the respondent was asked to recall all the words that he or
she could remember. After completing a series of other cognitive assessments and questions, the
respondent was again asked to recall as many words as possible from the list. The mean number
of correctly recalled words across the immediate and delayed iterations of the assessment was
used to gauge verbal memory performance in the present analysis. Change in verbal memory
performance was calculated as the difference in performance between the baseline (2008/2010)
and follow-up assessments (2012/2014).
Working memory. A subtraction task (known colloquially as the “serial 7s” task) was
used to assess working memory. Respondents were asked to subtract 7 from 100, 7 from 93, 7
from 86, 7 from 79, and 7 from 72. The outcome for this measure was the number of correct
59

subtractions across the five iterations (range 0—5). As with verbal memory performance,
change in working memory was calculated as the difference in performance between the baseline
and follow-up assessments.
3C(c-3). Covariates. The sociodemographic covariates of age, sex, race, and education
were incorporated into the present analyses as potential cognitive performance covariates. Age
and education were treated as continuous variables, while sex and race were coded
dichotomously (i.e., as male/female and as white/nonwhite).
3C(c-4). Mediators. The following variables were investigated as potential mediators of
generative activity—cognition associations, in keeping with the pathways outlined in section
1D(c) above: (1) health status (including major disease burden, body mass index, and total
impairment in activities of daily living); (2) physical activity; (3) affective well-being (including
positive and negative affect, self-perceptions of aging, and subjective age); (4) social activity
(including perceived social support and social contact with others); and (5) cognitive stimulation.
Body mass index. Body mass index (BMI) was calculated as the respondents’ weight in
pounds divided by his or her height in inches squared, multiplied by a factor of 703 (per Centers
for Disease Control and Prevention, 2015b). Change in body mass index over the follow-up
period was calculated by subtracting participants’ BMI values at the baseline assessment
(2008/2010) from BMI values at the follow-up assessment (2012/2014).
ADL impairment. Level of impairment in activities of daily living (ADLs) was
quantified according to the respondents’ total reported difficulty with walking, dressing, bathing,
eating, toileting, and getting in and out of bed (1 = does experience difficulty, 2 = no difficulty;
“no” responses were recoded as zero values to give a total range of 0—6). Change in total ADL
60

impairment was calculated as the difference in impairment between the baseline and follow-up
assessments.
Chronic/major disease burden. The number of major diseases reported by participants
represented a summed measure of participant responses (1 = yes; 2 = no; “no” responses were
recoded as zero) for the following health conditions diagnosed at any point in the past:
hypertension, diabetes, cancer, lung disease, heart condition (including angina, heart attack, and
congestive heart failure), and stroke. As with the two health status mediators above, change in
total disease burden was calculated as the difference in the number of reported conditions
between the baseline and follow-up assessments.
Physical activity. Engagement in vigorous (e.g., running, cycling, swimming) and
moderate (e.g., walking, cleaning, gardening) physical activity were queried with a 5-point
ordinal scale ranging from 1 (“Every day”) to 5 (“Hardly ever or never”). These measures were
reverse coded and summed to create a composite rating of physical activity. In creating this
composite measure, vigorous activity was weighted by a factor of 1.5 (as in Gruenewald et al.,
2012). Change in physical activity was calculated by subtracting total physical activity at
baseline from that at follow-up.
Positive affect. Respondents’ overall positive affect was quantified as the extent to
which they indicated feeling happy, interested, excited, enthusiastic, proud, alert, inspired,
determined, content, calm, hopeful, attentive, and active (1 = “Very much”; 5 = “Not at all”;
values were reverse coded). An aggregate mean rating across each of these items was produced
for both the baseline and follow-up waves, giving a possible range of 1—5. Change in mean
positive affect was calculated by subtracting positive affect scores at baseline from those at
follow-up.
61

Negative affect. Respondents’ overall negative affect was quantified as the extent to
which they indicated feeling sad, afraid, upset, guilty, scared, frustrated, hostile, jittery, ashamed,
nervous, and distressed (1 = “Very much”; 5 = “Not at all”; values were reverse coded). As with
positive affect, a mean negative affect score was produced for both the baseline and follow-up
assessments, and change in negative affect was calculated as the difference in these two scores.
Self-perceptions of aging. Scores on an abbreviated 8-item version of the Philadelphia
Geriatric Center Morale Scale (PGCMS; Lawton, 1975) were used to represent participants’
perceptions of their aging experience. The PGCMS queries individuals’ satisfaction with aging
(e.g., “I am as happy now as I was when I was younger”; “The older I get, the more useless I
feel”) and contains both positively- and negatively-phrased items. Negatively phrased items
were reverse coded such that higher scores are indicative of more favorable views of one’s aging.
Scores across all eight items were summed to give a composite representation of aging
satisfaction (possible range 8—48). Change in one’s aging satisfaction was calculated by
subtracting PGCMS scores at the follow-up assessment from those at baseline.
Subjective age. The difference in respondents’ felt (or “subjective”) age and actual age
was calculated by subtracting chronological age from felt age. Those who reported a subjective
age which is lower than their chronological age thus had negative age discrepancy values, while
those with higher subjective ages had positive ones. Change in these values from baseline to
follow-up was calculated and used to represent fluctuations in participants’ felt age across the
survey period.
Social support received from others. The total amount of perceived available social
support was quantified as the mean of participant responses on a series of three questions for
each of four social partner types (spouse/partner, children, other family members, friends): (1)
62

_____ understands the way you feel; (2) Can rely on _____ in the event of a serious problem; (3)
Can open up to _____ if you need to talk about worries or concerns. These items were each
measured on a 4-point scale ranging from 1 (“A lot”) to 4 (“Not at all”), and responses were
reverse coded such that higher values are indicative of greater support. Change in perceived
support for each partner type was respectively calculated as the difference in support between the
baseline and follow-up assessment waves.
Social contact with others. The total amount of social interaction with others was
operationalized as the amount of time per year which the respondent spent in interactions with
children, family members, and friends, whether in person, via writing/email, or via phone (1 =
“Three or more times a week”; 6 = “Less than once a year or never”; values were reverse coded).
Change in total contact for the respective partners was calculated as the difference in contact
between the baseline and follow-up assessment waves.
Cognitive stimulation. Respondents’ level of cognitive stimulation was operationalized
as the sum of two items in the HRS survey which assess overall motivation and interest in life.
The first of these indicates the extent to which the respondent feels that his or her daily activities
are trivial and unimportant (1 = “Strongly disagree”; 6 = “Strongly agree; scale reverse coded).
The second asks respondents to indicate the extent to which they have felt bored over the past 30
days (1 = “Very much”; 5 = “Not at all”). The response scales were summed to produce a
composite measure of overall interest and stimulation (possible range 2—11), and a change score
was produced to reflect change in cognitive stimulation from baseline to follow-up.
3C(d). Analyses
Using Mplus (v. 7.4), full path analyses with maximum likelihood estimation were
conducted to examine the associations of change in generative activity (i.e., volunteering with
63

children/youth; other forms of volunteering/charity work; and caregiving) from baseline
(2008/2010) to follow-up (2012/2014) with change in cognition in the respective domains of
verbal memory and working memory across that same time period. Analyses controlled for
baseline levels of generative activity, baseline cognition, baseline levels of the individual
mediators, and the sociodemographic characteristics listed in section 3C(c-3) above. In addition,
tests of multiple mediation were incorporated into each generative activity—cognition analysis
as shown in Figure 3-1. Maximum likelihood estimation with robust standard errors was used to
address the small degree of incomplete data (0.02%—14.23%) identified for the majority of
study variables.
3C(d-1). Supplementary analyses. Several sets of supplementary analyses were
conducted in order to further contextualize the results of the primary analyses described above.
First, for each generative activity—cognition path analysis, a reverse path analysis was carried
out so as to examine the converse association between change in cognition as a predictor of
change in generative activity (see sections 3E(b-1a), 3E(b-2a), 3E(b-3a), 3E(b-4a), 3E(b-5a), and
3E(b-6a)). Additional supplementary analyses were conducted to examine the association of
change in aggregate generative activity engagement (accounting for participants’ additive
participation in volunteering with youth, other types of volunteer/charity work, and caregiving)
with change in cognition (see section 3E(b-7)), as well as the association of various types of
generative activity overlap (e.g., volunteering in combination with caregiving) with cognition
(see section 3E(b-8)).
64

Figure 3-1. Outline of path modeling strategy utilized in Study 2 to assess generative activity—
cognition associations.

3D. Aims & Hypotheses

Using the analytic strategy described above, Study 2 examined the following specific
aims and hypotheses:
(1) Associations between change in frequency of engagement in volunteering with youth
were examined as a predictor of change in verbal memory (Aim 1), with the
expectation that longitudinal generative activity—cognition associations would be
significant and positive, such that more positive change in elders’ engagement in
volunteering with youth over the follow-up period would predict more positive
change in verbal memory performance (Hypothesis 1).
65

(2) Associations between change in frequency of engagement in volunteering with youth
were examined as a predictor of change in working memory (Aim 2), with the
expectation that longitudinal generative activity—cognition associations would be
significant and positive, such that more positive change in elders’ engagement in
volunteering with youth over the follow-up period would predict more positive
change in working memory performance (Hypothesis 2).
(3) Associations between change in frequency of engagement in other forms of
volunteering/charity work were examined as a predictor of change in verbal memory
(Aim 3), with the expectation that longitudinal generative activity—cognition
associations would be significant and positive, such that more positive change in
elders’ engagement in other forms of volunteering/charity work over the follow-
up period would predict more positive change in verbal memory performance
(Hypothesis 3).
(4) Associations between change in frequency of engagement in other forms of
volunteering/charity work were examined as a predictor of change in working
memory (Aim 4), with the expectation that longitudinal generative activity—cognition
associations would be significant and positive, such that more positive change in
elders’ engagement in other forms of volunteering/charity work over the follow-
up period would predict more positive change in working memory performance
(Hypothesis 4).
(5) Associations between change in frequency of engagement in caregiving to a sick or
disabled spouse were examined as a predictor of change in verbal memory (Aim 5),
with the expectation that longitudinal generative activity—cognition associations
66

would be significant and negative, such that more positive change in elders’
engagement in caregiving over the follow-up period would predict more negative
change in verbal memory performance (Hypothesis 5).
(6) Associations between change in frequency of engagement in caregiving to a sick or
disabled spouse were examined as a predictor of change in working memory (Aim 6),
with the expectation that longitudinal generative activity—cognition associations
would be significant and negative, such that more positive change in elders’
engagement in caregiving over the follow-up period would predict more negative
change in working memory performance (Hypothesis 6).
(7) Indicators from the five hypothesized mediating pathways identified in Figure 3-1
(health status, physical activity, affective well-being, social activity, and cognitive
stimulation) were explored as mediators of generative activity—cognition
associations (Aim 7). No a priori hypotheses were specified with regard to the
relative strength of these mediators in terms of their associations with elders’
generativity and cognitive function. Instead, differences in the extent to which they
individually mediate these associations were investigated in an exploratory manner.
3E. Results
3E(a). Descriptive Results
Descriptive statistics for all Study 2 variables are shown in Table 3-1. The average age
of participants in the composite HRS sample was approximately 74 across the range of 60—101.
The majority of participants were female (59%) and Caucasian (86%). Average educational
attainment was 13 years.
67

In terms of health status at baseline, participants reported approximately 2 major health
conditions on average, a mean BMI value of 28, and a mean ADL impairment value of 0.48
across the range of 0—6 impairments. These values tended to increase slightly over the follow-
up period from 2008/2010 to 2012/2014, with the exception of BMI values, which decreased.
The sample demonstrated mid- to high-range levels of total physical activity, with slight
decreases in physical activity over follow-up.
In terms of social relationships with others, the sample showed mid-range levels of
contact with children, friends, and other family members at baseline. Each of these contact
frequencies showed very slight decreases over the follow-up period. Perceived social support
from one’s spouse, children, friends, and other family members likewise hovered near the
midpoint of the respective measurement scales. Participants reported slight decreases in these
parameters from baseline to follow-up as well.
In terms of affective well-being, participants reported relatively high levels of positive
affect and relatively low levels of negative affect at baseline. Change in these measures over the
follow-up period was very slight, with decreases observed for positive affect and increases for
negative affect. Baseline subjective age—chronological age discrepancy was approximately -12
years on average, and this measure showed slight increases over follow-up. Most participants
showed mid- to high-range levels of satisfaction with aging, and these values tended to decline
on average across the follow-up period.
In terms of cognitive stimulation, most participants reported feeling relatively highly
stimulated in their daily lives, with slight decreases in this parameter emerging over follow-up.
Participants’ average levels of generative activity were quite low, a finding which was
unexpected given the favorable patterns of health and well-being noted above. Across the
68

ordinal scale of 0—5, average frequency of volunteering with children/youth was 0.32 (SD =
0.94), with 87% of the sample reporting no engagement. Average frequency of engagement in
other forms of volunteer/charity work was 0.80 (SD = 1.33), with 66% of the sample reporting
no engagement. Average caregiving frequency was 0.61 (SD = 1.47), with 82% of the sample
reporting no engagement. Across the follow-up period from 2008/2010 to 2012/2014, the
sample showed very slight decreases in these generativity measures, an effect which was largely
driven by the proportion of the sample which showed no change in activity (83% for
volunteering with children/youth, 67% for other volunteer/charity work, and 77% for caregiving;
these percentage values not shown in Table 3-1). In terms of aggregate generative activity, the
sample reported engaging in 0.64 activities (SD = 0.78) on average across the range of 0—3
activities, with the majority of the sample engaging in no activity. In terms of generative
activity overlap, most participants reported engaging in no activity at baseline (53.7%), with
lesser percentages reporting exclusive engagement in volunteer/charity work (20.1%), caregiving
(8.9%), and volunteering with youth specifically (2.5%). Small percentages of the sample
reported taking part in combinations of two or more activities at baseline, as shown in Table 3-1.
Over the follow-up period, there was no activity overlap for the largest proportion of the sample,
as this percentage of participants (43%) reported engaging in no generative activity. Meanwhile,
25 percent of the sample trended towards participation in volunteering only, 17 percent trended
towards giving up all activity, 8 percent trended towards participation in caregiving only, and 7
percent trended towards participation in both volunteering and caregiving.
Finally, the sample showed mid-range levels of cognitive performance on the verbal
memory (M = 4.80; SD = 1.52) and working memory (M = 3.13; SD = 1.87) assessments.
69

Functioning in these domains tended to decline over the follow-up period from 2008/2010 to
2012/2014 (verbal memory: M = -0.45; SD = 1.47; working memory: M = -0.22; SD = 1.93).

Table 3-1. Study 2 Descriptive Statistics; Health & Retirement Study (HRS) Leave-
Behind Sample (2008/2012 & 2010/2014 Participants; n = 6,189)

Generative activity variables n Mean (SD) Range
Freq. of engagement in volunteering with
children or young people at baseline:
5,898

0.32 (0.94) 0—5
Never/not in the last month (0) 5,106 (86.6%)
At least once a month (1) 279 (4.7%)
Several times a month (2) 168 (2.8%)
Once a week (3) 172 (2.9%)
Several times a week (4) 107 (1.8%)
Daily (5) 66 (1.1%)
Change in freq. of volunteering with
children or young people
5,784 -0.07 (1.01) -5—5

Freq. of engagement in other types of
volunteer/charity work at baseline:
5,922 0.80 (1.33) 0—5
Never/not in the last month (0) 3,931 (66.4%)
At least once a month (1) 648 (10.9%)
Several times a month (2) 474 (8.0%)
Once a week (3) 420 (7.1%)
Several times a week (4) 370 (6.2%)
Daily (5) 79 (1.3%)
Change in freq. of other types of
volunteer/charity work
5,828 -0.11 (1.18) -5—5

Freq. of engagement in caring for a sick
or disabled adult at baseline:
5,904 0.61 (1.47) 0—5
Never/not in the last month (0) 4,867 (82.4%)
At least once a month (1) 190 (3.2%)
Several times a month (2) 132 (2.2%)
Once a week (3) 114 (1.9%)
Several times a week (4) 198 (3.4%)
Daily (5) 403 (6.8%)
Change in freq. of caring for a sick or
disabled adult 5,793 -0.07 (1.72) -5—5

Aggregate generative activity: 6,012 0.64 (0.78) 0—3
70

0 activities
3,175 (52.8%)
1 activity
1,999 (33.3%)
2 activities
693 (11.5%)
3 activities
145 (2.4%)
Change in aggregate generative activity
5,969 -0.10 (0.78) -3—3

Generative activity overlap categories at
baseline: 5,777
Participation in no activity
3,101 (53.7%)
Participation in all activities
145 (2.5%)
Volunteering with youth only
201 (3.5%)
Other volunteer/charity work only
1,162 (20.1%)
Caregiving only
514 (8.9%)
Volunteering with youth + other vol.
344 (6.0%)
Volunteering with youth + caregiving
61 (1.1%)
Other vol. + caregiving
249 (4.3%)
Change in generative activity overlap:
5,576
No activity at baseline or follow-up
2,395 (43.0%)
Gave up all activity by follow-up
948 (17.0%)
Only caregiving by follow-up
459 (8.2%)
Only volunteering (both types) by
follow-up 1,408 (25.3%)

Both volunteering and caregiving by
follow-up 366 (6.6%)

Sociodemographic variables
n Mean (SD) Range
Age (at baseline)
6,189 73.52 (6.22) 60—101
Gender:
6,189
Female
3,658 (59.1%)
Male
2,531 (40.9%)
Race:
6,189
White
5,311 (85.8%)
Non-White
878 (14.2%)

Education
6,188 12.69 (2.89) 0—17

Health status variables
n Mean (SD) Range
# Major health conditions at baseline
6,140 1.53 (1.10) 0—6
Change in # major health conditions 6,067
0.23 (0.64) -4—4
Body mass index at baseline 6,098
28.13 (5.53) 15.44—74.72
Change in body mass index 6,031
-0.38 (2.65) -40.22—44.84
# ADL impairments at baseline 3,324
0.48 (1.00) 0—6
Change in # ADL impairments 2,820
0.24 (1.15) -5—6
71

Physical activity variables
n Mean (SD) Range
Physical activity level at baseline
6,182 6.17 (2.77) 2.50—12.50
Change in physical activity level
6,140 -0.42 (2.62) -10—10

Social activity variables
n Mean (SD) Range
Freq. of contact with children at baseline
5,655 4.03 (1.00) 1—6
Change in freq. of contact with children
5,491 -0.10 (0.92) -5—5
Freq. of contact with friends at baseline
5,798 3.82 (1.06) 1—6
Change in freq. of contact with friends
5,414 -0.12 (0.99) -4—4
Freq. of contact with other family
members at baseline 5,738 3.38 (1.09) 1—6
Change in freq. of contact with other
family members 5,309 -0.06 (1.11) -4.33—5
Perceived emotional support from one’s
spouse at baseline 4,037 2.60 (0.34) 1—4
Change in emotional support from one’s
spouse 3,429 -0.00 (0.38) -2.43—3
Perceived emotional support from one’s
children at baseline 5,674 2.35 (0.36) 1—4
Change in emotional support from one’s
children 5,537 -0.01 (0.36) -2.14—2.67
Perceived emotional support from one’s
friends at baseline 5,773 2.09 (0.40) 1—4
Change in emotional support from one’s
friends 5,402 -0.02 (0.40) -3—1.71
Perceived emotional support from other
family members at baseline 5,769 2.09 (0.46) 1—4
Change in emotional support from other
family members 5,353 -0.02 (0.46) -2.67—71

Affective well-being variables
n Mean (SD) Range

Baseline positive affect
6,137 3.63 (0.78) 1—5
Change in positive affect
6,054 -0.11 (0.69) -3.91—3.92
Baseline negative affect
6,135 1.63 (0.56) 1—4.82
Change in negative affect
6,052 0.03 (0.53) -3.09—3.55
Baseline subjective age—chronological
age discrepancy 5,851 -11.94 (11.27) -86—35
Change in subjective age—chronological
age discrepancy 5,477 0.72 (11.80) -74—86
Baseline self-perceptions of aging
5,941 31.27 (8.01) 8—48
Change in self-perceptions of aging
5,664 -1.52 (6.84) -34—35

72

Cognitive stimulation variables
n Mean (SD) Range

Baseline cognitive stimulation
6,155 8.43 (2.16) 1—11
Change in cognitive stimulation
6,109 -0.34 (2.31) -10—10

Cognitive performance variables
n Mean (SD) Range

Baseline verbal memory
6,155 4.80 (1.52) 0—10
Change in verbal memory
6,134 -0.45 (1.47) -8.00—6.50
Baseline working memory
5,908 3.13 (1.87) 0—5
Change in working memory
5,730 -0.22 (1.93) -5—5

3E(b). Results of Path Analyses

3E(b-1). Hypothesis 1: Associations between change in frequency of engagement in
volunteering with youth and change in verbal memory will be significant and positive. As
illustrated in Figure 3-2, change in frequency of volunteering with children/young people did not
significantly predict change in verbal memory performance over the follow-up period ( β = 0.003;
B = 0.005; p = 0.829), leaving Hypothesis 1 unsupported. Though not shown in Figure 3-2,
baseline frequencies of volunteering with children/young people did not predict change in verbal
memory, either ( β = -0.001; B = -0.002; p = 0.948)
In order to differentially test volunteering—verbal memory associations, the original
ordinal change predictor was converted into a categorical variable, then into a series of dummy
variables (e.g., no change in volunteering with youth vs. increases/decreases in same), and then
entered into path analyses for evaluation. This procedure likewise revealed no significant
associations between change in volunteering with children/young people and change in verbal
memory.
3E(b-1a). Supplemental analysis of reverse change associations. In addition, reverse
prediction of change in volunteering with youth as a function of change in verbal memory
73

performance was examined in order to evaluate potential reverse associations between
volunteering and cognition. This analysis revealed no direct effect of change in verbal memory
ability on change in frequency of volunteering with youth ( β = 0.002; B = 0.002; p = 0.844).

