STATE FLORIDA UNIVERSITY COLLEGE OF SOCIAL SCIENCES
LATER LIFE SOCIAL ENGAGEMENT AND HEALTH
By MANACY PAI
A Dissertation submitted to the Department of Sociology in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Degree Awarded: Summer Semester 2008
The members of the Committee approve the Dissertation of Manacy Pai defended on May 12, 2008. ______________________________ Anne E. Barrett Professor Directing Dissertation ______________________________ Mary A. Gerend Outside Committee Member ______________________________ Jill Quadagno Committee Member ______________________________ Robin W. Simon Committee Member Approved: ______________________________________ Irene Padavic, Chair, Department of Sociology
_____________________________________________ David W. Rasmussen, Dean, College of Social Sciences The office of Graduate Studies has verified and approved the above named committee members.
ii
ACKNOWLEDGEMENTS My thanks go first and foremost to Dr. Anne Barrett, my mentor and teacher, whose constant support, encouragement, and patience have culminated in the completion of this dissertation. Her confidence in me is unwavering and her commitment towards my advancement remains relentless. She is someone I can count on someone I trust, respect, and look forward to learning from throughout my academic career. Completion of this journey would not have been possible without the generous support of my committee members, Drs. Robin Simon, Jill Quadagno, and Mary Gerend. Dr. Simon has been like a second mentor to me. Her constant nodding and smiles during my first ASA presentation and practice job talk were nothing but a sheer source of strength that carried me through with poise. I am particularly indebted to her for having recommended and introduced me to so many wonderful people at Kent State. I also am thankful to Dr.Quadagno for her guidance and willingness to share her knowledge and expertise throughout my time here at FSU. Her down-to-earth approach and an unassuming nature have inspired me along every step of the way. I also am grateful to her for having given me my own office, which has helped me complete in a timely fashion not only this dissertation, but several other lingering projects. I want to offer a special thank you to Dr.Gerend, who as a non-sociologist has pushed me to think outside the box. I want to thank her particularly for having taken the time to listen patiently and offer invaluable suggestions on my second practice job talk presentation. I also am grateful to Dr. John Reynolds, Matt Gayman, and Matt Dutton for their willingness to answer my endless quantitative queries and help me figure out the constantly emerging statistical conundrums. Despite not being on my committee, Dr.Reynolds has been very generous with his time and guidance on both my masters and dissertation projects. He has been an amazing nextdoor neighbor in Pepper and I cannot thank him enough for his support during the several ups and downs of the job market process. I would be remiss if I did not mention Jamie Yeargan and Sharon Chapman, whose assistance with various administrative details of completing my dissertation has been invaluable. With their help, they have helped me stay on the top of everything. My special thanks to Dr. Lessan and Juanito who have been there for me through the ups and downs of graduate school. Over the years, Dr.Lessan has taught me how to look at the glass as half full instead of half empty an attitude that has served me well during these years away from home. Dr.Lessan and Juanito have become my family over the years and their friendship is something that I will cherish throughout my life. The one person who kept me sane through this entire process of graduate school is my roommate and best friend, Pallavi. Her ability to withstand my temper and listen patiently to my countless sociological and feminist arguments day in and day out continues to amaze me. Pallavi has not only socially but emotionally and instrumentally supported me throughout my time here at FSU. Her constant support and encouragement truly have enriched both my personal and professional endeavors.
iii
Finally, I would like to thank my parents for encouraging me to pursue my goals. Without their unconditional support and unwavering confidence, this dissertation and more importantly, this journey would not have been possible. I would like to thank them for their countless prayers, never-ending patience, and love that is beyond compare. More importantly, I would like to thank them for being that one constant factor in my life I can always count on. How they managed to stay away from their only child and support with such overwhelming enthusiasm my decision to pursue a life, which is this far is beyond me. They are the true definition of a blessing; they are a gift that is unearned and invaluable. To me, not even God is worthy of the worship that my parents are.
iv
TABLE OF CONTENTS List of Tables Abstract ........vi .....vii
1. Introduction...1 2. Theoretical Overview5 3. Review of Empirical Literature.....9 4. Methods.25 5. Descriptive Results35 6. Single Roles and Role Count.45 7. Role Combinations56 8. SES and Social Engagement.71 9. Race and Social Engagement87 10. Summary of Findings..98 11. Discussion...107 12. Conclusion...123 APPENDICES...126 REFERENCES..132 BIOGRAPHICAL SKETCH.148
v
LIST OF TABLES Table 1 Table 2a Table 2b Table 3a Table 3b Table 3c Table 3d Table 3e Table 3f Table 4a Table 4b Table 5 Table 6 Table 7 Table 8 Table 9 Table 10 Table 11 Table 12 Table 13 Table 14 Description of the Study Variables36 Single Role and Role Count Distribution by Race.37 Role Combination Distribution by Race................37 Single Role and Role Count Distribution by Education................38 Single Role and Role Count Distribution by Income................39 Single Role and Role Count Distribution by Assets..39 Role Combination Distribution by Education...............39 Role Combination Distribution by Income..............40 Role Combination Distribution by Assets...............40 Mean Levels of Depressive Symptoms by Type of Social Roles....41 Mean Levels of Functional Impairment by Type of Social Roles...41 Correlation Matrix of Study Variables....43 OLS Regression of Wave 2 Mental Health on Single Roles...48 OLS Regression of Wave 2 Physical Health on Single Roles.51 OLS Regression of Wave 2 Mental Health on Multiple Roles...53 OLS Regression of Wave 2 Physical Health on Multiple Roles.54 OLS Regression of Wave 2 Mental Health on Role Combinations59 OLS Regression of Wave 2 Physical Health on Role Combinations..67 OLS Regression of Wave 2 Mental Health on Single Roles and Education.75 OLS Regression of Wave 2 Mental Health on Multiple Roles and Education, Income, and Assets77 OLS Regression of Wave 2 Mental Health on Role Combination and Education...78 vi
Table 15a Table 15b Table 16 Table 17 Table 18 Table 19 Table 20 Table 21 Table 22 Table 23 Table 24
OLS Regression of Wave 2 Physical Health on Single Roles and Income81 OLS Regression of Wave 2 Physical Health on Single Roles and Assets.83 OLS Regression of Wave 2 Physical Health on Multiple Roles and Education, Income, and Assets..85 OLS Regression of Wave 2 Mental Health on Single Roles and Race.89 OLS Regression of Wave 2 Mental Health on Multiple Roles and Race..91 OLS Regression of Wave 2 Physical Health on Multiple Roles and Race93 OLS Regression of Wave 2 Physical Health on Multiple Roles and Race95 Summary of the Mental and Physical Health Effects of Single Roles and Multiple Roles98 Summary of Mental and Physical Health Effects of Role Combinations..100 Summary of SES Variations in the Mental and Physical Health Effects of Social Engagement.101 Summary of Race Differences in Mental and Physical Health Effects of Social Engagement.105
vii
ABSTRACT The topic of later life social engagement is of growing importance worldwide given the confluence of large demographic, economic, and sociocultural trends. However, empirical research in this area faces some serious theoretical and methodological limitations. The purpose of this study was to address these limitations by: (1) examining the later life mental and physical health effects of five social roles, namely paid work, formal volunteering, informal helping, caregiving, and religious participation; (2) assessing the health impact of combining various social roles among older adults; and (3) evaluating whether the health effects of social engagement differ for older adults of varying race and SES groups. These questions were examined using data from two waves (1986 and 1989) of the Americans Changing Lives Study (ACL); this nationally representative data, subset to adults aged 50 and over (N = 1,614), are among the best available for research on later life social engagement. Ordinary Least Square Regression is employed to estimate multivariate linear regression models that assess relationships among aspects of social engagement, social structure, and mental and physical health. Bivariate analyses reveal race and SES variations in social engagement. In particular, while White and upper SES older adults have higher rates of participation in paid work, formal volunteering, and caregiving, a significantly greater proportion of Black and lower SES elders are engaged in activities that are less formal, including informal helping and religious participation. Multivariate analyses indicate that certain roles and role clusters are especially health-enhancing. Specifically, three roles, including paid work, formal volunteering, and informal helping, result in improvements in mental health. Neither caregiving nor religious participation is found to affect changes in mental health in this sample of older adults. In terms of physical health, all roles except caregiving are significantly predictive of improvements in functional health. In addition to single roles, multiple role performance is found to positively impact both mental and physical health in later life. Findings related to role combinations revealed that being involved in volunteer work, informal helping, and religious activities enhances mental health among those who are caregivers. Results also indicated that the impact of being an informal helper is greater for those involved in religious activities than their peers who are not religious participants. Two role combinations emerged as statistically significant predictors of physical health. In particular, both volunteering and informal helping enhanced the physical health benefits of paid work in later life. In other words, while working for pay results in improved physical health, the benefits attached to this role are greater for those who also are involved in helping activities, such as formal volunteering and informal helping. Analyses related to SES reveal that lower SES older adults benefit more from occupying several social roles compared to their upper SES counterparts. In particular, while the mental health benefits of working for pay, formal volunteering, and informal helping are greater for older adults with lower educational attainment, the physical health gains of paid work, volunteering, and informal helping are greater for the elderly with low levels of income and assets. Unlike SES, however, overall findings revealed that White and Black older adults are more similar than different in terms of the health benefits they derive from being socially engaged. There are only viii
two instances, where race differentials emerge. First, I find older Blacks but not Whites to benefit from paid work; and second, results revealed older African caregivers to report greater functional impairment than their White peers. The present study on social engagement in later life has important implications for social policy and social work practice and research. A solid understanding of the consequences associated with late life social involvement is critical to policymakers and practitioners as they attempt to better address the challenges and capitalize upon the enormous growth of elders. As these findings suggest, older adults do benefit from social engagement. However, the mental and physical health benefits of certain roles vary by older adults race and SES. These group variations may indicate that the meaning, importance, and health effects of social engagement may be socially and economically driven. Specifically, my study points out that our efforts to promote social engagement should include our efforts to identify the differences within the elderly population.
ix
CHAPTER ONE INTRODUCTION Social engagement and its relevance to mental and physical health have been of central interest to social scientists for over a century. Durkheims (1951 [1897]) classic theory linking social engagement to suicide sparked this concern, with researchers over a century showing a sustained interest in the topic. Over the past several decades, gerontologists also have shown an unrelenting interest in this issue, with early research suggestive of the health salutary effects of being socially active among older adults (Havighurst 1957). Activity, continuity and role theories all underscore connections between social engagement and health for the elderly. Theorists and researchers have argued that social engagement is a unique form of activity and an essential component of not only maintaining health (Rowe and Kahn 1997), but also recovering from illness and impairment in later life (Svanborg 2001). Moreover, the empirical literature has overwhelmingly indicated a strong link between later life social engagement and health. Social engagement has been measured in many ways; but, one of the key indicators of social involvement and determinants of well-being has been participation in social roles. Occupying multiple social roles such as paid worker, formal volunteer, and religious participant has been positively associated with health (Berkman and Breslow 1983; House, Landis, and Umberson 1988; Moen, Dempster-McClain, and Williams 1989; Moen, Dempster-McClain, and Williams 1992). In addition to the multiple roles perspective, which emerged from social psychology, a theory developed by gerontologists - activity theory - argues that being involved in social roles confers upon individuals diverse personal, social, and psychological resources, all of which are likely to result in greater mental and physical well-being (Gubrium 1972, 1973; Havighurst 1957; Herzog and House 1991). Research stemming from each of these perspectives suggests that the more roles individuals occupy, the better their mental and physical health. However, most prior research on health consequences of later life social engagement has involved examining a limited range of social roles, such as formal volunteer, caregiving, and religious participant. Other important social roles occupied by older adults, such as paid work and informal helping, remain largely under examined. For instance, while extensive research exists on retirement and its health effects, far less is known about the health consequences of paid work. 1
Similarly, while it is well-established that volunteering involves both formal and informal activities, studies examining this role have focused mainly on formal volunteering. Social engagement, as indicated by participation in a wide range of roles, has remained underexamined because older adults are often assumed to be socially disengaged. Yet, if the definition of social and productive engagement is broadened to include activities that are unpaid, informal, and less visible, we find that a significant proportion of individuals do remain socially and productively engaged until later in life. Their contributions as paid workers and informal helpers have important and underappreciated benefits for their own well-being and that of their families and communities. Failing to consider these social roles results in neglect of the full range of work done by older adults especially those who belong to minority groups. For example, while retirement may be normative for upper SES elderly, lower SES older adults may be forced to work until later in life. Neglecting the role of paid work in later life has resulted in little understanding of its effects on the socioeconomically disadvantaged segments of our society. Similarly, by excluding informal helping, researchers have left out the oldest-old individuals who are less likely to be engaged in formal volunteering given health limitations and transportation issues (Fischer and Shaffer 1993; Rochester and Hutchinson 2002), but may be actively providing help to neighbors, friends, and family living in close geographic proximity. In sum, by overlooking these roles, researchers essentially have overlooked the important and ongoing work performed by the marginalized groups of our population. Further, much of the existing research on social engagement as defined as performance of meaningful social roles has focused on the quantity of roles older adults occupy, overlooking potential variation in the effects of combining diverse social roles. Role combinations are important to consider because failing to do so inevitably leads to an assumption that all roles are similar in their meaning, quality, and effects. However, prior research has shown that some roles, such as caregiving, are more health deleterious than health enhancing. In contrast, volunteering is found to positively influence older peoples well-being. Furthermore, role identity researchers would argue that because some roles are more central to ones self, they are more likely than the less central ones to have strong effects on ones well-being. In addition to the centrality of roles to 2
identity, successful fulfillment of some roles (e.g., paid worker role) may require more time, energy, and commitment than certain other roles (e.g., volunteer role). Finally, it is reasonable to assume that while most roles may be taken on willingly, some may involve an involuntary incumbency. For instance, while volunteering may involve ones own will, caregiving may represent an involuntary transition. Role combinations, therefore, are worth considering because roles can offset the effects of one another and the combining of some roles may be health enhancing while other combinations are health deleterious. Finally, the premise that social engagement benefits all older adults is essentially problematic given that it has been a privilege for only certain segments of society. Critics of such concepts as social and productive engagement and successful aging assert that not all older adults may choose to remain socially engaged. Put simply, elders who are women, poor, or members of racial and ethnic minorities may find themselves obliged to be productive for two reasons. First, they may have considerable economic needs compelling them to work or personal commitments that demand their time be spent in a particular social role (e.g., caregiving). Second, growing attention to the concepts of social productivity and successful aging may pressure older adults into assuming social roles even though their personal preferences may point in the opposite direction. It seems unlikely, therefore, that older adults who remain actively engaged either out of necessity or societal pressure would fully benefit from social engagement. Unfortunately, there also is relatively little empirical evidence to support any linkage between social engagement and health for different subgroups of elderly. Prior studies in this area have examined the aging population as a whole, without assessing how structural factors, such as gender, race, and socioeconomic status, affect the opportunities for and benefits from social engagement. With limited attention given to the influence of structural contexts on the protective effects of later life social engagement, we have little evidence upon which to claim that social engagement benefits all segments of the population. The present study addresses this limitation by examining more closely the link between later life social engagement and health within the context of race and socioeconomic status. Using data from the first two waves (1986 and 1989) of the Americans Changing Lives (ACL) study, a nationally representative sample (N = 3,617) of adults 24 years and older who 3
lived in the United States, this research examines five social roles paid worker, formal volunteer, informal helper, caregiver, and religious participant and their main effects and effects of combining them on older adults mental and physical health. These roles were chosen because they are considered as activities that are personally and socially meaningful to those in later life. As such, while the present study limits its focus to paid work, volunteering, informal helping, caregiving, and religious participation, future studies should consider other types of social engagement, including spouse, parent, and grandparent roles, which are likely to be just as valuable to older adults. To summarize, the purpose of the present study is to extend the current literature pertaining to social engagement and health among older adults. In particular, I address the following three questions: (1) How do certain underexamined roles, particularly paid work, and informal helping affect mental and physical health of the elderly? (2) How do various combinations of social roles affect well-being? (For example, do caregivers who also volunteer have better health than those who are only involved in caregiving?) (3) Is the relationship between later life social engagement and health contingent upon the race and socioeconomic status of the role occupants? (For instance, are the benefits of formal volunteering greater for those belonging to any one race or socioeconomic status?)
4
CHAPTER TWO REVIEW OF THEORETICAL LITERATURE Overview Within the past two decades, gerontologists have developed the concepts of social engagement, productive aging, and successful aging as desired outcomes to the aging process for two reasons. First, they view aging in a positive light; and, second, they highlight the promise and potential of an aging population (Baltes and Baltes 1990; Butler and Gleason 1985; Caro, Bass, and Chen 1993; Rowe and Kahn 1997). These concepts, therefore, pose as desirable alternatives to the stereotypical view of older adults as mere consumers and dependent members of our society. Two theories that have been most effective in explaining the underlying links between social engagement and successful aging (i.e., enhanced mental and physical health) are activity and role theories. Of these two, activity theory relates most directly to the present study. I review it here, and then highlight the basic tenets of role theory, a theory that is more central to sociology and provides the foundation for most of the questions developed in this study.
Activity Theory of Aging Based on ideas first articulated by Havighurst and Albrecht (1953) that posited a connection between activity and health and well-being in later life, Lemon and colleagues (1972) constructed the activity theory of aging. This theory included clearly identified characteristics and mediators explaining the link between social and physical activity and life satisfaction. Since these early conceptualizations, the relation between activity and well-being has been explored extensively. A recent version of activity theory, presented in Rowe and Kahns (1997) definition of successful aging, is similar to earlier conceptualizations (Havighurst and Albrecht 1953) of the theory. More specifically, active social engagement continues to be considered a fundamental component of successful aging. On the whole, the premise underlying activity theory is that being actively engaged, such as in social roles, leads to greater life satisfaction and enhanced mental and physical health (Lemon et al. 1972; Rowe and Kahn 1997). Activity theory proposes that both the frequency with which people participate in activities and their level of intimacy and salience are vital for well-being. That is, the greater the frequency 5
and salience of activity, the greater the benefits accruing from it. These two tenets led to the more specific hypothesis that individuals involved in informal social activities are likely to be more satisfied with life as compared with peers engaged in formal activity. However, those involved in formal activity are likely to fare better when compared with people who are engaged in isolated and solitary pursuits. Most prior research supports the first of the two tenets: Higher levels of activity, measured either in terms of range of activities individuals engage in or the frequency with which they engage in different activities, positively predict well-being (e.g., Beck & Page, 1988; Herzog, Franks, Markus, & Holmberg, 1998; Lawton, Winter, Kleban, & Ruckdeschel, 1999; Markides & Martin, 1979; Menec & Chipperfield, 1997). The second tenet of the activity theory regarding the importance of intimacy, however, has marshaled less empirical support. Although research generally supports the view that social activity is positively related to well-being (e.g., Garfein & Herzog, 1995; Litwin, 2000; Morgan et al., 1991), there is little empirical research indicating that informal social activities confer more benefits than formal social pursuits (Okun et al., 1984). On the whole, there is ample evidence for the activity theory in social gerontology and research in the sociology of aging and the life course (e.g., Dorfman, Kohout, & Heckert, 1985; George, 1990; Kilty & Behling, 1985; MacLean, 1982; McPherson & Guppy, 1979). What remains less well-understood are the types and combinations of activities and roles that are either beneficial or damaging to later life health. Are formal roles, such as paid work and formal volunteering, more health-enhancing than less formal roles, including informal helping and religious participation? Moreover, are the benefits of occupying various social roles similar or different for subgroups of the elderly population? These are the questions that remain underexamined in prior research on social engagement and health. The present study will address these questions.
Role Theory Another theory commonly used to explain social activity in later life is role theory. Concepts and principles of this theory are widely used in the specific discussions of role occupancy in later life and its propositions are highly useful to the study of the impact of 6
engagement on well-being, especially as they relate to multiple role occupancy (e.g., Adelmann 1994; Hinterlong et al. 2001; Hinterlong et al. 2007; Moen et al., 1989, 1992; Van Willigen 2000). Role theory yields two competing hypotheses regarding the impact of multiple roles on the individual: role strain (Goode 1960) and role enhancement (Sieber 1974). The role strain hypothesis is primarily based on structural-functionalism (Goode 1960; Merton 1957; Parsons 1951), which focuses on the part played by social structure in directing human behavior. The basic assumption underlying structural functionalism is that roles and statuses consist of a set of rules and norms that guide behavior (Biddle 1986). Structural functionalists view individuals as actors who are assigned specific scripts for the roles they occupy, having to learn the appropriate behavior and then act accordingly in a given role. The role strain hypothesis, which is based on the structural-functionalist perspective, posits that occupying multiple roles negatively affects well-being. Goode (1960, p.483) defined role strain as the difficulty in fulfilling role obligations. Sieber (1974) further pointed out that the notion of role strain is comprised of two overlapping concepts, namely role overload and role conflict. The role overload perspective assumes that people have only a limited amount of time, energy, and resources and therefore lack sufficient means to successfully fulfill the obligations of diverse social roles simultaneously. The failure to successfully fulfill these obligations leads to psychological strain on the individual. In contrast with the role overload perspective, the role conflict perspective posits that it is not as much about resources as it about the conflicting demands emerging from different social roles at any given time (Merton 1957). For instance, employed women in a society with scarce alternative long-term care arrangements face role conflict because the pressures arising in the paid work role are often incompatible with the pressures that emerge within the caregiver role. Although both role overload and role conflict perspectives offer differing explanations for role strain, they share the common view that people with multiple roles suffer from role strain, ultimately resulting in poor health (Stryker and Statham 1985). An alternative viewpoint, the role enhancement perspective, posits that involvement in multiple social roles is greatly beneficial to individuals health. This perspective is embedded in symbolic interactionist theory, which considers social structure merely as a broad guide for interactions and activities and sees human beings as having agency and control over their roles, 7
statuses, and behaviors. For instance, Thoits (1987, p. 12) argued that human beings are not just robots programmed by society. They are willful actors, capable of choosing nonconformity and altering social structure if they so wish. Role enhancement theory thus views an individual not only as an actor but also the director of a given social role, having the power to negotiate, bargain, and balance between and among differing role obligations. The role enhancement perspective does not consider role overload and role conflict to be essential consequences of multiple role occupancy but instead suggests that the costs associated with occupying multiple roles are often outweighed by the benefits attached to them. For instance, Sieber (1974, p. 569) argued that role accumulation offers the individual four types of positive outcomes: role privileges, overall status security, resources for status enhancement, and enrichment of the personality and ego gratification. Similarly, drawing from identity theory and using the concept of role as equivalent to the concept of identity, Thoits (1983, 1986) asserts that multiple identities (based on multiple roles) provide individuals with existential security, purpose, and meaning, all of which ultimately reduce feelings of distress and promote well-being. Although both role strain and role enhancement perspectives have received attention in empirical research, the latter has gathered more support. That said, researchers over time have suggested that the relationship between multiple roles and health is far more complex that it may initially appear. For instance, while multiple role perfomance has consistently found to predict better mental and physical health, this relationship is contingent on several interrelated factors, including quality of the role, role meaning, and role satisfaction. One factor that has received less attention in this research is the configuration of social roles, which raises the possibility that one role may influence the benefits or costs associated with other roles. Stated otherwise, we are less aware of how combining different social roles affects the health of older adults. Are certain role clusters more health enhancing or health deletrious in later life? Are the health effects of combining various social roles similar across subgroups of the elderly population? The current study will employ role theory to address these questions.
8
CHAPTER THREE REVIEW OF EMPIRICAL LITERATURE Social Roles and Well-being Social roles, in general, are considered beneficial to ones health. The assumption underlying this thought is that being involved in a social role provides individuals with power, prestige, and resources -- all of which are likely to positively impact their mental as well as physical health. Below, I review findings related to the health effects of each of the five social roles that this study examines: paid worker, formal volunteer, informal helper, caregiver, and religious participant.
Paid Work and Health In a society that greatly values material success and productivity, paid work plays an integral part of peoples everyday lives, providing them with financial security, personal identity, and an opportunity for meaningful involvement in the community. Given the significance of the paid work role, researchers have extensively examined the link between working for pay and health. Empirical results suggest that employment, leads to both mental and physical well-being. A considerable body of research documents the substantial and varied health consequences of paid work. In studies examining the mental health effects of employment, paid work is found to predict lower levels of distress and depression, as well as greater life satisfaction and higher selfesteem (Aneshensel, Fredericks, and Clark 1981; Gore and Mangione 1983; Haw 1982; Mackie 1983; Menaghan 1989; Radloff 1975; Reskin and Coverman 1985; Thoits 1986). Similarly, research examining physical well-being shows a strong, positive association between paid work and various measures of health, including self-perceived health (e.g., Ross and Bird 1994), functional health (e.g., Herold and Waldron 1985; Linn, Sandifer, and Stein 1985), and rates of serious disease, hospitalization, and mortality (Passannante and Nathanson 1985; Romelsjo et al. 1992; Verbrugge 1983, 1986).
