A Life-Span Developmental Perspective on Social Status and Health
- 1Department of Sociology and Population Research Institute
- 2Department of Biobehavioral Health and Gerontology Center, The Pennsylvania State University, University Park.
- Address correspondence to Duane F. Alwin, PhD, Director, Center on Population Health and Aging, The Pennsylvania State University, 326-A Pond Laboratory, University Park, PA 16802. E-mail: dalwin{at}pop.psu.edu
Abstract
This article presents a life-span developmental (LSD) perspective on the linkages between social status and health. The objective is to develop a conceptual framework that is useful in understanding why people are differentially exposed to risks of disease or protective factors and the social conditions that link the effects of risk and protective factors to the social environment over the life span. The discussion distinguishes between the complementary concepts of “life span,” “life cycle,” and “life course,” critical theoretical distinctions that may help refine hypotheses about the relationship between health and social status. We argue that life-cycle and life-course concepts can be viewed as embedded in a more general LSD perspective. Using the theoretical principles derived from this perspective, the review examines (a) gender differences, (b) race–ethnic experiences, (c) childhood experiences, (d) educational levels, (e) socioeconomic differences, and (f) age differences. The emphasis in the review is to highlight the value of a broader LSD perspective in the study of health inequalities. The article ends with a brief summary of where future research is headed and novel developments in the study of social status and health.
Social inequalities in power, privilege, certification, and economic well-being have always been part of society, as we know it, and these inequalities are as prevalent in today's society as they ever were. Sociologists and epidemiologists have long known about the consequences of social inequality for the health and well-being of individuals, although for some reason this issue has until recently been ignored by the medical community. Fortunately, this connection was rediscovered by researchers in the early 1990s, and as a consequence, it is now a prominent focus of requests for research by the National Institutes of Health. One of the important features of the scholarly terrain in this area is that it is multidisciplinary, with researchers using many different theoretical perspectives to contribute to the accumulation of knowledge. Even the popular press has discovered that social status is related to health (Cohen, 2004).
As David Mechanic (2000) pointed out in a recent review essay, the connection between social inequality and health has been known for centuries. In his words, “by the mid-19th century there were already careful, detailed inquiries in England, France, Germany, and the United States on how the conditions of the poor cut life short” (Mechanic, 2000, p. 269). Some of these inquiries included Edwin Chadwick's (1842) Report on the Sanitary Condition of the Labouring Population of Great Britain and John Griscom's (1845) The Sanitary Condition of the Laboring Population of New York: With Suggestions for Its Improvement. In the early 1960s, there was a resurgence of interest in the topic when Michael Harrington's book The Other America (1962) pointed out that mass poverty continued to exist in America despite the unprecedented prosperity of the nation and the social welfare legislation of Roosevelt's New Deal. Whether among the racial and ethnic minorities in the urban ghetto or the populations of rural America, the renewed awareness of pockets of poverty led to the widely acclaimed policy of the Johnson Administration—The War on Poverty—and a renewed focus on social inequalities in health and illness. During this era, Saxon Graham (1958) documented the greater prevalence of certain chronic conditions among the socially disadvantaged, and the classic demographic analysis of mortality by Kitagawa and Hauser (1964) established the importance of race and socioeconomic factors in health and well-being. A few years later, John Kosa, Aaron Antonovsky, and Irving Kenneth Zola's (1969) edited volume Poverty and Health: A Sociological Analysis included an exhaustive investigation of social inequalities in health and illness, concluding that “Whatever aspect of health, whatever stage of the morbid episode is examined, the [less socioeconomically advantaged] are at a [greater] disadvantage” (1969, p. 325). In other words, the basic idea that social inequalities are related to health and disease has been around for a long time.
With these origins in the 1960s, studies of social status and health grew exponentially beginning in the 1990s, driven partly by the increasing numbers of disciplines that started to focus on the issue and partly by federal mandates that called eventually for research aimed at eliminating health disparities, e.g., Healthy People 2010. In an important review article, Oakes and Rossi (2002) plotted the trends in the number of articles on the relationship between socioeconomic status (SES) and health in health journals, documenting that between 1990 and 1999, the mean number of published articles with SES keywords in title fields was 175.6 per year or a cumulative total of 3,544 articles!