Figure 3-2. Path model of associations between change in frequency of volunteering with
children/young people and change in verbal memory performance. Model adjusted for age, sex,
race, education, baseline activity, baseline levels of the individual mediators, and baseline
memory. Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

3E(b-2). Hypothesis 2: Associations between change in frequency of engagement in
volunteering with youth and change in working memory will be significant and positive.
As with the outcome of verbal memory performance, change in frequency of volunteering with
children/young people did not significantly predict change in working memory performance over
the follow-up period ( β = 0.014; B = 0.026; p = 0.358), as shown in Figure 3-3. This finding
74

fails to provide support for Hypothesis 2. Though not shown in Figure 3-3, baseline frequencies
of volunteering with children/young people likewise did not predict change in working memory
( β = 0.002; B = 0.003; p = 0.915).
Once again, in order to differentially test volunteering—memory associations, dummy
variables representing increases and decreases in volunteering with youth were entered into path
analyses as predictors of change in working memory performance and compared with the
referent of no change in volunteering. These analyses also showed no pattern of significant
association.
3E(b-2a). Supplemental analysis of reverse change associations. Finally, a reverse
analysis of change in working memory performance as a predictor of change in frequency of
volunteering with youth was conducted and revealed no significant association between the two
( β = 0.011; B = 0.006; p = 0.377).
75

Figure 3-3. Path model of associations between change in frequency of volunteering with
children/young people and change in working memory performance. Model adjusted for age,
sex, race, education, baseline activity, baseline levels of the individual mediators, and baseline
memory. Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

3E(b-3). Hypothesis 3: Associations between change in frequency of engagement in
other forms of volunteering/charity work and change in verbal memory will be significant
and positive. As illustrated in Figure 3-4, change in frequency of other types of
volunteer/charity work did not independently predict change in verbal memory performance
across the follow-up period ( β = 0.015; B = 0.019; p = 0.279), thus failing to support Hypothesis
3. Though not shown in Figure 3-4, baseline frequencies of other volunteering and charity work
did not predict change in verbal memory, either ( β = 0.014; B = 0.016; p = 0.352).
76

As with the previous two sets of path analyses, differential assessments of other
volunteering/charity work—memory associations with categorical predictors showed no pattern
of significant association.
3E(b-3a). Supplemental analysis of reverse change associations. In addition, reverse
investigations of change in verbal memory performance as a predictor of change in frequency of
other types of volunteering showed no significant associations between these variables ( β =
0.014; B = 0.011; p = 0.289).

Figure 3-4. Path model of associations between change in frequency of other forms of
volunteering/charity work and change in verbal memory performance. Model adjusted for age,
sex, race, education, baseline activity, baseline levels of the individual mediators, and baseline
memory. Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

77

3E(b-4). Hypothesis 4: Associations between change in frequency of engagement in
other forms of volunteering/charity work and change in working memory will be
significant and positive. Unlike the previous sets of analyses, the path analysis investigating
change in other types of volunteering and charity work as a predictor of change in working
memory performance (see Figure 3-5) did reveal a small positive association ( β = 0.031, B =
0.050; p = 0.021). This finding provides modest support for the expectation articulated in
Hypothesis 4. Though not shown in Figure 3-5, baseline frequencies of other types of
volunteering/charity work also predicted slight increases in working memory ( β = 0.030, B =
0.043; p = 0.030).
Though small in magnitude, the significant direct effect of change in volunteering/charity
work on change in working memory function was mediated by the variables of change in contact
with one’s children ( β = 0.002, B = 0.003; p = 0.040) and change in one’s satisfaction with aging
( β = 0.003, B = 0.005; p = 0.014).
Differential assessments of volunteering—memory associations incorporating categorical
predictors revealed a very similar patterns of results, with increases in volunteering activity in
particular predicting enhanced memory performance relative to the referent of no change in
volunteering over follow-up.
3E(b-4a). Supplemental analysis of reverse change associations. The results described
above must be interpreted with the caveat that reverse analysis of change in working memory
performance as a predictor of change in frequency of volunteering/charity work revealed a direct
effect of working memory ability on engagement in volunteering/charity work over time ( β =
0.029, B = 0.018; p = 0.033).
78

Figure 3-5. Path model of associations between change in frequency of other forms of
volunteering/charity work and change in working memory performance. Model adjusted for age,
sex, race, education, baseline activity, baseline levels of the individual mediators, and baseline
memory. Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

3E(b-5). Hypothesis 5: Associations between change in caregiving frequency and
change in verbal memory will be significant and negative. As illustrated in Figure 3-6,
change in the frequency of caregiving was not significantly associated with change in verbal
memory performance ( β = 0.008, B = 0.006; p = 0.621), leaving Hypothesis 5 unsupported.
Though not shown in Figure 3-6, baseline frequencies of caregiving did not predict change in
verbal memory, either ( β = 0.021, B = 0.021; p = 0.169).
Differential assessments of caregiving—verbal memory associations were conducted
using dummy-coded categorical caregiving predictors. The analyses likewise revealed no pattern
79

of significant association between change in caregiving frequency and change in verbal memory
performance.
3E(b-5a). Supplemental analysis of reverse change associations. In addition, reverse
investigations of change in verbal memory performance as a predictor of change in caregiving
frequency revealed no significant associations between the two ( β = 0.006, B = 0.007; p =
0.600).

Figure 3-6. Path model of associations between change in caregiving frequency and change in
verbal memory performance. Model adjusted for age, sex, race, education, baseline activity,
baseline levels of the individual mediators, and baseline memory. Standardized coefficients
shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

3E(b-6). Hypothesis 6: Associations between change in caregiving frequency and
change in working memory will be significant and negative. As shown in Figure 3-7, the
80

path analysis investigating change in caregiving frequency as a predictor of change in working
memory performance revealed a slight positive association ( β = 0.031, B = 0.035; p = 0.034),
thus failing to support Hypothesis 6. Though not shown in Figure 3-7, baseline frequencies of
caregiving were not significantly associated with change in working memory performance ( β =
0.003, B = 0.004; p = 0.818). There were no significant indirect effects for any of the mediators
examined in the current analysis.
Differential assessments of volunteering—memory associations incorporating categorical
predictors revealed a similar pattern of results, with increases in caregiving predicting enhanced
memory performance relative to those whose frequency of caregiving did not change over the
follow-up period.
3E(b-6a). Supplemental analysis of reverse change associations. Notably, reverse
analysis of change in working memory performance as a predictor of change in caregiving
frequency showed a positive direct effect of working memory ability on engagement in
caregiving over time ( β = 0.027, B = 0.024; p = 0.036).

81

Figure 3-7. Path model of associations between change in caregiving frequency and change in
working memory performance. Model adjusted for age, sex, race, education, baseline activity,
baseline levels of the individual mediators, and baseline memory. Standardized coefficients
shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

The observed direct effects of the generative activity predictors investigated in sections
3E(b-1) through 3E(b-6) are summarized in Table 3-2 below:
Table 3-2. Direct Effects of Generative Activity Predictors on Cognitive Function; HRS
Leave-Behind Sample (2008/2012 & 2010/2014 Participants; n = 6,189)
Generative Activity Predictors

Change in
Verbal
Memory
Performance
Change in
Working
Memory
Performance

Change in freq. of volunteering with children/young people
0.003 (0.005) 0.014 (0.026)
Baseline freq. of volunteering with children/young people
-0.001 (-0.002) 0.002 (0.003)

82

Change in freq. of other volunteer/charity work
0.015 (0.019) 0.031 (0.050)*
Baseline freq. of other volunteer/charity work
0.014 (0.016) 0.030 (0.043)*

Change in freq. of caregiving
0.008 (0.006) 0.031 (0.035)*
Baseline freq. of caregiving
0.021 (0.021) 0.003 (0.004)

Note. Unstandardized coefficients in parentheses; * p < 0.05

3E(b-7). Supplementary analysis: Associations between change in aggregate
generative activity engagement and change in cognition. Analyses of change in aggregate
generative activity engagement as a predictor of change in verbal memory performance showed
no significant associations between the two ( β = 0.014, B = 0.026; p = 0.337), nor any significant
association between baseline aggregate generative activity engagement and change in verbal
memory performance, either ( β = 0.041, B = 0.022; p = 0.150). However, positive change in
aggregate generative activity engagement was significantly predictive of slight increases in
working memory performance ( β = 0.032, B = 0.078; p = 0.024), as were baseline levels of
aggregate generative activity engagement ( β = 0.031, B = 0.077; p = 0.033).
3E(b-8). Supplementary analysis: Generative activity overlap as a predictor of
cognitive performance. Two sets of dummy variables were created in order to assess
associations between patterns of generative activity overlap and cognitive performance. The first
represented generative activity overlap at baseline, and the second represented change in patterns
of generative activity overlap across the follow-up period. Both sets and their corresponding
frequencies are shown in Table 3-1 and were reported in section 3E(a) above. Relative to the
referent of no participation in any of the three generative activities at baseline, only the dummy
reflecting exclusive participation in caregiving showed a significant association with change in
verbal memory performance across the follow-up period ( β = 0.035, B = 0.180; p = 0.003).
83

None of the generative activity overlap dummies was significantly predictive of change in
working memory performance. Meanwhile, for the dummies representing change in activity
overlap from baseline to follow-up, only the dummies reflecting trends towards exclusive
participation in caregiving and volunteering, respectively, tended to show associations with
change in cognitive performance. Relative to the referent of no participation at either baseline or
follow-up, caregiving was significantly associated with change in verbal memory ( β = 0.029, B =
0.156; p = 0.015) and marginally significantly associated with working memory ( β = 0.023, B =
0.164; p = 0.059). Relative to the referent of no participation at either baseline or follow-up,
volunteering was significantly associated with both verbal memory ( β = 0.044, B = 0.150; p =
0.002) and working memory ( β = 0.031, B = 0.137; p = 0.022).
3F. Discussion
The present investigation sought to trace longitudinal associations between change in
frequency of generative activity (including volunteering with children/youth, other types of
volunteer/charity work, and caregiving) and change in cognitive performance (including verbal
memory and working memory, respectively) among American elders from the Health and
Retirement Study. Contrary to expectations, engagement in the generative activities noted above
was generally not predictive of cognitive performance, and in cases in which significant
associations were observed (as with the volunteering/charity work—working memory
association and the caregiving—working memory association), they were small in magnitude.
These results were surprising for a number of reasons, and they will be discussed in detail below.
First, however, the significant findings which did emerge from the study will be examined.
To begin, elders’ engagement in volunteering/charity work showed a positive association
with enhancements in working memory over the four-year follow-up period. The magnitude of
84

this association, as well as the magnitude of the associated indirect effects, was small, and there
was also evidence of a significant reverse association between working memory and engagement
in volunteering/charity work, suggesting that more favorable profiles of working memory may
select cognitively advantaged older adults for participation in the generative activity of
volunteering. Together, these findings imply that though there may be slight cognitive benefits
of participation in volunteering among elders, these benefits may only be available to those
whose cognitive abilities are robust from the start.
The second significant finding of the present analysis was that engagement in caregiving
predicted better working memory performance. Once again, the magnitude of this association
was quite small, and there was a significant reverse association between change in working
memory function and change in caregiving frequency over the follow-up period. While the
discovery of the latter finding was not particularly surprising given that caregivers must be
relatively adept in terms of their physical and cognitive function in order to provide care to a sick
or disabled individual (thus providing additional support for the healthy caregiver hypothesis;
Bertrand et al., 2012), it was somewhat unexpected to learn that the healthy caregiver hypothesis
was upheld in the context of the converse caregiving—working memory association, rather than
its counterpart, the stress process model of caregiver burden (Aneshensel et al., 1995). This
suggests that caregiving represents a generative activity which may actually bolster working
memory performance in later life, though those who are selected to participate in caregiving may
be predisposed to better cognitive functioning at baseline relative to their peers.
The third set of significant findings from the analyses presented above involves the
supplementary findings presented in sections 3E(b-7) and 3E(b-8). For the former analysis
investigating change in aggregate generative activity as a predictor of change in cognitive
85

performance, only the outcome of working memory performance was shown to be associated
with change in and baseline levels of aggregate generative activity. This may suggest that
increased engagement in contributory activity promotes specific enhancements in working
memory, the cognitive system responsible for holding information in awareness in order for it to
be used and manipulated across a variety of complex tasks. However, findings from the latter
supplementary analysis which investigated generative activity overlap (see section 3E(b-8)) do
not provide support for the idea that additive participation in contributory activity yields benefits
for cognition. The overlap analysis which examined change in patterns of activity overlap across
the follow-up period from 2008/2010 to 2012/2014 showed that combined participation in two or
more generative activities was not significantly associated with cognition, whereas exclusive
participation in either caregiving or volunteering was associated with slight enhancements in
both verbal memory and working memory. While these two sets of findings are as of yet
inconclusive on the value of participation in multiple generative activities, the specificity of the
activity overlap analysis in combination with the primary results cited above provides more
compelling support for the idea that singular participation in one contributory activity or another
may promote enhanced cognition in later life relative to participation in multiple activities.
In considering the remaining findings from the present study, one particularly surprising
result emerged in the low proportion of HRS respondents (13%) who reported engaging in the
generative activity of volunteering with children. This result prompts consideration of Erikson’s
(1950) foundational work on generativity, which posited that helping and nurturing young people
is a central component of this developmental task. The observed low proportion of older adults
in the HRS sample who reported engaging in volunteering with youth may suggest that the desire
to contribute to young people in an active and deliberate manner wanes somewhat in later life,
86

leading to lesser participation in this activity among the general population of American elders.
Moreover, the current results showed no significant associations between engagement in
volunteering with children/youth and cognitive performance. This finding fails to align with the
results of the Experience Corps program regarding the beneficial effects of generative activity
engagement with children for both executive function and memory (Carlson et al., 2008, 2009).
This discrepancy is likely due to the fact that the average frequency of engagement in volunteer
activity with children among HRS respondents (0.32 across the possible range of 0—5, reflecting
engagement less than once per month) was quite low relative to the high frequencies of
engagement (at least 15 hours per week) required of Experience Corps volunteers (Fried et al.,
2004). It is also possible that the formal experimental structure of the Experience Corps program
set up different expectations regarding the generative meaning and purpose of the volunteer
program than would less systematic forms of volunteering which are available to the general
public. This difference, in turn, could have implications for downstream cognitive functioning.
Perhaps the most unexpected finding in the current analysis was the very low percentage
of older individuals who reported engagement in generative activity. As noted in the results
section and above, only 13 percent of the sample reported engaging in volunteering with children
at baseline, 34 percent in other forms of volunteering/charity work, and 18 percent in care
provision. These results are surprising given the relative good health and functional status of the
sample, and they suggest some degree of disengagement from contributory activity, especially
volunteering, at the level of national trends. These findings may also suggest that efforts to
engage elders in generative activity may currently be lacking in our society. At the societal
level, such efforts will be critical in the coming years as the population of older adults grows
from 50 million to nearly 90 million (U.S. Census Bureau, 2015). Without formal programmatic
87

efforts to encourage generative engagement among elders, the generative potential of our older
adult population will remain untapped, and an important opportunity to enhance societal vitality
will be lost.
Though the results presented in the current study were not indicative of strong
associations between elders’ engagement in generative activity and cognitive function over time,
there were several other notable patterns of results when generative activity was viewed as an
independent predictor of health and well-being. For instance, both volunteering predictors
(volunteering with young people and other types of volunteer/charity work) were associated with
decreases in the number of chronic conditions reported by participants, lesser ADL impairment,
higher levels of physical activity, greater positive affect, more favorable self-perceptions of
aging, greater social connectedness and social support, and higher levels of cognitive
stimulation. This body of results thus supports what has been reported in the literature regarding
the benefits of generativity for both physical and psychological health in later life (see section
1C) and affirms the theoretical perspective that generativity may promote successful aging
(Fisher, 1995; Villar, 2012).
Finally, it is necessary to highlight the limitations of the present study. First among these
was the fact that only a small number of generative activities were surveyed in the Health and
Retirement Study, thereby limiting opportunities for participants to indicate engagement in the
various types of generative activity in which they may take part and which may hold personal
meaning for them. As noted in section 3A above, generativity in later life can manifest in many
different types of activities, and it is very possible that the generative activity survey items
incorporated in HRS were not expansive enough to capture this range, thus leading to a failure to
capture potential generativity—cognition associations that could arise from other forms of
88

generative activity engagement. As a solution, future empirical analyses should take advantage
of qualitative research designs in order to directly query the types of generative activities in
which older Americans engage. Another limitation of the present study is that it could not
account for the motivations which may underlie generative activity. Like most large-scale
surveys which assess older Americans’ time usage, HRS does not query motivations for activity
engagement, although such motivations may strongly influence patterns of engagement and, by
extension, health outcomes such as cognition. For those who study contributory activity, this is
problematic, particularly given that the motivations which impel such activity may be even more
consequential for cognitive performance than is the actual frequency of that activity. For this
reason, scholars of generative activity should advocate for the inclusion of measures querying
motivation in future large-scale investigations of time usage and activity engagement. A third
important limitation of the current analysis was that only 2 waves of baseline and follow-up data
are currently available for the HRS leave-behind questionnaire incorporating the generative
activity measures of interest, with a period of only 4 years separating the respective waves. This
follow-up period may have been too brief to observe appreciable associations between generative
activity and cognition. Lastly, the HRS sample investigated in the present study was quite
healthy at baseline, a finding which may indicate a selection effect preventing the inclusion of
less healthy adults for whom engagement in generative activity could hold cognitive functioning
benefits.
Though the current results do not provide conclusive support for generative activity as a
means of promoting later life cognitive function, they do provide important information
regarding normative patterns of contributory engagement among American elders. This
information suggests that there is substantial room for growth in generative contributory activity
89

among older adults in the United States and that such increases could foster enhancements in
health and well-being for both individuals and society. As a potential societal intervention for
enhancing cognitive fitness, generative activity may still constitute an apt choice, but further
observational and experimental analysis is necessary to validate its utility in this regard.