9
Much of the existing research related to paid work and health, however, is limited to young and middle-aged workers with little attention given to the health consequences of employment among older adults. Studies that have focused on older workers have mainly examined retirement and job loss, which are related to, but not the same as, examining the health effects of paid work. Although findings related to retirement have consistently shown the health benefits of paid work in the lives of the elderly, few studies have directly examined the health consequences of working for pay in late life. For example, the workplace is found to be an important source of friendships, and social support for elders (Boss et al. 1990). Participating in more than 100 hours of paid work after retirement has also been shown to positively impact older peoples self-perceptions and well-being (Luoh and Herzog 2002). Researchers have noted, however, that conditions of work often modify the relationship between work and health. That is, those with more autonomy, creativity, and occupational self-direction reap greater benefits from paid work as opposed to their counterparts who work in less autonomous work environments (Ross and Van Willigen 1997). Older workers especially are less likely to have jobs with such desired characteristics. Although few studies have examined how employment affects the health of older adults, an even more striking absence in this literature centers on the health effects of combining paid work with unpaid work roles. An exception to this is research on paid work and caregiving, which suggests that while caregivers experience heavier burden and greater job-related stress than noncaregivers, working outside represents an opportunity for social interaction (Goldsmith and Goldsmith 1995). What remains unknown are the health effects of combining paid work with other unpaid roles, such as volunteering and informal helping. Further, we have yet to examine whether the health effects of combining paid work with other roles in later life vary by race and socioeconomic status. By not extensively examining the effects of employment in later life, researchers have overlooked the lives of several specific subgroups of our elderly population, including women, race-ethnic minorities, and the poor. More specifically, a higher proportion of racial minorities and the poor as opposed to Whites and the rich are forced by economic necessity to work until late in life. Moreover, it is these underprivileged groups of people who are most likely to suffer from poor physical health and work conditions. A comprehensive examination of paid work and health, therefore, should include paid work performed at later life and explore variation in the health effects of this role based on older adults race and socioeconomic position. 10
Formal Volunteering and Health Empirical research over the last three decades indicates that volunteering has strong positive effects on the psychological and physical health of older adults (Fengler 1984; Glass et al. 1995; Havighurst, Neugarten, and Tobin 1968; Herzog, Kahn, Morgan, Jackson, and Antonucci 1989; Wilson and Musick 1999; Van Willigen 2000). Researchers have offered explanations for the health-enhancing effects of volunteering in later life; and, the most important reason is that in a period of life that is typically characterized by role reduction, volunteering represents a unique opportunity for social activity and integration. In addition to social integration, providing help to others is found to be a self-validating experience coupled with feelings of trust, belief, and reciprocity (Riessman 1965). Similarly, being able to provide instrumental help to others in need is likely to instill in older adults confidence about their own abilities, which in turn is likely to reduce negative affect and foster feelings of personal control and self-esteem. The relationship between volunteering and well-being has generally been explored using the role theory concept of role enhancement. According to the role enhancement perspective, involvement in a social and productive role, such as volunteering, provides older adults with personal, social, and psychological resources (Moen et al. 1992), all of which lead to better health. For instance, Moen and colleagues (1992) found that of the different social roles used in their study, volunteering had an especially health salutary effect in later life. Similarly, Van Willigen (2000) found that being a volunteer is associated with increased life satisfaction; and, Thoits and Hewitt (2001) found increased life satisfaction, happiness, self-esteem, and self-reported physical health and reduced levels of depression in their sample of older volunteers. Morrow-Howell and colleagues (2003) found that volunteering positively affected self-reported health and improved functional health. Studies also have found that compared to non-volunteers, those involved in volunteering have lower rates of mortality (Musick et al. 1999; Oman et al. 1999; Luoh and Herzog 2002). Studies that have looked at hours of volunteering and their health effects have found that even a small amount of volunteering is associated with reduced distress and enhanced psychological well-being among older adults (Morrow-Howell et al. 2003; Musick et al. 1999; Musick and Wilson 2003; Van Willigen 2000), suggesting that not necessarily the amount of 11
volunteering but rather the mere involvement in volunteer work can spark positive spirits among older volunteers. Research on the mental health effects of volunteering also has examined the effects of the number of organizations, although the overall findings remain mixed. For instance, Oman and colleagues (1999) find that being a volunteer for two or more organizations positively impacts mental well-being; however, volunteering for just one organization has no impact on mental health. In contrast, Van Willigen (2000) found that each additional organization leads to greater positive effect; that is, being a volunteer for two organizations has a greater impact over being a volunteer for one organization, and being a volunteer for three organizations has an even greater impact than participating in two organizations. Research on the health benefits of volunteering also has considered to some extent the part that role context plays in conditioning the effects of being a volunteer. The concept of role context suggests that volunteering may not be equally advantageous to all. For example, while Fengler (1984) found disadvantaged older adults with poor health to report greater life satisfaction from volunteering as compared to the advantaged and more physically fit elderly, Musick and colleagues (1999) found the protective effect of being a volunteer to be strongest among those with lower social integration. In sum, there are numerous prior studies that demonstrate the health salutary effects of late life volunteering. However, less is known about the impact of volunteering in combination with other social roles. Studies should examine whether the health salutary effects of being a volunteer persist when volunteering is combined with more stressful roles, such as caregiving. It is logical to assume that stress emerging from caregiving can offset the benefits that accrue from being a volunteer. Alternatively, being involved in two helping activities may lead to enhanced selfperceptions and health among older adults. Whether or not these effects of combining volunteering with other roles differ between subgroups of older adults is yet another question that remains largely missing from previous research. The present study examines these issues.
Informal Helping and Health Other than a few descriptive reports (e.g., Hodgkinson and Weitzman 1996; Niyazi 1996), research has paid little attention to informal helping. Informal helping (or giving) includes 12
informal ways of helping out, such as running errands for a neighbor, baby-sitting, and driving a friend to a destination. Although informal volunteering activities such as these can be assumed to confer positive mental and physical health benefits for all elderly, there is little to no empirical investigation of this social role. However, we do know that providing help or service to others without the expectation of private gain (altruism) can be beneficial to peoples mental health (Midlarsky and Kahana 1994). Research reveals that altruistic acts predict higher levels of self-esteem (Krause and Shaw 2000), personal growth (Dulin et al. 2001; Fengler 1984), and psychological well-being (Krause, Herzog, and Baker 1992). Several explanations have been put forth to explain the link between informal helping and mental health among older adults. One is that, as human beings, we are naturally predisposed to helping others in need and by doing that we are likely to experience increased selfesteem and life satisfaction (Adler 1937). Similarly, Reissmans (1965) research on self-help groups suggests that helping someone who is in need especially a non-family individual results in enhanced perceptions of self, sense of mattering, and a feeling of self-validation all of which are associated with increased psychological well-being. Another theory proposed to explain the link between helping others and emotional well-being suggests that helping others in need helps older adults to cope with their own problems and troubles (Vaillant 2002). In addition to altruism, if we were to extend the scope of informal helping to informal emotional support, we find a few studies suggesting that being emotionally present for someone positively impacts ones mental well-being. For instance, Krause and Shaw (2000) found that helping others is associated with greater levels of self-esteem. Although this positive influence of providing emotional support persisted mainly for those belonging to upper social class, overall results indicated the health enhancing effects of being an emotional supporter. The role of informal helper/supporter is important to examine because ignoring this role results in neglecting the work done by several marginalized groups of society. Findings on the distribution of volunteering in the general population indicates that in contrast to all types of formal volunteering, women and minorities are generally more likely than men and Whites to be involved in informal helping (Hodgkinson and Weitzman 1996; Niyazi 1996). For instance, Blacks 13
and people from ethnic minority groups are actively engaged in informal helping at the community level. One reason for their lower rates of participation in formal volunteering is simply that they do not get asked to volunteer (Ferree et al. 1998; Hodgkinson 1995). Formal volunteering in our culture is predominantly a White, and middle-class activity, with non-Whites and other minority groups being less frequently asked to participate. Moreover, one might infer that given the cultural stereotypes and prejudices, race and ethnic minority people face barriers, including alienation, tokenism, and racism in their attempts to join formal groups (Akpeki 1995; Niyazi 1996). This suggests that both the poor and race and ethnic minorities are actively and consistently engaged in providing help to others around them. By overlooking the role of an informal helper, we essentially overlook the important and ongoing socially productive, and often personally rewarding, activities of the marginalized groups of our population. The limited research on informal helping/giving suggests that this role is linked with positive mental health. Future studies should examine whether these benefits extend to physical health. Another logical extension is to explore how informal helping and formal volunteering interact to influence well-being. Is one type of volunteering more beneficial than the other? Are the effects of one role stronger for any one subgroup of the population? How do each interact with other roles, such as caregiving and religious participation? This study addresses these issues by examining the individual effects of informal helping, as well as, the effects of combining informal helping with other roles on mental and physical health.
Caregiving and Health Unlike paid work, volunteering, and informal helping, much of the existing research related to caregiving finds this role to be associated with increased stress, burden, and psychological distress (Aneshensel et al. 1993; George and Gwyther 1986; Hoyert and Seltzer 1992; Schulz and Williamson 1991; Strawbridge et al. 1997; Marks and Lambert 1999). Caregiver stress -- a commonly employed concept in this literature -- is grounded in the broader framework of the stress process (Pearlin et al. 1990). According to this process, caregiver stress typically emerges out of two types of stressors, namely primary and secondary stressors. Primary stressors 14
include objective conditions related to the patients illness (e.g., cognitive impairment, problematic behavior, and/or physical dependencies) as well as more subjective reactions to the caregiving role, such as feelings of overload and captivity. Secondary stressors also emerge as a result of caregiving; however, they extend beyond the boundaries of the caregiving role reaching, for instance, into the family, work, and wider social network settings. In addition to the help they provide to care recipients, caregivers are often spouses, parents, and paid workers. Although these family and work-related roles are normative for most adults, pressures within each role are likely to increase when the role of a caregiver is added to the existing roles. For example, the chronic and often progressive makeup of the caregiving tasks is likely to fill up available time, leaving less room for oneself, fewer opportunities for social activities, and more difficulty in fulfilling other role obligations (Skaff and Pearlin 1992; Stoller and Pugliesi 1989). Combining such time and energy commitments can lead to role strain, negatively affecting both mental and physical well-being. Despite most research on the health consequences of caregiving emphasizing its negative aspects, some studies (Brody 1985; Krause, Herzog, and Baker 1992; Motenko 1989) have highlighted some of the more positive effects of providing care. Moreover, recent research focusing on both the negative and positive effects of caregiving, suggests that caregivers not only have adverse experiences, but also positive feelings related to their caregiver role (Kramer 1997). In general, the enjoyable aspects of caregiving reported by many caregivers include an enhanced relationship with the care recipient, personal growth, self-validation, and increased understanding of the aging process (Kramer 1997; Lawton et al. 1991). Although research reveals some positive effects of caregiving, the vast majority of studies suggest that the caregiver role has negative health impacts. Additionally, studies examining caregiving in conjunction with other family and work-related roles suggest that caregiving results in an added burden, interfering with the responsibilities attached to other roles. However, studies involving caregiving in the context of other roles have limited their examination to spouse, parent, and paid worker roles. What remains underexamined is how caregiving when combined with other social roles -- such as formal volunteering and informal helping -- affects mental and physical health. For example, do those who are engaged in caregiving and religious activities fare better, in 15
terms of their health than those who are only involved in caregiving? This pattern may emerge because the emotional and spiritual support drawn from religious congregations may diminish the stress that emerges from caregiving. Although prior research alludes to some of these issues, a more thorough analysis of them awaits future research.
Religious Participation and Health Research on the impact of religious involvement on health indicates a strong and positive link between religious participation and both mental and physical health. Theorists and researchers have discussed a range of possible reasons why religious participation may be predictive of enhanced mental and physical well-being. One perspective suggests that engagement in religious activity both reduces the exposure to and acts as a buffer against chronic and acute stressors. For instance, religious involvement is likely to foster healthy individual behaviors and lifestyle choices (Ellison 1994a, 1994b; George, Ellison, and Larson 2002; Grasmick, Bursik, and Cochran 1991), which are found to result in improved mental and physical health. Another perspective suggests that the salubrious effects of religious involvement on health result partly from the social resources afforded within religious communities. In other words, religious involvement provides a solid and consistent source of social integration which, in turn is a strong predictor of both mental and physical well-being. In particular, religious activity provides settings in which like-minded individuals who tend to share basic beliefs, values, interests, and traditions meet on a regular basis. Further, members of church may provide emotional support (e.g., companionship) as well as material aid (e.g., goods, services) to one another informally (Hatch 1991; Ellison and George 1994; Wuthnow 1994); and, social and instrumental support have been found consistenly to predict better mental and physical health (Antonnuci and Akiyama 1997; Berkman, Leo-Summers, and Horwitz 1992). Finally, psychological resources are another reason why, according to some studies, religious involvement predicts better health. For instance, ample research finds that religious involvement predicts better health partly because religious individuals enjoy higher levels of selfesteem and mastery than their peers who are not involved in religious activity. Religious groups 16
also may foster positive reflected appraisals through constructive feedback and partnership, thereby enhancing the self-esteem of their members (Ellison 1994). Although the foregoing suggests some consistently found health salutary effects of religious participation, we know little about how this relationship varies by important social contexts, such as SES. Further, little is known about how combining the religious participant role with other social roles affects the mental and physical health of older adults. For instance, are those involved in religious activity likely to benefit more from volunteering and informal helping as compared to their counterparts who are only engaged in religious activity? Given that service to others is a highly valued tenet in all religious groups, service-oriented roles, such as volunteering, informal helping, and caregiving, are likely to enhance both mental and physical health among religious participants. Religious participation, alternatively, may help those involved in roles, such as caregiving, by reducing stress that emerges from being a caregiver. These issues are far from examined in previous research and is the subject of investigation in the present study.
Role Combinations and Health As this literature review suggests, the link between social engagement and health is far more complex than it initially appears. Specifically, while empirical findings do suggest that social roles are generally health enhancing, they also indicate that some roles result in more negative than positive effects. Despite this variation in the effects of roles, a review of the previous research reveals that most studies have primarily attended to the mere quantity of social roles and not examined in detail how they may offset the effects of one another and how combining them affects people occupying diverse social locations. Role combinations are important to consider because failing to do so inevitably leads to an assumption that all roles are similar in their meaning, quality, and health effects (Menaghan 1989, 1991). However, prior research has shown that some roles, such are caregiving, can be more health deleterious than health enhancing. With the exception of a few studies that acknowledge some potentially beneficial effects of caregiving (Brody 1985; Krause, Herzog, and Baker 1992; Motenko 1989), most research suggests that providing care to an ailing spouse, parent, child, friend, or relative can be highly stressful, ultimately negatively impacting ones mental and 17
physical health (Aneshensel et al. 1993; George and Gwyther 1986; Hoyert and Seltzer 1992; Schulz and Williamson 1991; Strawbridge et al. 1997; Marks and Lambert 1999). In contrast, volunteering is found to have a more positive influence on well-being. Volunteering allows individuals to involve themselves in a productive social role, leading to increased psychological resources, such as mastery and self-esteem, a larger social network, and higher social integration (Moen, Dempster-McClain, and Williams 1992), all of which positively shape their mental as well as physical well-being (Fengler 1984; Glass et al. 1995; Havighurst, Neugarten, and Tobin 1968; Herzog, Kahn, Morgan, Jackson, and Antonucci 1989; Wilson and Musick 1999). In addition to differences in the effects of roles, role identity researchers (e.g., Simon 1995, 1997; Thoits 1983), argue that because some roles are more central to ones sense of self, they are more likely than the less central roles to strongly affect well-being. For instance, social roles, such as paid worker and caregiver, tend to be more central to ones identity than roles such as that of an informal helper and/or volunteer. Consequently, strain or pressures emerging in the employee and caregiver roles compared to the informal helper and volunteer roles, are likely to result in greater psychological and physical distress. Thus, based on the role centrality perspective, one might infer that some roles may leave people more vulnerable to distress than others. Alternatively, roles that are central to ones identity may be more health enhancing than the less central roles. Related to the role centrality position is the argument that successful fulfillment of some roles as compared to others may require higher levels of human and cultural capital. For instance, unlike informal helping, formal volunteering often demands acquired resources, such as verbal, writing, and administrative skills. In addition to skill sets, some roles, such as the paid worker role, may simply require more time, energy, and commitment than others. For example, paid work often includes everyday obligations, such as producing deliverable and time stipulated tasks and maintaining relations with co-workers. In contrast, volunteering, for instance, may involve more choice in terms of timing, hours, and chores. This suggests that while some roles (e.g., the paid worker role) may lead to greater pressures, others (e.g., volunteering) may involve less strain and subsequently, lower mental and physical distress. Finally, yet another reason to consider role combinations is that while most roles may be taken on willingly, some may involve an involuntary incumbency. For instance, while 18
volunteering may involve ones own will, caregiving may be an involuntary transition. Similarly, while some roles, such as paid worker, are normative in nature, others, for instance, caregiving may represent the unexpected and less normative role. In sum, roles vary on a number of different dimensions and their health effects may vary along these dimensions. Role combinations are, therefore, worth considering because roles can offset the effects of one another; and, some combinations of roles may be either health enhancing or health deleterious. Furthermore, not all role combinations are likely to similarly affect all people. For instance, while formal roles, including formal volunteering and paid work may positively influence the health of White Americans, this may not be the case with African Americans. Lastly, the same combination of roles may have differential effects depending on whether we are examining mental or physical health. The present study explores these issues by examining separate effects as well the effects of combining roles on older adults mental and physical health. Finally, I will examine whether and how the links between role combinations and health vary by peoples social structural position in particular, by race and socioeconomic status.
Social Engagement, Social Structural Location, and Health Sociologists have extensively argued that individual circumstances and the resulting health consequences do not exist in vacuum but instead are inextricably tied to ones position in society. Put simply, social structural locations, such as age, gender, race, and SES, matter in terms of organization of roles, quality of role experiences, role satisfaction, and consequently the psychological and physical health effects of occupying a given social role. Below, I discuss the relevance of two structurally relevant contexts race and socioeconomic status in the relationship between social engagement and health.
Race, Social Engagement and Health While social engagement in general appears to be beneficial for later life health, potential race differences in the health effects of social roles remain less examined. Most prior research on 19
roles and health is restricted to samples of White adults (Thompson and Brown 1980; Verbrugge 1983, 1987; Welch and Booth 1977). Some studies are focused mainly on African Americans (Broman 1991; Coleman, Antonucci, Adelmann, and Crohan 1987), while others include separate analyses for Whites and Blacks (Rushing, Ritter, and Burton 1992; Waldron and Jacobs 1989). However, most do not assess whether the impact of later life social engagement on health differs for White Americans and their African American counterparts. Studies that have examined such differences have limited their definition of social engagement to roles that are characteristic of middle age, namely paid worker, parent, and spouse (Coleman, Antonucci, and Adelmann 1987; Hibbard and Pope 1991; Jennings, Mazaik, and McKinlay 1984; Kotler and Wingard 1989; Verbrugge 1983; Waldron and Jacobs 1989). However, there are reasons to believe that the health effects of other roles, such as formal volunteering, informal helping, caregiving, and religious participation vary between White and Black elders. Formal volunteering, for instance, primarily is a White and middle class activity, with African Americans being less frequently asked to participate (Ferree, Barry, and Manno 1998; Hodgkinson 1995). Moreover, given this cultural bias, it is logical to assume that African Americans feel less wanted and experience barriers during volunteering, including alienation and perceived discrimination (Akpeki 1995; Niyazi 1996). Under such circumstances then, it is unlikely that the health benefits of volunteering for African Americans are comparable to those accruing to their White peers. In contrast to formal volunteering, informal helping and religious activities are less structured and may allow for unique opportunities for African American elders to meaningfully engage in their community. Being involved in these informal roles also may represent an opportunity for Black elders to meet people with similar life circumstances, which could result in a strong sense of mutual acceptance, trust, and community. Less formal roles such as these may be especially relevant for the cohort of older adults considered in the present study. These individuals were born in the early 20th century and thus likely experienced for much of their lives, instances of widespread racial prejudice and discrimination in all social roles. Consequently, it is likely that 20
both the meaning and the effect of roles differ for these individuals based on whether they are Black or White. Like volunteering, informal helping, and religious participation, caregiving also is expected to result in different health outcomes for White and Black older adults. Given the race differences in distribution of economic resources to provide care (Logan and Spitze 1994), it is likely that Black caregivers experience higher levels of stress and burden than their White counterparts. In addition to economic resources, research also shows racial disparities in the health resources of caregivers. African American caregivers are far more likely to be physically ill themselves, affecting their ability to provide care and excacerbating the stress they experience from caregiving (Dautzenberg et al. 1999; Williams et al. 2003). These reasons lead us to assume that caregiving may result in higher levels of both mental and physical distress for African American caregivers than White American care providers. While the foregoing provides theoretical rationales for why social roles may differentially affect Whites and African Americans, empirical research is needed to evaluate the proposed links between race, social engagement, and later life well-being. Moreover, unlike previous studies that focused mainly on either single roles or multiple role occupancy, future research needs to consider how combining of various roles affects mental and physical health differently for White and Black older adults. Is a combination of formal roles, including paid work and formal volunteering less beneficial to Black older adults than their White peers? Do Black older adults benefit more from combining informal social activities, such as informal helping and religious participation? The present study examines these questions.
Socioeconomic Status, Social Engagement and Health Like race, socioeconomic status also is likely to shape the link between social engagement and health. Most prior studies related to later life social engagement, however, are limited to samples of middle-class individuals with little attention given to potential variations in the health effects of roles and role combinations by SES. The protective effects of social roles are likely to vary by SES because SES tends to affect both social and psychological resources that are needed to successfully occupy and maintain a given social role. 21
Empirical research shows that SES, as measured by education, affects peoples level of social skills and competence. For instance, a study conducted by Hogg and Keller (1990) found that older women with higher educational attainment fared significantly better on measures of social competence, including empathy and the ability to role-take, than those with lower educational achievements. Social and interpersonal skills are important not only in role occupancy but also for maintaining effectively the roles that are occupied. Thoits (1987) argued that those with multiple roles belonging to lower SES may be less able to negotiate satisfying relations with role partners and, therefore, may experience more strain and distress within those roles, compared to their higher SES counterparts with same set of roles. Moreover, financial strain is likely to result in declining quality in all roles leading to distress and limiting the mental and physical health benefits accruing from role occupancy. In addition to social competence, socioeconomic status also has been linked with ones levels of perceived social support. For instance, those belonging to lower SES are more likely to be surrounded by friends and family members of relatively similar socioeconomic position, which may prevent high levels of perceived social support. For example, Ross and Mirowsky (1989) found that those with more education have higher levels of social support when compared to those with fewer years of education. Finally, socioeconomic status also has been found to influence psychological resources, such as personal control and feelings of self-esteem, both of which influence the quality of the role, role satisfaction, and health effects of occupying it. Compared to their higher SES counterparts, lower SES individuals may be more likely to experience feelings of personal inadequacy, resulting in a lower sense of personal control and lower levels of self-esteem. Sense of control and self-esteem are essential in the effective functioning of individuals in social scenarios because they enable people to cope actively with problems instead of passively accepting or ruminating over them (Mirowsky and Ross 1989; Wheaton 1983). Financial strain or economic hardship can interfere with demands, expectations, and obligations related to the different social roles by lowering peoples psychological resources, limiting their ability to benefit from them.
22
The foregoing suggests that there are reasons why the health effects of various social roles may vary by SES. Empirical research needs to examine the link between SES, social engagement and health by examining whether roles held in later life affect differentially older adults of upper and lower socioeconomic position. Are certain roles and role clusters more beneficial to those at the lower end of the SES ladder? These questions remain far from examined in prior research on social engagement and the present study aims to examine them.
Summary and Research Aims There are several conclusions to be drawn from previous research on later life social engagement and health. Although most research indicates that being socially and productively engaged in later life (i.e., occupying multiple social roles) is beneficial to peoples psychological and physical well-being, empirical evidence related to individual roles suggests that not all roles/activities are equally beneficial to ones health. However, most prior work has focused on either the effects of individual social roles or the number of roles on the health. Effects of combining various social roles are less well understood. Furthermore, research on later life social engagement has systematically excluded roles that are less formal and visible, such as informal helping and thereby neglected work done mostly by the marginalized segments of society. Taking a more detailed look at roles and health is important as an increasing proportion of people today are actively engaged in diverse formal and informal, private and public social roles, until later in life. As sociologists, we are faced with a growing need to uncover how different roles, role combinations, and social activities affect older adults mental and physical well-being. Moreover, we need to examine whether the benefits and costs associated with occupying and combining different roles in later life are distributed equally among people of different race and SES groups. Do some people benefit less from occupying multiple roles? Are formal roles, such as that of paid worker and volunteer, more health enhancing than less formal roles, such as informal helper? In an attempt to address these underexamined questions and more comprehensively estimate the mental and physical health consequences of later life social engagement, the proposed study will examine the effects of individual roles and the consequences of combining them on peoples psychological and physical well-being. More specifically, this study will pursue the following aims: 23
A1: To examine the health effects of each of the following five social roles: paid worker, formal volunteer, informal helper, caregiver, and religious participant. A2: To explore the effects of various combinations of these five roles on mental and physical health. A3: To assess whether the individual and combined effects of these roles vary by older adults race and socioeconomic status.
Research Hypotheses H1: Paid work, formal volunteering, informal helping, and religious participation all reduce Wave 2 depressive symptoms and functional impairment among older adults. H2: Caregiving increases levels of depressive symptoms and functional impairment at Wave 2. H3: Having multiple roles results in fewer depressive symptoms and lower functional impairment. H4: Lower SES older adults benefit less from being socially engaged compared to their upper SES counterparts. H5: Lower SES older adults benefit more from informal activities, such as informal helping, and religious participation. In contrast, upper SES elders benefit more from engaging in formal social roles, including paid work and formal volunteering H6: African American elders benefit less from being socially engaged compared to their White American peers. H7: Black older adults benefit more from informal roles, namely informal helping and religious participation. In contrast, White American elderly derieve more benefits from formal activities, such as paid work and formal volunteering.
24
CHAPTER FOUR METHODS Sample The data for this research come from two Waves of the Americans Changing Lives (ACL) study (House 1989a, 1989b). ACL is a nationally representative sample (N = 3,617) of adults 24 years of age or older who lived in the United States. The data were collected, in part, to examine the involvement of individuals in activities broadly considered productive and to document the sociocultural variations in nature, meaning, determinants, and consequences of productive activity (House 1989a, 1989b). They include indicators of roles typically not found in large, nationally representative samples, making these data ideal for the proposed research. A multistage stratified area probability sampling was conducted to obtain the study sample. The baseline data were collected in 1986 (N = 3,617) and include an oversample of black adults (n = 1,174) and persons 60 years of age or older (n = 1,669). Two follow-up interviews were conducted in 1989 (N = 2,867) and 1994 (N = 2,348). All surveys included face-to-face interviews conducted in the home of the respondent. A sub-sample of white and black individuals aged 50 to 96 at W1 was extracted from the original data set for use in this study with 1,614 cases available at Wave 2. At the initial interview, correcting for oversampling, fifty-nine percent of this sub-sample was female, ten percent were black, and sixty-two percent were married. These respondents reported an average of of 10.9 years of formal education (SD = 2.53). The large sample size provides sufficient statistical power for these analyses, and reduces the likelihood of Type II errors. The dataset includes a series of weight variables designed to account for the complex sample design and oversampling used to create the initial panel data. When used in analyses, these weights produce standard errors that more accurately and adequately address the design effects that arise from the sampling technique employed. The suitable weight for the analyses in this study is v1860, referred to as the final centered post-stratification weight, and accounts for nonresponse as well as the sample design.
25
Prior to proceeding to the description of measures, addressed in brief are the effects of sample attrition. As in the case of most panel data surveys, the current study experienced some loss of participants between the two Waves. The loss of participants can bias study results if those who remain differ significantly from the population they are supposed to represent. Although it is difficult to determine decisively the extent of this bias, basic understanding of the problem was gauged by conducting the following procedure. First, a binary outcome measure was constructed by assigning a score of 1 to all those who were lost to follow-up and a score of 0 to all those who continued to participate in the second Wave. Second, using logistic regression analyses, the binary variable was regressed on Wave 1 measures of age, gender, race, education, income, marital status, indicators of social engagement, and mental and physical health. Findings from this analysis indicate that the loss of subjects was not a random event. In particular, I found that respondents who did not participate in the second Wave were more likely to be older, male, and non-white, and have less education and income. Moreover, the results reveal that while there were no differences between the two groups in their social engagement status, there were significant differences with respect to physical health. While these findings are consistent with patterns found in other similar panel data studies, the potential biasing effects of this non-random attrition should be kept in mind as the study findings are reviewed.