While decades of studies provide evidence of the relationship between social conditions and health, the “reinvigoration” of the focus on this relationship has brought both renewed awareness of “health disparities” and an increased understanding of some of their origins. Recent studies have focused on attempts to understand the social processes underlying the observed differences and the mechanisms that mediate these processes (e.g., Williams, 1990; Link & Phelan, 1995). Recent research has contributed important conceptual, methodological, and empirical strategies that further illuminate these relationships. New conceptualizations have stressed the multidimensional nature of social status, the importance of a life-span perspective on inequality—including the use of life-course concepts—and behavioral mechanisms of transmission. The recent literature has also emphasized the increasing importance and availability of longitudinal data to address issues of stability and change in the analysis of the effects of social inequalities on health, and new analytic techniques (such as event history analysis, latent growth models, and multilevel contextual models) have enhanced the ability of researchers to articulate the nature of the role of social status in health and disease.
A Theoretical Perspective
This section briefly reviews a life-span developmental (LSD) perspective on the relationship between social status and health. This is a framework that can be used to understand why people are differentially exposed to risks of disease and/or protective factors as well as the social conditions that link the effects of risk and protective factors to the social environment (Link & Phelan, 1995). This perspective takes the long view by emphasizing the need to understand the role of inequalities in health and disease across the entire life span—from birth to death—and includes the consideration of life-course factors (or social pathways) in understanding linkages between social status and health. This perspective, elaborated on below, sensitizes the researcher to the existence of multiple life-cycle periods of relevance (e.g., childhood, adolescence, midlife, and old age) and issues of the timing of influences of social inequalities over the life span. The emphasis on the life course (as distinct from life-cycle and life-span concepts) draws attention to the importance of understanding the potential consequences of role transitions, trajectories, and sequences of roles for health. This perspective also emphasizes the historical embeddedness of social experience and the potential role of cohort and period factors in disease incidence and prevalence. In the following discussion we define our key concepts and introduce several theoretical principles that will help guide our brief reconnaissance of the research literature.
What Is Social Status?
When the term “social status” (or even the more restrictive term SES) is used, it is often treated as if it refers to one thing. It does not! Social status is multidimensional and includes both ascribed statuses (statuses attached to the individual at birth based on biological or group attributes) and achieved statuses (statuses that are gained through access to opportunity and the exercise of individual volition). The most commonly studied ascribed statuses include gender, race–ethnicity, and age; research on achieved statuses typically focus on SES, defined as “income, education and occupational status” (Adler, Boyce, Chesney, Cohen, Folkman, Kahn, & Syme, 1994). However, income, education, and occupational status are not necessarily alternative indicators of the same latent concept of SES, but refer to different institutional structures and are, in fact, causally ordered (Blau & Duncan, 1967). Education is arguably causally prior to occupation, that is, level of schooling is a criterion for access to jobs, and occupational status is linked causally to wage rates and earnings. Moreover, as studies of social mobility have shown, although there may be linkage between parental statuses and one's own, they are not necessarily the same. Thus, an LSD perspective would argue that to focus only on adult achieved statuses without any attention to family background ignores what may be an even more important source of social inequality that has consequences for health in adulthood (e.g., see Elo & Preston, 1992; Preston, Hill, & Drevenstedt, 1998; Wadsworth, 1997).
More importantly, it should be recognized that global concepts such as “socioeconomic status” are in themselves very limiting, because they are not specific about what aspect of social status is important for a particular outcome, and they exclude consideration of other statuses, particularly ascribed statuses. Without a complete specification of the inequalities experienced across the entire life span, it is difficult to infer the direct impacts of status characteristics. Recent developments have emphasized the value of specifying more concretely the meaning of social inequality and the relationship of components of various forms of “capital” to health and well-being. For example, O'Rand (2001, pp. 200–202) distinguishes between “human capital” (the stock of productive knowledge and skills), “social capital” (the stock of direct and indirect social relationships, i.e., “strong and weak ties”), “personal capital” (resiliency, positive affect, self-confidence, and control), and “psychophysical capital” (physical and mental health). Similarly, Oakes and Rossi (2002) emphasize the need to develop direct measures of what is meant by these various forms of social inequalities, including “material capital” (owned resources such as houses, cars; earned resources, such as income, savings; and expected wealth, e.g., inheritances), “human capital” (fixed endowments of an individual, such as the schooling, occupational skills, cognitive functioning, psychological motivation, specialized knowledge), “social capital” (resources associated with the individual, family, or community's role in the social system, including social networks, group memberships, obligations, information channels, and norms), and “cultural capital” (values, beliefs, and orientations of an individual, family, or community, such as those linked to health behaviors).