90

CHAPTER 4: STUDY 3

4A. Background
As the pace of population aging increases over the next 30 years, the growing body of
older adults in the United States and in countries around the world will set a standard for how
future generations conceptualize aging. The way in which these individuals regard their own
aging experience, as well as the way that society reciprocally regards their aging, will have
important and far-reaching implications for our collective social welfare and vitality. For
generativity, the manner in which individuals conceive of their contributory value and develop it
through generative activity may represent an important prognosticator of future health and well-
being across older adulthood. These individual self-perceptions and patterns of expressed
generativity may also inform younger generations’ understanding of what it means to contribute
meaningfully and productively to the world around them. At the societal level there exists a
congruent responsibility to appropriately esteem and leverage the resource of generativity among
older individuals. However, in order to most productively leverage generativity and to reap its
potential benefits for health, it is important to understand which components of generativity may
be impactful for functional outcomes in later life, including that of cognition.
As highlighted in the preceding three chapters, both one’s generative self-concept and
one’s generative activities show compelling associations with health and well-being in later life.
The extent to which they are associated with enhancements in cognitive function is as of yet
unconfirmed, but there is a significant possibility that meaningful, purposeful, and sustained
expressions of generativity may promote better cognitive performance in later life, even as fluid
abilities are expected to decline (Baltes, 1993; Craik & Bialystok, 2006). From an empirical
standpoint, research on generativity has yet to comparatively evaluate whether it is one’s
91

generative activities or one’s generative self-concept which is more consequential for cognitive
functioning during older adulthood. It is possible that the cognitive effects of older adults’
generative self-concept may actually exceed those of generative activity. Evidence for this
perspective derives from recent examinations of demographic and health trends among older
Americans. Such investigations indicate that older adults in the United States are enjoying
longer life expectancies, but with concurrently higher levels of morbidity and physical disability
(Crimmins & Beltrán-Sánchez, 2011). This finding suggests that elders’ capacity to engage in
generative activity may decline in older adulthood given such expansions of morbidity and
functional impairment, but that a considerable proportion of the lifespan may remain thereafter in
which one’s generative self-concept could be nurtured as a means of sustaining cognitive
functioning. Alternatively, it is also possible that generative activity may serve as a means of
intervention to engage sedentary older adults who suffer from chronic disease and to thereby
enhance their cognitive function, as experimental assessments of such activity engagement in the
Experience Corps program have shown (Fried et al., 2004; Carlson et al., 2008, 2009). A more
likely possibility, however, is that one’s generative self-concept and generative activities
mutually reinforce one another to promote cognitive function in later life. This possibility will
be investigated in Study 3 below.
4B. Significance
In order to promote a vital and productively engaged society of elders, gerontologists and
policymakers must be equipped with knowledge of generativity as it operates within the growing
population of older adults. In particular, these leaders must be acquainted with the relative
magnitude of generative self-concept—cognition and generative activity—cognition
associations, as such knowledge is indispensable to the appropriate design of generativity-based
92

societal interventions that are believed to promote successful aging (Villar, 2012). In order to be
maximally effective, these interventions must be designed with a fundamental understanding of
the relative and mutually reinforcing effects of generative self-concept and generative activity on
cognitive performance. An initial step towards achieving such understanding is represented
within Study 3. The unique contribution of this investigation is that it facilitates comparisons
regarding the relative strength of generative activity—cognition associations versus generative
self-concept—cognition associations. In so doing, Study 3 provides an answer to the central
question of the dissertation, namely whether generative actions or generative self-perceptions are
more strongly associated with cognitive function over time. Additional detail regarding this
study, its methodologies, and its specific aims and hypotheses will be outlined in section 4C
below.
4C. Methods
4C(a). Dataset
Like Study 1, Study 3 capitalized on the availability of robust generativity and cognitive
performance measures within the National Study of Midlife Development (MIDUS). However,
Study 3 examined data from the second and third waves of the MIDUS investigation, known as
MIDUS II (Ryff et al., 2012; Ryff & Lachman, 2013) and MIDUS III (Ryff et al., 2017;
University of Wisconsin Survey Center, 2015), respectively, whereas Study 1 utilized data from
the first and second waves of MIDUS. Importantly, cognitive performance data were not
collected at MIDUS I, which prevented the incorporation of a baseline control for cognition in
Study 1. The use of MIDUS II and III data in Study 3 facilitates the inclusion of such a control
and may thus represent an important, and perhaps critical, advancement over the survey
methodology utilized in Study 1. This may be particularly true given that elders display a
93

considerable level of interindividual variability in measures of fluid intelligence in later life
(Glisky, 2007), a finding which suggests that individuals’ baseline levels of cognition should be
taken into account in order to most accurately predict cognitive performance at later points in
older adulthood.
As noted in section 2C(a), MIDUS is a longitudinal investigation of health and well-
being across the adult life course which currently spans three waves: MIDUS I (1995-1996; n =
7,108); MIDUS II (2004-2006; n = 4,963), and MIDUS III (2013-2014; n = 3,294). MIDUS II
and III differ from MIDUS I in that they incorporated a number of substudies designed to assess
health and well-being in several specific domains, including that of cognitive function.
Cognitive performance data were collected at both MIDUS II and III via the Brief Test of Adult
Cognition by Telephone, or BTACT (Lachman & Tun, 2008; Tun & Lachman, 2005; Tun &
Lachman, 2006), which is the source of the cognitive performance measures used in the present
study.
4C(b). Participants
As with the previous two investigations, participants 50 years of age and older at the
baseline assessment wave were selected for inclusion in the analytic sample. The sample was
also limited to those participants who completed both the MIDUS general survey at waves II and
III and the cognitive substudy at waves II and III, producing an aggregate sample of 1,499
individuals.
4C(c). Measures
4C(c-1). Generative self-concept. As in Study 1, individuals’ generative self-concept
was dually conceptualized as self-ratings of contributory behavior toward others (e.g., self-
perceived generative contributions) and self-ratings of generative character (e.g., self-perceived
94

generative characteristics). The operationalization of these constructs within the MIDUS survey
is reiterated below.
Self-perceived generative contributions. The variable of self-rated generative
contributions reflects respondents’ perceptions of their current contributions to the well-being of
others in terms of emotional, instrumental, financial, and material support given. This measure
was surveyed at both the baseline (MIDUS II) and 10-year follow-up (MIDUS III) waves.
Respondents were asked to rate their present level of contributions to others on an 11-point scale,
with a score of 0 representing one’s “worst possible contribution” and a score of 10 representing
one’s “best possible contribution.” Change in self-reported generative contributions was
computed by subtracting MIDUS II scores from MIDUS III scores, yielding a continuous change
score ranging from -10 to +10.
Self-perceived generative characteristics. The variable of self-rated generative
characteristics reflects respondents’ perceptions of their own generative character (e.g., their
generative traits and achievements). This measure is an abbreviated six-item version of the
original 20-item Loyola Generativity Scale, which was developed in 1992 by McAdams and de
St. Aubin. Respondents were asked to indicate their agreement (1 = “A lot”; 4 = “Not at all”)
with each of the following six statements regarding their generative character: (1) “Others would
say that you have made unique contributions to society”; (2) “You have important skills you can
pass along to others”; (3) “Many people come to you for advice”; (4) “You feel that other people
need you”; (5) “You have had a good influence on the lives of many people”; and (6) “You like
to teach things to people” These six items were reverse coded and summed to produce a scale
measure of self-perceived generative character ranging from 6 (lowest self-perception of
generative character) to 24 (highest self-perception of generative character). The six-item
95

Loyola measure has demonstrated high internal consistency amongst MIDUS I, II, and III
participants, with reliability coefficients of 0.84, 0.85, and 0.85 respectively (Brim et al., 2009a;
Brim et al., 2009b; University of Wisconsin Institute on Aging, 2015). As with the generative
contributions variable, change in self-perceived generative characteristics was calculated by
subtracting MIDUS II scores from MIDUS III scores.
4C(c-2). Generative activity. Three measures of generative activity engagement were
incorporated into the present study as predictors of cognitive function. These included
volunteering, providing emotional support to others, and providing instrumental support to
others. Note that though continuous measures of activity in each of these domains (representing
total hourly engagement per month) were utilized in the statistical analyses presented below,
ordinal distributions of the variables were also tested as predictors of cognition and yielded no
differential effects when incorporated into the analyses.
Volunteering. Volunteering was operationalized in the current analysis as participants’
total hours of reported monthly engagement in volunteering in hospitals or health care settings,
in schools, for political causes, and for other organizations or charities. A sum measure of
aggregate volunteer hours was created for both MIDUS II and MIDUS III, and change in total
volunteering was calculated by subtracting the total hours of engagement at MIDUS II from that
at MIDUS III.
Emotional support provision. Frequency of emotional support provision was assessed as
the total number of hours spent giving such support (e.g., comforting, listening, giving advice)
each month to a spouse/partner, parent, in-laws, children/grandchildren, and other family
members/friends. The total number of hours of engagement in each of these four domains was
summed to produce a composite representation of aggregate emotional support provision to
96

others. Change in this parameter was assessed as the difference in hourly support provision
between MIDUS II and III.
Instrumental support provision. Instrumental support provision was assessed as the total
number of hours spent per month giving unpaid assistance to others (e.g., for needs such as
functional assistance, household chores, and transportation) to a parent, in-laws,
children/grandchildren, and other family members/friends. As with emotional support provision,
the total number of hours of engagement in each of these domains was summed to produce a
composite representation of aggregate instrumental support provision to others. Change in this
parameter was again assessed as the difference in hourly support provision between MIDUS II
and III.
4C(c-3). Cognition. Cognitive function was assessed at MIDUS II and III with the Brief
Test of Adult Cognition by Telephone (BTACT). This short cognitive battery includes six
subtests which assess working memory, episodic verbal memory, verbal fluency, reasoning,
processing speed, and task switching ability via telephone (Lachman & Tun, 2008; Tun &
Lachman, 2005; Tun & Lachman, 2006). In comparison with in-person cognitive assessments,
the BTACT demonstrates criterion validity, and it has also shown good test-test reliability
(Lachman et al., 2014). Factor analyses with the BTACT subtests revealed that a two-factor
solution fit the observed measurement loadings best, producing two distinct factors: (1) episodic
verbal memory; and (2) executive function (Lachman et al., 2014). Consistent with this factor
structure and with previous investigations utilizing the BTACT (Agrigoroaei & Lachman, 2011;
Seeman et al., 2011; Stephan et al., 2013), the present analyses combined participant scores from
the six BTACT subtests into composite measures of executive function and episodic memory.
Detailed descriptions of the subtests that comprise these measures are provided in the BTACT
97

documentation (Tun & Lachman, 2005; Lachman et al., 2014); brief summaries of each are
provided below.
Episodic memory composite score. The episodic memory composite score incorporates
standardized scores from the immediate and delayed iterations of the Rey Auditory-Verbal
Learning Test (RAVLT) as administered during the BTACT. During each administration of the
RAVLT, 15 words are read aloud to the participant at a pace of one word per second, and then
the participant is given one minute to recall as many words as possible. This assessment has a
possible range of 0—15 correctly recalled words. Scores from the immediate and delayed
administrations were standardized and averaged to produce z-scored episodic memory composite
measures for both MIDUS II and III, and a change score was calculated from these by
subtracting MIDUS II values from those at MIDUS III.
Executive function composite score. The executive function composite score was
calculated as the average of standardized scores from the working memory, verbal fluency,
reasoning, processing speed, and task alternation assessments, giving a z-scored measure of
cognitive performance in this domain. Working memory was assessed with a backwards digits
test in which the participant listens to a list of numbers and is then asked to recall the list in
reverse order. The number of digits in the list gradually increases with each iteration of the test
(from two to eight digits), and the longest span of digits that is correctly recalled represents the
final score (range: 0—8). Verbal fluency was assessed with a listing assessment in which the
respondent is asked to list in one minute as many examples as possible from the category of
animals. The range of responses for this measure was restricted only by the time limit.
Reasoning was assessed using the number of correctly completed problem sets in a number
series completion test. In this test, participants are presented with a list of numbers and are asked
98

to provide the subsequent number in the series. A total of five lists of varying difficulty are
presented to the participant, and one point is awarded for each correctly completed number series
(range: 0—5). Processing speed was assessed using scores on a backwards counting test in
which the participant is asked to count backwards from 100. The final score for this measure
represents the number of correct responses reported. Task alternation was measured using a
stop/go test in which the participant is asked to respond with the word “stop” upon hearing the
word “red” and to respond with the word “go” upon hearing the word “green.” After this
exercise, participants are then prompted to give the reverse response (i.e., to say “go” when
presented with the word “red”). Finally, participants were tested on a composite task alternation
exercise via a prompt to switch back and forth between these congruent and incongruent
response modes. Task alternation was ultimately assessed by calculating the mean of response
times for the switching and non-switching trials. As with the episodic memory outcome above,
change in executive function across the follow-up period from MIDUS II to III was calculated by
subtracting MIDUS II executive function composite scores from MIDUS III executive function
composite scores.
4C(c-4). Covariates. The following variables were included as covariates in models of
associations between generativity and the cognitive function measures described above.
Variables which were examined as prospective mediators of generativity—cognition associations
are noted in the summaries below. Importantly, assessments of generative self-concept—
cognition associations incorporated generative activity variables as mediators, while assessments
of generative activity—cognition associations incorporated generative self-concept variables as
mediators.
99

Sociodemographic factors. Age, education, race, and sex were included in analytic
models as sociodemographic covariates. Age was represented as a continuous variable, while
sex, race, and level of educational attainment were represented categorically (i.e., male/female;
white/non-white; high school degree or less/some college or more).
Health status. Body mass index (BMI), the number of major health conditions, and the
number of impairments in activities of daily living (ADLs) were included as health status
covariates and as potential mediators of generativity—cognition associations. BMI was
calculated as the respondent’s weight in kilograms divided by height in meters squared. The
number of major health conditions was calculated as in previous MIDUS investigations (e.g.,
Gruenewald et al., 2012) as the sum of nine health conditions, including AIDS, autoimmune
disorders, cancer, diabetes, heart disease, hypertension, lung problems, neurological disorders,
and stroke. The number of major conditions was topcoded at 5 conditions. The number of ADL
impairments was calculated as the sum of limitations in the following activities on a reverse-
coded scale of 1 (“Not at all”) to 4 (“A lot”): Lifting or carrying groceries; bathing or dressing;
climbing several flights of stairs; bending, kneeling, or stooping; walking several blocks; and
moderate-intensity activities such as household chores. Change in these variables was calculated
as the difference in scores from the MIDUS II assessment to the MIDUS III assessment ten years
later.
Health behavior. Level of physical activity was included in the current analyses as a
health behavior mediator, while smoking status was included as a prospective covariate.
Physical activity at both MIDUS II and MIDUS III was calculated as the sum of moderate and
vigorous activity, both of which were measured on a reverse-coded scale of 1 (“Never”) to 6
(“Several times a week or more”). These measures were summed to give composite measures of
100

overall physical activity, with vigorous physical activity weighted by a factor of 1.5 (as in
Gruenewald et al., 2012). Ten-year change in physical activity was established by subtracting
MIDUS II physical activity level from that at MIDUS III. Change in smoking status from
MIDUS II to MIDUS III was determined by creating status categories (e.g., current smoker,
former smoker, never smoked) and generating dummy variables from these categorizations
reflecting those who identified themselves as consistent non-smokers at both MIDUS II and III,
those who identified as former smokers by MIDUS III, and those who identified as consistent
smokers at both MIDUS II and III.
Social contact. Several potential mediators were investigated in the domain of social
contact, including frequency of contact with (1) family, (2) friends, and (3) neighbors. Contact
with family members and friends was measured on a reverse-coded 8-point scale (1 = “Never or
hardly ever”; 8 = “Several times a day”), while contact with neighbors was measured on reverse-
coded 6-point scale (1 = “Never or hardly ever”; 6 = “Almost every day”). For each of these
parameters, ten-year change in the level of contact was assessed by subtracting MIDUS II
contact frequencies from MIDUS III contact frequencies.
Productive engagement. Change in work status (i.e., paid employment status) from
MIDUS II to MIDUS III was included in the present analyses as a potential covariate. This
variable was constructed by creating status categories reflecting ongoing employment from
baseline to follow-up, no employment during this period, dropping employment during this
period, and adding employment during this period. Dummy variables were then created to
reflect change in work status from baseline to follow-up.
Affective well-being. Affective well-being covariates included positive and negative
affect, both of which were represented as six-item scale measures of the respective affective
101

dimensions. Positive affect was assessed as how often participants indicated feeling “cheerful,”
“in good spirits,” “extremely happy,” “calm and peaceful,” “satisfied,” and “full of life” over the
past 30 days, while negative affect was assessed as how often participants indicated feeling “so
sad nothing could cheer you up,” “nervous,” “restless or fidgety,” “hopeless,” “that everything
was an effort,” and “worthless” over that same period. Scores for each of the six items were
reverse coded and summed, producing a 5-point scale score (1 = “None of the time”; 5 = “All of
the time”). Respective change in positive and negative affect from baseline to follow-up were
calculated by subtracting affect scores at MIDUS II from affect scores at MIDUS III.
4C(d). Analyses
Study 3 utilized full path analyses with maximum likelihood estimation to model
associations between change in the following generative self-concept and generative activity
predictors from MIDUS II to MIDUS III and change in cognitive function in the following
performance domains over that same time period:
Generative self-concept:
1) Self-perceived generative contributions and episodic memory
2) Self-perceived generative contributions and executive function
3) Self-perceived generative characteristics and episodic memory
4) Self-perceived generative characteristics and executive function
Generative activity:
5) Volunteering and episodic memory
6) Volunteering and executive function
7) Emotional support and episodic memory
8) Emotional support and executive function
102

9) Instrumental support and episodic memory
10) Instrumental support and executive function
For each of the ten sets of path analyses, regression models of generativity—cognition
associations were produced using the statistical tools available in MPlus (version 7.4), with tests
of multiple mediation incorporated for each analysis as shown in Figure 4-1. Note that each
analysis controlled for baseline generativity, baseline cognition, baseline levels of the respective
mediators (see again Figure 4-1 as well as the methodological documentation above), age, sex,
race, education, work status, partner status, and smoking behavior. Maximum likelihood
estimation with robust standard errors was used to address the small degree of incomplete data
(0.07%—8.27%) for the study variables.

103

Figure 4-1. Outline of path modeling strategy utilized in Study 3 to assess (a) generative self-
concept—cognition associations and (b) generative activity—cognition associations.