Description of Measures The independent variables of primary interest (i.e., social role indicators) and sociodemographic controls are measured at Wave 1. Controls for initial levels of mental and physical well-being also are measured at the first Wave. Finally, the two dependent variables (i.e., depressive symptoms and functional impairment) are measured at Wave 2. Dependent Variables Two dependent variables (measured at Wave 2) will be examined in this study: depressive symptoms and functional impairment. Depressive symptoms are chosen primarily because it is one of the most common psychological problems and is experienced by everyone to certain degree at some point over the life course (Pearlin and Johnson 1977). Secondly, ACL contains well-tested 26
measures of depression that have been widely used in previous studies on social engagement and health, making the current study comparable to the earlier ones. Depressive symptoms are measured by a standardized scale (=.82) constructed using eleven items from the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977). The CES-D has become a standard in community mental health surveys over the past two decades (e.g., Eaton and Kessler 1981; Fredrichs, Aneshensel, and Clark 1981; Lennon 1982; Ross, Mirowsky, and Huber 1983). Additionally, it includes both psychological and physiological components of depression, which is preferable given the fact that many other conditions give rise to depressive physiological symptoms, particularly among the elderly (Johnson and Meile 1981; Mirowsky and Ross 1992). In ACL, respondents were asked to consider each of the following statements and choose the category that best described how often in the past week they felt that way: I felt depressed; I felt that everything I did was an effort; My sleep was restless; I was happy; I felt lonely; People were unfriendly; I enjoyed life; I did not feel like eating; I was full of energy; I felt sad; I felt that people disliked me; I could not get 'going'; and I felt that people cared about me. Responses were hardly ever, some of the time, or most of the time. Functional Impairment (=.84) is the physical health measure used in this study. Given the focus of this study on individuals aged 50 and above, functional impairment serves as an ideal outcome variable. It is an index created from 12 items, assessing the respondents ability to perform ADLs and IADLs. ADLs (activities of daily living) are things that are done normally in daily living, including any daily activity we perform for self-care (e.g., feeding ourselves, bathing, dressing, grooming, etc). IADLs (instrumental activities of daily living) are tasks that, in addition to activities of daily living, one must be able to perform in order to live independently (without the assistance or substantial supervision of another person). Examples include grocery shopping, meal preparation, using the telephone, laundry, light housekeeping, paying bills, and managing medications. For the purpose of this study, I used the imputed version of functional impairment constructed by the ACL staff, which ranges from 1 to 4, with 4 representing greater levels of functional impairment.
27
Independent Variables Although the main purpose of this study is to assess the mental and physical health effects of role combinations, I will first examine the separate and individual health effects of each of the five social roles held by the respondents. The next step will be to assess the link between the number of roles1 and mental and physical health. The third and the most important step will include examining the effects of combining various roles on mental and physical well-being.
Individual Social Roles at Wave 1 The paid worker role is assessed by the respondents current employment status. Respondents were asked the following question: We would like to know about what you do -- are you working for pay, looking for work, retired, keeping house, a student, or something else? Responses are coded such that all those who reported being employed are coded as 1 and all others coded as 0. Thirty-four percent of the respondents over the age of 50 were engaged in paid work. The formal volunteer role is examined using the following question: Did you do volunteer work in the last year for a church, synagogue, or other religious organization? for a school or educational organization? for a political group or labor union? for a senior citizens group or related organization? for any other national or local organization, including United Fund, hospital, and the like? Responses to each of these probes were coded 1 indicating yes and 0 indicating no. In the current study, a dichotomous variable is created where respondents reporting having worked for at least one of the groups are coded 1 and considered as having done volunteer work and those reporting not having worked for any of the groups coded 0. Thirty-eight percent of all those over 50 reported having participated in formal volunteering.
1This step is taken because previous studies have consistently used role count when examining the health effects of multiple roles. Thoits (1983a, 1986) suggested that multiple roles should be measured as the sum of social positions held by the individual, based on the notion that role-identities are reciprocal relationships, dependent upon recurrent interaction between role partners (1986, p.259).
28
To measure the informal helper role, interviewers said the following: Now let's talk about help you may have given in the last year to friends, neighbors, or relatives who did not live with you. We are interested in help you provided during the last 12 months for which you did not receive pay. During the last 12 months, did you provide transportation, shop or run errands for friends, neighbors or relatives who did not live with you? Did you help others with their housework or with the upkeep of their house, car or other things? Did you do childcare without pay for persons not living in your household? Did you do any other things to help neighbors, friends or relatives who did not live with you? Responses to each of these probes were coded 1 indicating yes and 0 indicating no. In the current study, a dichotomous variable is created where respondents reporting having helped in at least one of the above ways are coded 1 and be considered informal helpers and those reporting not having helped in any way are coded 0. Sixtyeight percent of this sample reported being involved in the role of informal helping. The caregiver role is assessed by responses given to the following item: Now I would like to talk with you about friends and relatives who have trouble taking care of themselves because of physical or mental illness, disability, or for some other reason. Are you currently involved in helping someone like this by caring for them directly or arranging for their care by others? Responses are coded such as those saying yes to the above question are coded as 1 and considered occupying a caregiving role and those saying no coded as 0 and considered noncaregivers. Forty-five percent of older adults in this sample said they were caregivers. To measure religious participation, interviewers asked the following question: How often do you usually attend religious services? (Would you say more than once a week, once a week, 2 or 3 times a month, about once a month, less than once a month, or never?) For the purpose of this study, a dichotomous variable was created with all those selecting either of the first five options coded as 1 and those saying they never attend religious services being used as a reference group2. Sixty-seven percent of the present sample participated in religious activities.
2
I also used an alternate coding, where all those selecting either of the first four options were coded as 1 and those choosing either of the last two options (i.e., less than once a month or never) were coded as 0. Analyses based on this more stringent selection criteria revealed few differences from those that were conducted using the originally coded religious participation variable. The results of these supplementary analyses are added as a footnote in Chapter 6, which contains findings pertaining to the mental and physical health effects of single roles.
29
Role Count The role count variable is constructed by counting the total number of roles occupied by each individual from among five roles, including paid worker, formal volunteer, informal helper, caregiver, and religious participant. The variable ranges from 0 to 5, with an average of 2.36 roles.
Role Combinations Ten role combinations are assessed by employing interactions between pairs of social roles. These combinations are driven by my interest in examining how each of the five roles interacts with other roles to affect later life mental and physical health. More specifically, the following ten combinations are examined in the present study: paid work and volunteering, paid work and informal helping, paid work and caregiving, paid work and religious participation, volunteering and informal helping, volunteering and caregiving, volunteering and religious participation, informal helping and caregiving, informal helping and religious participation, and caregiving and religious participation.
Wave 1 Health Variables Wave 1 mental and physical health, measured by depressive symptoms and functional impairment respectively, will be used to address the broader issue of social causation versus social selection. Social causation assumes that social engagement (occupying social roles) influences health. Participating in various social roles provides people with power, prestige, status and emotional gratification, all which result in better mental and physical health. By contrast, the social selection standpoint posits that healthy people are the ones most likely to take on and maintain multiple social roles. That is, adequate mental and physical health resources are needed to both undertake and sustain involvement in social roles (Stearns, Stearns, and Hollis 1996). A third possibility is that both mechanisms operate: social engagement leads to and results from higher levels of health. In this way, both social causation and social selection may simultaneously produce patterns of social engagement among older adults (Verbrugge 1983).
30
The present study, which uses longitudinal panel data addresses this issue by controlling for Wave 1 mental and physical health. As in the case of Wave 2 health measures, mental and physical health at Wave 1 are measured by depressive symptoms and functional impairment, respectively. If after controlling for Wave 1 mental and physical health, social engagement continues to positively predict health outcomes at Wave 2, the results would suggest (not conclude) that social selection does not explain all of the links between role occupancy and health; rather, it is that both selection and causation are under operation.
Sociodemographic Variables Race and socioeconomic status are considered possible moderators of the relationship between later life social engagement and health. Race is coded 1 for Blacks and 0 for Whites. Members of other racial/ethnic groups are dropped as they make up only 5.8 percent of the sample (each group is made up no more than 2.2 percent of the sample). SES is measured using three indicators: education, household income, and assets. Education is coded intervally to represent the highest year of schooling completed and ranged from 0 to 17. Household income is the gross income of the respondent and spouse. Income was originally coded in categories ranging from (1) less than $5,000 to (10) over $80,000. I used the imputed income measure calculated by the ACL staff, which assigns income values to the midpoint of each category. Assets was the total amount of liquid cash from checking and savings account, stocks, bonds, and real estate. This variable was originally coded in categories ranging from (1) less than $5,000 to (7) $500,000 or more. I used the imputed assets measure calculated by the ACL staff, which assigns asset values to the midpoint of each category. Missing values (a total of 8% for all three measures of SES combined) are coded to the mean. Age is a continuous measure ranging from 50-96 years. Additionally, age squared3 is employed to test for curvilinear effects in the link between social engagement and health. Gender is coded 1 for women and 0 for men.
I used age squared variable to test for curvilinear effects in the relationship between social engagement and health. Findings revealed neither of the three indicators of social engagement (i.e., single roles, multiple role occupancy, and role combinations) to be curvilinearly associated with either mental or physical health.
3
31
Data Analysis Plan Univariate and bivariate analyses are conducted first, prior to analyzing each of the proposed research questions. Providing descriptive statistics addresses the following: (1) the distribution of mental and physical health resources by role occupancy; and (2) the distribution of role occupancy by race and socioeconomic status. Main effects of the independent variables (social roles, number of roles, and role combinations at W1) on the dependent variables (depressive symptoms, and functional impairment at W2) are explored using both bivariate and multivariate techniques. The bivariate associations are estimated using correlational and means testing analyses. The relationship between indicators of social engagement and mental and physical health (i.e., depressive symptoms and functional impairment, respectively) at the multivariate level are assessed using ordinary least squares regression. Bivariate Analyses Correlation coefficients, t-tests, and one-way Anova are used to assess bivariate associations between variables. Correlational analyses using the Pearson Product Moment Correlation coefficient determine the nature and the direction of the relationship between various pairs of variables. Conducting t-tests enables me to assess whether the means of two groups are statistically significant from each other. More specifically, it determines the average level mental and physical health between individuals who do and do not occupy the five social roles. One-way Anova determines the mean level of differences in health based on the various role combinations and the number of roles. More generally, the bivariate analyses inform us of the associations between two variables of interest without the simultaneous influence of other related factors. Multivariate Analyses These sets of analyses are divided into three parts: (1) Individual roles and health, (2) number of roles (role count) and health, and (3) role combinations and health. Below I list the specific analyses pertinent to each of the main study questions.
32
Individual Social Roles and Health After examining the bivariate associations, the next step was investigating the influence of each of the individual social roles on the dependent variables, while controlling for other independent variables. This step determined whether or not the association between social roles and depression is a spurious one. Specifically, using OLS regression, depression was regressed on each of the five social roles and control variables, including sociodemographic factors, and Wave 1 indicators of mental and physical health. The same set of analyses was conducted using the second dependent variable functional impairment. After examining the relationship between social roles and health, I tested several interaction terms to determine whether the benefits of social engagement varies by race and SES (e.g., paid work*SES reported at W1; informal helping*race reported at W1). Specifically, I addressed whether the protective effects of social roles on health vary by individuals race and socioeconomic status. Interaction effects for each social role (dichotomous variable) and race (dichotomous variable) were tested by simply adding the product of the two to the regression models. A statistically significant interaction term suggests that the main effect of a social role on well-being varies by race. Interaction terms, including socioeconomic status (intervally measured variable) and social roles (dichotomous variables) were tested using different approaches based on whether the effect of the independent variable (social role) on the dependent variables is altered by the moderator(s) in a linear, quadratic, or stepwise way (Baron and Kenny 1986).
Number of Roles (Role Count) and Health After testing the link between individual social roles and health, I assessed the association between the number of social roles and mental and physical well-being. Although not the focus of this study, this step was taken because previous studies have consistently used role count when examining the health effects of multiple roles. The role count variable ranges from 0 to 5 roles, with 0 representing people with no role involvement and 5 corresponding to people involved in all the five roles. First, each of the two dependent variables was regressed using OLS on the role count variable, while controlling for the sociodemographic factors and Wave 1 measures of mental 33
and physical health. After analyzing the main association between number of roles and health, the final set of analyses included adding one at a time interaction terms to the regression models; these consisted of interactions between the role count variable and race and socioeconomic status.
Role Combinations and Health After exploring the link between individual social roles, the number of roles, and health, the last but the most important step was examining how combinations of various roles affect mental and physical well-being. The link between these role combinations and well-being was examined by regressing each of the two health outcomes on each of the above mentioned (and additional) role mixes. All analyses controlled for sociodemographic factors including age, gender, race, education, income and assets at W1. Additionally, mental and physical well-being (W1) were controlled in all multivariate analyses. Next, several interaction terms (e.g., formal volunteer and informal helper*race at W1; caregiver and informal helper*SES) were entered one at a time to determine whether the effects of combining different roles vary by older adults social structural context.
34
CHAPTER FIVE DESCRIPTIVE RESULTS
Overview In this chapter, I provide results pertaining to the following four descriptive analyses. First, I present the means (or proportions) and standard deviations for all study variables. Second, I examine the overall role distributions among older adults and across the specific race and SES groups. Third, I report the average levels of depressive symptoms and functional impairment by older adults social engagement status. Finally, a correlation matrix demonstrating the relationships between all of the study measures is presented. Means, (or Proportions) and Standard Deviations This section presents the means (or proportions) and standard deviations for all the variables used in the study. Table 1 shows that thirty-four percent of the respondents over the age of 50 had engaged in paid work during the past year when interviewed in 1986. Thirty-eight percent of the elderly had been involved in some type of volunteer work and 68 percent of the older adults reported having participated in informal helping. Finally, results show forty-five percent of the older adults to be caregivers and sixty-seven percent of them to have participated in some sort of religious activity. Statistics related to the demographic characteristics of the sample show that sixty-seven percent of the elderly is female, thirty percent is Black, and a little over half (i.e., 55 percent) of the individuals are married. The average age is sixty-six years and the mean level of education for this sample is 10 years. Finally, the average annual income for this sample is $20,653 and the mean level of assets is $20,450.
35
Table 1. Description of Variables (n = 1,614)
Variable Dependent Variables (W2) Depressive Symptoms (CES-D) Functional Impairment Independent Variables (W1) Social Engagement Variables Paid Work Formal Volunteering Informal Helping Caregiving Religious Participation Prior Health Variables (W1) Depressive Symptoms (CES-D) Functional Impairment Sociodemographic Variables Age Gender (1 = Female) Race (1 = Black) Marital Status (1 = married) Years of Education Total Income (thousand dollars) Assets (thousand dollars) Range -1.13 to 4.25 14 Mean (SD) .010 (.99) 1.64 (.99)
0, 1 0, 1 0, 1 0, 1 0, 1 -1.18 to 4.25 14 50 96 0, 1 0, 1 0, 1 0 17 2,500 110,000 5,000 - >500,000
.34 .38 .68 .45 .67 -.007 (.99) 1.58 (.93) 66.33 .67 .30 .55 10.67 20,658 20,450 (8.6)
(3.6) (21,130) (17,000)
Notes: *p<.05 **p < .01 ***p < .001; standardized variables; standard deviations are in parentheses; t-tests are used; Americans Changing Lives Data (1986, 1989).
Race and SES Variations in Social Engagement Here, I present single role, role count, and role combination distributions across both race and SES groups. Tables 2a and 2b present results for race, whereas Tables 3a through 3f offer findings related to SES. Results for single role occupancy (see Table 2a) show that White older adults had significantly higher rates of paid work, volunteering, and caregiving than their African American peers. African American elders, however, reported higher rates of participation in both informal helping and religious activity. In terms of average number of roles occupied, African American older adults reported fewer activities at Wave 1 than their White counterparts, but this difference was not statistically significant.
36
Table 2b shows us race differences in role combinations. Of the ten role combinations, statistically significant race differences emerge in five of them. These include the following: paid work and caregiving, paid work and religious participation, volunteering and informal helping, volunteering and caregiving, and informal helping and caregiving. In each of these, except in the case of paid work and religious participation, the rates of White participation exceed engagement by African Americans.
Table 2a. Statistics for Role Distribution by Race among Older Adults (Means, Standard Deviations and/or Percentages) --------------------------------------------------------------------------------------------------------------------Variable Whites Blacks (n = 1,135) (n = 479) --------------------------------------------------------------------------------------------------------------------Paid Worker (%) .37 .33*** Formal Volunteer (%) .43 .31*** Informal Helper (%) .50 .66*** Caregiver (%) .48 .37*** Religious Participant (%) .59 .78*** Role Count M (SD) 2.58 (1.3) 2.39 (1.2) --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; t-tests are used to test for significant race differences.
Table 2b. Role Combination Distribution by Race among Older Adults (Percentages) -----------------------------------------------------------------------------------------------------------Role Combination Whites Blacks (n = 1,135) (n = 479) -----------------------------------------------------------------------------------------------------------Paid Work and Volunteering .16 .13 Paid Work and Informal Helping .27 .24 Paid Work and Caregiving .17 .13* Paid Work and Religious Participation .21 .27** Volunteering and Informal Helping Volunteering and Caregiving Volunteering and Religious Participation Informal Helping and Caregiving Informal Helping and Religious Participation .36 .24 .31 .39 .47 .26*** .13*** .28 .24*** .46
Caregiving and Religious Participation .31 .30 -----------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; t-tests are used to test for significant race differences.
37
The SES related patterns are presented in Tables 3a through 3f. Tables 3a through 3c show that a greater proportion of the higher SES older adults compared to lower SES elderly are involved in four of the five social roles. Specifically, Table 3a shows that a greater proportion of the high educated are engaged in paid work, formal volunteering, and caregiving. Similarly, tables 3b and 3c indicate that a greater proportion of the high income and high assets elders are involved in these three roles. In contrast, a higher percentage of low SES elderly as compared to their upper SES peers are engaged in informal helping, and religious activity. In particular, tables 3a through 3c show that a greater proportion of the low educated, low income and those with low assets are involved in regular religious activities. In terms of the average number of roles occupied, lower SES individuals reported fewer activities at Wave 1. In other words, elders with lower education, income, and assets reported fewer social roles than their peers with greater socioeconomic resources. In terms of differences in role combinations by educational level, results show higher participation rates at higher levels of education. Specifically, except for the combination of paid work and religious participation, all other two role combinations show higher rates of participation by upper SES older adults than lower SES elders. Similarly, in terms of income and asset differences in role combinations, results show a higher proportion of upper SES elders compared to lower SES older adults to be engaged in most role combinations. For only two role combinations, namely, volunteering and religious participation, and informal helping and religious participation, we see an opposite pattern.
Table 3a. Statistics for Role Distribution by Educational Level among Older Adults (Means, Standard Deviations and/or Percentages) --------------------------------------------------------------------------------------------------------------------Variable < High School High School College Degree N = 767 N = 448 N = 399 --------------------------------------------------------------------------------------------------------------------Paid Worker (%) .22 .39 .40*** Formal Volunteer (%) .23 .38 .54*** Informal Helper (%) .78 .72 .52*** Caregiver (%) .39 .45 .54*** Religious Participant (%) .69 .64 .61 Role Count M (SD) 2.02 (1.2) 2.56 (1.3) 2.87 (1.2) --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; one-way anovas are used to test for significant education differences.
38
Table 3b. Statistics for Role Distribution by Income Level among Older Adults (Means, Standard Deviations and/or Percentages) --------------------------------------------------------------------------------------------------------------------Variable < $20,000 $20,000-30,000 > $30,000 N = 1043 N = 365 N = 206 --------------------------------------------------------------------------------------------------------------------Paid Worker (%) .24 .33 .43*** Formal Volunteer (%) .20 .35 .50*** Informal Helper (%) .72 .67 .48*** Caregiver (%) .41 .50 .53*** Religious Participant (%) .68 .65 .58*** Role Count M (SD) 2.21 (1.0) 2.43 (1.5) 2.65 (1.5) --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; one-way anovas are used to test for significant income differences. Table 3c. Statistics for Role Distribution by Asset Level among Older Adults (Means, Standard Deviations and/or Percentages) --------------------------------------------------------------------------------------------------------------------Variable < $20,000 $20,000 50,000 > $50,000 N = 965 N = 401 N = 248 --------------------------------------------------------------------------------------------------------------------Paid Worker (%) .27 .35 .38*** Formal Volunteer (%) .25 .44 .51*** Informal Helper (%) .76 .73 .56*** Caregiver (%) .39 .49 .50*** Religious Participant (%) .66 .63 .61** Role Count M (SD) 2.14 (1.2) 2.65 (1.3) 2.78 (1.2) --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; one-way anovas are used to test for significant asset differences.
Table 3d. Differences in Role Combination by Education Level among Older Adults (Means,
Standard Deviations and/or Percentages) --------------------------------------------------------------------------------------------------------------------Role Combination < High High College School School Degree N = 767 N = 448 N = 399 --------------------------------------------------------------------------------------------------------------------Paid Work and Volunteering .11 .19 .25*** Paid Work and Informal Helping .17 .34 .38*** Paid Work and Caregiving .10 .20 .23*** Paid Work and Religious Participation .17 .29 .26*** Volunteering and Informal Helping Volunteering and Caregiving Volunteering and Religious Participation .21 .13 .23 .37 .21 .33 .53*** .35*** .42***
Informal Helping and Caregiving .27 .39 .47*** Informal Helping and Religious Participation .40 .52 .53*** Caregiving and Religious Participation .27 .31 .37*** --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; one-way anovas are used to test for significant education differences.
39
Table 3e. Differences in Role Combination by Income Level among Older Adults (Percentages)
--------------------------------------------------------------------------------------------------------------------Role Combination < $20,000 $20,000-30,000 > $30,000 N = 1043 N = 365 N = 206 --------------------------------------------------------------------------------------------------------------------Paid Work and Volunteering .10 .22 .31*** Paid Work and Informal Helping .18 .38 .50*** Paid Work and Caregiving .10 .24 .31*** Paid Work and Religious Participation .17 .30 .36*** Volunteering and Informal Helping Volunteering and Caregiving Volunteering and Religious Participation Informal Helping and Caregiving Informal Helping and Religious Participation .27 .17 .27 .30 .43 .43 .26 .38 .44 .54 .45*** .29*** .36*** .47*** .50***
Caregiving and Religious Participation .29 .34 .30 --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; one-way anovas are used to test for significant income differences.
Table 3f. Differences in Role Combination by Asset Level among Older Adults (Means, Standard Deviations and/or Percentages) --------------------------------------------------------------------------------------------------------------------Role Combination < $20,000 $20,000 50,000 > $50,000 N = 965 N = 401 N = 248 --------------------------------------------------------------------------------------------------------------------Paid Work and Volunteering .12 .19 .22*** Paid Work and Informal Helping .23 .31 .32*** Paid Work and Caregiving .13 .21 .18*** Paid Work and Religious Participation .21 .24 .26
Volunteering and Informal Helping Volunteering and Caregiving Volunteering and Religious Participation Informal Helping and Caregiving Informal Helping and Religious Participation .25 .14 .22 .28 .43 .43 .28 .38 .43 .52 .47*** .30*** .43*** .45*** .54***
Caregiving and Religious Participation .27 .35 .34 --------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p < .01 ***p < .001; one-way anovas are used to test for significant asset differences.
40
Mean Levels of Depressive Symptoms and Functional Impairment by Social Engagement Next, mean levels of depressive symptoms and functional health for individuals occupying different social roles are reported in tables 4a and 4b. As expected, paid workers, volunteers, and informal helpers all report significantly fewer depressive symptoms and less functional impairment than their peers who are not involved in these roles. Caregivers, however, do not differ from non-caregivers in either mental or physical health. Finally, religious participants, while report lower levels of functional health than their non-religious involved peers, there is no difference in their mental health.
TABLE 4a. Mean Level of Depressive Symptoms by Social Role Occupancy Religious Participant 1 0 1 0 1 0 1 0 1 0 -----------------------------------------------------------------------------------------------------------------------------------------------Mean -.190 .114*** -.216 .151*** -.122 .299*** -.006 .024 .021 -.011 S. D. .88 1.0 .85 1.0 .91 1.0 1.0 .98 1.0 .98 -----------------------------------------------------------------------------------------------------------------------------------------------Paid Worker Volunteer Helper Caregiver
Note: *p<.05 **p < .01 ***p < .001; t-tests are used to test for significant differences by role occupancy.
TABLE 4b. Mean Level of Functional Impairment by Social Role Occupancy Religious Participant 1 0 1 0 1 0 1 0 1 0 -----------------------------------------------------------------------------------------------------------------------------------------------Mean 1.29 .114*** 1.41 1.79*** 1.47 2.01*** 1.62 1.66 1.60 1.73* S. D. .70 1.0 .81 1.0 .85 1.0 .96 1.0 .96 1.0 -----------------------------------------------------------------------------------------------------------------------------------------------Paid Worker Volunteer Helper Caregiver
Note: *p<.05 **p < .01 ***p < .001; t-tests are used to test for significant differences by role occupancy.
Correlation Matrix Finally, Table 5 reports Pearsons correlations across all study variables. The purpose of this exploration was twofold: (1) to understand the bivariate relationships among the variables used to empirically test the hypotheses under consideration; and (2) to test for the existence of
41
multicollinearity4 across the main outcome and independent variables. Bivariate relationships with correlation coefficients + .70 would indicate the presence of unacceptable levels of collinearity. Data analyzed here do not exhibit sufficiently strong bivariate relationships to cause concern. Findings from the correlation analyses (see Table 5) indicate a strong and negative association between three of the five indicators of social engagement and both mental and physical health. In particular, paid work, volunteering, and informal helping are all negatively and significantly related with depressive symptoms and functional health at Wave 2. Surprisingly, however, caregiving shows no correlation to either depressive symptoms or functional health. Finally, religious participation, while negatively correlated to functional health, does not correlate with depressive symptoms in this sample.
4
I also performed a test of collinearity using the coldiag2 function in STATA9, which is consistent with the diagnostic procedures of Harkness (Belsley, Kuh, and Welsch 1980). While Belsley et al. (1980) suggest that collinearity likely exists across a set of variables if a condition number produced exceeds 30.0 in this diagnostic method, the present study variables generated a condition number of 22.9. This indicates that collinearity will not be major influence in the results presented.