Life-Span Principles
We argue that it is useful to view the relationship between social inequalities and health from an LSD perspective. Human development and aging are lifelong processes, and a complete understanding of how processes of social stratification affect health requires a long-term perspective. We distinguish between the complementary concepts of “life span,” “life cycle,” and “life course” in discussing the effects of social inequalities and health. The importance of this distinction will become evident when we discuss the relevance of several theoretical “life-span principles” to research on social status and health in that they will help refine hypotheses about the relationship between social status and health. (Note that we refer to these as theoretical principles, rather than hypotheses. As theoretical principles, they orient us to the reality we study and provide a framework within which hypotheses can be formulated and tested.) It can be shown that the life-cycle and life-course concepts can be thought of as embedded in a more general LSD perspective. The life span—the length of life for an individual organism (see Olshansky, Carnes, & Brody, 2002)—draws attention to the temporal scope of inquiry, and an LSD perspective focuses therefore on the effects of social status differences on health across the entire life span. The LSD perspective conceptualizes human development as multidimensional and multidirectional processes of growth involving both gains and losses across the entire life span (e.g., Baltes, 1987; Baltes, Staudinger, & Lindenberger, 1999). This perspective is in no way based on purely ontological notions of development rooted in biology: We conceive development as embedded in multiple social contexts. In this sense, development is conceived as a dynamic process in which the ontogeny of development interacts with the social environment, a set of interconnected social settings, embedded in a multilayered social and cultural context (Bronfenbrenner, 1979; Dannefer, 1984).
There is no inherent incompatibility between the LSD perspective and the more recent introduction of notions of the life course. The life course refers to the social pathways defined by events and transitions experienced by individuals and the sequences of roles and experiences followed by individuals over particular phases of their lives (e.g., Elder, 1995; Elder & Johnson, 2003; Elder, Johnson, & Crosnoe, 2003; Elder & O'Rand, 1996; Ferraro, 2001). In addition, we distinguish between these two concepts and a third—life cycle. Historically, “life cycle” refers to “maturational and generational processes driven by mechanisms of reproduction in natural populations” (O'Rand & Krecker, 1990, p. 242). It refers to a fixed sequence of irreversible stages, tied specifically to sexual reproduction. The irreducible properties of the life cycle, therefore, were successive forms (stages), irreversible development (maturation), and the reproduction of form (generation). These elements of the life cycle define the bases of time and variation over the life span—they provide a context for the study of life course. In other words, the concept of “life course” supplements rather than replaces the concepts of “life cycle” and “life span,” and it is, thus, important to appreciate the distinctiveness of all three of these concepts and to avoid using them synonymously (see Alwin, 2005).
We can distinguish among these concepts with reference to Figure 1, which is an extension of Riley's (1973) heuristic scheme showing the intersection of biographic and historical time. The distinction between biographic and historical time is straightforward and we need not spend time on this, except to note that variations occurring within biographic time—life-cycle and life-course factors—can interact in important ways with those occurring in historic time (see Alwin, 1995). Variations in experience that occur in biographic time are often confused in part because they are confounded in time. We prefer to think of the biologically driven stages or phases of the life span in much broader time metric and life-course transitions as more narrowly construed within the social constructions derived from them. In other words, variations in the life course, i.e., in social pathways, occur within a much narrower definition of time, e.g., the transition to adulthood (see Rindfuss, 1991).
The first principle of a life-span perspective on social status and health is that social status can affect health at any point from birth (or even before) until death (Settersten, 2003). This principle is seldom evident in research on social status and health; instead, many researchers assume that social status is a constant feature of people's lives and that the circumstances of measurement (often occurring only once) reflect one's position over a lifetime (Hayward & Gorman, 2004, p. 88). By contrast, this principle asserts that social status factors have an impact on health at multiple points across the life span, and that the effects of social inequalities may accumulate over time (Dannefer, 1987, 1988; O'Rand, 1996).