4D. Aims & Hypotheses
Using the path modeling strategy described above, Study 3 sought to achieve the
following ten aims and tested the following corresponding hypotheses:
1) Change in self-perceptions of generative contributions to others from MIDUS II to III
was investigated as a predictor of change in episodic memory over that same time
period (Aim 1), with the expectation that generativity—cognition associations would
be significant and positive, such that more positive change in elders’ self-
perceptions of their generative contributory value would predict enhancements
in memory performance over the follow-up period from MIDUS II to III
(Hypothesis 1).
104

2) Change in self-perceptions of generative contributions to others from MIDUS II to III
was investigated as a predictor of executive function over that same time period (Aim
2), with the expectation that generativity—cognition associations would be significant
and positive, such that more positive change in elders’ self-perceptions of their
generative contributory value would predict enhancements in executive function
over the follow-up period from MIDUS II to III (Hypothesis 2).
3) Change in self-perceptions of generative characteristics from MIDUS II to III was
investigated as a predictor of episodic memory over that same time period (Aim 3),
with the expectation that generativity—cognition associations would be significant
and positive, such that more positive change in elders’ self-perceptions of their
generative characteristics would predict enhancements in memory performance
over the follow-up period from MIDUS II to III (Hypothesis 3).
4) Change in self-perceptions of generative characteristics from MIDUS II to III was
investigated as a predictor of executive function over that same time period (Aim 4),
with the expectation that generativity—cognition associations would be significant
and positive, such that more positive change in elders’ self-perceptions of their
generative characteristics would predict enhancements in executive function over
the follow-up period from MIDUS II to III (Hypothesis 4).
5) Change in frequency of volunteering from MIDUS II to III was investigated as a
predictor of change in episodic memory over that same time period (Aim 5), with the
expectation that generative activity—cognition associations would be significant and
positive, such that more positive change in volunteering frequency would predict
105

enhancements in memory performance over the follow-up period from MIDUS II
to III (Hypothesis 5).
6) Change in frequency of volunteering from MIDUS II to III was investigated as a
predictor of change in executive function over that same time period (Aim 6), with the
expectation that generative activity—cognition associations would be significant and
positive, such that more positive change in volunteering frequency would predict
enhancements in executive function over the follow-up period from MIDUS II to III
(Hypothesis 6).
7) Change in frequency of emotional support provision from MIDUS II to III was
investigated as a predictor of change in episodic memory over that same time period
(Aim 7), with the expectation that generative activity—cognition associations would
be significant and positive, such that more positive change in the frequency of
emotional support provision would predict enhancements in memory
performance over the follow-up period from MIDUS II to III (Hypothesis 7).
8) Change in frequency of emotional support provision from MIDUS II to III was
investigated as a predictor of change in executive function over that same time period
(Aim 8), with the expectation that generative activity—cognition associations would
be significant and positive, such that more positive change in the frequency of
emotional support provision would predict enhancements in executive function
over the follow-up period from MIDUS II to III (Hypothesis 8).
9) Change in frequency of instrumental support provision from MIDUS II to III was
investigated as a predictor of change in episodic memory over that same time period
(Aim 9), with the expectation that generative activity—cognition associations would
106

be significant and positive, such that more positive change in the frequency of
instrumental support provision would predict enhancements in memory
performance over the follow-up period from MIDUS II to III (Hypothesis 9).
10) Change in frequency of instrumental support provision from MIDUS II to III was
investigated as a predictor of change in executive function over that same time period
(Aim 10), with the expectation that generative activity—cognition associations would
be significant and positive, such that more positive change in the frequency of
instrumental support provision would predict enhancements in executive
function over the follow-up period from MIDUS II to III (Hypothesis 10).
11) Indicators from the five hypothesized mediating pathways identified in Figure 4-1
(health status, physical activity, affective well-being, social activity, and generative
activity/generative self-concept) were explored as mediators of generativity—
cognition associations (Aim 11). No a priori hypotheses were specified with regard to
the relative strength of these mediators in terms of their associations with elders’
generativity and cognitive function. Instead, differences in the extent to which they
individually mediate these associations were investigated in an exploratory manner.
4E. Results
4E(a). Descriptive Results
Descriptive statistics for all variables assessed in Study 3 are shown in Table 4-1. The
average age of respondents in the sample was approximately 62 years at the baseline MIDUS II
assessment. The majority of the sample was white (95%), female (55%), and had attained a high
school diploma or less (69%).
107

In terms of health status at baseline, the mean body mass index of the sample was 28 (SD
= 5.65), the mean number of major health conditions was 1 (SD = 1.01), and the mean number of
ADL impairments was 2 (SD = 0.73). Respondents tended to show slight increases in these
health status measures over the 10-year follow-up period, with the exception of BMI, which
decreased slightly. The sample demonstrated mid-range levels of average physical activity at
baseline and declined on average in total physical activity over follow-up. Fifty percent of
respondents reported being consistent non-smokers over the follow-up period, while 43 percent
and 7 percent reported being former smokers and consistent smokers, respectively, by follow-up.
In terms of affective well-being, most respondents reported relatively low levels of
negative affect and relatively high levels of positive affect at baseline, with very minor increases
in positive affect and very minor decreases in negative affect observed over the follow-up period.
The majority of the sample maintained a marital or cohabitating partner over the follow-
up period (65%). The sample also displayed moderate to high levels of average contact with
family, friends, and neighbors at baseline, values which remained relatively stable over the 10-
year follow-up period. Respondents provided low to mid-range levels of emotional and
instrumental support to others at baseline, and these levels of support provision tended to decline
over the follow-up period.
Most of the sample was not working for pay at either baseline or follow-up (52%), while
21 percent reported working at both baseline and follow-up, 23 percent had stopped working by
follow-up, and 4 percent had started working by follow-up.
In terms of generative self-concept, respondents reported mid-range levels of self-
perceived generative contributions to others at baseline (M = 6.73; SD = 2.09) and mid- to high-
range levels of self-perceived generative characteristics (M = 17.32; SD = 3.80). On average,
108

respondents tended to decrease slightly in both parameters over the follow-up period (self-
perceived contributions: M = -0.26; SD = 2.37; self-perceived characteristics: M = -0.63; SD =
3.19).
In terms of generative activity, respondents reported low- to mid-range levels of hourly
engagement per month in emotional support provision (M = 51.66; SD = 60.42), instrumental
support provision (M = 19.14; SD = 26.25), and volunteering (M = 9.07; SD = 17.67) at baseline.
In general, these values declined slightly over the follow-up period for the respective activities
(emotional support: M = -5.09; SD = 68.95; instrumental support: M = -2.51; SD = 30.72;
volunteering: M = -0.21; SD = 21.76).
Finally, respondents demonstrated mid-range scores for both the z-scored episodic
memory composite measure (M = 0.03; SD = 0.96) and the executive function composite
measure (M = 0.05; SD = 0.87) at the baseline MIDUS II assessment. Respondents tended to
decline in performance in both memory (M = -0.21; SD = 0.93) and executive function (M = -
0.29; SD = 0.64) over the follow-up period.

Table 4-1. Study 3 Descriptive Statistics; MIDUS II (2004—2006) & MIDUS III (2013—
2014); n = 1,499

Generative self-concept variables n Mean (SD) Range
Baseline self-perceived generative
contributions 1,444 6.73 (2.09) 0—10
Change in self-perceived generative
contributions 1,375 -0.26 (2.37) -10—10

Baseline self-perceived generative
characteristics 1,485 17.32 (3.80) 6—24
Change in self-perceived generative
characteristics 1,433 -0.63 (3.19) -14—12

Generative activity variables n Mean (SD) Range

109

Emotional support given at baseline 1,468 51.66 (60.42) 0—250
Change in emotional support given 1,469 -5.09 (68.95) -250—250

Instrumental support given at baseline 1,466 19.14 (26.25) 0—100
Change in instrumental support given 1,479 -2.51 (30.72) -100—100

Volunteering hours at baseline 1,466 9.07 (17.67) 0—205
Change in volunteering hours 1,403 -0.21 (21.76) -170—240

Sociodemographic variables n Mean (SD) Range
Age (at baseline) 1,499 61.68 (7.99) 50—83
Gender: 1,499

Female 830 (55.4%)

Male 669 (44.6%)

Race: 1,468

White 1,390 (94.7%)

Non-White 78 (5.3%)

Education: 1,497

High school or less 1,026 (68.5%)

Some college or more 471 (31.5%)

Health status variables n Mean (SD) Range
# Major health conditions at baseline 1,499 0.99 (1.01) 0—5
Change in # major health conditions
1,499
0.40 (0.97) -5—4
Body mass index at baseline
1,437
28.01 (5.65) 15.60—82.31
Change in body mass index
1,396
-0.10 (3.14) -22.79—15.20
# ADL impairments at baseline
1,491
1.59 (0.73) 1—4
Change in # ADL impairments
1,480
0.27 (0.69) -2.83—2.83

Health behavior variables n Mean (SD) Range
Physical activity level at baseline 1,481 7.50 (3.20) 2.50—15.00
Change in physical activity level 1,450 -0.12 (3.41) -12.50—11.83
Change in smoking status: 1,497
Consistent non-smoker 749 (50.0%)
Former smoker 638 (42.6%)
Consistent smoker 110 (7.4%)

Social contact variables n Mean (SD) Range
Change in partner status: 1,497
Partner at both 966 (64.5%)
No partner at either 320 (21.4%)
110

Gained partner 44 (2.9%)
Lost partner 167 (11.2%)
Freq. of contact with family at baseline 1,491 6.09 (1.45) 1—8
Change in freq. of contact with family
1,459 -0.02 (1.57) -7.00—6.00
Freq. of contact with friends at baseline
1,486 5.65 (1.63) 1—8
Change in freq. of contact with friends
1,456 -0.14 (1.76) -7.00—7.00
Freq. of contact with neighbors at baseline
1,489 5.00 (1.23) 1—6
Change in freq. of contact with neighbors 1,459 0.03 (1.30) -5.00—5.00

Productive engagement variables n Mean (SD) Range

Change in work status: 1,454
No work 760 (52.3%)
Added work 58 (4.0%)
Dropped work 333 (22.9%)
Working at both 303 (20.8%)

Affective well-being variables n Mean (SD) Range

Baseline negative affect 1,491 1.43 (0.50) 1—4.17
Change in negative affect 1,473 -0.03 (0.50) -3.00—3.00
Baseline positive affect 1,496 3.53 (0.67) 1—5
Change in positive affect 1,538 0.06 (0.65) -3.00—4.00

Cognitive performance variables n Mean (SD) Range

Episodic memory at baseline 1,498 0.03 (0.96) -2.16—3.83
Change in episodic memory 1,493 -0.21 (0.93) -4.88—3.32
Executive function at baseline 1,499 0.05 (0.87) -2.80—2.51
Change in executive function 1,499 -0.29 (0.64) -5.69—2.04

4E(b). Results of Path Analyses
4E(b-1). Hypothesis 1: Associations between change in self-perceived generative
contributions and change in episodic memory performance will be significant and positive.
As shown in Figure 4-2, there was no direct effect of change in self-perceived generative
contributions on change in memory performance ( β = 0.049, B = 0.019; p = 0.068), a finding
which fails to support Hypothesis 1. While the path analyses identified a significant total effect
111

of self-perceived contributions on episodic memory over time ( β = 0.067, B = 0.026; p = 0.013),
this was attributable not to the unique effect of individuals’ self-perceptions of their generative
contributory value, but to collective indirect effects of the mediating variables. Notably, only the
mediator of change in ADL impairment was significantly predictive of change in episodic
memory performance ( β = 0.010, B = 0.004; p = 0.022).
Though not shown in Figure 4-2, baseline self-perceptions of generative contributory
worth were predictive of change in episodic memory performance over follow-up ( β = 0.069, B
= 0.031; p = 0.012), whereas change in self-perceived contributory worth was not.
Ancillary findings from this set of path analyses demonstrate that increases in self-
perceived generative contributory value were, in and of themselves, predictive of several key
outcomes, including lesser ADL impairment, higher levels of physical activity, more frequent
contact with family members and friends, higher levels of emotional and instrumental support
provision to others, and greater overall engagement in volunteering (see Figure 4-2).
4E(b-1a). Supplemental analysis of reverse change associations. Lastly, reverse
analysis of change in memory performance as a predictor of change in self-perceived generative
contributions to others showed no significant direct association ( β = 0.044, B = 0.113; p =
0.087). Baseline episodic memory performance at MIDUS II likewise did not predict change in
self-perceived generative contributions over the follow-up period from MIDUS II to MIDUS III
( β = 0.030, B = 0.073; p = 0.196).
112

Figure 4-2. Path model of associations between change in self-perceived generative
contributions and change in episodic memory performance. Model adjusted for age, sex, race,
education, work status, partner status, smoking behavior, baseline generativity, baseline levels of
the individual mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, **
p < 0.01, *** p < 0.001.

4E(b-2). Hypothesis 2: Associations between change in self-perceived generative
contributions and change in executive function will be significant and positive. As shown in
Figure 4-3, there was no direct effect of change in self-perceived generative contributions on
change in executive function ( β = 0.007, B = 0.002; p = 0.834). This finding fails to corroborate
Hypothesis 2. Though not indicated in Figure 4-3, baseline self-perceptions of generative
contributory value were also not predictive of change in executive function over follow-up ( β = -
0.001, B = 0.000; p = 0.976).
Mirroring the findings presented above, supplementary assessments of the independent
associations of self-perceived generative contributory value with the respective mediators (see
113

Figure 4-3) showed that increases in self-perceived generative contributory value were
independently predictive of lesser ADL impairment, higher levels of physical activity, more
frequent contact with family members and friends, higher levels of emotional and instrumental
support provision to others, and greater overall engagement in volunteering over follow-up.
4E(b-2a). Supplemental analysis of reverse change associations. Reverse analysis of
change in executive function as a predictor of change in self-perceived generative contributions
to others showed no significant direct association ( β = 0.004, B = 0.015; p = 0.871). Baseline
executive function at MIDUS II likewise did not predict change in self-perceived generative
contributions over the follow-up period from MIDUS II to MIDUS III ( β = -0.002, B = -0.006; p
= 0.932).

Figure 4-3. Path model of associations between change in self-perceived generative
contributions and change in executive function. Model adjusted for age, sex, race, education,
work status, partner status, smoking behavior, baseline generativity, baseline levels of the
individual mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, ** p <
0.01, *** p < 0.001.
114

4E(b-3). Hypothesis 3: Associations between change in self-perceived generative
characteristics and change in episodic memory performance will be significant and
positive. As shown in Figure 4-4, change in self-perceived generative characteristics was
significantly and positively associated with change in episodic verbal memory over the follow-up
period ( β = 0.076, B = 0.022; p = 0.001), thus providing support for Hypothesis 3. The direct,
unmediated effect of self-perceived generative characteristics on memory performance occupied
a substantial proportion (86%) of the total effect, while there were no significant indirect effects
among the investigated mediators, nor an aggregate significant indirect effect.
Though not shown in Figure 4-4, baseline self-perceptions of generative character were
also predictive of change in episodic memory performance at follow-up ( β = 0.066, B = 0.016; p
= 0.012).
Ancillary findings from this set of path analyses demonstrate that increases in self-
perceived generative characteristics were, in and of themselves, predictive of several key
outcomes over the follow-up period, including higher levels of physical activity, more frequent
contact with family, friends, and neighbors, higher levels of emotional and instrumental support
provision to others, and greater overall engagement in volunteering (see Figure 4-4).
4E(b-3a). Supplemental analysis of reverse change associations. The findings above
must be interpreted with the caveat that reverse analysis of change in memory performance as a
predictor of change in self-perceived generative characteristics showed a significant positive
association ( β = 0.085, B = 0.293; p = 0.002). Baseline episodic memory, however, did not
emerge as a significant predictor of change in self-perceived generative characteristics ( β =
0.028, B = 0.094; p = 0.272).
115

Figure 4-4. Path model of associations between change in self-perceived generative
characteristics and change in episodic memory performance. Model adjusted for age, sex, race,
education, work status, partner status, smoking behavior, baseline generativity, baseline levels of
the individual mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, **
p < 0.01, *** p < 0.001.

4E(b-4). Hypothesis 4: Associations between change in self-perceived generative
characteristics and change in executive function will be significant and positive. Figure 4-5
shows the results of path analyses investigating the association between change in self-perceived
generative characteristics and change in executive function over the follow-up period. As
illustrated, there was a significant and positive direct effect of change in self-perceived
generative characteristics on change in executive function ( β = 0.063, B = 0.013; p = 0.049), thus
supporting Hypothesis 4. The direct, unmediated effect of self-perceived generative
characteristics on executive function occupied 65% of the total effect, with the remainder
116

attributable to the effects of the mediators, including the sole significant mediator, change in
frequency of volunteering ( β = 0.011, B = 0.002; p = 0.010).
Though not illustrated in Figure 4-5, baseline self-perceptions of generative
characteristics were not predictive of change in executive function over follow-up ( β = -0.006, B
= -0.001; p = 0.840).
As expected, increases in self-perceived generative characteristics were once again
independently predictive of higher levels of physical activity, more frequent contact with family,
friends, and neighbors, higher levels of emotional and instrumental support provision to others,
and greater overall engagement in volunteering (see Figure 4-5).
4E(b-4a). Supplemental analysis of reverse change associations. Reverse analysis of
change in executive function as a predictor of change in self-perceived generative characteristics
showed no significant direct association ( β = 0.058, B = 0.289; p = 0.057), though an indirect
effect of change in executive function on change in generative characteristics was observed ( β =
0.030, B = 0.152; p = 0.001). Baseline executive function at MIDUS II likewise did not predict
change in self-perceived generative characteristics over the follow-up period from MIDUS II to
MIDUS III ( β = -0.031, B = -0.114; p = 0.267), though another indirect effect of executive
function on change in generative characteristics was apparent ( β = 0.024, B = 0.087; p = 0.006).
117

Figure 4-5. Path model of associations between change in self-perceived generative
characteristics and change in executive function. Model adjusted for age, sex, race, education,
work status, partner status, smoking behavior, baseline generativity, baseline levels of the
individual mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, ** p <
0.01, *** p < 0.001.

4E(b-5). Hypothesis 5: Associations between change in frequency of volunteering
and change in episodic memory performance will be significant and positive. As shown in
Figure 4-6, change in frequency of volunteering did not significantly predict change in episodic
memory performance over the follow-up period ( β = 0.014, B = 0.001; p = 0.593), thus failing to
provide support for Hypothesis 5. In addition, baseline frequencies of volunteering did not
predict change in episodic memory performance at follow-up ( β = 0.029, B = 0.002; p = 0.341).
Ancillary findings from this set of path analyses demonstrate that increases in
volunteering frequency over the follow-up period independently predicted several notable
118

outcomes, including increases in contact with friends and increases in both self-perceived
generative contributory value and self-perceived generative characteristics (see Figure 4-6).
4E(b-5a). Supplemental analysis of reverse change associations. Reverse analysis of
change in memory performance as a predictor of change in volunteering frequency showed no
significant association ( β = 0.013, B = 0.299; p = 0.620), nor did assessments of baseline
episodic memory as a predictor of change in volunteering ( β = -0.021, B = -0.476; p = 0.431).

Figure 4-6. Path model of associations between change in frequency of volunteering and change
in episodic memory performance. Model adjusted for age, sex, race, education, work status,
partner status, smoking behavior, baseline generative activity, baseline levels of the individual
mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, ** p < 0.01, ***
p < 0.001.

4E(b-6). Hypothesis 6: Associations between change in frequency of volunteering
and change in executive function will be significant and positive. Figure 4-7 shows the
results of path analyses investigating the association between change in frequency of
119

volunteering and change in executive function over the follow-up period. As depicted, there was
a significant and positive direct effect of change in volunteering frequency on change in
executive function ( β = 0.083, B = 0.002; p = 0.002), thus supporting Hypothesis 6. There was
also a significant total effect ( β = 0.097, B = 0.003; p = 0.001), but no significant indirect effects.
Though not illustrated in Figure 4-7, baseline frequencies of volunteering were not
predictive of change in executive function over follow-up ( β = 0.028, B = 0.001; p = 0.309).
Once again, increases in volunteering frequency over the follow-up period independently
predicted increases in contact with friends and increases in one’s generative self-concept,
including both self-perceived generative contributory value and self-perceived generative
characteristics (see Figure 4-7).
4E(b-6a). Supplemental analysis of reverse change associations. Reverse analysis of
change in executive function as a predictor of change in volunteering revealed a significant direct
association ( β = 0.064, B = 2.186; p = 0.002), though baseline executive function at MIDUS II
did not predict change in volunteering frequency over follow-up ( β = 0.040, B = 1.006; p =
0.165).
120

Figure 4-7. Path model of associations between change in frequency of volunteering and change
in executive function. Model adjusted for age, sex, race, education, work status, partner status,
smoking behavior, baseline generative activity, baseline levels of the individual mediators, and
baseline memory. Standardized coefficients shown; * p < 0.05, ** p < 0.01, *** p < 0.001.