42
Table 5. Correlation Matrix -----------------------------------------------------------------------------------------------------------------------------------------------------------------W1 Dep W1 Hlth W2 Dep W2 Hlth Paid Volunteer Helper Caregiver Religious Work Participant -----------------------------------------------------------------------------------------------------------------------------------------------------------------W1 Dep 1.00 W1 Hlth W2 Dep W2 Hlth Paid Work Volunteer Helper Caregiver 0.35*** 0.53*** 0.30*** -0.11*** -0.15*** -0.19*** 0.01 1.00 0.30*** 0.62*** -0.28*** -0.17*** -0.27*** 0.03 1.00 0.40*** -0.14*** -0.17*** -0.19*** 0.01 1.00 -0.25*** -0.18*** -0.25*** 0.02 1.00 0.09*** 0.15*** 0.02 1.00 0.31*** 0.13*** 1.00 0.17*** 1.00
Religious -0.02 -0.03 -0.01 -0.06* -0.01 0.21*** 0.04* 0.01 1.00 Participant ----------------------------------------------------------------------------------------------------------------------------------------------------------------Notes: **p < .01 ***p < .001. Variables names are abbreviated. W1 Dep = Wave 1 Depressive Symptoms, W2 Dep = Wave 2 Depressive Symptoms, W1 Hlth = Wave 1 Functional Health, W2 Hlth = Wave 2 Functional Health.
43
Table 5. Correlation Matrix (contd.) ----------------------------------------------------------------------------------------------------------------------------------------------------------------Gender Race Age Marital Education Income Status ----------------------------------------------------------------------------------------------------------------------------------------------------------------Gender 1.00 Race Age Marital Status Education Income 0.01 0.08** -0.18*** 0.01 -0.16*** 1.00 -0.04 -0.17*** -0.32*** -0.25*** 1.00 -0.22*** -0.19*** -0.29*** 1.00 0.15*** 0.35*** 1.00 0.44*** 1.00
Assets -----------------------------------------------------------------------------------------------------------------------------------------------------------------Notes: **p < .01 ***p < .001. Variables names are abbreviated. W1 Dep = Wave 1 Depressive Symptoms, W2 Dep = Wave 2 Depressive Symptoms, W1 Hlth = Wave 1 Functional Health, W2 Hlth = Wave 2 Functional Health.
Summary To summarize these findings, correlational and bivariate analyses reveal that three out of five indicators of social engagement are associated with both mental and physical health among older adults. In particular, that paid work, volunteering, and informal helping are all significantly and negatively correlated with depressive symptoms and functional impairment. Surprisingly, caregiving
44
has no impact on either mental or physical health. Religious participation has no influence on mental health but is negatively correlated with older adults physical health. In terms of the bivariate associations between SES and social engagement, results show that except in the case of a few roles (i.e., informal helping and religious participation) and role combinations (i.e., volunteering and religious participation, and informal helping and religious participation), a greater proportion of elders from higher socioeconomic strata than their lower SES counterparts are involved in social engagement. Similarly, other than for a few roles (i.e., informal helping and religious participation) and role combinations (i.e., paid work and religious participation), results show higher rates of Whites than African Americans in terms of social engagement. On the whole, results reveal higher rates of participation by African Americans and lower SES elders in informal roles, such as informal helping and religious participation. While these findings advance current understanding of race and SES differences in later life social engagement, it is not clear how race, SES, and social engagement may work in tandem in the prediction of changes in mental and physical health. Moreover, given the observed elevation of distress and functional impairment among African Americans and individuals of lower SES7, it is of interest to examine whether the benefits of multiple role involvement and varied role combinations are limited for these two groups. In order to attend to these considerations, I now turn to the presentation of multivariate analyses.
7
Tables consisting of means and standard deviations for all variables by both race and SES are added to the Appendix.
45
CHAPTER SIX THE MENTAL AND PHYSICAL HEALTH CONSEQUENCES OF SINGLE ROLES AND NUMBER OF ROLES Overview Although the bivariate analyses reveal strong links between indicators of social engagement and health, it is unable to determine whether these associations are spurious in nature. In other words, we have yet to determine whether the effects of social engagement on health remain after adjusting for demographic characteristics and prior levels of mental and physical well-being. The purpose of the present chapter is twofold: (1) examining whether and how the type of role influences later life mental and physical health, and (2) exploring the mental and physical health impact of occupying multiple social roles, as indicated by count of roles occupied. Analyses examining role combinations are presented in the subsequent chapter. To accomplish these tasks, I use OLS regression to regress both Wave 2 depressive symptoms and functional impairment first on single roles and then on indicator for multiple role occupancy. I control in all analyses for Wave 1 sociodemographic characteristics and prior levels of both mental and physical health. In the sections below, I first describe results for the mental health effects of single roles, followed by the effect of single role occupancy on physical health. Next, I present findings for the links between number of social roles and mental and physical health.
Single Roles and Health
Mental Health Results Table 6 presents the results of the regression of older adults mental health on five single social roles. Model 1 contains results only for relationships between the sociodemographic factors, prior health, and Wave 2 depressive symptoms. These results reveal that education is a significant predictor of changes in distress levels. In particular,
45
we find that older adults with higher levels of education report fewer depressive symptoms than their peers with lower educational attainment. Findings also demonstrate a strong and positive association between prior levels of both physical and mental health and subsequent change in mental health at Wave 2. That is, individuals with poor mental and physical health at Wave 1 report greater levels of psychological distress at Wave 2. Models 2-7 show results pertaining to the effects of social engagement on change in depressive symptoms from years 1986 to 1989. Models 2, 3, and 4 all show that being socially engaged in later life results in fewer depressive symptoms at Wave 2. Specifically, model 2 reveals that older adults who have occupied the role of paid worker report less distress at Wave 2 than those not involved in this role. Similarly, models 3 and 4 both show a strong and inverse association between formal volunteering and informal helping10 and change in depressive symptomology, respectively. That is, both volunteers and informal helpers report fewer depressive symptoms at Wave 2 than their peers who were not involved in these social roles. Moreover, these findings persist in models 2, 3, and 4 that include controls for prior levels of mental and physical health, which presents a stronger case for social causation, which posits that being socially engaged results in reduced distress and improved mental well-being. In contrast to paid work, formal volunteering, and informal helping, I found that neither caregiving nor religious participation11 significantly predicts change in distress
10
While both formal volunteering and informal helping result in decline in psychological distress at Wave 2, being a formal volunteer is a stronger predictor of changes in depressive symptoms compared to being an informal helper. Findings based on the alternate coding for religious participation revealed few differences from those presented using the main variable. In particular, I find that being involved in religious activities does not results in improvements in mental health but it does lead to enhanced physical health at Wave 2. While, neither race nor SES moderated the relationship between religious participation and mental health, income and assets shaped the link between religious involvement and physical health. In particular, the benefits of being involved in religious activities were greater for those with lower levels of income and assets.
11
46
levels. In particular, models 5 and 6 show us that caregivers and non-caregivers, religious participants and those who are not involved in religious activity do not differ in terms of changes in their mental health between 1986 and 1989. These findings persist even in models 5 and 6 that control for prior levels of physical and mental health. Of the sociodemographic factors entered in these models, educational status is the only significant predictor of a change in distress levels over time: Older adults with higher levels of education report lower distress at Wave 2 than their peers with low education. Finally, as expected, prior levels of mental and physical health are strong and independent predictors of depressive symptoms at Wave 2. Taken together, these results coincide with prior studies of social engagement, which suggest that being actively engaged predicts better mental health for all older adults. However, results from further analyses (presented in chapters 8 and 9) will show that the relationship between social engagement and mental health is more complex than it initially appears.
47
Table 6. OLS Regression of Depressive Symptoms at Wave 2 on Single Social Roles at Wave 1 (N = 1,614) Independent Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Sociodemographic Variables Age -.007 -.003 -.004 -.001 -.000 -.000 -.003 (.00) (.00) (.00) (.00) (.00) (.00) (.00) Gender (1 = female) .075 .061 .079 .074 .074 .074 .060 (.04) (.04) (.04) (.04) (.04) (.04) (.04) Race (1 = black) .083 .088 .086 .076 .084 .082 .079 (.04) (.04) (.04) (.04) (.04) (.05) (.04) Marital Status (1 = married) -.042 -.047 -.030 -.037 -.043 -.042 -.027 (.04) (.04) (.04) (.04) (.04) (.04) (.04) Education -.033*** -.032*** -.030*** -.031*** -.034*** -.033*** -.028*** (.00) (.00) (.00) (.00) (.00) (.00) (.00) Income -.002 -.001 -.002 -.002 -.002 -.002 -.001 (.00) (.00) (.00) (.00) (.00) (.00) (.00) Assets -.001 -.000 -.000 -.000 -.000 -.000 -.000 (.00) (.00) (.00) (.00) (.00) (.00) (.00) Wave 1 Health Variables Depressive Symptoms .435*** .434*** .435*** .432*** .435*** .435*** .427*** (.02) (.02) (.02) (.02) (.02) (.02) (.02) Functional Health .127*** .118*** .120*** .118*** .127*** .127*** .108*** (.02) (.02) (.02) (.02) (.02) (.02) (.02) Social Role Engagement Paid Worker (1 = yes) -----.102* -------------------.097* (.05) (.05) Formal Volunteer (1 = yes) ----------.119** ------------.114* (.04) (.04) Informal Helper (1 = yes) ----------------.093* -----------.070* (.04) (.05) Caregiver (1 = yes) -------------------.019 -----.038 (.04) (.04) Religious Participant (1 = yes) ----------------------.006 .038 (.04) (.04) R Squared .29 .30 .31 .30 .29 .29 .31 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Notes: *p<.05 **p < .01 ***p < .001; Data presented are unstandardized coefficients with standard errors in parenthesis.
48
Physical Health Results As in the case of mental health, I used OLS regression to assess whether social engagement resulted in better physical health among older adults. Table 7 reports results pertaining to the effects of single role occupancy on physical health. Model 1 shows the associations between sociodemographic factors and health variables and Wave 2 physical health. These results show that, as in the case of mental health, education significantly and inversely predicts change in physical health. Similarly, prior levels of physical and mental health also are predictive of subsequent changes in physical well-being. Models 2-7 reveal findings pertaining to the effects of paid work, volunteering, informal helping, caregiving, and religious participation on the physical health of older adults. On the whole, the patterns emerging here are similar to those pertaining to mental health. More specifically, I find four of the five social roles to strongly and positively predict changes in physical health among older adults. In particular, we see that those involved in paid work, formal volunteering, informal helping, and religious activity report significantly lower functional impairment at Wave 2 compared to their peers with no such social involvement. It is important to note however that once prior levels of physical health are controlled, the statistically significant effects of three of the four social roles, namely paid work, formal volunteering, and informal helping, disappear. This pattern suggests that while social engagement results in improvements in physical health over time, this relationship is stronger for those physically healthier at Wave 1. Once again, contrary to expectations, I find caregiving to bear no effect on the physical health status of older adults. This is surprising in the light of prior findings that caregiving is in general not only emotionally but also physically distressing. It is important to note that a significant relationship between caregiving and physical health fails to emerge even when I control for prior levels of mental and physical health.
49
Summary On the whole, findings presented thus far provide support to previous studies that suggest a health salutary effect of social engagement for older adults. Paid work, volunteering, and informal helping emerge as the strong and positive social roles in terms of both mental and physical health outcomes. Religious participation, however, is only predictive of physical health changes. Caregiving is one role that was found to predict neither mental nor physical health changes in this sample of older adults.
50
Table 7. OLS Regression of Functional Impairment at Wave 2 on Single Social Roles at Wave 1 (N = 1,614) Independent Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Sociodemographic Variables Age .019*** -.020*** -.025*** -.023*** -.025*** -.026*** -.019*** (.00) (.00) (.00) (.00) (.00) (.00) (.00) Gender (1 = female) .061 .066 .099* .088* .094* .117* .078 (.04) (.04) (.04) (.04) (.04) (.04) (.04) Race (1 = black) .069 .072 .061 .033 .058 .090 .080 (.05) (.04) (.05) (.05) (.05) (.05) (.05) Marital Status (1 = married) -.039 -.047 -.030 -.037 -.043 -.042 -.032 (.04) (.04) (.04) (.04) (.04) (.04) (.04) Education -.021*** -.023** -.019** -.018** -.025*** -.026** -.015* (.00) (.00) (.00) (.00) (.00) (.00) (.00) Income -.000 -.000 -.000 -.000 -.000 -.000 -.000 (.00) (.00) (.00) (.00) (.00) (.00) (.00) Assets -.000 -.000 -.000 -.000 -.000 -.000 -.000 (.00) (.00) (.00) (.00) (.00) (.00) (.00) Wave 1 Health Variables Depressive Symptoms .260*** .254*** .251*** .244*** .261*** .257*** .229*** (.02) (.02) (.02) (.02) (.02) (.02) (.02) Functional Health .571*** .578*** .576*** .574*** .582*** .577*** .564*** (.02) (.02) (.02) (.02) (.02) (.02) (.02) Social Role Engagement Paid Worker (1 = yes) -----.254* -------------------.227*** (.02) (.05) Formal Volunteer (1 = yes) ----------.214*** ------------.121* (.04) (.05) Informal Helper (1 = yes) ----------------.280*** -----------.237*** (.05) (.05) Caregiver (1 = yes) -------------------.009 -----.056 (.04) (.04) Religious Participant (1 = yes) -----------------------.195*** -.143** (.04) (.04) R Squared .42 .42 .42 .42 .42 .42 .43
---------------------------------------------------------------------------------------------------------------------------------------------------------------Notes: *p<.05 **p < .01 ***p < .001; Data presented are unstandardized coefficients, standard errors in parenthesis.
51
Role Count and Health Extensive literature (e.g., Thoits 1983a, 1986) on the health effects of social engagement has suggested that social engagement should be measured as the sum of social positions held by the individual, based on the notion that role-identities are reciprocal relationships, dependent upon recurrent interaction between role partners (Thoits 1986, p.259). Keeping in mind this tradition and the overwhelmingly strong empirical support for the link between multiple roles and health, the present study examined the effect of number of social roles on the mental and physical health of older adults. The findings pertaining to the mental health effects of multiple roles are presented first, followed by results reporting the association between multiple roles and physical health.
Mental Health Results Table 8 reports results for the mental health effects of multiple roles. In model 1, background characteristics of older adults are entered. Of these variables, only education was significantly associated with change in distress levels at Wave 2. As expected, older adults with higher educational attainment reported few depressive symptoms at Wave 2 than their peers with low levels of education. Model 1 also contained variables measuring prior levels of mental and physical health, and like in the case of previous analyses, I find prior levels of health to be strongly associated with subsequent levels of mental health. In model 2, the role count variable is added and consistent with prior research, I find that having multiple roles is significantly and negatively related to changes in depressive symptoms8. In other words, model 2 shows that the greater the number of roles older adults hold, the less
8
Prior research suggests that specific social roles may differentially affect health. In particular, I wanted to explore the unique nature of caregiving because unlike other roles, research indicates this role to have negative effect on both mental and physical health. I re-ran the analyses with caregiving excluded from the role count variable. This did not change the substantive findings, although the relationship between role number and functional health did strengthen. There was no discernable impact on the coefficient for depressive symptoms. In both models, the amount of explained variation decreased slightly but did not alter the R-Squared associated with the inclusion of the engagement measures.
52
likely they are to feel depressed at Wave 2. This pattern persists in both models with and without controls added for prior levels of health. These results indicate that multiple roles preceded depression in causal ordering and had a significant influence on depression, indicating that in addition to social selection, causation is in effect as well.
Table 8. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Role Count Variable (N = 1,614) Independent Variable Model 1 Model 2 ----------------------------------------------------------------------------------------------------------------Role Count Variable ----.047* (.01) Sociodemographic Variables Age -.003 -.001 (.00) (.00) Gender .072 .085 (.04) (.04) Race .079 .096 (.05) (.05) Marital Status -.049 -.032 (.04) (.04) Education -.026*** -.024*** (.00) (.00) Income -.002 -.002 (.00) (.00) Assets -.000 -.000 (.00) (.00) Wave 1 Health Variables Depressive Symptoms .418*** .452*** (.02) (.02) Functional Health .128*** .120*** (.02) (.02) R Squared .29 .30 --------------------------------------------------------------------------------------------------------------------Notes: *p<.05 **p < .01 ***p < .001; Data presented here are unstandardized coefficients, with standard errors are in parenthesis.
Physical Health Results Analyses related to the physical health effects of multiple roles revealed similar patterns. Table 9 indicates that older adults having multiple roles report lower levels of functional impairment than their peers who are not involved in multiple roles. Stated otherwise, the greater the number of roles held by older adults, the less likely they were to experience decline in physical health between the two waves. Model 2 shows that these results emerge even after
53
controlling for prior levels of mental and physical well-being. Models not controlling for the initial levels of health (not shown) reveal similar findings. Education once again emerges as a strong predictor of physical health in later life, with older adults with higher levels of education reporting lower levels of functional impairment as compared to their peers with lower levels of educational attainment. Finally, prior levels of mental and physical well-being once again seem to be strongly connected with Wave 2 physical health.
Table 9. OLS Regression of Wave 2 Functional Health on Wave 1 Role Count Variable (N = 1,614) Independent Variable Model 1 Model 2 -----------------------------------------------------------------------------------------------------------Role Count Variable ----.049** (.01) Sociodemographic Variables Age -.013*** -.012*** (.00) (.00) Gender .049 .053 (.04) (.04) Race .002 .031 (.04) (.04) Marital Status -.041 -.031 (.04) (.04) Education -.017* -.013* (.00) (.00) Income -.000 -.000 (.00) (.00) Assets -.000 -.000 (.00) (.00) Wave 1 Health Variables Depressive Symptoms .082*** .080*** (.02) (.02) Functional Health .580*** .567*** (.02) (.02) R Squared .42 .44 -----------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses.
54
Summary The foregoing results provide testimony to the role accumulation hypothesis, which predicts a positive relationship between holding multiple roles and mental and physical wellbeing. In other words, present study finds that older adults having multiple social roles enjoy better mental and physical health at Wave 2 compared to their peers who are not engaged in multiple roles. These findings remain similar in models with and without controls for prior levels of mental and physical health.
55
CHAPTER SEVEN ROLE COMBINATIONS AND HEALTH Overview Existing research on the health effects of social engagement suggests that while some roles are health enhancing, others provide fewer health benefits or in fact, are health deleterious. This means that merely examining the number of roles held by individuals leads us to assume that the effect of any one role on health is the same regardless of the other simultaneous role enactments. A more inclusive treatment of this issue demands that in addition to single roles and multiple roles, we also examine the health effects of role combinations. The present study took this step by assessing the effects of two role combinations on the mental and physical health of older adults. The current chapter presents findings related to theses analyses. The first section of the chapter examines the effect of role combinations on changes in psychological health and the second section explores the impact of combining roles on changes in older adults physical health. Mental Health Results
Table 10 presents results for the mental health effects of combining social roles. I used OLS regression to examine the mental health impact of the following ten work role combinations: paid work and volunteering, paid work and informal helping, paid work and caregiving, paid work and religious participation, volunteering and informal helping, volunteering and caregiving, volunteering and religious participation, informal helping and caregiving, informal helping and religious participation, and caregiving and religious participation. Of the ten role combinations tested, only four emerged as statistically significant in predicting changes in distress levels among older adults. Below, I first report the significant role combinations, followed by a brief discussion on those role combinations that failed to attain statistical significance.
56
Formal Volunteer and Caregiver
Models 1 and 2 in Table 10 present results for the effect of combining formal volunteering and caregiving on mental health. The main effect analyses presented in model 1 show that while caregiving has no effect on mental health, formal volunteering predicts positive changes in mental health among older adults. The interaction analyses reveal that the effects of formal volunteering are greater among caregivers than non-caregivers.
Informal Helper and Caregiver
Models 3 and 4 in Table 10 present findings pertaining to the combination of the informal helper and caregiver roles and its influence on psychological health. The main effect analyses (see model 3) reveal that while caregiving has no effect on mental health, being an informal helper results in enhanced levels of mental well-being. More importantly, the interaction analyses presented in model 4 show that the impact of being an informal helper is greater for caregivers than their non-caregiver counterparts.
Informal Helper and Religious Participant
Models 5 and 6 in Table 10 offer findings for the mental health effects of combining informal helping and religious participation. Analyses involving the main effects of these two roles indicate that while religious participation has no impact on mental health, informal helping is positively associated with mental well-being. Further, an interaction between these two roles demonstrates that the effects of being an informal helper are greater for those involved in religious activities than their peers who are not religious participants.
57
Caregiver and Religious Participant
Another role combination that significantly predicts changes in distress levels is that of caregiving and religious participation. The main effect analyses (see model 7) reveal that neither caregiving nor religious participation significantly predict changes in levels of distress among older adults. That said, subsequent analyses (see model 8) involving interaction between these two roles reveal that religious participation positively impacts those who are involved in the role of caregiving. In other words, being involved in religious activities reduces depressive symptoms for older adults who are also caregivers.
Non-Significant Role Combinations
In addition to the above four combinations, I examined six other combinations of roles: paid work and volunteering, paid work and informal helping, paid work and caregiving, paid work and religious participation, volunteering and informal helping, and volunteering and religious participation. Models containing data for each of these combinations suggest that while most of these roles (i.e., paid work, volunteering, and informal helping) are independently associated with change in depressive symptoms, combining them does not significantly either decrease or increase levels of psychological distress between the two waves.
58
Table 10. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Role Combinations (N =1,614) ---------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 ---------------------------------------------------------------------------------------------------------------------------------------------------------Volunteer -.268*** -.351* (.03) (.04) Caregiver .286*** .399* (.03) (.05) Volunteer X -----.201 Caregiver (.07) .30 .30 R-Square --------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.249*** -.219** (.04) (.05) Caregiver .023 .190 (.03) (.08) Informal Helper X -----.212* Caregiver (.09) R-Square .29 .30 --------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.111** -.018 (.04) (.06) Religious Participant -.029 -.144 (.03) (.06) Informal Helper X ----.149* Religious Participant (.07) R-Square .29 .30 ---------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
59
Table 10. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Role Combinations (N =1,614) (contd.) ---------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M7 M8 M9 M10 M11 M12 --------------------------------------------------------------------------------------------------------------------------------------------------------Caregiver .255 .445 (.04) (.07) Religious Participant - .277*** -.373* (.04) (.05) Caregiver X ----244* Religious Participant (.09) R-Squared .29 .29 ---------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.087* -.092 (.05) (.06) Volunteer -.122** -.126* (.04) (.05) Paid Worker X ---.012 Volunteer (.08) R-Squared .30 .30 --------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.090* -.066 (.05) (.09) Informal Helper -.084* -.075 (.04) (.05) Paid Worker X ----.030 Informal Helper (.10) R-Squared .30 .30 --------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
60
Table 10. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Role Combinations (N =1,614) (contd.) ---------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M13 M14 M15 M16 M17 M18 ---------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.090* -.093 (.05) (.06) Caregiver .010 .008 (.04) (.05) Paid Worker X ---.006 Caregiver (.08) .30 .30 R-Squared --------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.090* -.061 (.05) (.08) Religious Participant -.010 -.024 (.04) (.05) Paid Worker X ----.042 Religious Participant (.09) R-Squared .29 .29 -------------------------------------------------------------------------------------------------------------------------------------------------------Volunteer -.111** -.101 (.04) (.10) Informal Helper -.055 -.052 (.04) (.05) Volunteer X ----.012 Informal Helper (.11) .30 .30 R-Squared ------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
61
Table 10. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Role Combinations (N =1,614) (contd.) -----------------------------------------------------------------------------------------------------------------------------------------Independent Variables M19 M20 ----------------------------------------------------------------------------------------------------------------------------------------Volunteer -.134** -.150 (.04) (.08) Religious Participant - .041 .035 (.04) (.05) Volunteer X ---.021 Religious Participant (.09) .30 .30 R-Squared ------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
62
Physical Health Results Table 11 presents results for the physical health effects of combining social roles. As in the case of mental health, I used OLS regression to examine the mental health impact of the following ten work role combinations: paid work and volunteering, paid work and informal helping, paid work and caregiving, paid work and religious participation, volunteering and informal helping, volunteering and caregiving, volunteering and religious participation, informal helping and caregiving, informal helping and religious participation, and caregiving and religious participation. Of the ten role combinations tested, only two were statistically significant in predicting changes in functional impairment among older adults. Below, I first report the significant role combinations, followed by a brief discussion on those role combinations that failed to attain statistical significance.
Paid Worker and Formal Volunteer
Models 1 and 2 in Table 11 present results for the paid worker and formal volunteer role combination. The main effect analyses (see model 1) reveal that both paid worker and formal volunteer roles predict positive changes in physical health among older adults. This is consistent with previous suggesting research that both of these roles provide a unique opportunity for individuals in a period that primarily characterized with role loss. More importantly, the interaction between these two roles (see model 2) suggests that being a formal volunteer enhances the physical health benefits of paid work. Stated otherwise, while paid work alone is health enhancing, the benefits of engaging in paid work are even greater for those involved in the role of informal helping.
Paid Worker and Informal Helper Models 3 and 4 in Table 11 present results for the paid worker and informal helper role combination. Analyses on the main effects of these two roles (see model 3) reveal that both paid work and informal helping result in improvements in physical health. This pattern coincides with previous findings that suggest that both these social roles confer upon elders diverse social and
64
psychological resources, which result in their enhanced physical well-being. Further, combining these two roles suggests that being an informal helper enhances the physical health benefits of paid work. In other words, the physical health benefits of working for pay in later life are even greater for those who engaged in informal helping.
Non-significant Findings As in the case of mental health analyses, I examined several other role combinations: paid work and caregiving, paid work and religious participation, volunteering and informal helping, volunteering and caregiving, volunteering and religious participation, informal helping and caregiving, informal helping and religious participation, and caregiving and religious participation. Results pertaining to these interaction analyses revealed that while individual roles continue to impact the change in physical health for older adults, combining these roles did not significantly either reduce or increase functional impairment at Wave 2. Particularly important is to note that the non-significance of these interactions persists in models that do not control for prior levels of mental and physical health.
65
Table 11. OLS Regression of Wave 2 Functional Impairment on Wave 1 Social Role Combinations (N =1,614) ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.558*** -.693* (.03) (.03) Volunteer -.226*** -.408* (.03) (.04) Paid Worker X ----.324** Volunteer (.06) .41 .42 R-Squared ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.512*** -.805* (.03) (.06) Informal Helper -.416*** -.565* (.03) (.04) Paid Worker X ----.377** Informal Helper (.07) R-Squared .40 .41 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.546** -.532* (.04) (.06) Caregiver .022 .055 (.03) (.04) Paid Worker X ----.093 Caregiver (.07) R-Squared .42 .42 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
67
Table 11. OLS Regression of Wave 2 Functional Impairment on Wave 1 Social Role Combinations (N =1,614) (contd.) ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M7 M8 M9 M10 M11 M12 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.594*** -.652*** (.03) (.04) Religious Participant -.063* -.114* (.03) (.04) Paid Worker X ---.092 Religious Participant (.06) R-Squared .41 .41 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Volunteer -.095* .008 (.04) (.09) Informal Helper -.075* -.045 (.04) (.05) Volunteer X ----.129 Informal Helper (.10) R-Squared .40 .40 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Volunteer -.116** -.150* (.04) (.05) Caregiver .034 .007 (.03) (.04) Volunteer X ---.069 Caregiver (.07) R-Squared .40 .40 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
68
Table 11. OLS Regression of Wave 2 Functional Impairment on Wave 1 Social Role Combinations (N =1,614) (contd.) ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M13 M14 M15 M16 M17 M18 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Volunteer -.089* -.092 (.04) (.07) Religious Participant -.097* -.098* (.04) (.04) Volunteer X ---.003 Religious Participant (.08) R-Squared .42 .42 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.107* -.076 (.04) (.05) Caregiver .037 .095 (.03) (.07) Informal Helper X -----.082 Caregiver (.08) R-Squared .41 .41 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.091* -.125* (.04) (.07) Religious Participant -.112** -.147* (.04) (.07) Informal HelperX ---.052 Religious Participant (.08) R-Squared .41 .37 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
69
Table 11. OLS Regression of Wave 2 Functional Impairment on Wave 1 Social Role Combinations (N =1,614) (contd.) -----------------------------------------------------------------------------------------------------------Independent Variables M19 M20 -----------------------------------------------------------------------------------------------------------Caregiver .027 -.016 (.03) (.06) Religious Participant -.120** -.149*** (.04) (.05) Caregiver X ---.065 Religious Participant (.08) R-Squared .42 .42 -----------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
SUMMARY In review, of the ten role combinations examined, four were found to significantly predict distress levels, whereas only two were significantly predictive of changes in physical health. The combinations emerging as significant in terms of mental health include volunteering and caregiving, informal helping and caregiving, informal helping and religious participation, and caregiving and religious participation. The two combinations to significantly predict physical health include paid work and volunteering, and paid work and informal helping.
70
CHAPTER EIGHT SOCIOECONOMIC STATUS, SOCIAL ENGAGEMENT, AND HEALTH Overview Chapters 6 and 7 suggest the health salutary effects of later life social engagement. This finding coincides with previous studies indicating that being involved in multiple social roles confers upon individuals diverse personal, social, and psychological resources, all of which result in greater mental and physical well-being (Gubrium 1972, 1973; Havighurst 1964; Herzog and House 1991). But what remains untested is whether the health salubrious effects of social engagement extend to all elderly regardless of their socioeconomic position. The present chapter focuses on this question. I first present results for the effects of the interaction between each of the three measures of social engagement (i.e., single roles, role count, and role combinations) and each of the three SES indicators (i.e., education, income, and assets) on the mental health of older adults. The second section consists of findings pertaining to the effect of social engagement and SES on physical health in later life. Mental Health Results
Socioeconomic Status, Single Roles, and Mental Health I used OLS regression models to first assess whether the relationship between single social roles and mental health varied by SES. I evaluated two-way interaction terms between each of the five roles (i.e., paid work, volunteering, informal helping, caregiving, and religious participation) and each of the three SES measures (i.e., education, income, and assets). A statistically significant interaction term indicates that the effect of a particular social role on mental health is contingent on socioeconomic status. Of the five interaction terms assessed using education, four emerged as statistically significant in predicting changes in distress levels. However, none of the five income or asset-related interactions were statistically significant in predicting mental health changes between the two waves. Below I first discuss the significant findings involving interactions between social engagement and education. Next, while I address
71
the non-significant findings involving social engagement and income and assets, this discussion is brief and tables for these findings are not included in this presentation. Table 12 consists of results for the main effects of single roles, and interactions between single roles and education. As stated previously, social engagement predicts better mental health among older adults. In particular, models 1, 3, and 5 show that paid work, volunteering, and informal helping all tend to strengthen mental health in later life. These data, however, reveal no statistically significant differences in mental health between caregivers and non-caregivers (see model 7) and those who do or do not participate in religious activity (see model 9). Taken together, these results suggest that social engagement predicts better mental health for all older adults but as the findings presented below will show, the mental health benefits of social engagement are greater for elders with lower educational attainment. Evidence of this may be found in models 2, 4, 6, and 10. These models indicate that the effect of social engagement on mental health varies by educational status of older adults. More specifically, model 2 reveals that older adults with low levels of education benefit more from engaging in paid work compared to their peers with higher levels of educational attainment. Similarly models 4, 6, and 8 show that older adults with low levels of education benefit more from being volunteers, informal helpers, and religious participants, compared to their more highly educated counterparts. That said, it is important to note that the statistically significant interactions dissipate once prior levels of physical health are controlled in the models. That is, while prior mental health does not play a role, prior levels of functional health mediate the function education plays in the link between social engagement and Wave 2 mental health. Taken together, the results thus far suggest that while being socially engaged helps all elderly, the mental health benefits of it are greater for those with lower educational attainment. In contrast to the results for education, findings related to income and assets reveal no statistically significant effect of these two financial resources on the link between social engagement and mental health. In other words, the effects of paid work,
72
volunteering, informal helping, caregiving, and religious participation on mental health do not differ by the income and asset levels of older adults.
SES, Multiple Roles, and Mental Health Table 13 presents results for the interaction analyses of the number of roles occupied by older adults and their socioeconomic status. As stated earlier, I find that the number of roles strongly and positively predicts mental health among older adults. Of the sociodemographic variables entered, results show education to be significantly and inversely associated with depressive symptoms. However, none of the other sociodemographic factors significantly predict mental health in this sample. On the whole, these results suggest that all older adults appear to benefit from being engaged in multiple social roles. Interaction analyses also provide support to the foregoing results. Evidence of the similar mental health effects of multiple roles for elders of varying SES levels, is found in models 2, 4, and 6. These models contain results of the interaction analyses (i.e., role count*education, role count*income, and role count*assets). The data in these models show that while the size of the additive effects of multiple role involvement appears to have decreased, statistically significant interaction effects between multiple role involvement and the three SES indicators fail to emerge. Stated otherwise, regardless of the socioeconomic position, all older adults seem to benefit from being engaged in multiple social roles.
SES, Role Combinations, and Mental Health Findings for the significant interactions between role combinations and SES on mental health are presented in Table 14. Of the 20 interactions assessed involving SES, only one significantly predicts change in levels of distress among older adults. This significant interaction is between paid work, informal helper and education. This interaction showed that if an older adult was engaged in paid work and informal helping, not having higher levels of education resulted in enhanced mental health benefits. In
73
contrast, for those not involved in paid work, their mental health status remained the same regardless of whether or not they were involved in informal helping. Moreover, if an older adult was not highly educated, the mental health effect of being an informal helper was greater regardless of her/his paid work status. The other examined interactions were between the remaining nine two role combinations and each of the three SES indicators. These results showed that while SES affected the link between single roles and health, the impact of combining various roles on health did on differ for older adults with varying education, income, and asset levels.
74
Table 12. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Roles and Education (N =1,614) ---------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 ---------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.238*** -.502*** (.05) (.17) Education -.057*** -.067*** (.00) (.00) Paid Worker X ---.034** Education (.01) .29 .30 R-Squared --------------------------------------------------------------------------------------------------------------------------------------------------------Formal Volunteer -.256*** -.421*** (.05) (.14) Education -.060*** -.087*** (.00) (.01) Formal Volunteer X .041* Education (.01) R-Squared .28 .30 -------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.245*** -.533*** (.05) (.15) Education -.055*** -.073*** (.00) (.01) Informal Helper X .027** Education (.01) .28 .29 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, income, assets, and prior levels of mental and physical health.
75
Table 12. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Roles and Education (N =1,614) (Contd.) ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M7 M8 M9 M10 ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------Caregiving .011 .065 (.04) (.13) Education -.028*** -.026*** (.00) (.00) Caregiving X -----.004 Education (.01) .28 .28 R-Squared ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Religious Participant .008 .314* (.04) (.14) Education -.033*** -.013 (.00) (.01) ----.027* Religious Participant X Education (.01) .28 .30 R-Squared -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, income, and prior levels of mental and physical health.
76
Table 13. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Role Count, Education, Income, and Assets (N = 1,614) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Role Count -.048* -.043 (.01) (.05) Education -.023** -.025 (.00) (.00) Role Count X Education ----.001 (.00) .29 .29 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Role Count -.048* -.016 (.01) (.02) Income -.002 -.000 (.00) (.00) Role Count X Income ----.001 (.00) .29 .29 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Role Count -.048* -.018 (.01) (.02) Assets -.000 -.000 (.00) (.00) Role Count X Assets -----.002 (.00) .29 .29 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, and prior levels of mental and physical health.
77
Table 14. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Social Role Combinations and SES (N =1,614) ---------------------------------------------------------------------------------------Independent Variables M1 M2 ---------------------------------------------------------------------------------------Paid Worker -.099* .148 (.05) (.29) Informal Helper -.082* -.283 (.04) (.17) Education -.029*** -.058*** (.00) (.01) Paid Worker X ----.084* Informal Helper X (.03) Education R-Squared .31 .36 ---------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, income, and prior levels of mental and physical health. Physical Health Results
Physical Health Effects of Single Social Roles and SES As in the case of mental health, I used OLS regression models to assess whether the link between social engagement and physical health varied by SES. Once again, I evaluated two-way interaction terms between each of the five social roles and each of the three SES measures. A statistically significant interaction term indicates that the effect of social engagement on physical health is contingent on socioeconomic status. Of the five interactions assessed using education, none emerged statistically significant. Of the five interactions involving income, two were found to significantly predict changes in physical health. Of the five interactions pertaining to single roles and assets, three significantly predicted changes in physical health. Table 15 present results pertaining to these analyses. Consistent with prior research, these findings (see models 1, 3, 5, and 9) reveal that paid work, formal volunteering, informal helping,
78
and religious participation all strongly and positively shape physical health in later life. However, as the findings presented below will show, these initial benefits of some roles extended to all elderly, will be limited to those with lower income and assets. Models 2, 4, 6, 8, and 10 of Table 15a all show results pertaining to the interactions between each of the five single roles and income. Here, I find that while being engaged in paid work, formal volunteering, informal helping and religious participation is still predictive of changes in physical health, the effects of paid work and formal volunteering differ for elders with varying income levels. In particular, I find that the physical health benefits of working for pay are greater for elderly with low income compared to their higher income peers (see model 2). Similarly, model 4 shows that low income older adults benefit more from volunteering than their counterparts with higher levels of income. Additional analyses also reveal that that these differences in physical health impact between those who are employed or not, volunteering or not, are greatest for those with low levels of income with this gap, in effect, closing as we move up the income ladder. In addition to income, assets also proved to be a significant moderator of the relationship between individual roles and physical health. In particular, models 2, 4, and 6 of Table 15b all show that while paid work, volunteering, and informal helping results in better physical health for all older adults, the benefits of engaging in these roles is particularly beneficial to those with few assets. As in the case of income, supplementary analyses revealed that the differences in changes in physical health between those who occupy and do not occupy these roles are greatest for those with low levels of assets. Results of these analyses show that this gap in the impact of social engagement on health narrows with increase in the level of assets. In contrast to income and assets, education failed to moderate the relationship between single roles and physical health. Stated otherwise, the effects of paid work, formal volunteering, informal helping, caregiving, and religious participation on physical health did not vary by the educational status of older adults.
79
Physical Health Effects of Role Count and Socioeconomic Status The next task was to examine whether the physical health benefits of multiple role involvement vary by socioeconomic status. As expected, Model 1 suggests that being involved in multiple roles reduces functional impairment but subsequent interaction analyses reveal that this relationship between multiple roles and physical health is contingent on SES. As in the case of single roles, the physical health benefits of multiple role performance are greater for elderly belonging to lower SES. Models 2, 3, and 4 of Table 16 indicate that the benefits of being involved in multiple roles are greater for those with lower levels of education, income, and assets. Stated differently, the lower the level of education, income and assets, the greater the benefit of being involved in multiple roles on changes in Wave 2 physical health.
SES, Role Combinations, and Physical Health In addition to single roles and role count, the present study also examined SES variations in the physical health effects of role combinations. To explore this, I examined interactions between each of the ten role combinations and each of the three indicators of SES. I expected lower SES older adults to benefit less from being involved in formal and more structured social roles, such as paid work and formal volunteering, and gain more from informal roles, including informal helping and religious activity. Contrary to these expectations, however, findings revealed no SES differences in any of the ten examined social role combinations. It is also important to note that patterns related to SES, role combinations, and physical health remained same regardless of whether prior levels of mental and physical health were controlled for in the models.
80
Table 15a. OLS Regression of Wave 2 Functional Impairment on Wave1 Social Roles and Income (N = 1,614) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.315*** -.400*** (.05) (.07) Income -.040*** -.044*** (.00) (.00) Paid Worker X ----.051* Income (.00) .31 .33 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Formal Volunteer Income Formal Volunteer X Income -.299*** (.04) -.038*** (.00) ----.404*** (.07) -.046*** (.00) .061** (.00)
R-Squared .30 .34 -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.457*** -.373*** (.06) (.06) Income -.008*** -.006** (.00) (.00) --.005 Informal Helper X Income (.00) .31 .32 R-Squared ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, and prior levels of mental and physical health.
81
Table 15a. OLS Regression of Wave 2 Functional Impairment on Wave1 Social Roles and Income (N = 1,614) (contd.) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M8 M9 M10 M11 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Caregiver .022 .009 (.06) (.05) Income -.003 -.001 (.00) (.00) Caregiver X ---.009 Income (.00) .31 .31 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Religious Participant -.249**** -.115* (.06) (.05) Income -.003* -.002 (.00) (.00) -.002 Religious Participant X ---Income (.00) .32 R-Squared .32 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, and prior levels of mental and physical health.
82
Table 15b. OLS Regression of Wave 2 Functional Impairment on Wave1 Social Roles and Assets (N = 1,614) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Worker -.669*** -.594*** (.09) (.09) Assets -.117*** -.089*** (.01) (.01) ---.101*** Paid Worker X Assets (.02) .36 .38 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Formal Volunteer Assets Formal Volunteer X Assets -.436*** (.09) -.094*** (.02) ----.357*** (.09) -.067*** (.02) .060* (.02)
.35 .36 R-Squared -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helper -.468*** -.375*** (.09) (.08) Assets -.114*** -.083** (.02) (.02) Informal Helper X ---.061* Assets (.03) R-Squared .35 .37 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
83
Table 15b. OLS Regression of Wave 2 Functional Impairment on Wave1 Social Roles and Assets (N = 1,614) (contd.) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M8 M9 M10 M11 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Caregiver .054 .010 (.08) (.07) Assets -.069** -.027 (.02) (.01) Caregiver X ---.011 Assets (.02) .35 .36 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Religious Participant -.208* .004 (.09) (.07) Assets -.048* .010 (.02) (.02) Religious Participant X -----.045 Assets (.02) R-Squared .36 .37 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
84
Table 16. OLS Regression of Wave 2 Functional Impairment on Wave 1 Role Count, Education, Income, and Assets (N = 1,614) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent Variables M1 M2 M3 M4 M5 M6 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Role Count -.126*** -.239*** (.01) (.05) Education -.016* -.041** (.00) (.01) Role Count X Education ----.010* (.00) .42 .42 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Role Count -.126*** -.181*** (.01) (.02) Income -.001 -.010** (.00) (.00) Role Count X Income ---.002** (.00) R-Squared .42 .43 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Role Count -.126*** -.149*** (.01) (.02) Assets -.000 -.000 (.00) (.00) Role Count X Assets -----.001* (.00) .42 .42 R-Squared -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, and prior levels of mental and physical health.
85
Summary This chapter presented results related to SES variations in the mental and physical health effects of later life social engagement. In particular, I examined whether the health impact of single roles, number of roles, and role combinations varied between lower and upper SES older adults. On the whole, findings revealed that while social engagement results in better health, the benefits of being socially engaged are greater for lower SES adults compared to their upper SES peers. Moreover, I found education, income, and assets to each play a unique role in the relationship between social engagement and health. Specifically, I found that education but not income and assets moderates the link between social engagement and mental health. I found that the benefits of roles, in particular paid work, formal volunteering, and informal helping are greater for older adults with less education. In terms of multiple roles, however, education did not play a role. That is, the benefits of having multiple social roles extended equally to all older adults regardless of their level of educational attainment. Finally, analyses involving role combinations revealed that education modified the link between paid work and informal helping and mental health. In terms of SES disparities in the physical health effects of social engagement, I found income and assets but not education to play a role. That is, social roles, namely paid work, volunteering, and informal helping all seemed to benefit older adults with low levels of income and assets to a greater extent than their socioeconomically advantaged peers. These two indicators of SES also influence the link between multiple roles and physical health. Specifically, I found that the positive impact of occupying multiple social roles was greater for elderly with few financial resources, specifically income and assets than older adults with high monetary resources. In terms of role combinations, however, as in the case of mental health, findings revealed no statistically significant SES differences.
86
CHAPTER NINE RACE, SOCIAL ENGAGEMENT, AND HEALTH Overview The SES variations in the mental and physical health impact of social engagement revealed thus far undermine the assumption that role experiences and resulting health effects are the same for individuals regardless of their positions in the social structure. SES, however, is not the only status distinguishing context that remains underexamined in prior research. Another social characteristic that has received inadequate attention in the research on later life social engagement is race. One of the aims of the present study was to address this shortcoming in the literature on social engagement and health. In particular, it examines how the effects of the type of role, number of roles, and role combinations vary between White Americans and their African American counterparts. In the sections below, I first present results related to the mental health effects of race and social engagement, followed by findings pertaining to the links between race, social engagement and physical health. Mental Health Results
Race, Single Roles, and Mental Health I used OLS regression models to assess whether the relationship between single social roles and mental health varied by race. I evaluated two-way interaction terms between each of the five social roles and the dichotomous race variable. A statistically significant interaction term indicates that the effect of a particular social role on mental health is contingent on race. Of the five interactions that were assessed, only one emerged statistically significant in predicting changes in mental health. Findings presented in Table 17 show that the only social role in which racial variations appear is paid work. In particular, model 2 indicates that the mental health benefits of being engaged in paid work are greater for African American older adults
87
compared to their White American counterparts. The statistically significant race difference remains even in models that control for prior levels of mental and physical health. This finding is surprising given that a greater proportion of Blacks than Whites tend to work in low paying jobs with poor work conditions factors that are shown to strongly and negatively influence mental health. Returning to Table 17, results for the interactions between other four roles and race are presented in models 4, 6, 8, and 10. On the whole, findings reveal more similarities than differences in the mental health effects of social engagement between White Americans and their African Americans counterparts. More specifically, models 4, 6, 8, and 10 show that volunteering, informal helping, caregiving, and religious participation all predict levels of distress similarly between White Americans and African Americans.
88
Table 17. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Single Roles and Race (N = 1,614) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent M1 M2 M3 M4 M5 M6 Variables -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Work -.586*** -.521** (.06) (.03) Race (1 = Black) .142 .255 (.03) (.04) Paid Work X ----.212** Race (.06) R-Squared .29 .29 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Formal Volunteering -.131** -.130* (.04) (.05) Race (1 = Black) .088 .089 (.05) (.06) Formal Volunteering X -----.004 Race (.10) R-Squared .29 .29 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helping -.229** -.241** (.05) (.06) Race (1 = Black) .080 .051 (.05) (.07) Informal Helping X ----.020 Race (.09) R-Squared .29 .29 --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
89
Table 17. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Single Roles and Race (N = 1,614) (contd.) ---------------------------------------------------------------------------------------------------------------------------------------------------------------Independent M7 M8 M9 M10 Variables ---------------------------------------------------------------------------------------------------------------------------------------------------------------Caregiving .026 .010 (.04) (.05) Race (1 = Black) .087 .065 (.05) (.06) Caregiving X ---.053 Race (.09) R-Squared .29 .29 --------------------------------------------------------------------------------------------------------------------------------------------------------------Religious Participation .016 .011 (.04) (.05) Race (1 = Black) .082 .063 (.05) (.09) Religious Participation X ---.022 Race (.11) R-Squared .29 .29 ---------------------------------------------------------------------------------------------------------------------------------------------------------------
Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
90
Race, Multiple Roles, and Mental Health The next task was to examine whether race modified the health effects of multiple social roles. Table 18 presents results for the interaction between number of social roles and mental health. Model 1 shows that the number of roles strongly and positively predicts mental health among all older adults. Evidence of this is found in models 2. In particular, this model indicates that while the size of the additive effects of multiple role involvement appears to have decreased slightly, statistically significant interaction effect between multiple role involvement and race fail to emerge. Stated otherwise, both White and African American older adults seem to benefit equally from being engaged in multiple social roles.
Table 18. OLS Regression of Wave 2 Depressive Symptoms on Wave 1 Role Count and Race (N = 1,614) -----------------------------------------------------------------------------------------------------------Independent M1 M2 Variables -----------------------------------------------------------------------------------------------------------Role Count -.048* -.047* (.01) (.02) Race (1 = Black) Role Count X Race .100* (.04) ---.106 (.10) -.002 (.03)
R-Squared .28 .28 -----------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, income, assets, and prior levels of mental and physical health.
Race, Role Combinations, and Mental Health In addition to single roles and role count, I also examined race variations in the mental health effects of role combinations. I expected African Americans to benefit less from a combination of formal roles, such as paid work and formal volunteering and obtain greater benefits from being engaged in informal roles, including informal helping and religious
91
participation. But contrary to expectations, the present data revealed no racial variation in the health impact of any of the ten role combinations (results not included in the presentation). Stated otherwise, findings revealed more similarities than differences in the health consequences of combining various social roles for White and Black older adults. Physical Health Results
Race, Single Roles, and Physical Health I used OLS regression models to assess whether the link between social engagement and physical health varied by race. Once again, I evaluated two-way interaction terms between each of the five social roles and race. A statistically significant interaction term indicates that the effect of social engagement on physical health is contingent on race. Of the five interactions involving race, only one was found to statistically predict changes in physical health. Overall findings indicate that as in the case of mental health, there is little racial discrepancy in the physical health impact of single social roles. The only case in which Blacks and Whites differ in the gain of physical benefits of social engagement is that of caregiving. Table 19 presents results pertaining to the race differences in physical health benefits of social engagement. Model 1 shows no independent effect of the caregiver role on physical health; however, the interaction analyses reveal a racial variation in the physical health effects of caregiving. Specifically, I find African American caregivers to report higher levels of functional impairment compared to their White peers involved in caregiving. Again, however, in the case of other social roles, Whites and African Americans remain similarly affected in terms of the physical health outcome.
92
Table 19. OLS Regression of Wave 2 Functional Impairment on Wave 1 Single Roles and Race (N = 1,614) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Independent M1 M2 M3 M4 M5 M6 Variables -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Caregiving .294 .290 (.04) (.04) Race (1 = Black) .034 .040 (.04) (.05) Caregiving X ----.156 Race (.09) R-Squared .42 .42 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Paid Work -.110** -.113* (.04) (.05) Race (1 = Black) .009 .006 (.04) (.05) Paid Work X ---.008 Race (.09) R-Squared .42 .42 -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------Formal Volunteering -.099* -.069 (.04) (.04) Race (1 = Black) .004 .043 (.04) (.05) Formal Volunteering X -----.108 Race (.09) R-Squared .42 .42 --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, and prior levels of mental and physical health.
93
Table 19. OLS Regression of Wave 2 Functional Impairment on Wave 1 Single Roles and Race (N = 1,614) (contd.) ---------------------------------------------------------------------------------------------------------------------------------------------------------Independent M7 M8 M9 M10 Variables --------------------------------------------------------------------------------------------------------------------------------------------------------Informal Helping -.077** -.046 (.04) (.05) Race (1 = Black) .004 .048 (.04) (.07) Informal Helping X ----.083 Race (.09) R-Squared .42 .42 ---------------------------------------------------------------------------------------------------------------------------------------------------------Religious Participation -.117** -.140** (.04) (.04) Race (1 = Black) .023 .048 (.04) (.08) Religious Participation X ---.096 Race (.10) R-Squared .42 .42 ------------------------------------------------------------------------------------------------------------------------------------------------------------
Note: *p<.05 **p<.01 ***p<.001. Data presented in the table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, education, income, assets, and prior levels of mental and physical health.
94
Race, Multiple Roles, and Physical Health Table 20 contains results of the analyses that were performed when the Wave 2 measure of physical health was regressed on the interaction between race and the role count variable. The data in the left-hand column of this table reveal that regardless of race, being engaged in multiple social roles is associated with lower levels of functional impairment at Wave 2. However, as the findings presented in Model 2 shows, the physical health effects of being involved in multiple social roles vary by race. Model 2 indicates the statistically significant interaction effect between race and multiple role involvement on changes in functional impairment. In particular, the physical health benefits of having multiple social roles are greater for White Americans than their African American counterparts.
Table 20. OLS Regression of Wave 2 Functional Health on Wave 1 Role Count and Race (N = 1,614) -----------------------------------------------------------------------------------------------------------Independent M1 M2 Variables -----------------------------------------------------------------------------------------------------------Role Count -.050** -.104*** (.01) (.02) Race (1 = Black) Role Count X Race .007 (.04) ---.261* (.11) .081* (.04)
R-Squared .42 .42 -----------------------------------------------------------------------------------------------------------Note: *p<.05 **p<.01 ***p<.001. Data presented in table are unstandardized coefficients with standard errors in parentheses. All models adjust for age, gender, race, marital status, income, assets, and prior levels of mental and physical health.
Race, Role Combinations, and Physical Health In addition to single roles and role count, I also examined race variations in the physical health effects of role combinations. As predictions surrounding mental health, I 95
expected African Americans to benefit less from a combination of formal roles, such as paid work and formal volunteering and obtain greater benefits from being engaged in informal roles, including informal helping and religious participation. But similar to the findings related to mental health, the present data revealed no race differences in the health impact of any of the ten role combinations (tables for these non-significant findings are not presented here). More specifically, non-significant interactions between race and role combinations reflect more resemblance than disparity in the health consequences of combining various social roles for White and Black older adults.
Summary This chapter presented results pertaining to the Black-White differences in the mental and physical health effects of later life social engagement. In particular, I examined whether the health impact of single roles, number of roles, and role combinations varied between White and African American older adults. Overall findings revealed few significant interactions between race and social engagement. In terms of single roles, I found race to moderate the association between paid work and mental health. Older African Americans benefitted more from the paid worker role than their White American peers. Race also influences the link between caregiving and physical health. In particular, African American caregivers reported greater declines in physical health compared to their White American counterparts. In addition to single roles, I tested for race variation in health effects of multiple roles. Findings indicated race to affect the impact of multiple roles on physical health but not mental health. Specifically, the physical health benefits of having multiple social roles are greater for White Americans than their African American counterparts. Finally, the main task of the study was to focus on role combinations and examine if the health benefits of combining varied roles differ between White and Black older
96
adults. Contrary to expectations, the present data revealed no race variations in the health impact of combining different social roles.