4E(b-7). Hypothesis 7: Associations between change in frequency of emotional
support provision and change in episodic memory performance will be significant and
positive. As shown in Figure 4-8, change in frequency of emotional support provision did not
significantly predict change in episodic memory performance over the follow-up period ( β =
0.024, B = 0.000; p = 0.461), a finding which failed to provide support for Hypothesis 7. In
addition, baseline frequencies of emotional support provision did not predict change in episodic
memory performance at follow-up, either ( β = -0.005, B = 0.000; p = 0.868).
Supplemental findings from this set of path analyses demonstrated that increases in
frequency of emotional support provision over the follow-up period independently predicted
121

increases in body mass index, increases in contact with family, increases in instrumental support
provision, and increases in both self-perceived generative contributory value and self-perceived
generative characteristics (see Figure 4-8).
4E(b-7a). Supplemental analysis of reverse change associations. Reverse analysis of
change in memory performance as a predictor of change in frequency of emotional support
provision showed no significant direct association ( β = 0.018, B = 1.377; p = 0.482) between the
two parameters. However, assessments of baseline episodic memory as a predictor of change in
emotional support provision showed a significant direct effect ( β = 0.081, B = 5.953; p = 0.002).

Figure 4-8. Path model of associations between change in frequency of emotional support
provision and change in episodic memory performance. Model adjusted for age, sex, race,
education, work status, partner status, smoking behavior, baseline generative activity, baseline
levels of the individual mediators, and baseline memory. Standardized coefficients shown; * p <
0.05, ** p < 0.01, *** p < 0.001.

122

4E(b-8). Hypothesis 8: Associations between change in frequency of emotional
support provision and change in executive function will be significant and positive. Figure
4-9 shows the results of path analyses investigating the association between change in frequency
of emotional support provision and change in executive function over the follow-up period. As
depicted, there was no significant direct effect of change in frequency of emotional support
provision on change in executive function ( β = -0.039, B = 0.000; p = 0.248), thus providing no
corroboration for Hypothesis 8.
In addition, baseline frequencies of emotional support provision to others were not
predictive of change in executive function over follow-up ( β = -0.032, B = 0.000; p = 0.337).
As with the path analyses presented in section 4E(b-7) above, increases in frequency of
emotional support provision over the follow-up period independently predicted increases in body
mass index, increases in contact with family, increases in instrumental support provision, and
both increases in self-perceived generative contributory value and self-perceived generative
characteristics (see Figure 4-9).
4E(b-8a). Supplemental analysis of reverse change associations. Finally, reverse
analysis of change in executive function as a predictor of change in frequency of emotional
support provision did not indicate a significant association between the two ( β = -0.027, B = -
2.993; p = 0.215), though baseline executive function at MIDUS II did predict change in
emotional support provision ( β = 0.050, B = 4.035; p = 0.032).
123

Figure 4-9. Path model of associations between change in frequency of emotional support
provision and change in executive function. Model adjusted for age, sex, race, education, work
status, partner status, smoking behavior, baseline generative activity, baseline levels of the
individual mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, ** p <
0.01, *** p < 0.001.

4E(b-9). Hypothesis 9: Associations between change in frequency of instrumental
support provision and change in episodic memory performance will be significant and
positive. As shown in Figure 4-10, change in frequency of instrumental support provision did
not significantly predict change in episodic memory performance over the follow-up period ( β =
-0.003, B = 0.000; p = 0.924), a finding which failed to corroborate Hypothesis 9. In addition,
baseline frequencies of instrumental support provision did not predict change in episodic
memory performance at follow-up, either ( β = -0.003, B = 0.001; p = 0.391).
Supplemental findings from this set of path analyses demonstrated that increases in
frequency of instrumental support provision over the follow-up period independently predicted
124

increases in body mass index, increases in contact with family, increases in emotional support
provision, and both increases in self-perceived generative contributory value and self-perceived
generative characteristics (see Figure 4-10).
4E(b-9a). Supplemental analysis of reverse change associations. Reverse analysis of
change in memory performance as a predictor of change in frequency of instrumental support
provision showed no significant direct association ( β = -0.003, B = -0.088; p = 0.908) between
the two parameters. However, assessments of baseline episodic memory as a predictor of change
in instrumental support provision showed a significant negative direct effect ( β = -0.055, B = -
1.759; p = 0.007).

Figure 4-10. Path model of associations between change in frequency of instrumental support
provision and change in episodic memory performance. Model adjusted for age, sex, race,
education, work status, partner status, smoking behavior, baseline generative activity, baseline
levels of the individual mediators, and baseline memory. Standardized coefficients shown; * p <
0.05, ** p < 0.01, *** p < 0.001.

125

4E(b-10). Hypothesis 10: Associations between change in frequency of instrumental
support provision and change in executive function will be significant and positive. Figure
4-11 shows the results of path analyses investigating the association between change in
frequency of instrumental support provision and change in executive function over the follow-up
period. As depicted, there was no significant direct effect of change in frequency of instrumental
support provision on change in executive function ( β = 0.047, B = 0.001; p = 0.223), thus
providing no corroboration for Hypothesis 10.
In addition, baseline frequencies of instrumental support provision to others were not
predictive of change in executive function over follow-up ( β = 0.015, B = 0.000; p = 0.674).
As with the path analyses presented in section 4E(b-9) above, increases in frequency of
instrumental support provision over the follow-up period independently predicted increases in
body mass index, increases in contact with family, increases in emotional support provision, and
both increases in self-perceived generative contributory value and self-perceived generative
characteristics (see Figure 4-11).
4E(b-10a). Supplemental analysis of reverse change associations. Finally, reverse
analysis of change in executive function as a predictor of change in frequency of instrumental
support provision did not indicate a significant association between the two ( β = 0.028, B =
1.350; p = 0.244), though baseline executive function at MIDUS II did predict change in
instrumental support provision ( β = -0.048, B = -1.685; p = 0.024).
126

Figure 4-11. Path model of associations between change in frequency of instrumental support
provision and change in executive function. Model adjusted for age, sex, race, education, work
status, partner status, smoking behavior, baseline generative activity, baseline levels of the
individual mediators, and baseline memory. Standardized coefficients shown; * p < 0.05, ** p <
0.01, *** p < 0.001.

The observed direct effects of the generative self-concept and generative activity
predictors investigated in sections 4E(b-1) through 4E(b-10) are summarized in Table 4-2 below:
Table 4-2. Direct Effects of Generative Self-Concept and Generative Activity Predictors
on Cognitive Function; MIDUS II (2004—2006) & MIDUS III (2013—2014); n = 1,499
Generative Self-Concept Predictors

Change in
Episodic
Memory
Performance
Change in
Executive
Function

Change in self-perceived generative contributions
0.049 (0.019)† 0.007 (0.002)
Baseline self-perceived generative contributions
0.069 (0.031)* -0.001 (0.000)

Change in self-perceived generative characteristics 0.076 (0.022)** 0.063 (0.013)*
Baseline self-perceived generative characteristics
0.066 (0.016)* -0.006 (-0.001)
127

Generative Activity Predictors

Change in
Episodic
Memory
Performance
Change in
Executive
Function

Change in frequency of volunteering
0.014 (0.001) 0.083 (0.002)**
Baseline frequency of volunteering
0.029 (0.002) 0.028 (0.001)

Change in frequency of emotional support provision
0.024 (0.000) -0.039 (0.000)
Baseline frequency of emotional support provision
-0.005 (0.000) -0.032 (0.000)

Change in frequency of instrumental support provision
-0.003 (0.000) 0.047 (0.001)
Baseline frequency of instrumental support provision
0.028 (0.001) 0.015 (0.000)

Note. Unstandardized coefficients in parentheses; † p < 0.07, * p < 0.05, ** p < 0.01

4F. Discussion
The central goal of the current investigation was to assess in a comparative manner
whether one’s generative self-concept (including self-perceptions of generative contributions to
others and self-perceptions of generative characteristics) or one’s generative activities (including
engagement in volunteering, emotional support provision, and instrumental support provision)
are more strongly associated with cognitive function over time in older adulthood. The results
showed comparable, albeit small, positive associations between change in one’s self-perceived
generative characteristics and change in cognition (for the outcomes of both episodic memory
and executive function) and change in frequency of volunteering and change in cognition (for
only the outcome of executive function) over the follow-up period from MIDUS II to MIDUS
III. Change in elders’ self-perceived generative contributions was not significantly predictive of
cognitive performance, a finding which was surprising given the pattern of results described in
Study 1. However, baseline levels of perceived generative contributions to others did show a
128

significant positive association with memory performance. In addition, neither change in
frequency of emotional support provision nor change in frequency of instrumental support
provision was significantly associated with cognitive functioning among the elders surveyed in
the MIDUS sample.
The significant findings highlighted above will be addressed in greater detail here. First,
the two significant generative self-concept associations between (1) self-perceived generative
characteristics and episodic memory; and (2) self-perceived generative characteristics and
executive function will be discussed. As shown in Figure 4-4, self-perceived generative
characteristics were directly predictive of enhancements in memory function over the follow-up
period, as were baseline self-perceptions of generative characteristics. These findings
collectively suggest that both previously established self-perceptions of one’s generative
character and changes in those perceptions appear to predict memory function as individuals
move further into older adulthood. In addition, no significant indirect effects emerged among
any of the hypothesized mediators, suggesting that elders’ conceptualizations of their generative
character may independently function to bolster memory performance in later life. In reviewing
these findings, it is important to note that there was evidence of a potential converse association
between change in episodic memory function and change in elders’ self-perceived generative
character. This result implies that robust memory ability in older adulthood may selectively
facilitate the development of one’s generative identity among those who enjoy high levels of
cognitive function at baseline. The second finding of a significant positive association between
change in self-perceptions of generative characteristics and change in executive function is
notable in that it was partially mediated by frequency of volunteer activity. This finding suggests
that elders’ generative character may promote enhancements in executive functioning by
129

prompting individuals to engage in the generative activity of volunteering. This result also
dovetails with another key finding from the current study, which was a significant positive
association between change in volunteering frequency and change in executive function across
the follow-up period from MIDUS II to MIDUS III.
The current results further indicate that elders’ generative self-concept and generative
activities may have mutually reinforcing effects on one another during older adulthood. As
shown in Figures 4-2 and 4-3, positive change in one’s self-perceived generative contributory
value was associated with increases in one’s engagement in emotional support provision,
instrumental support provision, and volunteering, with effects being relatively robust for the
latter two generative activities. This pattern of results held true for change in self-perceived
generative characteristics as well, as shown in Figures 4-4 and 4-5. Examination of the converse
associations (i.e., generative activity as a predictor of generative self-concept) demonstrated that
one’s generative activity engagement is associated with increases in both self-perceived
generative contributory value and self-perceived generative characteristics over time. Such
positive associations were observed for all three of the generative activity items examined in the
current study (see Figures 4-6 through 4-11). Notably, volunteering and instrumental support
were more strongly associated with enhancements in one’s generative self-concept than was
emotional support, which may speak to the importance of a sense of active and tangible
engagement in shaping one’s generative self-concept in later life. That is, elders who provide a
practical contributory service (e.g., engaging in instrumental helping behaviors such assisting a
disabled friend with household tasks) may feel a stronger sense of generative value than would
an elder whose contributions are limited to routine emotional support provision (e.g., providing
advice to an adult child over the telephone).
130

Several other findings from the present study are noteworthy and will be addressed
below. First, it is important to observe that though change in the generative contributions
measure did not predict change in either memory performance or executive function, baseline
self-perceptions of one’s generative contributory value were positively predictive of change in
memory performance over the follow-up period. This finding implies that elders’ prior
conceptualizations of their generative contributory value – regardless of subsequent change in
this parameter – may be impactful for cognition as older adults move into the latter stages of life.
This may also account for the discrepancy in findings between Studies 1 and 3 in terms of the
predictive value of elders’ self-perceptions of generative contributory worth for cognition. Study
1 revealed a significant positive association between change in self-perceptions of generative
contributory value and episodic memory, as well as a significant positive association between
baseline levels of perceived generative contributory worth and memory. The predominant
generative driver of memory performance, however, may have been baseline perceptions of
social contributory value, as the results of Study 3 suggest. As a second point of discussion, the
generativity indicators examined in Study 3 showed correlations with a number of notable health
and social outcomes (see Figures 4-2 through 4-11). Positive change in self-perceptions of
generative contributory value, for example, predicted lesser impairment in activities of daily
living, greater physical activity, and greater social connectedness with one’s family and friends.
Change in self-perceptions of one’s generative characteristics, meanwhile, were similarly
associated with higher levels of physical activity and with greater social contact with family,
friends, and neighbors. Change in the generative activity predictors of volunteering, emotional
support provision, and instrumental support provision tended to predict increases in social
contact only. In reviewing these findings, it is worthwhile to note that the generative self-
131

concept predictors of self-perceived generative contributory value and self-perceived generative
characteristics were more strongly associated with health and functional outcomes (e.g.,
decreased ADL impairment, increased physical activity) than were the generative activity
predictors per se. Further analyses will be necessary in order to understand how elders’
generative self-concept may promote physical well-being, but these results do support previous
investigations which showed that older adults’ self-perceptions of their generativity predict
enhanced physical health (Gruenewald et al., 2012).
Lastly, a number of limitations in the current analysis require acknowledgement here.
The first is the fact that the MIDUS study lacks complete longitudinal measures of cognitive
function across each of its existing survey waves. Currently, MIDUS offers only two waves of
cognitive performance data, which precluded the use of a third wave of cognition data in the
present analysis. This circumstance also prevented the use of more sophisticated statistical
modeling techniques which require three waves of data for both the predictor and outcome in
question (e.g., growth curve modeling). An additional limitation regarding the MIDUS data was
highlighted in Study 1 and concerns the lack of generalizability to all racial groups given the low
proportion of non-white respondents surveyed in MIDUS. This limitation also extends to the
current findings.
In aggregate, the results reported here provide modest support for the idea that
generativity may promote cognitive vitality in later life. Though the associations reported above
between change in generativity – both in the form of generative activity and in the form of one’s
generative self-concept – and change in cognitive function were small, they were nonetheless
appreciable and may point to an important opportunity to better acquaint older Americans with
their own generative potential in order to enhance health and well-being. On a related note, the
132

generative self-concept and generative activity measures assessed in the current study were
predictive of a number of indicators of physical and social well-being, which provides additional
support for the theoretical viewpoint articulated in section 2F that generativity may promote
successful aging (Fisher, 1995; Villar, 2012). The current results also suggest an approximately
equivalent magnitude of association between elders’ generative self-concept and cognition and
their generative activity and cognition, as well as a reciprocal reinforcing effect between the two.
These findings should be taken into consideration in the development of future societal
interventions which seek to capitalize on generativity as a means of promoting health and well-
being. Together, they suggest that successful generativity interventions among older adults
would ideally incorporate both generative activity components and generative self-concept
enrichment components in order to have maximal benefits for health.
In closing, the present study indicates that both older adults’ generative self-concept and
their generative activities may equivalently contribute to enhancements in cognitive functioning
over time. However, it also points to a need for further longitudinal assessments of
generativity—cognition associations, both in observational and experimental contexts, in order to
more accurately trace the relationship between change in generativity and change in cognitive
function across time. Only in continuing to evaluate generativity’s role in later life development
will researchers be positioned to take advantage of its potential enhancing effects in the realm of
health and functional outcomes, including that of cognition.

133

CHAPTER 5: SUMMARY & GENERAL DISCUSSION
The current set of dissertation studies sought to evaluate longitudinal associations
between generativity and cognitive functioning among older adults. More specifically, the
analyses endeavored to ascertain how generativity in the form of elders’ generative self-concept
and in the form of elders’ generative activities may be associated with cognition over time
(Studies 1 and 2, respectively). Another central objective of the research presented above was to
evaluate the relative magnitude of generative self-concept—cognition and generative activity—
cognition associations (Study 3). The investigations summarized below capitalized on the
availability of large-scale, nationally representative survey data on generativity and cognitive
performance in the National Survey of Midlife Development (MIDUS) and the Health and
Retirement Study (HRS). Summaries of the current results will be reviewed briefly below,
following by general discussion of the aggregate findings.
5A. Study 1
Study 1 aimed to assess longitudinal associations between older Americans’ generative
self-concept (specifically their self-perceived generative contributions and self-perceived
generative characteristics) and their cognitive performance. This investigation used data from
the National Survey of Midlife Development, or MIDUS, to assess associations between change
in these generativity parameters over the ten-year follow-up period between MIDUS I and II and
cognition performance at the ten-year follow-up. Full path analyses with maximum likelihood
estimation and accompanying tests of multiple mediation were used to assess associations
between (1) self-perceived generative contributions and episodic memory; (2) self-perceived
generative contributions and executive function; (3) self-perceived generative characteristics and
episodic memory; and (4) self-perceived generative characteristics and executive function.
134

Significant positive associations were anticipated for all four predictor—outcome pairings. The
results of this analysis showed that only change in self-perceived generative contributory value
from MIDUS I to MIDUS II was associated with cognition. Change in this parameter was
significantly and positively associated with both episodic memory and executive function at the
ten-year follow-up, though the observed associations were not especially large in magnitude.
Associations between change in self-perceived generative contributions and memory were
slightly larger than were those between self-perceived generative contributions and executive
function, suggesting that one’s perceptions of his or her contributory value may have stronger
effects on memory function than on executive processes. No significant associations were
observed between ten-year change in older adults’ self-perceived generative characteristics and
either episodic memory or executive function. In addition, baseline self-perceptions of one’s
generative contributions to others also predicted memory performance at follow-up, suggesting
that one’s prior perceptions of generative contributory value may also contribute to memory
function at later points in older adulthood. Notably, positive change in self-perceived generative
contributions to others among MIDUS respondents not only predicted enhanced cognitive
functioning in the domains of memory and executive function, but also predicted a number of
other indicators of successful aging, including lower risk of disability (as assessed with change in
ADL impairment) and greater engagement with life (as assessed with change in frequency of
volunteering, emotional support provision, and instrumental support provision).
5B. Study 2
Study 2 investigated associations between frequency of engagement in generative activity
and cognitive performance over time among American elders in the Health and Retirement Study
(HRS). As with Study 1, Study 2 utilized full path analyses with accompanying tests of multiple
135

mediation to assess associations between (1) change in frequency of engagement in volunteering
with youth and verbal memory; (2) change in frequency of engagement in volunteering with
youth and working memory; (3) change in frequency of engagement in other types of
volunteer/charity work and verbal memory; (4) change in frequency of engagement in other
types of volunteer/charity work and working memory; (5) change in frequency of engagement in
caregiving and verbal memory; and (6) change in frequency of engagement in caregiving and
working memory. With the exception of caregiving, all associations were expected to be
significant and positive. Caregiving associations were expected to be significant and negative.
The results of this investigation, however, differed from these expectations in that engagement in
the generative activities noted above was generally not associated with cognitive performance,
and in instances in which significant associations were observed (e.g., for the
volunteering/charity work—working memory association and the caregiving—working memory
association), they were small in magnitude. One of the most surprising findings from this
analysis was the low percentage of older individuals who reported engagement in generative
activity (13% for volunteering with children/youth; 34% for other forms of volunteering/charity
work; and 18% for caregiving). This finding was valuable in and of itself and may indicate a
need for societal interventions to increasingly engage older adults in purposeful and potentially
health-enhancing generative activities.
5C. Study 3
Lastly, Study 3 sought to evaluate in a comparative context whether elders’ generative
self-concept or their actual generative activity engagement is most strongly associated with
cognitive functioning over time in later life. This investigation thus endeavored to provide an
answer to the core question of the dissertation by evaluating the relative strength of the
136

generative self-concept—cognition and generative activity—cognition associations. Study 3 also
utilized data from the MIDUS study, but capitalized on the availability of dual generativity and
cognition data in both the second and third waves of MIDUS, thus facilitating the prediction of
change in cognitive performance along with the incorporation of a baseline cognitive control. As
with the prior two studies, full path analyses with accompanying tests of multiple mediation were
used to assess associations between change in the following generative self-concept/generative
activity predictors and change in the following cognitive functioning outcomes: (1) self-
perceived generative contributions and episodic memory; (2) self-perceived generative
contributions and executive function; (3) self-perceived generative characteristics and episodic
memory; (4) self-perceived generative characteristics and executive function; (5) volunteering
and episodic memory; (6) volunteering and executive function; (7) emotional support and
episodic memory; (8) emotional support and executive function; (9) instrumental support and
episodic memory; and (10) instrumental support and executive function. Of these ten sets of
path analyses, significant and positive associations were observed only for analysis (3), (4), and
(6). In addition, baseline levels of perceived generative contributory value were found to
positively predict episodic memory as well.
In general, the respective generative self-concept—cognition and generative activity—
cognition associations were small, positive, and approximately equivalent in magnitude,
suggesting that both parameters may contribute modestly to enhancements in cognitive function
in later life. Additionally, Study 3 indicated that elders’ generative self-concept and generative
activities may have mutually reinforcing effects on one another during older adulthood, as self-
perceptions of generative contributory value and of generative characteristics were found to
predict increases in generative activity engagement, and vice versa.
137