97
CHAPTER TEN SUMMARY OF FINDINGS Below, I present a brief summary of findings for the following four analyses (1) the health effects of individual social roles and role count (2) the health impact of combining social roles, (3) SES variations in the health effects of later life social engagement, and (4) race differences in the health consequences of being socially engaged. Single Roles and Role Count Table 21 presents a summary of results pertaining to single roles and their effect on mental and physical health among older adults. Three roles, namely paid work, formal volunteering, and informal helping result in improvements in mental health. Neither caregiving nor religious participation is found to affect changes in mental health in this sample of older adults. In terms of physical health, all roles except caregiving, are significantly predictive of improvements in functional impairment. In addition to single roles, multiple role performance is found to positively impact both mental and physical health in later life. Table 21. Summary of the Mental and Physical Health Effects of Single Roles and Multiple Roles Single Roles Paid Work Formal Volunteering Informal Helping Caregiving Religious Participation Mental Health Physical Health
+ + + ns ns
+ + + ns +
98
Table 21. Summary of the Mental and Physical Health Effects of Single Roles and Multiple Roles (contd.) Single Roles Role Count Mental Health Physical Health +
+
Notes: + indicates that role occupancy predicts an improvement in health; indicates that role occupancy predicts a decline in health; and, ns indicates a statistically nonsignificant relationship between role occupancy and health. Role Combinations Findings from the role combination analyses (for a summary, see Table 22) indicate that of the ten role combinations assessed, only four emerged statistically significant in predicting changes in mental health. These include caregiving and volunteering, caregiving and informal helping, informal helping and religious participation, and caregiving and religious participation. These findings indicate that being involved in volunteer work, informal helping, and religious activities enhances mental health among those who are caregivers. Stated differently, being involved in these other social roles results in improved psychological well-being among those involved in caregiving. In terms of physical health, two of the ten role combinations emerged as statistically significant. In particular, paid work and volunteering and paid work and informal helping were found to be significant. These patterns indicated that both volunteering and informal helping enhanced physical health benefits of paid work in later life. In other words, while working for pay results in improved physical health, the benefits attached to this role are even greater for those who also are involved in helping activities, such as formal volunteering and informal helping.
99
Table 22. Summary of the Mental and Physical Health Effects of Role Combinations Role Combinations Paid Work X Formal Volunteering Paid Work X Informal Helping Paid Work X Caregiving Paid Work X Religious Participation Formal Volunteering X Informal Helping Formal Volunteering X Caregiving Formal Volunteering X Religious Participation Informal Helping X Caregiving Informal Helping X Religious Participation Caregiving X Religious Participation Mental Health Physical Health
ns ns ns ns ns + ns + + +
+ + ns ns ns ns ns ns ns ns
Notes: + indicates that role occupancy predicts an improvement in health; indicates that role occupancy predicts a decline in health; and, ns indicates a statistically nonsignificant relationship between role occupancy and health. Socioeconomic Status and Social Engagement Table 23 presents summary of results for SES variations in the mental and physical health effects of single roles, multiple roles, and role combinations. In terms of mental health, one indicator of SES, namely education emerged to be statistically significant in influencing the link between single roles and mental health. In particular, I found that the mental health benefits of paid work, formal volunteering, and informal helping were greater for older adults with low levels of education compared to their
100
higher educated peers. None of the three SES indicators were found to moderate the relationship between number of social roles and mental health. In regards to physical health, two indicators of SES, in particular income and assets were found to shape the relationship between single roles and physical health. In particular, I found those with low income to benefit more from being involved in paid work, and formal volunteering than their peers with high levels of income. Similarly, results showed that elders with low levels of assets derived more physical health benefits from being involved in paid work, formal volunteering, and informal helping compared to their peers with high levels of assets. In addition to single roles, I found all three indicators of SES (i.e., education, income, and assets) to moderate the link between multiple roles and physical health. More specifically, the physical health benefits of occupying multiple roles were greater for elders with lower levels of education, income, and assets compared to their peers with higher SES attainment. Table 23. Summary of SES Variations in the Mental and Physical Health Effects of Social Engagement Single Roles X SES Paid Work X Education Formal Volunteering X Education Informal Helping X Education Caregiving X Education Religious Participation X Education Paid Work X Income Formal Volunteering X Income Informal Helping X Income Mental Health Physical Health
+ + + ns + ns ns ns
ns ns ns ns ns + + ns
101
Table 23. Summary of SES Variations in the Mental and Physical Health Effects of Social Engagement (contd.) Single Roles X SES Caregiving X Income Religious Participation X Income Paid Work X Assets Formal Volunteering X Assets Informal Helping X Assets Caregiving X Assets Religious Participation X Assets Role Count X Education Role Count X Income Role Count X Assets Role Combinations X SES Paid Work X Formal Volunteering X Education Paid Work X Informal Helping X Education Paid Work X Caregiving X Education Paid Work X Religious Participation X Education Formal Volunteering X Informal Helping X Education Mental Health Physical Health
ns ns ns ns ns ns ns ns ns ns
ns ns + + + ns ns + + +
ns + ns ns ns
ns ns ns ns ns
102
Table 23. Summary of SES Variations in the Mental and Physical Health Effects of Social Engagement (contd.) Single Roles X SES Formal Volunteering X Caregiving X Education Formal Volunteering X Religious Participation X Education Informal Helping X Caregiving X Education Informal Helping X Religious Participation X Education Caregiving X Religious Participation X Education Paid Work X Formal Volunteering X Income Paid Work X Informal Helping X Income Paid Work X Caregiving X Income Paid Work X Religious Participation X Income Formal Volunteering X Informal Helping X Income Formal Volunteering X Caregiving X Income Formal Volunteering X Religious Participation X Income Informal Helping X Caregiving X Income Informal Helping X Religious Participation X Income Caregiving X Religious Participation X Income Paid Work X Formal Volunteering X Assets Mental Health Physical Health
ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns
ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns
103
Table 23. Summary of SES Variations in the Mental and Physical Health Effects of Social Engagement (contd.) Single Roles X SES Paid Work X Informal Helping X Assets Paid Work X Caregiving X Assets Paid Work X Religious Participation X Assets Paid Work X Informal Helping X Assets Paid Work X Caregiving X Assets Paid Work X Religious Participation X Assets Formal Volunteering X Informal Helping X Assets Formal Volunteering X Caregiving X Assets Formal Volunteering X Religious Participation X Assets Informal Helping X Caregiving X Assets Informal Helping X Religious Participation X Assets Caregiving X Religious Participation X Assets Mental Health Physical Health
ns ns ns ns ns ns ns ns ns ns ns ns
ns ns ns ns ns ns ns ns ns ns ns ns
Notes: + indicates that role occupancy predicts an improvement in health; indicates that role occupancy predicts a decline in health; and, ns indicates a statistically nonsignificant relationship between role occupancy and health. Race and Social Engagement Race-related findings showed that African Americans and White Americans are more similar than different in terms of the mental and physical health effects of later life
104
social engagement. I found only two instances where the two groups differed in the health effects of being socially involved. First, while paid work benefitted both Whites and blacks, the mental health enhancing impact of this role was greater for African Americans than their White peers. Second, the role of caregiving resulted in a decline in physical health for African Americans but not their White American counterparts. In addition to single roles, I found race difference in the physical health impact of multiple roles. In particular, findings revealed that occupying multiple social roles results in improved physical health among White elders but not for their African American peers. Table 24. Summary of Race Differences in Mental and Physical Health Effects of Social Engagement Single Roles and Race Mental Health Physical Health Paid Work X Black Formal Volunteering X Black Informal Helping X Black Caregiving X Black Religious Participation X Black Role Count X Black
+
ns ns ns ns ns
ns ns ns
ns
-
Role Combinations X Race Paid Work X Formal Volunteering X Black Paid Work X Informal Helping X Black Paid Work X Caregiving X Black
ns ns ns
ns ns ns
105
Table 24. Summary of Race Differences in Mental and Physical Health Effects of Social Engagement (contd.) Single Roles and Race Paid Work X Religious Participation X Black Formal Volunteering X Informal Helping X Black Formal Volunteering X Caregiving X Black Formal Volunteering X Religious Participation X Black Informal Helping X Caregiving X Black Informal Helping X Religious Participation X Black Caregiving X Religious Participation X Black Mental Health Physical Health
ns ns ns ns ns ns ns
ns ns ns ns ns ns ns
Notes: + indicates that role occupancy predicts an improvement in health; indicates that role occupancy predicts a decline in health; and, ns indicates a statistically nonsignificant relationship between role occupancy and health.
106
CHAPTER ELEVEN DISCUSSION Social engagement in later life has received growing attention over the past two decades with the emergence of gerontological concepts such as productive and successful aging. These concepts suggest that social engagement is a distinct form of activity and an important element of maintaining health and even recovering from illness in later life (Rowe and Kahn 1997; Svanborg 2001). Although empirical research generally supports these views, findings presented in most studies are based on a limited range of social roles mainly formal volunteering, caregiving, and religious participation. Other roles, including paid work and informal helping, have received less attention. Furthermore, studies examining links between social engagement and health have focused on number of roles, giving less attention to the health impact of role combinations. Role combinations are important to consider because, as the findings of the present study reveal, some roles can enhance the effects of other roles. Finally, there is an assumption in the social and productive aging literature that social engagement, in and of itself, is a good thing. This assumption is essentially problematic because research has yet to examine whether the protective effects of social engagement extend to different subgroups of the elderly population. The present study addresses these limitations and complements past research on social engagement and health in several ways. Unlike prior research, it provides the distribution of social engagement among elders of specific race and SES groups. The studys contribution also stems from its focus on a combination of both paid and unpaid social roles, including paid work, formal volunteering, informal helping, caregiving, and religious participation. Existing research explores the relationship between some of these roles to mental and physical well-being, but I investigate the effects of combining these roles in later life. Moreover, unlike existing studies that focus on the elderly population as a whole, the present study examined the process of later life social engagement under the context of two structurally relevant contexts, namely race and socioeconomic status.
107
In the sections below, I discuss patterns related to the following four topics: (1) the distribution of roles among older adults, (2) health effects of individual roles and role count, (3) role combinations and their health effects, and (4) the relevance of race and socioeconomic status in the relationship between social engagement and mental and physical health. Role Distribution among Older Adults In prior research, the emphasis on the concept of social engagement exists partly to contradict the stereotypes of old age and reverse the social devaluation of the elderly. While the focus of the present study is not to authenticate that older adults are in fact useful to society because they are productive, it does extend prior findings related to social involvement in later life. I find that more than a third of the elderly population work for pay, more than a third volunteer, more than two-thirds are involved in informal helping activities, forty-five percent are engaged in caregiving and sixty-seven percent are involved in regular religious activity. These patterns are consistent with prior studies on social engagement. The present study, however, adds to existing research by assessing race and SES variations in types, number, and combinations of five social roles. Overall findings reveal that White and higher SES older adults have significantly higher rates of participation in roles such as paid work, formal volunteering, and caregiving than their African American and lower SES counterparts. In contrast with the patterns for these roles, I find that African American and lower SES elderly report higher rates of involvement in informal helping and religious activity. These findings are not surprising given that a higher proportion of African Americans and lower SES older adults may be prevented from engaging in paid work due to illness or disability (Kelly-Moore and Ferraro 2004). In terms of volunteering, two reasons could explain the lower rates of participation by African American and lower SES elders. First, both African American and lower SES individuals are likely to lack
108
cultural resources that facilitate gaining information about volunteer work, organizing it, and furnishing social support and backing. Second, being asked to volunteer is a fundamental determinant of volunteering; and, prior studies have revealed that both African Americans and lower SES individuals are less likely to be asked to participate in volunteer activities (Ferree, Barry, and Manno 1998; Hodgkinson 1995). In contrast, roles such as informal helping and religious participation represent unique opportunities for African American and lower SES individuals to remain socially and meaningfully engaged and provide service to those in their community. In addition, these roles also may enable African American and lower SES elderly to meet people from similar walks of life, enabling a sense of mutual understanding, trust, and ultimately a strong and meaningful social network. This suggests that the focus in the literature solely on formal volunteering overlooks both the less visible segments of the elderly population and the less formal means of civic engagement. In sum, these findings suggest that while a substantial proportion of the elderly population is engaged in social roles, the involvement in social engagement varies by individuals social structural location. Further analyses also reveal that the benefits associated with social engagement vary by older adults race and socioeconomic status. I review findings related to this in the subsequent sections but first I discuss results for the health effects of individual social roles. Single Roles and Health Findings pertaining to single roles show that social engagement does indeed strongly and positively impact changes in both mental and physical health. I find that three of the five social roles, namely paid work, formal volunteering, and informal helping, positively predict mental and physical health for older adults. This is consistent with several prior studies that report these three social roles provide elders with effective social and psychological resources to not only maintain physical and mental health but also recover from illness. Involvement in these roles provides individuals with personal,
109
social, and psychological resources, all which protect against mental and physical distress and result in overall well-being (House, Landis, and Umberson 1988; Moen, DempsterMcClain, and Williams 1989; Moen, Dempster-McClain, and Williams 1992). While both formal volunteering and informal helping positively shape later life health, being a volunteer has a stronger effect on both mental and physical health than being an informal helper. This pattern may have emerged because compared to informal helping, formal volunteering represents a more socially and culturally recognized position in society. In turn, being involved in a role that is formally distinguished may give older adults a greater sense of accomplishment and recognition than engaging in a less socially visible activity, such as informal helping. In contrast to paid work, formal volunteering, and informal helping, caregiving and religious participation fail to emerge as significant predictors of health in this sample; and, this is surprising given the often found links between these two social roles and health. The lack of health impact of caregiving in this sample may reflect similar socioeconomic circumstances between caregivers and non-caregivers. Prior research suggests that while caregiving often negatively affects health, this relationship may be contingent upon several interrelated factors, including socioeconomic resources available to caregivers and level of illness and disability among care recipients (Lim and Zebrack 2004). While the latter cannot be examined using the ACL data, to examine the function economic resources play in determining the impact of caregiving, I estimated models including interactions between caregiving and SES. The analyses did not show SES to affect the link between caregiving and either of the two health outcomes. Finally, contrary to prior studies, the present study did not find religious participation to affect mental health among older adults. The lack of significant mental health differences between elders who participate in religious activities and those who do not, may be reflective of the latter being engaged in spiritual activities. While substantive differences may exist between religiosity and spirituality, those involved in either of these activities may derive positive self-perceptions, support and control, which is
110
mirrored in the comparable health effects between those engaged and not engaged in religious participation. While the present study was unable to control for spirituality, future studies should consider its role in the relationship between religious involvement and psychological well-being. Multiple Roles and Health In addition to single roles, the present study also examined the impact of the number of roles on later life health. The results revealed that being involved in multiple roles reduces depressive symptoms and functional impairment and leads to better mental and physical health. These patterns match those in previous studies, which suggest that engagement in several different social roles provides older adults with a unique opportunity to maintain social and psychological resources (Hinterlong et al. 2007; Moen et al. 1992; Morrow-Howell 2002; Thoits 1983). In addition to remaining integrated in the community, being involved in varied roles allows older adults to hold positive perceptions of social and instrumental support, both of which are found to be strongly connected with better health outcomes (Antonnuci and Akiyama 1997; Berkman et al. 1992; Glynn, Christenfeld, and Gerin 1999). Moreover, multiple role involvement also results in enhanced levels of psychological resources, such as mastery and self-esteem, both of which are critical in predicting later life well-being. Role Combinations and Health While examining multiple role performance shows us the relationship between the number of roles occupied and health, it does not tell us which combinations of roles are particularly health enhancing as well as those that may be harmful to health. Taking this into account, the main task of this study included examining how combining different social roles affects mental and physical health in later life. I argue that role combinations are important to consider because failing to do so inevitably leads us to assume that all roles are similar in their meaning, quality, and effects. Prior research as well as findings presented here suggests that different roles have different health effects. To explore the
111
effect each role has on the other, I examined ten two role combinations and of these, four emerged as statistically significant predictors of changes in mental health and two predicted changes in physical health. The four combinations that were found to predict changes in mental health include formal volunteering and caregiving, informal helping and caregiving, informal helping and religious participation, and caregiving and religious participation. Caregiving, surprisingly, has no impact on the mental health of older adults. Yet, combining caregiving with either volunteering or informal helping yields positive mental health results. More specifically, volunteering and informal helping both enhance mental health for those involved in caregiving. While existing research has not examined the association between caregiving roles (an often obligatory activity) and volunteer roles (a discretionary activity), there are reasons to expect that volunteer and helping behaviors may positively affect caregivers. In particular, caregiving is not carried out in social isolation. Caregivers often receive formal and/or informal assistance from organizations, families, and individuals within their community, sometimes at low or no cost, which makes the task of providing care more viable. As Burr and colleagues (2005) note, these caregivers often are motivated to give back to the organizations and people that provide them assistance in their caregiving tasks. Being able to establish this social network of care within the community, where caregivers are both recipients and providers of help may positively impact their mental well-being. Moreover, caregivers who also volunteer or are involved in informal help giving in the community may develop enhanced perceptions of social support. That is, involving oneself in multiple care/help activities not just within the family but broader community may lead caregivers to perceive strong and constant social support amidst their own stressful life transitions. In turn, these positive perceptions of social support in later life may boost mental health among caregivers. Finally, being able to provide assistance to non-kin members and volunteering for organizations despite having
112
competing family caregiving responsibilities may only mean greater sense of resilience, control, and self-validation all of which are reflected in enhanced psychological health for caregivers. Results also show that combining religious participation and informal helping positively impacts mental health. In particular, I find that informal helping tends to foster mental health among religious participants. This finding is interesting yet expected given the common forces that bind people in these two social roles. In other words, those involved in religious activity often are bound by values of responsibility towards the community, and a moral obligation to help those in need. Informal helping, therefore, represents an opportunity for those involved in religious activity to help others and carry out their social and more importantly moral obligation towards others resulting in an improved mental well-being. In terms of physical health, results showed that two role combinations, namely paid work and volunteering, and paid work and informal helping, are significant predictors of improvements in functional impairment. Both volunteering and informal helping led to lower levels of functional impairment among paid workers. A reason that may explain this finding is that those working for pay are likely to reap greater social and psychological resources if they also participate in help-related activities, such as volunteering and informal helping. Being able to engage in roles that enable helping others in the community despite the competing paid work commitments may result in enhanced self-perceptions, which in turn may positively shape physical health. While my study highlights the relevance of examining role combinations, it also shows us that several role combinations do not have a health impact over and above their individual effects as single roles. For instance, while formal volunteering and informal helping individually and positively influence both mental and physical health, combining these two roles does not yield a statistically significant health effect. Similarly, in contrast to prior findings that show caregivers to experience greater job-related stress than non-
113
caregivers (Goldsmith and Goldsmith 1995; Scharlach 1994), present study reveals no health effect of combining these two roles. SES, Social Engagement and Health On the whole, findings discussed thus far coincide with prior studies suggesting that being socially engaged enhances mental and physical health among older adults. What existing research fails to explain, however, is whether the positive influence of social engagement extends to all elderly regardless of their social structural location. This study assessed links between SES, social engagement and health; and, results revealed that the impact of being socially engaged differs among elders of varying SES levels. Overall findings reveal that lower SES older adults benefit more from occupying several social roles compared to their upper SES counterparts. In particular, while the mental health benefits of working for pay, formal volunteering, and informal helping are greater for older adults with lower educational attainment, the physical health gains of paid work, volunteering, and informal helping are greater for the elderly with low levels of income and assets. These findings stand contrary to expectations given the potential SES differences in the quality of roles occupied. Besides the differences in role quality, research also documents variations in life course capital (Elman and ORand 2004; House et al. 1994), which is needed to successfully carry out role obligations. In other words, I expected lower SES elders to benefit less from social engagement because they lack adequate social, cultural, and psychological resources that often are needed to effectively manage social roles. Nevertheless, the present finding that lower SES older adults benefit more from being socially engaged is encouraging and demands that we explore alternate explanations. One such explanation is that lower SES individuals may reap greater sense of control, mattering, and independence than upper SES peers from being able to successfully occupy socially and economically meaningful roles until late in their lives.
114
For instance, volunteering and helping others in the community probably may lead to a greater sense of mattering and mastery among those with competing life stressors compared to those who are more advantaged socioeconomically. Being able to help others despite their own problems may result in a sense of pride, self-validation, and selfefficacy among lower SES elderly. In comparison, elders occupying the higher end of the SES ladder may derive their sense of control not from their recent or current engagement status, but rather from their years of stable and socially acknowledged role histories. The foregoing suggests that the relationship between social engagement and health is far more complex that it initially appears. Even more striking are the unique roles that education, income, and assets play in the link between later life social engagement and health. Education, which tends to occur early in the life course, plays a critical role in influencing the relationship between social engagement and mental health. Income and assets, alternatively, shape the link between social engagement and physical health. This difference is likely to emerge because education is found to matter more in predicting psychosocial resources, such as selfesteem, sense of control, and mental well-being, whereas income and assets are more predictive of the physical maintenance of self. Education, in and of itself without adequate financial resources, such as income and assets may do little to maintain physical health or recover from illness (Herd, Goesling, and House 2007). It is clear that socioeconomic factors play an important role in the relationship between later life social engagement and health. Yet existing studies, including the present one, are focused on a limited range of SES indicators. Future research should consider expanding the conceptualization of SES. That is, in addition to broad measures of SES (e.g., education, income, and assets), forthcoming studies should investigate specific factors that are distinct from, but often reflect SES, such as social and cultural capital, neighborhood characteristics, access to health care, and availability of transportation. Doing so will provide a more detailed picture of the ways in which socioeconomic advantages or alternatively, disadvantages translate into opportunities or barriers related to social engagement and its health benefits in later life.
115
Race, Social Engagement and Health In addition to socioeconomic status, the present study assessed whether the effect of social engagement on health varies by race. Unlike SES, however, overall findings revealed little variation in the health impact of being socially engaged between White and Black elderly. Stated otherwise, based on findings presented here, it appears that White and Black older adults are more similar than dissimilar in terms of the health benefits they derive from being socially engaged. There are only two instances, however, where race differentials do emerge. One of the race differences that emerged indicated that older Black caregivers report greater functional impairment than their elderly White counterparts. Race differences in the distribution of resources to provide care and the level of impairment among care recipients may explain this finding. For instance, research shows that Black caregivers are less likely to use formal services. This is partly because of cultural resistance to the use of such services, but another reason for this difference is that some formal services may be less accessible to racial minorities (Schoenberg et al. 1998; Wallace et al. 1998). In addition to economic resources, research also shows variations in the health resources between White and Black caregivers. In particular, African American caregivers are far more likely to be in poor health themselves, affecting their ability to provide care and heightening the stress that emerges from being a caregiver (Dautzenberg et al. 1999; Williams et al. 2003). Finally, besides caregiver-related factors, aspects related to care recipients come into play as well. For example, we know that Black care recipients endure higher levels of physical and cognitive impairment (Calderon and Tennstedt 1998), which may result in greater physical stressors among Black caregivers compared to their White peers involved in caregiving. In contrast to the physical distress emerging from caregiving, older Blacks report enhanced mental health in case of one social role: paid work. Specifically, I find older African American men and women to reap greater benefits from paid work compared to their White elderly peers. This is surprising given that a greater proportion of African
116
Americans tend to hold low paying jobs and work under poor employment conditions. Further, research shows African Americans to be at greater risk for physical and cognitive impairment and at a much earlier age in life as compared to White elders (Clark and Maddox 1992; Ferraro, Farmer, and Wybraniec 1997). What then explains the finding that African Americans report better mental health from working in late life? This pattern may be the result of African Americans lower savings over the life course and a greater need of income in later life than their White peers. Thus, being able to work for pay remain self-reliant may actually serve as a source of strength to them, which is reflected reflecting in their psychological well-being. Alternatively, this counterintuitive finding may be a result of race patterns in selection into paid work in late life. In other words, it is plausible that older Blacks who were in worst health had to stop working early, leaving the relatively healthy African Americans engaged in paid work. Moreover, it is likely that older African Americans who remain in paid work are the ones with not only good health but also superior work conditions that are comparable to those enjoyed by their White peers. To explore this possibility, I re-estimated models employing interactions between paid work, race, and SES; however, results from these analyses failed to support the alternative explanation. In other words, I found no variation in the mental health effects of paid work between White and Black elders of varying SES levels. In addition to single roles, I examined whether multiple roles and role combinations affected health differently for White and Black older adults. Results revealed no racial variation in the health effects of combining social roles. I also did not find race differences in the mental health impact of having multiple roles in later life. However, race did influence the relationship between multiple role performance and physical health. In particular, I find White elders to gain more from occupying multiple roles as opposed to their African American counterparts. This pattern may be the result of greater functional and cognitive impairment among Black older adults compared to White elders (Hayward and Heron 1999; Manton, Patrick, and Johnson 1987). In other words, while individual roles may affect health similarly for the two groups, occupying
117
multiple roles may represent greater physical strain for African American elders, limiting the physical health benefits accruing from multiple role occupancy. Lack of variation in the health effects of other social roles and role combinations is surprising given the uneven distribution of resources between Blacks and Whites to effectively occupy and maintain social roles. In particular, I anticipated Blacks to report fewer benefits from formal volunteering given that formal volunteering is mainly a White, middle-class activity, which likely poses practical barriers and difficulties, including lack of commitment by mainstream organizations to involve Black and ethnic minorities, conflicting cultural expectations, tokenism, and racial discrimination (Akpeki 1995; Niyazi 1996). The lack of difference in the benefits of volunteering between Black and White older adults, therefore, stands contrary to the assumption that racial minorities benefit less from roles that are more formal and traditionally intended for White Americans. The lack of overall race differences in the present study should not be deemed as a definitive statement about Black and White older adults involved in later life social activity. Future studies, instead, should continue examining how being engaged in various social roles affects similarly or/and differently older Black and White elders. One way to do so is to examine multiple facets of mental and physical health while exploring racial variations in social engagement and later life well-being. Prior research suggests that the expression of both stress and well-being is likely by race (Vega and Rumbaut 1991; Williams and Harris-Reid 1999). For instance, studies employing clinical measures suggest that African Americans have lower levels of major depression, despite consistently reporting depressive symptoms in studies utilizing indicators of non-clinical depression. Additionally, the stress literature suggests that while African Americans are exposed to more stress than their White peers, they are less likely to experience substance abuse problems. In terms of physical health, however, Blacks report worse outcomes compared to Whites. These seemingly paradoxical findings in existing research underscore the relevance of developing research on later life social engagement that includes varied aspects of mental and physical health outcomes. Doing so is important
118
because it would enable us to identify the specific mental and physical reactions experienced by distinct social groups to later life social engagement. In sum, while the present study extends prior research by examining the influence of socioeconomic status and race on the link between social engagement and health, a more comprehensive research agenda would include assessing the interactive effects of SES, race and other structural factors, including gender and age. In addition to the structural level factors, future research also should consider how older adults perceive social roles and sustain types of social engagement over time. In particular, forthcoming studies should consider how individuals come to enter and exit social roles in later life. Are specific segments of the elderly population more or less likely to make certain role transitions? What are the social, economic, political, and cultural factors that influence older adults decisions remain or alternatively, not remain socially engaged? Findings from the present study identify a unique relationship between social engagement and health; however, its theoretical underpinnings are not examined in this study. Extant evidence indicates that role quality, role identity, and the centrality and satisfaction associated with roles shape the link between social engagement and health (Simon 1995, 1997; Thoits 1983, 1992). While the data used in this study does not provide measures for these dimensions of role performance across the activities, future efforts should include employing other panel studies to examine these issues.
Limitations and Future Directions Although this study makes an important contribution by examining several of the previously underexamined issues related to later life social engagement and health, it has several important limitations. The primary limitation of this study is its inability to demonstrate conclusively a causal link between social engagement and health. While attempts were made to adjust for temporality, there was not adequate control of unobserved factors and, as such, I am unable to ascertain the direction of the observed associations. One way to assess the extent of reciprocal links between social engagement
119
and health was to re-estimate models9 regressing social engagement at Wave 2 on mental and physical health at Wave 1. The findings of these supplementary analyses revealed that while prior levels of mental health did not influence subsequent role involvement, participation in four of the five social roles was contingent on Wave 1 physical health. In particular, I found that those with physical limitations at Wave 1 were less likely to be engaged in paid work, formal volunteering, informal helping, and religious activities at Wave 2. Caregiving, however, was one role that was not affected by prior levels of functional impairment. This pattern may reflect the often obligatory nature of caregiving, which compels individuals to meet their caregiving responsibilities despite their competing physical limitations. Future studies on the links between social engagement and health must more rigorously evaluate the causal assumptions presented here using more sophisticated modeling techniques, including SEM and latent growth curve procedures (Kessler and Greenberg 1981). Using these techniques will allow us to determine more decisively the causal linkages between social engagement and mental and physical health. Moreover, it will enable us to uncover the stages at which differences in health effects of social roles based on race and SES peak or become most pronounced. In addition to these methodological limitations, theoretical weaknesses of the present study deserve attention. Specifically, while I examine the otherwise underexamined links between social structure, social engagement, and later life health, the explanations offered to account for these associations need to be empirically evaluated. For instance, I argue that lower SES older adults are likely to be confronted by chronic stressors, which limits the health benefits accruing from social engagement. However, measures of chronic strain experienced by older adults were not obtained in the ACL survey. Similarly, the data used in this study also lack measures of the institutional constructs, including role quality, which is known to be an important predictor of both mental and physical health (Adelmann 1987; Baruch and Barnett 1986). For paid work,
These analyses examining the selection effects controlled for all Wave 1 sociodemographic variables, including age, gender, race, marital status, education, income, and assets.
9
120
two items assess the individuals perceptions of various job characteristics (i.e., psychosocial demands of the job and job decision latitude). However, there are no similar measures of quality for the other social roles. Upcoming studies should employ other panel studies that enable assessing role quality as a factor in the relationship between later life social engagement and health. Future research also should examine other roles in addition to the ones examined in the present study. For example, family roles, such as spouse, parent, and grandparent may be equally meaningful in later life and consequently affect mental and physical health among older adults. Moreover, it would be interesting to examine whether findings related to configuration of these roles for young and middle-aged adults extend to those in later life. For instance, how does combining the spousal and parental roles in later life compare to a combination of these roles in earlier stages of the life course? And, do the health benefits or costs associated to these family roles and role combinations vary by older adults social structural location? These are questions that should prominently feature in future studies on later life social engagement and health. Besides examining other roles, researchers also should consider additional health outcomes. For instance, while existing research examines the mental and physical health consequences of social engagement, future research should assess whether being socially engaged affects older adults cognitively. Social disengagement is suggested to result in a cognitive decline among older adults (Bassuk, Glass, and Berkman 1999; Hultsch, Hammer, and Small 1993; Rogers, Meyer, and Mortel 1990); however, what remains less well-understood is how engaging in socially meaningful roles, such as formal volunteering and informal helping influences cognitive competency in later life. Examining additional health outcomes is important if we want to capture the full extent of the impact of social engagement in later life. Another aspect to be considered in future studies on social engagement is the distinction between current role occupancy and histories of role involvement. An examination of this type would enable us to assess the more fine-grained SES and race
121
differences in social engagement. For instance, we may find that while lower SES elders benefit socially and psychologically from current role participation, those belonging to a higher socioeconomic status may reap such benefits from their years of stable and socially valued role histories. Differentiating between present and past role involvements also would enable us to distinguish between the relevance of current resources versus resources accumulated over the life course for later life health and well-being. Finally, while SES emerges as an important factor shaping later life effects of social engagement, other social structural factors need to be considered as well. For instance, gender is an important social construct that is inextricably tied to both SES and health outcomes. Evidence of this lies in research suggesting that the beneficial effects of education are lower for women as compared to their male counterparts. Women also are more likely to experience depression and chronic conditions than their male counterparts. Taken as a whole, this area of research suggests that focusing on the intersection of SES, race, and gender is likely to lead to a more complete understanding of the impact of social engagement in later life.
122
CHAPTER TWELVE CONCLUSION The present study builds on prior research in several important ways. Results indicate that certain social roles and role clusters are particularly health enhancing. More importantly, results also reveal that while social engagement is indeed health-enhancing, the mental and physical health benefits of certain roles vary by older adults race and SES. These group variations may indicate that the meaning, importance, and health effects of social engagement may be socially and economically driven. That is, variations in social and economic resources are likely to determine role meaning, role quality, role satisfaction, and ultimately benefits that accrue from occupying a given social role. Specifically, my study points out that our efforts to promote social engagement should include our efforts to identify the differences within the elderly population. While my findings expand our current understanding of the relationships between later life social engagement and health, they also raise some questions concerning the meaning and measurement of social engagement and the commonly held assumption that social engagement in later life is, in and of itself, a beneficial thing. Regarding what is defined as social engagement, the present study argues in favor of a broader definition, that includes not only paid work and formal volunteering but also less visible roles, such as informal helping, caregiving, and religious participation. Future research should expand on the current study by examining other roles held by elders, including spousal, parent, and grandparent roles. While previously considered as unproductive in nature, roles such as these allow for meaningful interaction not only for older adults but also for families and communities more broadly. For instance, caring for grandchildren contributes to the well-being of the extended family. Similarly, helping out children or providing them with emotional and financial support decreases older adults dependence in later life. Such activities can thus be a part of a larger exchange across generations, wherein they provide meaning to not only individuals but also to families and society in general.
123
In addition to the definition of social engagement, it also is important to take into account the ways in which social participation is measured. The present study contributes to this cause by examining the health impact of combining social roles. Future research should extend this effort by exploring how older adults perceive different social roles. For instance, what does it mean to be a volunteer, informal helper and/or a caregiver in later life? This line of research suggests that the meanings individuals attach to roles are a function of differences in the ways they are socialized into roles. It also allows us to explore the differences and similarities in the meanings of roles based on the age, gender, class, and race of individuals. For instance, a compelling body of research reveals that certain communities of African Americans consider parenting as an act of volunteering (Newman 1994). For them, children are privately owned public goods, who if raised well can lead to a conscientious and productive nation. Variations such as these in the very meaning of social roles can ultimately lead to variations in the health benefits accruing from them. Finally, results of the current study underscore the importance of examining the diversity of experiences in social engagement among older adults. In particular, findings indicate that the benefits of social engagement vary by SES. Lower SES older adults are likely to benefit more from being socially engaged compared to their higher SES counterparts. This is an unexpected finding but it demands that future research examines the role played by social structure in shaping the link between social engagement and health in later life. By doing so, we know who benefits from being socially engaged but more importantly, we also learn about those who cannot or do not fulfill these roles and expectations, and consequences they experience as a result of social disengagement. These issues are of growing importance as we find ourselves in the midst of a longevity revolution. How we spend time in these prolonged years of our lives is yet to be determined and the possibilities are endless. We need to identify and build on these possibilities based on empirical findings of what enhances both mental and physical health of our large and growing aging population. By looking at the links between the
124
types, and combinations of social activity and health outcomes for the elderly, the present study takes a step towards that goal.
125
APPENDIX A
MEANS AND STANDARD DEVIATIONS OF VARIABLES BY RACE Whites (n = 1,135) W2 Depressive Symptoms (CES-D) W2 Functional Impairment W1 Depressive Symptoms (CES-D) W1 Functional Impairment Women (percent) Age Years of Education Household Income Total Assets -.091 (.93) 1.58 (.92) -.096 (.93) 1.51 (.89) .67 66.57 (.25) 11.44 (3.2) 24.15 (22.4) 30.03 (17.9) Blacks (n = 479) .251 (1.0)*** 1.80 (1.1)*** .227 (1.0)*** 1.73 (1.0)*** .68 65.76 (.40) 8.84 (3.8)*** 12.37 (14.4)*** 15.37 (11.1)***
Notes: *p<.05 **p < .01 ***p < .001; Standard deviations are in parentheses; t-tests are used; W1 refers to Wave 1, W2 refers to Wave 2
126
APPENDIX B MEANS AND STANDARD DEVIATIONS OF VARIABLES BY EDUCATION --------------------------------------------------------------------------------------------------------------------Variable < High School High School College Degree N = 767 N = 448 N = 399 ---------------------------------------------------------------------------------------------------------------------
W2 Depressive Symptoms (CES-D) W2 Functional Impairment W1 Depressive Symptoms (CES-D) W1 Functional Impairment Women (percent) Black (percent) Age Household Income
.257 (1.0) 1.86 (1.0) .218 (1.0) 1.80 (1.0) .66 .44 67.92 (8.6) 11.26 (11.6)
-.167 (.89) 1.44 (.83) -.174 (.88) 1.35 (.76) .72 .16 64.63 (8.3) 24.72 (21.3)
-.264 (.80)*** 1.35 (.82)*** -.124 (.82)*** 1.30 (.79)*** .63 .15*** 65.18 (8.45) 34.1 (25.)***
Total Assets 17.7 (12.9) 30.1 (17.5) 36.6 (1.8)*** -----------------------------------------------------------------------------------------------------------Notes: *p<.05 **p < .01 ***p < .001; Standard deviations are in parentheses; t-tests are used; W1 refers to Wave 1, W2 refers to Wave 2
127
APPENDIX C
MEANS AND STANDARD DEVIATIONS OF VARIABLES BY INCOME --------------------------------------------------------------------------------------------------------------------Variable < $20,000 $20,000-30,000 > $30,000 N = 1043 N = 365 N = 206 ---------------------------------------------------------------------------------------------------------------------
W2 Depressive Symptoms (CES-D) W2 Functional Impairment W1 Depressive Symptoms (CES-D) W1 Functional Impairment Women (percent) Black (percent) Age Years of Education
.175 (1.0) 1.82 (1.0) .146 (1.0) 1.74 (1.0) .73 .38 68.31 (8.5) 9.51 (3.5)
-.270 (.79) 1.36 (.74) -.223 (.87) 1.30 (.73) .58 .13 63.53 (7.7) 12.20 (2.7)
-.327 (.84)*** 1.27 (.62)*** -.349 (.75)*** 1.21 (.56)*** .52*** .12*** 61.28 (7.1)*** 13.80 (2.4)***
Total Assets 18.30 (12.4) 35.14 (16.3) 47.71 (17.0)*** -----------------------------------------------------------------------------------------------------------Notes: *p<.05 **p < .01 ***p < .001; Standard deviations are in parentheses; t-tests are used; W1 refers to Wave 1, W2 refers to Wave 2
128
APPENDIX D MEANS AND STANDARD DEVIATIONS OF VARIABLES BY ASSETS --------------------------------------------------------------------------------------------------------------------Variable < $20,000 $20,000-50,000 > $50,000 N = 965 N = 401 N = 248 ---------------------------------------------------------------------------------------------------------------------
W2 Depressive Symptoms (CES-D) W2 Functional Impairment W1 Depressive Symptoms (CES-D) W1 Functional Impairment Women (percent) Black (percent) Age Years of Education
.196 (1.0) 1.83 (1.0) .215 (1.0) 1.76 (1.0) .61 .46 66.56 (8.9) 9.29 (3.5)
-.132 (.84) 1.44 (.82) -.225 (.87) 1.36 (.79) .58 .15 66.45 (8.5) 12.03 (2.9)
-.368 (.78)*** 1.38 (.74)*** -.340 (.75)*** 1.33 (.72)*** .47*** .04*** 64.71 (7.7) 12.92 (2.6)***
Total Income 11.39 (11.0) 24.04 (17.5) 44.83 (28.9)*** -----------------------------------------------------------------------------------------------------------Notes: *p<.05 **p < .01 ***p < .001; Standard deviations are in parentheses; t-tests are used; W1 refers to Wave 1, W2 refers to Wave 2
129
APPENDIX E
BINARY LOGISTIC REGRESSION OF WAVE 2 ROLE OCCUPANCY ON WAVE 1 MENTAL AND PHYSICAL HEALTH AND SOCIODEMOGRAPHIC VARIABLES Informal Religious Paid Work ( Volunteering helping (yes = Caregiving Participation yes = 1) (yes = 1) (yes = 1) 1) (yes = 1) Model 1 Model 2 Model 3 Model 4 Model 5 Wave 1 Health Variables Depressive Symptoms (CES-D) Functional Impairment Sociodemographic Variables Age Women Black Married Years of Education Household Income Total Assets -.134 *** (.01) -.683 *** (.14) .165 (.17) -.275 (.15) .094 *** (.02) .013 *** (.00) -.156 *** (.05) .006 (.00) .230 (.12) .254 (.14) .438 *** (.12) .147 *** (.02) -.006 (.00) .173 *** (.04) .10 -881.01 -.073 *** (.00) -.135 (.14) -.568 *** (.14) .055 (.13) .134 *** (.02) -.007 (.00) .093 (.05) .17 -764.10 -.028 *** (.00) .334 ** (.11) -.345 (.13) .428 *** (.12) .063 *** (.01) -.003 (.00) .016 (.04) .05 -964.89 .005 (.00) .708 *** (.12) .970 *** (.14) .197 (.12) -.014 (.01) -.003 (.00) .090 * (.04) .05 -899.26 -.104 (.08) -.668 *** (.10) -.127 (.06) -.269 *** (.07) -.166 * (.06) -.270 *** (.06) .082 (.06) -.012 (.06) -.073 (.06) -.182 * (.06)
Psuedo R-Squared .26 Log Likelihood -637.02 Notes: *p<.05 **p < .01 ***p < .001;
130
APPENDIX F FSU HUMAN SUBJECTS APPROVAL LETTER
131
REFERENCES Adelmann, Pamela K. 1994. Multiple Roles and Psychological Well-being in a National Sample of Older Adults. Journal of Gerontology: Social Sciences 49: S277-S285. Adler, A. 1937. Understanding Human Nature. Hazelden: Center City, MN. Akpeki. 1995. Black on Board: Report of a conference held at NCVO on 30 November 1994, NCVO. Aneshensel, Carol S., Ralph R. Fredericks, and Virginia A. Clark. 1981. Family Roles and Sex Differences in Depression. Journal of Health and Social Behavior 22:379-93. Aneshensel, Carol. S., Pearlin, Leonard. I., and Schuler, Roberleigh. H. 1993. Stress, Role Captivity, and the Cessation of Caregiving. Journal of Health and Social Behavior 34: 54-70. Antonucci, Toni. C., and Akiyama, H. 1997. Social Support and the Maintenance of Competence. In S. Willis and K. W. Schaie (Eds.), Societal Mechanisms for Maintaining Competence in Old Age. Springer Publishing Co. Baltes, Margaret. M., Wahl, H. W., and Schmid-Furtoss, U. 1990. The Daily Life of Elderly Germans: Activity Patterns, Personal Control, and Functional Health. Journal of Gerontology 45: 173-179. Baron, Reuben. M., and Kenny, David. A. 1986. The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations. Journal of Personality and Social Psychology 51: 1173-1182. Barnett, Rosalind. C., and Baruch, Grace K. 1985. Women's Involvement in Multiple Roles and Psychological Distress. Journal of Personality and Social Psychology 49: 13545. Baruch, Grace. K., and Barnett, Rosalind. C. 1986. Role quality, Multiple Role Involvement, and Psychological Well-Being in Midlife Women. Journal of Personality and Social Psychology 51: 578585. Bassuk, Shari. S., Glass, Thomas. A., and Berkman, Lisa. F. 1999. Social Disengagement and Incident Cognitive Decline. Annals of Internal Medicine 131: 165173. Beck, Scott. H., and Page, Joe. W. 1988. Involvement in Activities and the Psychological Well Being of Retired Men. Activities, Adaptation, and Aging 11: 3147. Berkman, Lisa. F., and Breslow, Lester. 1983. Health and Ways of Living: The Alameda County Study. New York: Oxford University Press.
132
Berkman, Lisa. F., Leo-Summers, L., and Horwitz, R. I. 1992. Emotional Support and Survival after Myocardial Infarction. Annals of Internal Medicine 117:1003-1009. Belsley, David. A., Kuh, Edwin. and Welsch, Roy. E. 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New York: John Wiley. Biddle, Bruce J. 1986. Recent Developments in Role Theory. Annual Review of Sociology 12: 67-92. Boss, P., Caron, W., Horbal, J. and Mortimer, J. 1990. Predictors of Depression in Caregivers of Dementia Patients: Boundary Ambiguity and Mastery. Family Process 29: 245254. Britton, N.J. 1999. Recruiting and Retaining Black Volunteers: A Study of Black Voluntary Organisations. Voluntary Action 1: 9-23. Brody, Elaine. M. 1985. Parent Care as a Normative Family Stress. The Gerontologist 25:1929. Broman, Clifford L. 1991. Gender, Work-Family Roles, and Psychological Well-being of Blacks. Journal of Marriage and the Family 53:509-520. Brown, William. M., Consedine, Nathan., and Magai, Carol. 2005. Altruism Relates to Health in an Ethnically Diverse Sample of Older Adults. The Journal of Gerontology: Psychological and Social Sciences 60: P143-152. Burgess, Ernest. W. 1954. Social Relations, Activities, and Personal Adjustment. American Journal of Sociology 59: 352360. Burr, Jeffrey A., Choi, Namkee G., Mutchler, Jan E., and Garo, Francis G. 2005. Caregiving and Volunteering: Are Private and Public Helping Behaviors Linked? Journals of Gerontology 60: S247-S256. Burton, Russell. 1998. Global Integrative Meaning as a Mediating Factor in the Relationship between Social Roles and Psychological Distress. Journal of Health and Social Behavior 39: 20115. Butler, Robert. N and Gleason, Herbert. P. 1985. Productive aging. New York: Springer Publishing Company. Calasanti, Toni. M., and Slevin, Kathleen. F. 2001. Gender, Social Inequalities, and aging. CA: Alta Mira Press. Campbell, Angus., Converse, Philip and Rogers, Willard L. 1976. The Quality of American Life. New York: Sage.
133
Caro, Francis. G., Bass, Scott. A., and Chen, Yung-Ping. 1993. Introduction: Achieving a Productive Aging Society. In Bass, S. A, Caro, F.G., and Chen, Y (eds.), Achieving a Productive Aging Society, pp. 3-26. Westport, CT: Auburn House. Cohen, Sheldon. 1988. Psychosocial Models of the Role of Social Support in the Etiology of Physical Disease. Health Psychology 7: 269-297. Coleman, Lerita. M., Antonucci, Toni. C., and Adelmann, Pamela. K. 1987. Role Involvement, Gender, and Well-being. In Crosby, Faye. J (eds.) Spouse, Parent, Worker: On Gender and Multiple Roles, pp. 138-153. New Haven, CT: Yale University Press. Coleman, Lerita. M., Antonucci, Toni. C., Adelmann, Pamela. K., and Crohan, S. E. 1987. Social Roles in the Lives of Middle-Aged and Older Black Women. Journal of Marriage and the Family 49: 761-771. Collins, John Gary. 1988. Prevalence of Selected Chronic Conditions, United States, 1983-85. NCHS Advance Data 155: l-16. Davies, Allyson. R., and Ware, John. E. 1981. Measuring Health Perceptions in the Health Insurance Experiment. Santa Monica: Rand. Dorfman, Lorraine. T., Kohout, Frank. J., and Heckert, D. A. 1985. Retirement Satisfaction in the Rural Elderly. Research on Aging 7: 577599. Dulin, P., Hill, R. D., Anderson, J., and Rasmussen, D. 2001. Altruism as a Predictor of Life Satisfaction in a Sample of Low-income Older Adult Service Providers. Journal of Mental Health and Aging 7: 349359. Durkheim, Emile. 1951 [1897]. Suicide. New York: Free Press. Eaton, William., and Kessler, Larry. G. 1981. Rates of Symptoms of Depression in a National Sample. American Journal of Epidemiology 114: 528-538. Elder, Glen. H., Jr. 1995. The Life Course Paradigm: Social Change and Individual Development. In Phyllis Moen, Glen H. Elder, Jr., and Kurt Lscher (eds.), Examining Lives in Context: Perspectives on the Ecology of Human Development, pp. 101-139. (APA Science Volumes) Washington: APA Press. Ellison, Christopher G. 1994a. Religion and Health. In Encyclopedia of Sociology, edited by George Ritzer. London: Blackwell. Ellison, Christopher G. 1994b. Religion, the Life Stress Paradigm, and the Study of Depression. In Jeffrey S Levin (eds.) Religion in Aging and Health: Theoretical Foundations and Methodological Frontiers, pp. 78-121. Thousand Oaks, CA: Sage Publications.
134
Elman, Cheryl and ORand, Angela. 2004. The Race is to the Swift: Socioeconomic Origins, Adult Education and Economic Attainment. American Journal of Sociology 110: 123-160. Fengler, Alfred. P. 1984. Life Satisfaction of Subpopulations of Elderly: The Comparative Effects of Volunteerism, Employment, and Meal Site Participation. Research on Aging 6: 189-212. Ferree, G. D., Barry, J., and Manno, B. 1998. The National Survey of Philanthropy and Civic Renewal. National Commission on Philanthropy and Civic Renewal. Fischer, L. R., and Schaffer, K. B. 1993. Older Volunteers. Newbury Park, CA: Sage Publications, Inc. Frerichs Ralph. R., Aneshensel Carol. S., and Clark Virginia. A. 1981. Prevalence of Depression in Los Angeles County. American Journal of Epidemiology 113: 691-699. Garfein, Adam. J., and Herzog, A. R. 1995. Robust Aging among the Young-Old, OldOld, and Oldes-Old. Journal of Gerontology 50: S77-S87. George, Linda. K. 1990. Social Structure, Social Processes, and Social-Psychological States. In R. H. Binstock and L. K. George (Eds.), Handbook of aging and the social sciences (3rd ed., pp. 205226). San Diego, CA:Academic Press. George, Linda. K., and Gwyther, L. P. 1986. Caregiver Well-being: A Multidimensional Examination of Family Caregivers of Demented Adults. The Gerontologist 26: 253-259. George, Linda. K., Ellison, Christopher. G., and Larson, David. B. 2002. Explaining the Relationships Between Religious Involvement and Health. Psychological Inquiry 13:190-204. Glass, Thomas A., Seeman, Teresa E., Herzog, A. R., Kahn, Robert and Berkman, Lisa F. 1995. Change in Productive Activity in Late Adulthood: MacArthur Studies of Successful Aging. Journal of Gerontology: Social Sciences 50: S65-S76. Glynn, Laura. M., Christenfeld, Nicholas., and Gerin, William. 1999. Social support, Gender and Cardiovascular Responses to Stress. Psychosomatic Medicine 61: 234-242. Goldsmith, Elizabeth. B. and Goldsmith, Ronald. E. 1995. Full-time employees as caregivers to the elderly. Journal of Social Behaviour and Personality 10: 71930. Goode, William. J. 1960. A Theory of Role Strain. American Sociological Review 25:483-496.
135
Gore, Susan., and Mangione, Thomas. W. 1983. Social Roles, Sex Roles and Psychological Distress: Additive and Interactive Models of Sex Differences. Journal of Health and Social Behavior 24: 300-312. Gove, Walter. R., and Geerken, Michael, R. 1977. The Effect of Children and Employment on the Mental Health of Married Men and Women. Social Forces 56: 6676. Gove, Walter. R., Hughes, Michael and Briggs Style, Carolyn. 1983. Does Marriage Have Positive Effects on the Psychological Well-being of the Individual? Journal of Health and Social Behavior 24: 122-131. Gove, W. R., and Tudor, J. F. 1973. Adult Gender Roles and Mental Illness. The American Journal of Sociology 78: 812-835. Grasmick, H. G., Bursik, R. J. Jr., and Cochran, J. 1991. Render unto Caesar what is Caesars: Religiosity and taxpayers inclinations to cheat. The Sociological Quarterly 32: 251-266. Gubrium, J. 1972. Toward a Socio-environmental Theory of Aging. The Gerontologist 12: 281284. Gubrium, J. 1973. The Myth of the Golden years: A Socio-environmental Theory of Aging. Springfield, IL: Charles C Thomas. Guttentag, M., S. Salasin, and D. Belle. 1980. The Mental Health of Women. New York: Academic Press. Haglund K. 2000. Parenting a second time around: An ethnography of AfricanAmerican grandmothers parenting grandchildren due to parental cocaine abuse. Journal of Family Nursing 6: 120135. Hartunian, Nelson S., Charles N. Smart, and Mark S. Thompson. 1981. The Incidence and Economic Costs of Major Healtlz Impairments: A Comparative Analysis of Cancer, Motor Vehicle Injuries, Coronary Heart Disease, and Stroke. Lexington, MA: Lexington Books. Hatch, L.R. 1991. Informal support patterns of older African American and white women. Research on Aging 13: 144-170. Havighurst, Robert. 1957. The Social Competence of Middle-Aged People. Genetic Psychological Monographs, 56: 297-375. Havighurst, Robert. J. and Ruth Albrecht. 1953. Older people. New York: Longmans, Green and Co.