These results, as well as those of Studies 1 and 2, will be collectively discussed in the
context of the broader generativity literature in section 5D below.
5D. General Discussion
Within the domain of gerontological research, later life generativity is a relatively new
area of inquiry, and few empirical studies exist on the subject with which to compare the current
results. Particularly few in number are those studies which evaluate generativity’s associations
with health outcomes such as cognition among older adults. These trends are due in large part to
a historical tendency to disregard the contributory potential of older individuals, not to mention
to discount the possibility that elders can improve their health and well-being through generative
contributory behavior and a sustained sense of generative contributory worth. Stereotypes
related to elders’ generative contributory value have perhaps hampered even scholarly thought
on the subject, as evidenced by the relative dearth of literature on later life generativity prior to
the launch of initiatives such as the Experience Corps program (Fried et al., 2004) and the
Encore movement (Encore, 2014; Freedman, 2007) in the late 1990s. This circumstance makes
empirical investigations such as those presented in the preceding chapters all the more critical to
the advancement of both scholarly and lay understanding of the generative potential of the elder
population.
5D(a). Generative Activity and Cognition
Despite the small number of existing studies on generativity—cognition associations, a
number of comparisons can be made between those investigations and the findings presented in
the studies above. First, the current body of dissertation research provides partial support for the
findings of Experience Corps researchers on the beneficial effects of generative activity
engagement for cognitive function. These investigators found that intensive engagement in the
138

generative activity of volunteering for no less than 15 hours per week led to enhancements in
both memory and executive function over a period of several months (Carlson et al., 2008).
Similarly, the research presented in Studies 2 and 3 above showed that longitudinal increases in
engagement in volunteering produced increases in both working memory (Study 2) and
composite measures of executive function (Study 3) across a period of several years. However,
Study 2 produced this finding in the context of generic forms of volunteering only, not in the
case of volunteering with children, which could suggest that volunteering with young people
does not confer any especial benefits for cognition. A more likely explanation, though, is that
Study 2 participants from the HRS sample did not take part in volunteering with children at high
frequencies, if at all (see again Table 3-1), therefore precluding the potential observation of
positive associations between this form of volunteering and cognition. The low frequencies of
engagement in volunteering in Study 2 (less than once per month on average for both
volunteering with children and other types of volunteering) were somewhat surprising, despite
the fact that the percentage of Study 2 participants reporting any level of engagement in
volunteering was consistent with national averages (approximately 25—30% among those 50
and older; Bureau of Labor Statistics, 2016). Discrepancies between engagement in volunteering
and anticipated enhancements in cognitive functioning may thus point to the idea that sporadic
volunteering is not enough to foster improved cognition among elder volunteers; higher
frequency activity may be necessary in order to elicit these enhancements. On this basis, further
development of moderate to high-intensity generative activity interventions such as Experience
Corps may be warranted in order to achieve the cognitive performance benefits observed among
Experience Corps volunteers, including significantly improved memory and executive function
(Carlson et al., 2008), enhanced cortical plasticity and cognitive reserve (Carlson et al., 2009),
139

and increased cortical volume in areas of the brain which are particularly susceptible to the
effects of dementia (Carlson et al., 2015).
Intriguingly, the three remaining forms of generative activity surveyed in Studies 2 and 3
– caregiving, emotional support provision, and instrumental support provision – were either
associated with small enhancements in cognitive function (as in the case of caregiving; Study 2)
or were not associated with cognition at all (as in the case of emotional and instrumental support
provision; Study 3). Evidence of better cognitive performance among caregivers is consistent
with the healthy caregiver hypothesis (Bertrand et al., 2012), as opposed to the more widely
accepted stress process model of caregiving (Aneshensel et al., 1995), which has garnered
support from studies of elderly caregivers who demonstrate high levels of stress and
subsequently elevated risk of cognitive impairment (Lee, Kawachi, & Grodstein, 2004;
Vitaliano, Murphy, Young, Echeverria, & Borson, 2011). It is possible that the positive
caregiving—cognition association observed in Study 2 is due to the low average frequencies of
caregiving (i.e., less than one instance of caregiving per month) reported among HRS volunteers
and that more intensive engagement in caregiving would not have elicited positive associations
with cognitive function. Future studies of caregiving—cognition relationships must proactively
account for precise frequencies of engagement in order to assess this possibility. The finding of
nonsignificant associations between both emotional and instrumental support provision and
cognitive performance also warrants discussion here. First, research on the health effects of
support provision as opposed to support receipt is a understudied phenomenon, but preliminary
work suggests that giving social support to others evokes particular benefits for health, including
decreased mortality risk (Brown, Nesse, Vinokur, & Smith, 2003) and higher levels of
psychological well-being (Wang & Gruenewald, 2017). However, no studies to date have
140

specifically evaluated the effect of social support provision on cognitive performance. The
nonsignificant results of the support provision—cognition associations presented in Study 3 do
not align with the more general investigations of support provision and health cited above in that
they did not show positive associations between engagement in support provision and cognitive
performance. This may have been due to the fact that emotional and instrumental support
provision is often considered to be highly normative during the second half of life based on
longstanding patterns of social exchange across the life course (Berkman & Glass, 2000) and
may thus rely on established cognitive circuitry at the level of affective activation and
subsequent cognitive relay processes. This prompts an important consideration regarding all
forms of generative activity – that is, consideration of the type of motivation which compels
engagement in contributory behavior, as well as the context in which such behavior occurs.
When generative activity is motivated by obligation, compulsion, or even ambivalence, it may
fail to embody the sense of affective significance that is hypothesized to drive generativity’s
associations with cognition (see again section 1D(c-3)). As noted previously, the motivations
which prompt elders to engage in contributory activity may be even more consequential for
cognitive function than is the frequency of engagement in such activity. When motivation data
regarding generative activity engagement is lacking, as in the HRS and MIDUS datasets used in
Studies 2 and 3, analyses may fail to capture generativity—cognition associations which would
otherwise be manifest. As a future direction, the incorporation of assessments of motivation in
generativity surveys will be critical.
5D(b). Generative Self-Concept and Cognition
It is also important to reflect upon prior generativity—cognition research as it relates to
the findings presented in Studies 1 and 3 concerning elders’ generative self-concept. In
141

aggregate, these findings suggest that positive change in older adults’ sense of their generative
contributory worth and generative character may promote enhancements in cognitive functioning
in later life. These results provide support for the research presented in Hagood and Gruenewald
(2016), which showed that experimental exposure of elders to positive messages regarding their
generative contributory value led to enhanced memory performance relative to exposure to
negative messages regarding the same. While the analyses presented in Studies 1 and 3 were
observational rather than experimental in nature, they nonetheless provide important insights on
how elders’ generative self-concept may shape cognition in the natural context of American
culture. The consideration of culture is vital to the study of generativity and is an important
attribute of the studies presented above, as cultural factors are believed to strongly shape
generative orientations and impulses, as well as perceptions of generative potential at both the
individual and societal levels (see Alexander, Rubinstein, Goodman, & Luborsky, 1991; de St.
Aubin, 2004; de St. Aubin & Bach, 2015; Kotre, 2004). In particular, the extent to which a
culture is individualistic or collectivist in orientation is thought to impact generativity. More
individualistic societies are believed to embrace generativity to a lesser degree, which,
intriguingly, may be consequential for cognitive function. Evidence suggests that the
maintenance of individualistic versus collectivistic ideologies leads to differential patterns of
neural activity in the brain, as well as to differential downstream cognitive performance
outcomes (Chiao & Harada, 2008; Chiao et al., 2009). American culture has often been
identified as highly individualistic in nature, and as such, it may present a barrier to the
development of generativity and to favorable generative self-concept—cognition associations.
However, even among American elders who have come of age in this social context, is it
possible to cultivate more positive self-perceptions of generative value and generative
142

characteristics through interventions such as Experience Corps (Gruenewald et al., 2016).
Engagement in such interventions may then positively influence older adults’ cognitive
functioning downstream. This possibility, however, will be substantiated only through continued
evaluation of the developmental role of generative self-concept in both observational and
experimental research studies.
5D(c). Generativity and Health
Beyond the core generativity—cognition associations documented above, it is essential to
note that the findings presented in this manuscript provide added support for prior theoretical
conceptualizations of generativity as a facilitator of successful aging (Fisher, 1995; Villar, 2012).
Studies 1 and 3 showed that increases in older adults’ generative self-concept were associated
with lesser risk of impairment in activities of daily living, higher levels of physical activity,
greater participation in volunteering, greater engagement in social support provision, and higher
levels of social contact with family, friends, and neighbors. Increased generative activity,
meanwhile, was associated in Studies 2 and 3 with decreases in the number of chronic conditions
reported by participants, lesser impairment in activities of daily living, greater overall physical
activity, higher levels of positive affect, more favorable self-perceptions of aging, greater social
connectedness and support, higher levels of cognitive stimulation, and higher levels of
engagement in volunteering and support provision. These findings overwhelmingly support the
research presented in section 1C regarding generativity’s correlations with favorable profiles of
physical, social, and psychological well-being in later life. Moreover, the findings also fit within
the model of successful aging proposed by Rowe and Kahn in 1997. That is, elders’ generativity
showed consistent associations with the three hypothesized elements of successful aging in Rowe
and Kahn’s conceptualization: (1) prevention of disability and disease, (2) maintenance of
143

physical and cognitive abilities, and (3) sustained engagement with life. Importantly, both
generative activity and generative self-concept appear to predict enhancements in these domains,
and measures of generative self-concept actually emerged in Study 3 as stronger predictors of
physical functioning outcomes (e.g., lower levels of ADL impairment and higher physical
activity) than did generative activity. This suggests the intriguing possibility that one’s mental
conceptualizations of generativity may act to promote physical health and well-being above and
beyond the effects of actual physical engagement in contributory activity. From the perspective
of psychoneuroimmunology, or the branch of science which investigates interactions between the
processes of the mind and the processes of the body, this premise is not at all unusual, but to
those unfamiliar with mind—body connections, it may be surprising. However, this finding
builds upon an established and growing body of research which demonstrates that self-
perceptions of generativity are strongly predictive of health outcomes, including reduced risk of
disability, reduced risk of institutionalization in nursing facilities, and reduced mortality risk
(Grand et al., 1988, 1990; Gruenewald et al., 2007, 2009, 2012; Okamoto & Tanaka, 2004;
Pitkala et al., 2004). While the research presented in the prior chapters could not conclusively
determine whether one’s contributory self-concept or contributory activity is more strongly
associated with change in cognitive performance across older adulthood, it did suggest that these
two constructs mutually reinforce one another, with elders’ self-perceptions of generativity
predicting increases in generative activity, and vice versa. Future investigations of generativity,
therefore, will be best equipped for success if they incorporate measures of both generative self-
concept and generative activity and if they specifically ascertain how these parameters may
interact with one another to promote more favorable trajectories of health and functioning with
advancing age.
144

5D(d). Limitations & Future Directions
The current set of dissertation studies demonstrates a number of limitations which should
be addressed in future investigations of generativity—cognition associations. First, the limited
number of pertinent waves of data in both the MIDUS and HRS datasets ( ≤ 2 for each of the
three studies) precluded the use of more sophisticated statistical modeling techniques (e.g.,
growth curve modeling) which might have provided greater insight into generativity’s potential
impact on cognitive functioning over time in older adulthood. This impediment will only be
addressed through continued collection of generativity measures in large-scale surveys of
psychosocial development in later life. Moreover, generativity measures must be surveyed and
collected in conjunction with measures of health and functioning in order for researchers to
evaluate generativity—health associations in a robust interdisciplinary context. A second
limitation of the current work concerns the observational data utilized in Studies 1—3. To the
extent of present knowledge, these studies represent the first observational assessments of
generativity—cognition associations. While the accompanying observational data is invaluable
in terms of its statistical power and ability to illuminate national trends in elders’ generativity, it
lacks the capacity of experimental data (such as that obtained from programmatic interventions
like Experience Corps) to test the impact of generativity interventions on cognitive functioning.
In the future, those who study generativity must concertedly carry out both observational and
experimental research in order to allow both forms of investigation to inform and complement
one another. For example, the results of the Experience Corps studies on generativity—
cognition associations prompted the observational research conducted here, which then led to the
discovery of relatively low frequencies of participation in contributory activity among American
older adults (see again Study 2), which should prompt experimental investigation of the effects
145

of heightened frequencies of generative activity engagement among the broader population of
American elders. In addition, the small number of generative activities surveyed in both HRS
and MIDUS may have prevented those individuals who engaged in other forms of contributory
behavior from endorsing their participation. As a means of resolving this issue, future studies of
generativity should accommodate qualitative designs in order to more fully capture the possible
range of generative activities and feelings which older adults endorse. Lastly, the generative
self-concept measures included in Studies 1 and 3 (particularly the abbreviated form of the
Loyola Generativity Scale used to assess generative characteristics) may not appropriately gauge
generativity among older adults, as other scholars have previously noted (Schoklitsch &
Baumann, 2012). Generativity researchers must seek to test and deploy new measures and scales
of elders’ generative self-concept in order to most precisely target those dimensions of
contributory worth and characteristics which hold meaning in later life.
5D(e). Implications
While research on associations between generativity and cognition is as of yet
preliminary, the studies presented in the previous chapters nevertheless provide important
information regarding the potential implications of later life generativity as it relates to
opportunities for improving cognition and bolstering societal health and vitality. First, the
research presented here collectively suggests that both generative activity engagement and tools
for improving conceptualizations of generative value and character may represent a route to
enhanced cognitive functioning in the second half of life. While further investigation is needed
to corroborate this claim at the scholarly level, generativity can be freely explored by both
individuals and communities as a means of promoting cognitive health. To this end, individuals
can seek out existing community programs which offer opportunities for contributory
146

engagement and enrichment of one’s contributory self-concept, while civic groups and
organizations should look to expand their service offerings so as to incorporate generative
activities which may yield cognitive benefits. Similarly, the current results also suggest a need
to develop a greater number of programs like Experience Corps which offer opportunities for
moderate to high-intensity generative engagement. Such engagement would hypothetically
prevent age-related cognitive decline (per Carlson et al., 2008, 2009, 2015), would enhance
elders’ sense of their own generative worth and character (per Gruenewald et al., 2016), and
would potentially lead to a positive cascade of enhancements in other domains of health and
well-being, as the present findings suggest. Importantly, such interventions would ideally
incorporate both generative activity and generative self-concept enrichment components, given
that Study 3 suggests that the two forms of generativity have mutually reinforcing effects on one
another, as well as independent positive associations with cognitive function. Lastly, the present
findings indicate that positive shifts in the way that we as a society regard the generative
contributory potential of older adults may give rise to benefits for elders’ health and functioning.
Aging stereotypes regarding older adults’ lack of contributory value may already be doing
concerted harm to older adults’ cognitive health, as prior work suggests (Hagood & Gruenewald,
2016). The current research, especially that presented in Studies 1 and 3, evokes the possibility
that increases in elders’ sense of their generative contributory worth and character may promote
cognitive vitality in later life. As a potential antidote for societal ageism concerning older adults’
contributory value, both anti-ageism campaigns and one-on-one coaching designed to enhance
elders’ feelings of generative worth should be explored. Such interventions may not only yield
their intended effect of reducing the prevalence of aging stereotypes and enhancing generativity,
147

but they may also give rise to global enhancements in health and well-being, including in the
domain of cognition.
5E. Conclusion
In sum, the studies presented in this dissertation represent an important step forward in
scholarly understanding of associations between generativity and cognition in older adulthood.
While this focal area is still in its infancy, studies such as those detailed above provide critical
insights into the role that later life generativity may play in shaping cognitive function. Among
its contributions, this body of work is the first of its kind to probe longitudinal associations
between generativity and cognitive performance among older adults using large and nationally
representative survey data. This research also significantly extends our understanding of
generativity’s potential to promote healthy aging in the domain of cognition and beyond. As a
preliminary step in this sphere of empirical investigation, the studies presented here also
illuminate the further work that must be done to more clearly elucidate the relationship between
generativity and cognitive function among older individuals.
In closing, generativity represents a compelling area of research within contemporary
gerontology, but its true capacity to enhance health in later life may not yet be realized among
those who study human development given the relatively small number of interdisciplinary
studies on the subject. In order to continue to develop scholarly knowledge of associations
between generativity and healthy aging, investigations such as those presented in the current
manuscript must be undertaken. In so doing, the true impact of generativity can be understood
and enacted for the good of both current and future generations of elders.