136
Havighurst, Robert., Bernice. L. Neugarten, and Sheldon S. Tobin. 1968. Disengagement and Patterns of Aging. In Bernice Neugarten (ed), Middle Age and Aging: A Reader in Social Psychology, pp. 161-172. Chicago: University of Chicago Press. Haw, Mary Ann. 1982. Women, Work, and Stress: A Review and Agenda for the Future.Journal of Health and Social Behavior 23: 132-144. Herd, Pamela., Goesling, Brian and House, James S. Socioeconomic Position and Health: The Differential Effects of Education Versus Income on the Onset versus Progression of Health Problems. Journal of Health and Social Behavior 48: 223-238. 2007. Herold, Joan and Ingrid Waldron. 1985. Part-Time Employment and Women's Health. Journal of Occupational Medicine 27: 405-412. Herzog, A. Regula., Franks, M. M., Markus, H. R., and Holmberg, D. 1998. Activities and Well Being in Older Age: Effects of Self-Concept and Educational Attainment. Psychology of Aging 13: 179-185. Herzog, A. Regula., and House, James. S. 1991. Productive Activities and Aging Well. Generations 15: 4954. Herzog, A. Regula., Kahn, Robert, J., Morgan, James, Jackson, James S and Antonucci, Toni C. 1989. Age Differences in Productive Activities. Journal of Gerontology 44: S129-S138. Hibbard, Judith H. and Pope, Clyde R. 1991. Effect of Domestic and Occupational Roles on Morbidity and Mortality. Social Science and Medicine 32: 805-811. Hinterlong, James. E., Morrow-Howell, Nancy., and Sherraden, Michael. 2001. Productive Aging: Principles and Perspectives. In Morrow-Howell, N, Hinterlong, J, E, and Sherraden, M (eds.) Productive Aging: Concepts and Controversies, pp. 4-17. Baltimore: Johns Hopkins University Press. Hinterlong, James. E., Morrow-Howell, Nancy., and Rozario, P. A. 2007. Productive Engagement and Late Life Physical and Mental Health: Findings from a Nationally Representative Panel Study. Research on Aging 29: 348-370. Hodgkinson, V. A., and Weitzman, M. S. 1996. Giving and volunteering in the United States: Findings from a national survey. Washington, DC: Independent Sector. House, James. S. 1989a. Americans Changing Lives, Waves I and II, 1986 and 1989 (ICPSR version). [Computer file]. Ann Arbor: University of Michigan, Survey Research Center [Producer], 1994; Ann Arbor, MI: Inter-University Consortium for Political and Social Research [Distributor], 1997.
137
House, James. S. 1989b. Americans Changing Lives Codebook, Waves I and II, 1986 and 1989 (ICPSR version). Ann Arbor: University of Michigan, Survey Research Center [Producer]; Ann Arbor, MI: Inter- University Consortium for Political and Social Research [Distributor]. House James. S., Landis, Karl. R., and Umberson, Debra. 1988. Social Relationships and Health. Science 241: 540-545. House, James. S., Leprowski, J. M., Kinney, A. M., Mero, R. P., Kessler, Ronald C. and Herzog, A. Regula. 1994. The Social Stratification of Aging and Health. Journal of Health and Social Behavior 35: 213-234. Hoyert, Donna. L., and Seltzer, Marsha. M. 1992. Factors Related to the Well-being and Life Activities of Family Caregivers. Family Relations 74-81. Hughes, Michael., and Gove, Walter. 1981. Living Alone, Social Integration and Mental Health. American Journal of Sociology 87: 48-74. Hultsch, D. F., Hammer, M., and Small, B. J. 1993. Age Difference in Cognitive Performance in Later Life: Relationships to Self-Reported Health and Activity Life Style. Journal of Gerontology 48: P111. Idler, Ellen. L., and Kasl, Stanislav. V. 1991. Health Perceptions and Survival: Do Global Evaluations of Health Status Really Predict Mortality? Journal of Gerontology 46: S55S65. Jackson, Pamela. 1997. Role Occupancy and Minority Mental Health. Journal of Health and Social Behavior 38: 237-255. Janevic, Mary. R., and Connell, Cathleen. M. 2001. Racial, Ethnic, and Cultural Differences in the Dementia Caregiving Experience: Recent Findings. The Gerontologist 41: 334-347. Jennings, Susan., Mazaik, Cheryl., and Mckinlay, Sonja. 1984. Women and Work: An Investigation of the Association between Health and Employment Status in Middle-Aged Women. Social Science and Medicine 19: 423-431. Johnson, David. R., Meile, Richard. L. 1981. Does Dimensionality Bias in Langner's 22item Index Affect the Validity of Social Status Comparisons? Journal of Health and Social. Behavior 22:415-33 Kandel, Denise. B., Davies, Mark and Raveis, Victoria H. 1985. The Stressfulness of Daily Social Roles for Women: Marital, Occupational and Household Roles. Journal of Health and Social Behavior 26: 64-78.
138
Kelley S.J. and Damato E.G. 1995. Grandparents as Primary Caregivers. MCN: The American Journal of Maternal/Child Nursing 20: 326332. Kelly Moore, Jessica. A., and Ferraro, Kenneth. F. 2004. The Black/White Disability Gap: Persistent Inequality in Later Life? Journal of Gerontology 59: S34S43. Kessler, Ronald. C. and Greenberg, David. 1981. Linear Panel Analysis: Models of Quantitative Change. New York: Academic Press. Kilty, K. M., and Behling, J. H. 1985. Predicting the Retirement Intentions and Attitudes of Professional Workers. The Journal of Gerontology 40: 219227. Kim, Jungmeen. E., and Moen, Phyllis. 2002. Retirement Transitions, Gender, and Psychological Well-being: A Life-Course, Ecological Model. The Journal of Gerontology 57:P212-P222. Kohn, Melvin and Schooler, Carmi. 1973. Occupational Experience and Psychological Functioning: An Assessment of Reciprocal Effects. American Sociological Review, 38: 97-118. Koropeckyj-Cox, Tanya. 1998. Loneliness and Depression in Middle and Old Age: Are the Childless More Vulnerable? Journal of Gerontology: Social Sciences 53B: S302 S312. Kosloski, Karl., Ekerdt, David., and De Viney, Stanley. 2001. The Role of Job-Related Rewards in Retirement Planning. Journal of Gerontology 56: P160-P169. Kotler, Pamela and Wingard, Deborah L. 1989. The Effect of Occupational, Marital and Parental Roles on Marital Quality: The Almeda County Study. American Journal of Public Health 79:607-613. Kramer, Betty. J. 1997. Gain in the Caregiving Experience: Where Are We? What Next? The Gerontologist 37: 218-232. Krause, Neal., and Shaw, Benjamin. A. 2000. Giving Support to Others, Socioeconomic Status, and Changes in Self-Esteem in Late Life. Journal of Gerontology 55: S323S333. Krause, Neal., Herzog, E. Regula., and Baker, Elizabeth. 1992. Providing Support to Others and Well-being in Later life. Journal of Gerontology: Psychological Sciences 47: P300P311. Lachman, Margie E. 1986. Personal Control in Later Life: Stability, Change, and Cognitive Correlates. In Margaret M. Baltes and Paul B. Baltes (ed), The Psychology of Control and Aging, pp. 207-236. Hillsdale, NJ: Erlbaum.
139
Lawton, M. P., Winter, L., Kleban, M. H., and Ruckdeschel, K. 1999. Affect and Quality of Life: Objective and Subjective. Journal of Aging and Health 11: 169198. Lemon, Bruce. W., Bengston, Vern. L., and Peterson, James. A. 1972. An Exploration of the Activity Theory of Aging: Activity Types and Life Satisfaction among In-movers to a RetirementCommunity. Journal of Gerontology, 27: 511-523. Lennon, Mary Clare. 1982. The Psychological Consequences of Menopause: The Importance of Timing of a Life Stage Event. Journal of Health and Social Behavior 23:35346. Lim, Jung-won and Zebrack, Brad. 2004. Caring for Family Members with Chronic Physical Illness: A Critical Review of Caregiver literature. Health and Quality of Life Outcomes 2: 1-9. Linn, M. W., Sandifer, R., and Stein, S. 1985. Effects of Unemployment on Mental and Physical Health. American Journal of Public Health 75: 502-506. Linville, Patricia. W. 1987. Self-complexity as a Cognitive Buffer Against Stress Related Illness and Depression. Journal of Personality and Social Psychology 52: 663 676. Litwin, Howard. 2000. Activity, Social Network, and Well-being: An Empirical Examination. Canadian Journal on Aging 19: 343362. Leucken, L. J., Suarez, E. C., Khun, C., Barefoot, S., Blumenthal, J. A., Siegler, I. C., and Williams, R. B. 1997. Stress in employed women: Impact of marital status and children at home on neurohormone output and homestrain. Psychosomatic Medicine 59: 352 359. Logan, John and Spitze, Glenna. 1994. Informal support and the use of formal services by older Americans. The Journals of Gerontology: Social Sciences 49: S25-S34. Luoh, Ming-Ching and A. Regula Herzog. 2002. Individual Consequences of Volunteer and Paid Work in Old Age: Health and Mortality. Journal of Health and Social Behavior 43: 490-509. Maddox, George L., and Elizabeth B. Douglass. 1973. Self-Assessment of Health: A Longitudinal Study of Elderly Subjects. Journal of Health and Social Behavior 14: 8793. Mackie, Marlene. 1983. The Domestication of Self: Gender Comparisons of SelfImagery and Self-Esteem. Social Psychology Quarterly 46: 343-50. MacLean, M. J. 1982. Differences Between Adjustment to and Enjoyment of Retirement. Canadian Journal on Aging 2:38.
140
Marcussen, Kristen. and Piatt, Liz. 2005. Race Differences in the Relationship between Role Experiences and Well-being. Health 9: 379-402. Markides, K. S., and Martin, H. W. 1979. A Causal Model of Life Satisfaction among the Elderly. Journal of Gerontology, 34: 86-93. Marks, N. F., and Lambert, J. D. 1999. Transitions to Caregiving, Gender, and Psychological Well-being: Prospective Evidence from the National Survey of Families and Households. NSFH Working Paper 82. Center for Demography and Ecology, University of Wisconsin, Madison, WI McMullin, J. A., and Marshall, Victor. W. 1996. Family, Friends, Stress and Wellbeing: Does Childlessness Make a Difference? Canadian Journal on Aging 15: 355 373. McPherson, B., and Guppy, N. 1979. Pre-Retirement Life-Style and the Degree of Planning for Retirement. Journal of Gerontology, 34: 254-263. Menaghan, Elizabeth G. 1989. Role Changes and Psychological Well-Being: Variations in Effects by Gender and Role Repertoire. Social Forces 67: 693-714. Menaghan, E. G. 1991. Work Experiences and Family Interaction Processes: The Long Reach of the Job? Annual Review of Sociology 17:419-444. Menec, V. H., and Chipperfield, J. G. 1997. The Interactive Effect of Perceived Control and Functional Status on Health and Mortality among Young-Old and Old-Old Adults. Journal of Gerontology, 52: 118-126. Merton, R. K. 1957. The Role-Set: Problems in Sociological Theory. The British Journal of Sociology, 8:106-120. Midlarsky, E., and Kahana, E. 1994. Altruism in later life. Thousand Oaks, CA: Sage. Minkler M., Fuller-Thomson E., Miller D. and Driver D. 1997. Depression in Grandparents Raising Grandchildren: Results of a National Longitudinal Study. Archives of Family Medicine 6: 445452. Mirowsky, J. and C.E. Ross. 1989. Social Causes of Psychological Distress. New York, NY: Aldine de Gruyter. Mirowsky, J., and Ross, C. E 1992. Age and Depression. Journal of Health and Social Behavior 33:187-205 Mirowsky, John. 2003. Social Causes of Psychological Distress. 2nd Edition. New York: Aldine de Gruyter
141
Moen, Phyllis., Donna Dempster-McClain, Robin Williams. 1989. Social integration and longevity: An event history analysis of women's roles and resilience. American Sociological Review 54: 635-47. Moen, Phyllis., Donna Dempster-McClain, Robin Williams. 1992. Successful Aging: A Life Course Perspective on Womens Multiple Roles and Health. American Journal of Sociology 97: 1612-1638. Morgan, K., Dallosso, H., Bassey, E. J., Ebrahim, S., Fentem, P. H., and Arie, T. H. D. 1991. Customary Physical Activity, Psychological Wellbeing, and Successful Aging. Ageing and Society 11: 399415. Morrow-Howell, Nancy., Jim Hinterlog, Philip A. Rozario, and Fengyan Tang. 2003. Effects of Volunteering on the Well-Being of Older Adults. Journal of Gerontology: Social Sciences 58B: S137-S145. Mossey J. M, Shapiro E. 1982. Self-rated health: a predictor of mortality among the elderly. American Journal of Public Health 72:800-808 Motenko A. K. 1989. The frustrations, gratifications, and well-being of dementia caregivers. The Gerontologist 29:166-172. Musick, M., Herzog, A. R., and House, J. S. 1999. Volunteering and Mortality among Older Adults: Findings From a National Sample. Journal of Gerontology: Social Sciences 54B: S173S180. Musick, Marc A. and John Wilson. 2003. Volunteering and Depression: The Role of Psychological and Social Resources in Different Age Groups. Social Science and Medicine 56: 259-269. Niyazi, F. 1996. Volunteering by Unemployed People: A Route to Opportunity and Volunteering by Disabled People: A Route to Opportunity. National Centre for Volunteering. Okun, M. A., Stock, W. A., Haring, M. J., and Witter, R. A. 1984. Health and Subjective Well being: A Meta-Analysis. International Journal of Aging, 19: 111-132. Oman, D., C. E. Thoresen, and K. McMahon. 1999. Volunteerism and Mortality Among the Community-Dwelling Elderly. Journal of Health Psychology 4: 301-16. Parsons, T. 1951. The Social System. Routledge, London Passannante, Marian R. and Constance A. Nathanson. 1985. Female Labor Force Participation and Female Mortality in Wisconsin, 1974-1978. Social Science and Medicine 21: 655-65.
142
Pearlin, L. I., and Johnson, J. S. 1977. Marital status, life-strains, and depression. American Sociological Review 42: 704715. Pearlin, L. I., Lieberman, M. A., Menaghan, E. G., and Mullan, J. T. 1981. The stress process. Journal of Health and Social Behavior 22: 337356. Pearlin L. I, Mullan J. T, Semple S. J, and Skaff M. M. 1990. Caregiving and The Stress Process: An Overview of Concepts and their Measures. Gerontologist 30: 583594. Pearlin, L. I., and Schooler, C. 1978. The structure of coping. Journal of Health and Social Behavior 24: 215. Poe L. M. 1992. Black Grandparents as Parents. Author, Berkeley,CA. Radloff, Leonne S. 1975. Sex Differences in Depression: The Effects of Occupation and Marital Status. Sex Roles 1 :249-65. Radloff, Leonne S. 1977. The CES-D scale: A self-report depression scale for research in the general population. Application of Psychological Measurement 1: 385-401. Rapp, S. R., and Chao, D. 2000. Appraisals of Strain and Gain: Effects on Psychological Well being of Caregivers of Dementia Patients. Aging and Mental Health 4: 142147. Rempel, J. 1985. Childless Elderly: What Are They Missing? Journal of Marriage and the Family, 47: 343-348. Reskin, Barbara A. and Shelley Coverman. 1985. Sex and Race in the Determinants of Psychological Distress: A Reappraisal of the Sex Role Hypothesis. Social Forces 63: 1038-59. Riessman, F. 1965. The helper therapy principle. Social Work 10: 27-32. Rochester, C. and R. Hutchison. 2002. Realising the Potential of Older Volunteers? An Evaluation of the Home Office Older Volunteers Initiative. London: Home Office Research Series. Roff, L. L., Burgio, L. D., Gitlin, L., Nichols, L., Chaplin, W., and Hardin, M. 2004. Positive Aspects of Alzheimers Caregiving: The Role of Race. The Journal of Gerontology: Psychological and Social Sciences 59: P185-P190. Rogers, R.L., Meyers, J.S., and Mortel, K.F. 1990. After reaching retirement age physical activity sustains cerebral perfusion and cognition. Journal American Geriatric Society 38: 123-128.
143
Romelsjo, Anders, George A. Kaplan, Richard D. Cohen, Peter Allebeck and Sven Andreasson. 1992. Protective Factors and Social Risk Factors for Hospitalization and Mortality among Young Men. American Journal of Epidemiology 135: 649-58. Ross, C. E., and C. E. Bird. 1994. Sex Stratification and Health Lifestyle: Consequences for Mens and Womens Perceived Health. Journal of Health and Social Behavior 35: 161-178. Ross, C. E., and Mirowsky, J. 1989. Explaining the Social Patterns of Depression Control and Problem-Solving Or Support and Talking. Journal of Health and Social Behavior 30: 206-219. Ross, C. E., J. Mirowsky and K. Goldsteen. 1990. The Impact of the Family on Health: The Decade in Review. Journal of Marriage and the Family 52:1059-1078. Ross, C. E., Mirowsky, J., and Huber, J. 1983. Dividing Work, Sharing Work, and InBetween: Marriage Patterns and Depression. American Sociological Review, 48: 809823. Ross, C. E. and M. Van Willigen. 1997. Education and the Subjective Quality of Life. Journal of Health and Social Behavior 38: 275-297. Rowe, J. W., and Kahn, R. L. 1997. Successful Aging. The Gerontologist, 37: 433-440. Rushing, B., Ritter, C. and Burton, R. P. 1992. Race Differences in Effects of Multiple Roles on Health: Longitudinal Evidence From a National Sample of Older Men. Journal Of Health and Social Behavior 33:126-39. Scharlach, Andrew. 1994. Caregiving and Employment: Competing or Complementary Roles. The Gerontologist 34: 378-385. Schulz, R., and Williamson, G. M. 1991. A Two-year Longitudinal Study of Depression among Alzheimer's Caregivers.Psychology and Aging 6: 569-578. Sieber, S. D. 1974. Toward a theory of role accumulation. American Sociological Review 39: 567578 Simon, Robin W. 1995. Gender, Multiple Roles, Role Meaning, and Mental Health. Journal of Health and Social Behavior 36: 182-194. Simon, Robin. 1997. The Meaning Individuals Attach to Role Identities and Their Implications for Mental Health. Journal of Health and Social Behavior 38: 256274. Skaff, M.M. and Pearlin, L.I. 1992. Caregiving: Role engulfment and the loss of self. The Gerontologist 32:656:664.
144
Stearns, Anthony A., Stearns, Harvey L. and Hollis, Lisa A. 1996. The Productivity and Functional Limitations of Older Adult Workers. In William Crown (eds.) Handbook on Employment and the Elderly. Westport, CT: Greenwood Press. Stoller, E and Pugliesi, K. 1989. Other Roles of Caregivers: Competing Responsibilities or Supportive Resources. Journal of Gerontology 44: S231-S238. Strawbridge W. J., Wallhagen M. I., Shema S. J. and Kaplan G. A. 1997. New burdens or more of the same? Comparing grandparent, spouse, and adult-child caregivers. The Gerontologist 37: 505 510. Stryker, Sheldon, and Anne Statham. 1985. Symbolic Interaction and Role Theory. In Gardner Lindzey and Elliot Aronson (3rd Ed), Handbook of Social Psychology, pp. 311378. NY: Random House. Svanborg, Alvar. 2001. Biomedical Perspectives on Productive Aging. In N. Morrow, J. Hinterlong, and M. Sherraden (eds.), Productive Aging: Concepts and Challenges, pp. 80-101. Baltimore: Johns Hopkins University Press. Thoits, Peggy. A. 1983. Multiple Identities and Psychological Well-being: A Reformulation and Test of the Social Isolation Hypothesis. American Sociological Review 48: 174187. Thoits, Peggy A. 1986. Multiple Identities: Examining Gender and Marital Status Differences in Distress. American Sociological Review 51: 259-272. Thoits, Peggy A. 1987. Negotiating Roles. In Faye J. Crosby (ed.), Spouse, Parent, Worker: On Gender and Multiple Roles, pp. 11-22. New Haven, CT: Yale University Press. Thoits, Peggy A. 1992. Identity Structures and Psychological Well-Being: Gender and Marital Status Comparisons. Social Psychology Quarterly 55: 236-256. Thoits, Peggy A. and L. Hewitt. 2001. Volunteer Work and Well-being. Journal of Health and Social Behavior 42: 115-131. Thompson, E. H, Jr. 2000. Gendered Caregiving of Husbands and Sons. In Elizabeth W. Markson and Lisa Ann Hollis-Sawyer, (eds) Intersections of Aging: Readings in Social Gerontology, pp. 333-344. Los Angeles: Roxbury Publishing Company. Thompson, M. K., and Brown, J. S. 1980. Feminine Roles and Variations in Womens Illness Behaviors. Pacific Sociological Review 23:405-422. Turner, R. J., and Noh, S. 1983. Class and Psychological Vulnerability among Women: The Significance of Social Support and Personal Control. Journal of Health and Social Behavior, 24: 2-15.
145
Vaillant, G.eorge. 2002. Aging Well. Little, Brown and Company. Van Willigen, Marieke. 2000. Differential Benefits of Volunteering Across the Life Course. Journal of Gerontology: Social Sciences 55B: S308-S318. Verbrugge, Lois. 1983. Multiple Roles and Physical Health of Men and Women. Journal of Health and Social Behavior 24:16-30. Verbugge, Lois M. 1986. Role Burdens and Physical Health of Women and Men. Women and Health 11: 47-77. Verbrugge, Lois. 1987. Role Responsibilities, Role Burdens, and Physical Health. In Crosby, F. J. (eds.) Spouse, Parent, Worker: On Gender and Multiple Roles, pp. 154-166. New Haven, CT: Yale University Press. Vitaliano, P. P., Zhang, J., and Scanlan, J. M. 2003. Is Caregiving Hazardous to Ones Physical Health? A Meta-Analysis. Psychological Bulletin, 129: 946-972. Waldron, Ingrid. 1983. Sex Differences in Illness Incidence, Prognosis and Mortality: Issues and Evidence. Social Science and Medicine 17: 107-23. Waldron, Ingrid and Jacobs, Jerry A. 1989. Effects of Multiple Roles on Womens Health: Evidence from a National Longitudinal Study. Women and Health 15: 3-19. Welch, Susan and Booth, Alan. 1977. Employment and Health among Married Women with Children. Sex Roles 3:385-397. Wheaton, B. 1983. Stress, Personal Coping Resources, and Psychiatric Symptoms: An Investigation of Interactive Models. Journal of Health and Social Behavior, 24: 208229. White, T. M., Townsend, A. L., and Stephens, M. A. 2000. Comparisons of African American and White Women in the Parent Care Role. The Gerontologist, 40: 718-728. Williams, S., Dilworth-Anderson, P., Goodwin, P. 2003. Caregiver role strain: the contribution of multiple roles and available resources in African-American women. Aging and Mental Health 7: 103-112. Wilson, W. J. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. University of Chicago Press. Wilson, John and Marc Musick. 1999. The Effects of Volunteering on the Volunteer. Law and Contemporary Problems 62: 141-168. Yee, J. L., and Schulz, R. 2000. Gender Differences in Psychiatric Morbidity among Family Caregivers: A Review and Analysis. The Gerontologist, 40: 147-164.
146
Zimmer, Z., Hickey, T., and Searle, M. S. 1995. Activity Participation and Well-being among Older People with Arthritis. The Gerontologist, 35: 463-471.
147
BIOGRAPHICAL SKETCH Manacy Pai EDUCATION Expected May 2008 Doctor of Philosophy, Sociology Florida State University, Tallahassee, FL Dissertation: Later Life Social Engagement and Health. Committee: Professor Anne Barrett (chair), Jill Quadagno, Robin Simon, Mary Gerend Qualifying Exam: Aging and the Life Course Minor Area: Inequality and Social Justice January 2008 June 2005 Paid Work Histories and Mental Health in Widowhood. Committee: Professor Anne Barrett (chair), Jill Quadagno, Robin Simon October 2002 Bachelor of Arts, Anthropology and Sociology (major) Knox College, Galesburg, IL Senior Research Project: Public Housing and The Elderly: The People of Mary Allen West. Research Project: Charles E. Broughton Certification in Aging Studies, Sociology Florida State University, Tallahassee, FL Master of Science, Sociology Florida State University, Tallahassee, FL Masters Paper: Long-term Payoffs of Work? Womens
PUBLICATIONS AND PRESENTATIONS Publications: Pai, Manacy and Anne E. Barrett. 2007. Long-term Payoffs of Work? Womens Paid Work Histories and Mental Health in Widowhood. Research on Aging, 29(5): 436-456. Barrett, Anne. E. and Manacy Pai. Sketches in Cyberspace: Using Student Drawings in an Online Sociology of Aging Course. Gerontology and Geriatrics Education (In Press). Gerend, Mary and Manacy Pai. Social Determinants of Black-White Disparities in Breast Cancer Mortality. Cancer Epidemiology Biomarkers & Prevention (Forthcoming). Papers under Review: Pai, Manacy and Deborah Carr. Do Personality Traits Moderate the Effects of Late-Life
148
Spousal Loss on Psychological Adjustment? (Revise and Resubmit at Journal of Health and Social Behavior). Papers in Progress: Kim, Joongbaeck and Pai, Manacy. Volunteering and Trajectories of Depression in Different Age Groups. Pai, Manacy. Informal Helping and Mental Health: Variations by Socioeconomic Status. Pai, Manacy. Later Life Civic Engagement and Perceptions of Economic Well-being. Pai, Manacy. Caregiving and Volunteering: Mental and Physical Health Effects of Combining these Two Activities in Later Life. HONORS AND AWARDS 2008 Claude and Mildred Pepper Dissertation Fellowship, Florida State University. ($11,000). 2007-2008 Outstanding Teaching Assistant Award Nominee, Florida State University. 2007 Graduate Student Mentorship Nominee, Department of Sociology, Florida State University. 2006-2007 Allen-Klar Best Graduate Student Research Paper Award (Long-term Payoffs of Work? Womens Paid Work Histories and Mental Health in Widowhood). RESEARCH AND TEACHING INTERESTS Aging and the Life Course, Medical Sociology, Sociology of Families, Social Structure and Personality PROFESSIONAL AND DEPARTMENTAL SERVICE Reviewer for The Sociological Quarterly and Journal of Nervous and Mental Disease. Student representative, Honors and Awards Committee, Department of Sociology, Florida State University, 2006-2007. Elected student representative, Admissions and Financial Aid Committee, Department of Sociology, Florida State University, 2006-2007.
149
Elected student representative, Departmental Policy Committee, Department of Sociology, Florida State University, 2005-2006. PROFESSIONAL MEMBERSHIPS American Sociological Association American Society on Aging Gerontological Society of America Society for the Study of Social Sciences
150