148

References
Agrigoroaei, S., & Lachman, M. E. (2011). Cognitive functioning in midlife and old age:
Combined effects of psychosocial and behavioral factors. The Journals of Gerontology,
Series B: Psychological Sciences and Social Sciences, 66 (Suppl 1), 130–140.
http://doi.org/10.1093/geronb/gbr017
Albinet, C. T., Boucard, G., Bouquet, C. A., & Audiffren, M. (2010). Increased heart rate
variability and executive performance after aerobic training in the elderly. European
Journal of Applied Physiology, 109 (4), 617-624. doi:10.1007/s00421-010-1393-y
Alexander, B. B., Rubinstein, R. L., Goodman, M., & Luborsky, M. (1991). Generativity in
culture context: The self, death and immortality as experienced by older American
women. Ageing and Society, 11 (4), 417-442. doi:10.1017/S0144686X00004384
Aneshensel, C. S., Pearlin, L, I., Mullan, J. T., Zarit, S. H., & Whitlatch, C. J. (1995). Profiles in
caregiving: The unexpected career. San Diego, CA: Academic Press.
Arvanitakis, Z., Wilson, R. S., Bienias, J. L., Evans, D. A., & Bennett, D. A. (2004). Diabetes
mellitus and risk of Alzheimer’s disease and decline in cognitive function. Journal of the
American Medical Association – Neurology, 61 (5), 661-666.
doi:10.1001/archneur.61.5.661
Baltes, P. B. (1993). The aging mind: Potential and limits. The Gerontologist, 33 (5), 580-594.
doi:10.1093/geront/33.5.580
Baltes, P. B., & Baltes, M. M. (1990). Psychological perspectives on successful aging: The
model of selective optimization with compensation. In P. B. Baltes & M. M. Baltes
(Eds.), Successful ageing: Perspectives from the behavioral sciences (1-34). Cambridge:
Cambridge University Press.
149

Barnes, L. L., Mendes de Leon, C. F., Wilson, R. S., Bienias, J. L., & Evans, D. A. (2004).
Social resources and cognitive decline in a population of older African Americans and
whites. Neurology, 63 (12), 2322-2326. doi:10.1212/01.WNL.0000147473.04043.B3
Becker, E. (1973). The denial of death. New York, NY: The Free Press.
Berkman, L. F., & Glass, T. (2000). Social integration, social networks, social support, and
health. In L. F. Berkman & I. Kawachi (Eds.), Social Epidemiology (137-173). New
York, NY: Oxford University Press.
Bertrand, R. M., Saczynski, J. S., Mezzacappa, C., Hulse, M., Ensrud, K., & Fredman, L. (2012).
Caregiving and cognitive function in older women: Evidence of the healthy caregiver
hypothesis. Journal of Aging and Health, 24 (1), 48-66. doi:10.1177/0898264311421367
Bherer, L., Erickson, K. I., Liu-Ambrose, T. (2013). A review of the effects of physical activity
and exercise on cognitive and brain functions in older adults. Journal of Aging Research,
2013, 1-8. doi:10.1155/2013/657508
Brim, O. G., Baltes, P. B., Bumpass, L. L., Cleary, P. D., Featherman, D. L., Hazzard, W. R., …
Shweder, R. A. (2009a). Documentation of scales and constructed variables in MIDUS 1.
In National Survey of Midlife Development in the United States (MIDUS), 1995-1996
(pp. 54-55). Retrieved from
https://www.icpsr.umich.edu/icpsrweb/ICPSR/series/00203/studies/2760?archive=ICPSR
&sortBy=7
Brim, O. G., Baltes, P. B., Bumpass, L. L., Cleary, P. D., Featherman, D. L., Hazzard, W. R., …
Shweder, R. A. (2009b). Documentation of psychosocial constructs and composite
variables in MIDUS II Project 1. In National Survey of Midlife Development in the
United States (MIDUS II), 2004-2006 (p. 61). Retrieved from
150

https://www.icpsr.umich.edu/icpsrweb/ICPSR/series/00203/studies/4652?archive=ICPSR
&sortBy=7
Brim, O. G., Baltes, P. B., Bumpass, L. L., Cleary, P. D., Featherman, D. L., Hazzard, W. R., …
Shweder, R. A. (2016). National Survey of Midlife Development in the United States
(MIDUS), 1995-1996. Ann Arbor, MI: Inter-university Consortium for Political and
Social Research. doi:10.3886/ICPSR02760.v11
Brown, S. L., Nesse, R. M., Vinokur, A. D., & Smith, D. M. (2003). Providing social support
may be more beneficial than receiving it: Results from a prospective study of
mortality. Psychological Science, 14 (4), 320-327. doi:10.1111/1467-9280.14461
Bureau of Labor Statistics, United States Department of Labor. (2014). Volunteering in the
United States – 2013. Retrieved from http://www.bls.gov/news.release/pdf/volun.pdf
Bureau of Labor Statistics, United States Department of Labor. (2016). Volunteers by selected
characteristics, September 2015. Retrieved from
https://www.bls.gov/news.release/volun.t01.htm
Burr, J. A., Tavares, J., & Mutchler, J. E. (2011). Volunteering and hypertension risk in later life.
Journal of Aging and Health, 23 (1), 24-51. doi:10.1177/0898264310388272
Butler, R. N. (1969). Age-ism: Another form of bigotry. The Gerontologist, 9 (4:1), 243-246.
doi:10.1093/geront/9.4_Part_1.243
Carlson, M. C., Erickson, K. I., Kramer, A. F., Voss, M. W., Bolea, N., Mielke, M., … Fried, L.
P. (2009). Evidence for neurocognitive plasticity in at-risk older adults: The Experience
Corps Program. Journal of Gerontology, Series A: Biological Sciences and Medical
Sciences, 64 (12), 1275-1282. doi:10.1093/gerona/glp117
151

Carlson, M. C., Kuo, J. H., Chuang, Y. –F., Varma, V. R., Harris, G., Albert, M. S., … Fried, L.
P. (2015). Impact of the Baltimore Experience Corps trial on cortical and hippocampal
volumes. Alzheimer’s & Dementia, 11 (11), 1340-1348. doi:10.1016/j.jalz.2014.12.005
Carlson, M. C., Saczynski, J. S., Rebok, G.W., Seeman, T., Glass, T. A., McGill, S., … Fried, L.
P. (2008). Exploring the effects of an “everday” activity program on executive function
and memory in older adults: Experience Corps. The Gerontologist, 48 (6), 793-801.
doi:10.1093/geront/48.6.793
Carpenter, S. M., Peters, E., Västfjäll, D., & Isen, A. M. (2013). Positive feelings facilitate
working memory and complex decision making among older adults. Cognition and
Emotion, 27 (1), 184-192. doi:10.1080/02699931.2012.698251
Carstensen, L. L. (1995). Evidence for a life-span theory of socioemotional selectivity. Current
Directions in Psychological Science, 4 (5), 151-156. doi:10.1111/1467-8721.ep11512261
Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory of
socioemotional selectivity. American Psychologist, 54 (3), 165-181. doi:10.1037/0003-
066X.54.3.165
Carstensen, L. L., & Mikels, J. A. (2005). At the intersection of emotion and cognition: Aging
and the positivity effect. Current Directions in Psychological Science, 14 (3), 117-121.
doi:10.1111/j.0963-7214.2005.00348.x
Centers for Disease Control and Prevention. (2011, February). Cognitive impairment: A call for
action, now! Retrieved from
https://www.cdc.gov/aging/pdf/cognitive_impairment/cogimp_poilicy_final.pdf
152

Centers for Disease Control and Prevention. (2014). National diabetes statistics report:
Estimates of diabetes and its burden in the United States, 2014. Atlanta, GA: U.S.
Department of Health and Human Services.
Centers for Disease Control and Prevention. (2015a). Number (in millions) of civilian, non-
institutionalized adults with diagnosed diabetes, 1980-2014. Retrieved from
https://www.cdc.gov/diabetes/statistics/prev/national/figadults.htm
Centers for Disease Control and Prevention. (2015b). About adult BMI. Retrieved from
http://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/index.html
Centers for Disease Control and Prevention. (2016, November 30). High blood pressure facts.
Retrieved from https://www.cdc.gov/bloodpressure/facts.htm
Chiao, J. Y., & Harada, T. (2008). Cultural neuroscience of consciousness: From visual
perception to self-awareness. Journal of Consciousness Studies, 15 (10-11), 58-69.
Chiao, J. Y., Harada, T., Komeda, H., Li, Z., Mano, Y., Saito, D., … Iidaka, T. (2009). Neural
basis of individualistic and collectivistic views of self. Human Brain Mapping, 30 (9),
2813-2820. doi:10.1002/hbm.20707
Corporation for National and Community Service. (2007). Baby Boomers and volunteering:
Findings from corporation research. Retrieved from
http://www.nationalservice.gov/sites/default/files/documents/boomer_research.pdf
Craik, F. I., & Bialystok, E. (2006). Cognition through the lifespan: Mechanisms of change.
Trends in Cognitive Sciences, 10 (3), 131-138. doi:10.1016/j.tics.2006.01.007
Crimmins, E. M., & Beltrán-Sánchez, H. (2011). Mortality and morbidity trends: Is there
compression of morbidity? Journal of Gerontology: Social Sciences, 66B, (5), 75-86.
doi:10.1093/geronb/gbq088
153

Crooks, V. C., Lubben, J., Petitti, D. B., Little, D., & Chiu, V. (2008). Social network, cognitive
function, and dementia incidence among elderly women. American Journal of Public
Health, 98 (7), 1221-1227. doi:10.2105/AJPH.2007.115923
Dávila, M. C., & Díaz-Morales, J. F. (2009). Age and motives for volunteering: Further
evidence. Europe’s Journal of Psychology, 5 (2), 82-95. doi:10.5964/ejop.v5i2.268
de St. Aubin, E. (2004). The propagation of genes and memes: Generativity through culture in
Japan and the United States. In E. de St. Aubin, D. P. McAdams, & T. -C. Kim (Eds.),
The generative society: Caring for future generations (pp. 63-82). Washington, DC:
American Psychological Association.
de St. Aubin, E., & Bach, M. (2015). Explorations in generativity and culture. In L. A. Jensen
(Ed.), The Oxford handbook of human development and culture: An interdisciplinary
perspective (pp. 653-665). New York, NY: Oxford University Press.
Driscoll, I., Hamilton, D. A., Petropoulos, H., Yeo, R. A., Brooks, W. M., Baumgartner, R. N., &
Sutherland, R. J. (2003). The aging hippocampus: Cognitive, biochemical, and structural
findings. Cerebral Cortex, 13 (12), 1344-1351. doi:10.1093/cercor/bhg081
Dulin, P. L., & Hill, R. D. (2003). Relationships between altruistic activity and positive and
negative affect among low-income older adult service providers. Aging and Mental
Health, 7 (4), 294-299. doi:10.1080/1360786031000120697
Eadie, B. D., Redila, V. A., & Christie, B. R. (2005). Voluntary exercise alters the
cytoarchitecture of the adult dentate gyrus by increasing cellular proliferation, dendritic
complexity, and spine density. Journal of Comparative Neurology, 486 (1), 39-47.
doi:10.1002/cne.20493
154

Elias, M. F., Wolf, P. A., D’Agostino, R. B., Cobb, J., & White, L. R. (1993). Untreated blood
pressure level is inversely related to cognitive functioning: The Framingham Study.
American Journal of Epidemiology, 138 (6), 353-364.
doi:10.1093/oxfordjournals.aje.a116868
Encore. (2014). Encore.org and the Encore vision. Retrieved from http://encore.org/who-we-
are/encore-org-encore-vision/
Erickson, K. I., & Kramer, A. F. (2009). Aerobic exercise effects on cognitive and neural
plasticity in older adults. British Journal of Sports Medicine, 43 (1), 22-24.
doi:10.1136/bjsm.2008.052498
Erickson, K. I., Voss, M. W., Prakash, R. S., Basak, C., Szabo, A., Chaddock, L., … Kramer, A.
F. (2010). Exercise training increases size of hippocampus and improves memory.
Proceedings of the National Academy of Sciences, 108 (7), 3017-3022.
doi:10.1073/pnas.1015950108
Erikson, E. H. (1950). Childhood and society. New York, NY: W.W. Norton & Company, Inc.
Erikson, E. H. (1963). Childhood and society (2
nd
ed.). New York, NY: W.W. Norton &
Company, Inc.
Erikson, E. H., Erikson, J. M., & Kivnick, H. Q. (1986). Vital involvement in old age. New York,
NY: W.W. Norton & Company, Inc.
Fairlie, H. (1988, March 28). Talkin’ bout my generation. The New Republic, 198 (13), 19-22.
Fisher, B. J. (1995). Successful aging, life satisfaction, and generativity in later life.
International Journal of Aging and Human Development, 41 (3), 239-250.
doi:10.2190/HA9X-H48D-9GYB-85XW
155

Fisher, G. G., McArdle, J. J., McCammon, R. J., Sonnega, A., & Weir, D. R. (2014, January 30).
New measures of fluid intelligence in the HRS (HRS Documentation Report DR-027).
Retrieved from http://hrsonline.isr.umich.edu/sitedocs/userg/dr-027b.pdf
Fredrickson, B. L. (2004). The broaden-and-build theory of positive emotions. Philosophical
Transactions of the Royal Society B, 359 (1449), 1367-1378. doi:10.1098/rstb.2004.1512
Freedman, M. (2002). Prime time: How Baby Boomers will revolutionize retirement and
transform America. New York, NY: Public Affairs.
Freedman, M. (2007). Encore: Finding work that matters in the second half of life. New York,
NY: Public Affairs.
Fried, L. P., Carlson, M.C., Freedman, M., Frick, K. D., Glass, T. A., Hill, J., … Zeger, S.
(2004). A social model for health promotion for an aging population: Initial evidence on
the Experience Corps model. Journal of Urban Health: Bulletin of the New York
Academy of Medicine, 81 (1), 64-78. doi:10.1093/jurban/jth094
Gill, T. M., Richardson, E. D., & Tinetti, M. E. (1995). Evaluating the risk of dependence in
activities of daily living among community-living older adults with mild to moderate
cognitive impairment. Journal of Gerontology, Series A: Biological Sciences and
Medical Sciences, 50A (5), M235-M241. doi:10.1093/gerona/50A.5.M235
Glisky, E. L. (2007). Changes in cognitive function in human aging. In D. R. Riddle (Ed.), Brain
aging: Models, methods, and mechanisms (3-20). Boca Raton, FL: CRC Press.
Grand, A., Grosclaude, P., Bocquet, H., Pous, J., & Albarede, J. L. (1988). Predictive value of
life events, psychosocial factors and self-rated health on disability in an elderly rural
French population. Social Science & Medicine, 27 (12), 1337-1342. doi:10.1016/0277-
9536(88)90198-0
156

Grand, A., Grosclaude, P., Bocquet, H., Pous, J., & Albarede, J. L. (1990). Disability,
psychosocial factors and mortality among the elderly in a rural French population.
Journal of Clinical Epidemiology, 43 (8), 773-782. doi:10.1016/0895-4356(90)90237-J
Gruenewald, T. L., Karlamangla, A. S., Greendale, G. A., Singer, B. H., & Seeman, T. E. (2007).
Feelings of usefulness to others, disability, and mortality in older adults: The MacArthur
Study of Successful Aging. The Journals of Gerontology, 62B (1), P28-P37.
doi:10.1093/geronb/62.1.P28
Gruenewald, T. L., Karlamangla, A. S., Greendale, G. A., Singer, B. H., & Seeman, T. E. (2009).
Increased mortality risk in older adults with persisitently low or declining feelings of
usefulness to others. Journal of Aging and Health, 21 (2), 398-425.
doi:10.1177/0898264308329023
Gruenewald, T. L., Liao, D. H., & Seeman, T. E. (2012). Contributing to others, contributing to
oneself: Perceptions of generativity and health in later life. The Journals of Gerontology,
Series B: Psychological Sciences and Social Sciences, 67 (6), 660-665.
doi:10.1093/geronb/gbs034
Gruenewald, T. L., Tanner, E. K., Fried, L. P., Carlson, M. C., Xue, Q. L., Parisi, J. M., …
Seeman, T. E. (2016). The Baltimore Experience Corps Trial: Enhancing generativity via
intergenerational activity engagement in later life. The Journals of Gerontology, Series B:
Psychological Sciences and Social Sciences, 71 (4), 661-670. doi:10.1093/geronb/gbv005
Gunstad, J., Paul, R. H., Cohen, R. A., Tate, D. F., Spitznagel, M. B., & Gordon, E. (2007).
Elevated body mass index is associated with executive dysfunction in otherwise healthy
adults. Comprehensive Psychiatry, 48 (1), 57-61. doi:10.1016/j.comppsych.2006.05.001
157

Hagood, E. W., & Gruenewald, T. L. (2016). Positive versus negative priming of older adults’
generative value: Do negative messages impair memory? Aging & Mental Health.
doi:10.1080/13607863.2016.1239063
Hauser, D. (2013). Manifestations of generativity during the last stage of life (Doctoral
dissertation). Retrieved from ProQuest Dissertations and Theses. (UMI No. 3604936).
Hedden, T., & Gabrieli, J. D. (2004). Insights into the ageing mind: A view from cognitive
neuroscience. Nature Reviews Neuroscience, 5 (2), 87-96. doi:10.1038/nrn1323
Hess, T. M., Auman, C., Colcombe, S. J., & Rahhal, T. A. (2003). The impact of stereotype
threat on age differences in memory performance. Journal of Gerontology: Psychological
Sciences, 58B (1), P3-P11. doi:10.1177/0956797613497023
Hess, T. M., Emery, L., & Queen, T. L. (2009). Task demands moderate stereotype threat effects
on memory performance. Journal of Gerontology: Psychological Sciences, 64B (4),
P482-P486. doi:10.1093/geronb/gbp044
Hess, T. M., & Hinson, J. T. (2006). Age-related variation in the influences of aging stereotypes
on memory in adulthood. Psychology and Aging, 21 (3), 621-625. doi: 10.1037/0882-
7974.21.3.621
Hess, T. M., Hinson, J. T., & Statham, J. A. (2004). Explicit and implicit stereotype activation
effects on memory: Do age and awareness moderate the impact of priming? Psychology
and Aging, 19 (3), 495-505. doi:10.1037/0882-7974.19.3.495
Holtzman, R. E., Rebok, G. W., Saczynski, J. S., Kouzis, A. C., Doyle, K. W., & Eaton, W. W.
(2004). Social network characteristics and cognition in middle-aged and older adults.
Journal of Gerontology: Psychological Sciences, 59B (6), P278-P284.
doi:10.1093/geronb/59.6.P278
158

Hong, S. I., & Morrow-Howell, N. (2010). Health outcomes of Experience Corps: A high-
commitment volunteer program. Social Science & Medicine, 71 (2), 414-420.
doi:10.1016/j.socscimed.2012.04.009
Hultsch, D. F., Hertzog, C., Small, B. J., & Dixon, R. A. (1999). Use it or lose it: Engaged
lifestyle as a buffer of cognitive decline in aging? Psychology and Aging, 14 (2), 245-
263. doi:10.1037/0882-7974.14.2.245
Huta, V., & Zuroff, D. C. (2007). Examining mediators of the link between generativity and
well-being. Journal of Adult Development, 14 (1), 47-52. doi:10.1007/s10804-007-9030-
7
Institute for Social Research. (2015). About the Health and Retirement Study. Retrieved from
http://hrsonline.isr.umich.edu/
Isaacs, K. R., Anderson, B. J., Alcantara, A. A., Black, J. E., & Greenough, W. T. (1992).
Exercise and the brain. Angiogenesis in the adult rat cerebellum after vigorous physical
activity and motor skill learning. Journal of Cerebral Blood Flow and Metabolism, 12
(1), 110-119. doi:10.1038/jcbfm.1991.14
James, B. D., Wilson, R. S., Barnes, L. L., & Bennett, D. A., (2011). Late-life social activity and
cognitive decline in old age. Journal of the International Neuropsychological Society, 17
(6), 998-1005. doi:10.1017/S1355617711000531
Kasper, J. D., Freedman, V. A., & Spillman, B. C. (2014, April 29). Disability and care needs of
older Americans by dementia status: An analysis of the 2011 National Health and Aging
Trends Study. Washington, DC: U.S. Department of Health and Human Services, Office
of the Assistant Secretary for Planning and Evaluation, and Office of Disability, Aging
159

and Long-Term Care Policy. Retrieved from
https://aspe.hhs.gov/system/files/pdf/77141/NHATS-DS.pdf
Kececi-Savan, D., Cengiz, M., Yavuzer, H., Yavuzer, S., Sulu, C., Doventas, A., & Beger, T.
(2016). Relation of ambulatory blood pressure measurement and cognitive functions in
hypertensive elderly patients. Aging Clinical and Experimental Research, 28 (4), 699-
704. doi:10.1007/s40520-015-0477-2
Keyes, C. L. M., & Ryff, C. D. (1998). Generativity in adult lives: Social structural contours and
quality of life consequences. In D. P. McAdams & E. de St. Aubin (Eds.), Generativity
and adult development: How and why we care for the next generation (pp. 227-263).
Washington, DC: American Psychological Association.
Kotre, J. (1984). Outliving the self: Generativity and the interpretation of lives. Baltimore, MD:
The Johns Hopkins University Press.
Kotre, J. (2004). Generativity and culture: What meaning can do. In E. de St. Aubin, D. P.
McAdams, & T. -C. Kim (Eds.), The generative society: Caring for future generations
(pp. 35-49). Washington, DC: American Psychological Association.
Lachman, M. E., Agrigoroaei, S., Tun, P. A., & Weaver, S. L. (2014). Monitoring cognitive
functioning: Psychometric properties of the Brief Test of Adult Cognition by Telephone
(BTACT). Assessment, 21 (4), 404-417. doi:10.1177/1073191113508807
Lachman, M. E., & Tun, P. A. (2008). Cognitive testing in large-scale surveys: Assessment by
telephone. In S. M. Hofer and, D. F. Alwin (Eds.), Handbook of cognitive aging:
Interdisciplinary perspectives (pp. 506-524). Thousand Oaks, CA: Sage Publications, Inc.
doi:10.4135/9781412976589.n30
160

Lang, F. R., & Carstensen, L. L. (2002). Time counts: Future time perspective, goals, and social
relationships. Psychology and Aging, 17 (1), 125-139. doi:10.1037/0882-7974.17.1.125
Lawton, M. P. (1975). The Philadelphia Geriatric Center Morale Scale: A revision. Journals of
Gerontology, 30 (1), 85-89. doi:10.1093/geronj/30.1.85
Lee, B. K., Glass, T. A., McAtee, M. J., Wand, G. S., Bandeen-Roche, K., Bolla, K. I., &
Schwartz, B. S. (2007). Associations of salivary cortisol with cognitive function in the
Baltimore memory study. Archives of General Psychiatry, 64 (7), 810-818.
doi:10.1001/archpsyc.64.7.810
Lee, S., Kawachi, I., & Grodstein, F. (2004). Does caregiving stress affect cognitive function in
older women? Journal of Nervous & Mental Disease, 192 (1), 51-57.
doi:10.1097/01.nmd.0000106000.02232.30
Levy, B. (1996). Improving memory in old age through implicit self-stereotyping. Journal of
Personality and Social Psychology, 71 (6), 1092-1107. doi:10.1037/0022-3514.71.6.1092
Levy, B. R., & Langer, E. J. (1994). Aging free from negative stereotypes: Successful memory in
China and among the American Deaf. Journal of Personality and Social Psychology, 66
(6), 989-997. doi:10.1037/0022-3514.66.6.989
Livingston, G. (2013, September 4). At grandmother’s house we stay. Retrieved from
http://www.pewsocialtrends.org/2013/09/04/at-grandmothers-house-we-stay/
Lum, T. Y., & Lightfoot, E. (2005). The effects of volunteering on the physical and mental
health of older people. Research on Aging, 27 (1), 31-55.
doi:10.1177/0164027504271349
161

Mather, M., & Carstensen, L. L. (2005). Aging and motivated cognition: The positivity effect in
attention and memory. Trends in Cognitive Science, 9 (10), 496-502.
doi:10.1016/j.tics.2005.08.005
Maxfield, M., Greenberg, J., Pyszczynski, T., Weise, D. R., Kosloff, S., Soenke, M., … Blatter,
J. (2014). Increases in generative concern among older adults following reminders of
mortality. International Journal of Aging and Human Development, 27 (1), 1-21.
doi:10.2190/AG.79.1.a
McAdams, D. P., & de St. Aubin, E. (1992). A theory of generativity and its assessment through
self-report, behavioral acts, and narrative themes in autobiography. Journal of
Personality and Social Psychology, 62 (6), 1003-1015. doi:10.1037/0022-3514.62.6.1003
McAdams, D. P., de St. Aubin, E., & Logan, R. L. (1993). Generativity among young, midlife,
and older adults. Psychology and Aging, 8 (2), 221-230. doi:10.1037/0882-7974.8.2.221
McAdams, D. P., & Logan, R. L. (2003). What is generativity? In E. de St. Aubin, D. P.
McAdams, & T. -C. Kim (Eds.), The generative society: Caring for future generations
(pp. 15-31). Washington, DC: American Psychological Association.
McEwen, B. S., & Sapolsky, R. M. (1995). Stress and cognitive function. Current Opinion in
Neurobiology, 5 (2), 205-216. doi:10.1016/0959-4388(95)80028-X
McGuire, A. P. (2005). Generativity in adolescents (Doctoral dissertation). Retrieved from
ProQuest Dissertations and Theses. (UMI No. 3170646).
Morrow-Howell, N., Hinterlong, J., Rozario, P. A., & Tang, F. (2003). Effects of volunteering on
the well-being of older adults. Journal of Gerontology: Social Sciences, 58B (3), S137-
S145. doi:10.1093/geronb/58.3.S137
162

National Alliance for Caregiving & AARP. (2004). Caregiving in the U.S. Retrieved from
http://www.caregiving.org/data/04finalreport.pdf
National Alliance for Caregiving & AARP. (2015). Caregiving in the U.S. Retrieved from
http://www.caregiving.org/wp-content/uploads/2015/05/2015_CaregivingintheUS_Final-
Report-June-4_WEB.pdf
Obisesan, T. O. (2009). Hypertension and cognitive function. Clinical Geriatric Medicine, 25
(2), 259-288. doi:10.1016/j.cger.2009.03.002
Obisesan, T. O., Obisesan, O. A., Martins, S., Alamgir, L., Bond, V., Maxwell, C., & Gillum, R.
F. (2008). High blood pressure, hypertension, and high pulse pressure are associated with
poorer cognitive function in persons aged 60 and older: The third National Health and
Nutrition Examination Survey. Journal of the American Geriatrics Society, 56 (3), 501–
509. doi:10.1111/j.1532-5415.2007.01592.x
Ofstedal, M. B., Fisher, G. G., & Herzog, A. R. (2005). Documentation of cognitive functioning
measures in the Health and Retirement Study (HRS Documentation Report DR-006).
Retrieved from http://hrsonline.isr.umich.edu/sitedocs/userg/dr-006.pdf
Okamoto, K., & Tanaka, Y. (2004). Subjective usefulness and 6-year mortality risks among
elderly persons in Japan. Journal of Gerontology: Psychological Sciences, 59B (5), P246-
P249. doi:10.1093/geronb/59.5.P246
Oman, D., Thoresen, C. E., & McMahon, K. (1999). Volunteerism and mortality among the
community-dwelling elderly. Journal of Health Psychology, 4 (3), 301-316.
doi:10.1177/135910539900400301
Pardo, J. V., Lee, J. T., Sheikh, S. A., Surerus-Johnson, C., Shah, H., Munch, K. R., … Dysken,
M. W. (2007). Where the brain grows old: Decline in anterior cingulate and medial
163

prefrontal function with normal aging. Neuroimage, 35 (3), 1231-1237.
doi:10.1016/j.neuroimage.2006.12.044
Park, D. C., & Schwarz, N. (Eds.). (2000). Cognitive aging: A primer. Philadelphia, PA:
Psychology Press.
Partnership for Solutions. (2004, September). Chronic conditions: Making the case for ongoing
care. Retrieved from http://partnershipforsolutions.org/DMS/files/chronicbook2004.pdf
Pessoa, L. (2009). How do emotion and motivation direct executive control? Trends in Cognitive
Science, 13 (4), 160-166. doi:10.1016/j.tics.2009.01.006
Peters, R. (2006). Ageing and the brain. Postgraduate Medical Journal, 82 (964), 84-88.
doi:10.1136/pgmj.2005.036665
Pitkala, K. H., Laakkonen, M. L., Strandberg, T. E., & Tilvis, R. S. (2004). Positive life
orientation as a predictor of 10-year outcome in an aged population. Journal of Clinical
Epidemiology, 57 (4), 409-414. doi:10.1016/j.jclinepi.2003.07.013
Plassman, B. L., Langa, K. M., Fisher, G. G., Heeringa, S. G., Weir, D. R., Ofstedal, M. B., …
Wallace, R. B. (2007). Prevalence of dementia in the United States: The Aging,
Demographics, and Memory Study. Neuroepidemiology, 29 (1-2), 125–132.
doi:10.1159/000109998
Post, S. G. (2005). Altruism, happiness, and health: It’s good to be good. International Journal
of Behavioral Medicine, 12 (2), 66-77. doi:10.1207/s15327558ijbm1202_4
Qiu, C., Winblad, B., Fratiglioni, L. (2005). The age-dependent relation of blood pressure to
cognitive function and dementia. The Lancet Neurology, 4 (8), 487-499.
doi:10.1016/S1474-4422(05)70141-1
164

Reed, A. E., & Carstensen, L. L. (2012). The theory behind the age-related positivity effect.
Frontiers of Psychology, 3 (3), 339. doi:10.3389/fpsyg.2012.00339
Reinhard, S. C., Feinberg, L. F., Choula, R., & Houser, A. (2015, July). Valuing the invaluable:
2015 update. Retrieved from http://www.aarp.org/content/dam/aarp/ppi/2015/valuing-
the-invaluable-2015-update-new.pdf
Roberts, R., & Knopman, D. S. (2013). Classification and epidemiology of MCI. Clinical
Geriatric Medicine, 29 (4), 753-72. doi:10.1016/j.cger.2013.07.003
Rothrauff, T., & Cooney, T. M. (2008). The role of generativity in psychological well-being:
Does it differ for childless adults and parents? The Journal of Adult Development, 15 (3),
148-159. doi:10.1007/s10804-008-9046-7
Rowe, J. W., & Kahn, R. L. (1997). Successful aging. The Gerontologist, 37 (4), 433-440.
doi:10.1093/geront/37.4.433
Ryff, C. D. (1995). Psychological well-being in adult life. Current Directions in Psychological
Science, 4 (4), 99-104. doi:10.1111/1467-8721.ep10772395
Ryff, C. D., Almeida, D., Ayanian, J., Binkley, N., Carr, D., Coe, C., … Williams, D. (2017).
National Survey of Midlife Development in the United States (MIDUS III), 2013-2014.
Ann Arbor, MI: Inter-university Consortium for Political and Social Research.
doi:10.3886/ICPSR36346.v5
Ryff, C. D., Almeida, D. M., Ayanian, J., Carr, D. S., Cleary, P. D., Coe, C., … Williams, D.
(2012). National Survey of Midlife Development in the United States (MIDUS II), 2004-
2006. Ann Arbor, MI: Inter-university Consortium for Political and Social Research.
doi:10.10.3886/ICPSR04652.v6
165

Ryff, C. D., & Lachman, M. E. (2013). National Survey of Midlife Development in the United
States (MIDUS II), Cognitive Project, 2004-2006. Ann Arbor, MI: Inter-university
Consortium for Political and Social Research. doi: 10.3886/ICPSR25281.v5
Salthouse, T. A. (1991). Theoretical perspectives on cognitive aging. Hillsdale, NJ: Lawrence
Erlbaum Associates, Inc.
Salthouse, T. A. (2006). Mental exercise and mental aging: Evaluating the validity of the “Use It
or Lose It” hypothesis. Perspectives on Psychological Science, 1 (1), 68-87.
doi:10.1111/j.1745-6916.2006.00005.x
Schoklitsch, A., & Baumann, U. (2012). Generativity and aging: A promising future research
topic? Journal of Aging Studies, 26, 262-272. doi:10.1016/j.jaging.2012.01.002
Schooler, C. (2007). Use it—and keep it, longer, probably: A reply to Salthouse (2006).
Perspectives on Psychological Science, 2 (1), 24-29. doi:10.1111/j.1745-
6916.2007.00026.x
Schwartz, C., Meisenhelder, J. B., Ma, Y., & Reed, G. (2003). Altruistic social interest behaviors
are associated with better mental health. Psychosomatic Medicine, 65 (5), 778-785.
doi:10.1097/01.psy.0000079378.39062.d4
Seeman, T. E., Lusignolo, T. M., Albert, M., & Berkman, L. (2001). Social relationships, social
support, and patterns of cognitive aging in healthy, high-functioning older adults:
MacArthur Studies of Successful Aging. Health Psychology, 20 (4), 243-255.
doi:10.1037/0278-6133.20.4.243
Seeman, T. E., Miller-Martinez, D. M., Merkin, S. S., Lachman, M. E., Tun, P. A., &
Karlamangla, A. S. (2011). Histories of social engagement and adult cognition: Midlife in
166

the U.S. Study. The Journals of Gerontology, Series B: Psychological Sciences and
Social Sciences, 66B (S1), 141-152. doi:10.1093/geronb/gbq091
Smith, J., Fisher, G., Ryan, L., Clarke, P., House, J., & Weir, D. (2013). Psychosocial and
lifestyle questionnaire 2006—2010 (HRS Documentation Report Core Section LB). Ann
Arbor, MI: Institute for Social Research, University of Michigan. Retrieved from
http://hrsonline.isr.umich.edu/sitedocs/userg/HRS2006-2010SAQdoc.pdf
Stephan, Y., Caudroit, J., Jaconelli, A., & Terracciano, A. (2013). Subjective age and cognitive
functioning: A 10-year prospective study. American Journal of Geriatric Psychiatry.
doi:10.1016/j.jagp.2013.03.007
Stewart, A. J., & Vandewater, E. A. (1998). The course of generativity. In D. P. McAdams & E.
de St. Aubin (Eds.), Generativity and adult development: How and why we care for the
next generation (pp. 75-100). Washington, DC: American Psychological Association.
Stine-Morrow, E. A. L., Parisi, J. M., Morrow, D. G., & Park, D. C. (2008). The effects of an
engaged lifestyle on cognitive vitality: A field experiment. Psychology and Aging, 23 (4),
778-786. doi:10.1037/a0014341
Swaab, D. F. (1991). Brain aging and Alzheimer's disease: “Wear and tear” versus “Use it or
lose it.” Neurobiology of Aging, 12 (4), 317-324. doi:10.1016/0197-4580(91)90008-8
Tang, F. (2009). Late-life volunteering and trajectories of physical health. Journal of Applied
Gerontology, 28 (4), 524-533. doi:10.1177/0733464808327454
Thoits, P. A., & Hewitt, L. N. (2001). Volunteer work and well-being. Journal of Health and
Social Behavior, 42 (2), 115-131. Retrieved from http://www.jstor.org/stable/3090173
167

Tun, P. A., & Lachman, M. E. (2005, January 21). Brief Test of Adult Cognition by Telephone
(BTACT): Technical report. Retrieved from
http://midus.wisc.edu/midus2/project3/TechReport_BTACT.pdf
Tun, P. A., & Lachman, M. E. (2006). Telephone assessment of cognitive function in adulthood:
The Brief Test of Adult Cognition by Telephone. Age and Ageing, 35 (6), 629-632.
doi:10.1093/ageing/afl095
Tuokko, H., Morris, C., & Ebert, P. (2005). Mild cognitive impairment and everyday functioning
in older adults. Neurocase, 11 (1), 40-47. doi:10.1080/13554790490896802
United States Census Bureau (2015, March). Projections of the size and composition of the U.S.
population: 2014 to 2016. Retrieved from
https://www.census.gov/content/dam/Census/library/publications/2015/demo/p25-
1143.pdf
University of Wisconsin Institute on Aging. (2015). Documentation of psychosocial constructs
and composite variables in MIDUS III, Project 1. Madison, WI: University of Wisconsin.
University of Wisconsin Survey Center. (2015, July). Midlife in the United States (MIDUS III)
Field Report. Madison, WI: University of Wisconsin.
Van Praag, H., Shubert, T., Zhao, C., & Gage, F. H. (2005). Exercise enhances learning and
hippocampal neurogenesis in aged mice. Journal of Neuroscience, 25 (38), 8680-8685.
doi:10.1523/jneurosci.1731-05.2005
Van Willigen, M. (2000). Differential benefits of volunteering across the life course. Journal of
Gerontology: Social Sciences, 55B (5), S308-S318. doi:10.1093/geronb/55.5.S308
Varma, V. R., Carlson, M. C., Parisi, J. M., Tanner, E. K., McGill, S., Fried, L. P., …
Gruenewald, T. L. (2014). Experience Corps Baltimore: Exploring the stressors and
168

rewards of high-intensity civic engagement. The Gerontologist, 55 (6), 1038-1049.
doi:10.1093/geront/gnu011
Villar, F. (2012). Successful ageing and development: The contribution of generativity in older
age. Ageing & Society, 32 (7), 1087-1105. doi:10.1017/S0144686X11000973
Vitaliano, P. P., Murphy, M., Young, H. M., Echeverria, D., & Borson, S. (2011). Does caring
for a spouse with dementia promote cognitive decline? A hypothesis and proposed
mechanisms. Journal of the American Geriatrics Society, 59 (5), 900-908.
doi:10.1111/j.1532-5415.2011.03368.x
Walther, K., Birdsill, A. C., Glisky, E. L., & Ryan, L. (2010). Structural brain differences and
cognitive functioning related to body mass index in older females. Human Brain
Mapping, 31 (7), 1052-1064. doi:10.1002/hbm.20916
Wang, D., & Gruenewald, T. (2017). The psychological costs of social support imbalance:
Variation across relationship context and age. Journal of Health Psychology.
doi:10.1177/1359105317692854
Zunzunegui, M. –V., Alvarado, B. E., Del Ser, T., & Otero, A. (2003). Social networks, social
integration, and social engagement determine cognitive decline in community-dwelling
Spanish older adults. Journal of Gerontology: Social Sciences, 58B (2), S93-S100.
doi:10.1093/geronb/58.2.S93

Abstract (if available)

Abstract The developmental phenomenon known as generativity collectively refers to those behaviors and affective states which originate from the desire to contribute in a positive and productive manner to the welfare of others. hough first identified as a psychosocial impulse of midlife, generativity has since been recognized for its developmental significance in older adulthood. Generativity has been shown to predict a number of desirable outcomes in later life, including those related to psychological well-being (e.g., greater autonomy, mastery, and purpose in life, as well as fewer depressive symptoms) and physical health (e.g., lesser risk of disability and mortality). Preliminary evidence suggests that generativity may also hold promise as a tool to enhance cognitive function in older adulthood, though only a small handful of studies have investigated this possibility in short-term, experimental contexts. In order to more robustly evaluate generativity’s potential to facilitate cognitive vitality across the second half of life, larger-scale longitudinal investigations of generativity—cognition associations are warranted. ❧ The current set of dissertation studies sought to examine longitudinal associations between both (1) generative contributory activity (encompassing volunteering, caregiving, and social support provision) and cognition

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Lifting up ourselves by lifting up others: examining cognitive-affective states linking generativity to well-being

PDF

Life transitions, leisure activity engagement, and cognition among older adults

PDF

Investigating the effectiveness of a social work intervention on reducing hospital readmissions among older adults

PDF

Using mixed methods to identify the characteristics of older fraud victims

PDF

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

PDF

Examination of the long-term psychosocial and functioning effects of a healthy living, life management behavior intervention for older adults

PDF

The effects of time and space on health status of older adults in China

PDF

Intergenerational transmission of values and behaviors over the family life course

PDF

Self-perceptions of Aging in the Context of Neighborhood and Their Interplay in Late-life Cognitive Health

PDF

Predicting cognitive decline and dementia in elderly twins from indicators of early life oral health

PDF

Translating two physical activity programs for older adults into home- and community-based settings: "Active Start" and "Healthy Moves for Aging Well"

PDF

Exploring end-of-life care experiences and disparities among Hispanic populations in the United States

PDF

Neuropsychological functioning in older adult long-term cancer survivors

PDF

Family caregiver well-being: generational differences and coping strategies

PDF

Cognitive health and self-rated memory in later life: linkages to race/ethnicity, multimorbidity, and survival

PDF

The dollars and sense of caregiving: exploring the financial financial realities of providing informal care to an older adult

PDF

Association of carotid artery stiffness with measures of cognition in older adults

PDF

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

PDF

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

PDF

Neuroinflammation and the behavioral consequences of air pollution over the life course

Asset Metadata

Creator Prickett, Elizabeth Hagood (author)

Core Title Thinking generatively versus acting generatively: exploring the associations of generative self-concept and generative activity with cognitive function among older adults

School Leonard Davis School of Gerontology

Degree Doctor of Philosophy

Degree Program Gerontology

Publication Date 08/02/2019

Defense Date 05/04/2017

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

Tag aging,cognition,generativity,OAI-PMH Harvest,older adults

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Enguidanos, Susan (committee chair), Abdou-Kamperveen, Cleopatra (committee member), Gruenewald, Tara (committee member), Wilber, Kate (committee member)

Creator Email hagood@usc.edu,lizhagood@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-422497

Unique identifier UC11265346

Identifier etd-PrickettEl-5677.pdf (filename),usctheses-c40-422497 (legacy record id)

Legacy Identifier etd-PrickettEl-5677.pdf

Dmrecord 422497

Document Type Dissertation

Rights Prickett, Elizabeth Hagood

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

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