The second principle of the life-span perspective, a corollary to the first, refers to the potential importance of the timing of the influences of social inequalities. If there are multiple life periods of relevance, which are the most important? Are there critical periods during childhood, adolescence, or adulthood in which social inequalities have their greatest impact on health? Early adversity may be overcome by later improvements in social circumstances, but may nonetheless leave the individual more vulnerable to the consequences of health risks in later life. Adolescent lifestyle patterns (e.g., risky health behaviors such as alcohol and tobacco use, or risky sexual behavior) linked to social status may have consequences for health over the remainder of the life span (Shanahan, 2000). Similarly, patterns of behavior in adulthood and old age linked to social status (e.g., nutrition and physical activity) may have direct consequences for disease prevention and/or mortality. Generally, the literature has developed in such a way that the question of which period of the life span is the most important for the impact of social inequalities is unresolved, but we assume that there are critical periods for the impact of social status on health. One implication of this theoretical principle is that if social status creates exposure to risk factors early in life, then research should look for the effects on health of early-life social inequalities rather than achieved social statuses in adult life.
A third principle of an LSD perspective involves the concept of life course, which states that the structure, sequence, and dynamics of events, transitions, and trajectories (social pathways) that take place within life-stage phases over the life span have consequences on health. The concept of life course is often misunderstood, in that it is often used as a synonym for “life cycle” or “life span,” but it refers to something quite different. Specifically, within demography, the concept of life course refers to trajectories of role transitions, and the social pathways followed by individuals over particular phases of life. Each phase or stage of the life span has a set of potential life-course patterns. Role transitions during the “transition to adulthood” (such as entering and leaving school, acquiring a full-time job, and marrying for the first time, for example) or in old age (the transition into retirement or into a long term care facility) are examples of the essence of the life-course concept. Each life-course transition is embedded in a life-cycle stage-specific trajectory that gives it specific form and meaning. Thus, during adulthood, work transitions are core elements of a work–life trajectory, and births are key markers along a parental trajectory. Multiple marriages and divorces are elements of a marital trajectory (see Elder et al., 2003). In terms of health outcomes, an interest in the life course hypothesizes how social pathways within specific life-cycle stages (such as adolescence) shape an individual's exposure to risk factors that promote or prevent healthy development.
Finally, one of the insights gained from the literature on the life-span perspective involves a fourth life-span principle—the historical embeddedness of social experience and the possibility of cohort differences in behavior—which is illustrated in Figure 1. Specifically, in studying the relationship between social status and health (cohort effects), because of the natural confounding of age and cohort in cross-sectional studies of health, it is often impossible to rule out the influence of cohort factors in looking at the relationship of social statuses and health. A sensitivity to the existence of cohort differences will also bring greater depth to the understanding of the potential for relationships to change with time (Alwin, 1995; Ryder, 1965).
Social Status and Health: A Review
In this section, we briefly review what is known about the linkage of several social statuses to health. Because of limitations of space, the review cannot be exhaustive and is by necessity incomplete. Rather, the emphasis in the review is to highlight the value of an LSD perspective and related theoretical principles for an understanding of the development of health inequalities.
Gender
Gender is embodied by a socially constructed set of expectations that differentiate the roles and attributes of the sexes and often accounts in part for a sexual division of labor. The gendered nature of life-course trajectories clearly structures and constrains the health-related resources and opportunities of men and women (e.g., Moen, 2001; Rossi, 1985). Those factors may work additively or they may interact with one another to produce the gender differences in health. Decades of research studies have provided evidence of differences between men and women in the prevalence of many diseases, disability, and mortality. Although speculation continues, there is currently little or no evidence that sex-linked biological (i.e., life cycle) differences underlie the greater mortality rates of men for almost all leading causes of death, the greater disability rates of women, or the varying risks for particular diseases. Instead, other factors in which men and women differ (e.g., social pathways or role-related factors) appear to be at work—participation in risky health behaviors (e.g., smoking, drinking, and substance abuse), personality characteristics (e.g., hostility, anger), levels of and responses to stressful situations, social support, and work stress. There is a considerable body of educational, sociological, and social–psychological literature on gender that has documented systemic gender inequalities that shape the life course at specific stages in ways that have consequences for health (e.g., Hochschild, 1989; Kanter, 1977; Sadker & Sadker, 1994).
Race–Ethnicity
A wide body of research studies also shows differences in disease, disability, and mortality across major race–ethnicity groups such that, on average, African American and Latino American adults evidence poorer health than do European Americans or Asian Americans, although some of these results remain controversial (National Research Council, 2003). Because wide differences are also found by achieved social status, many studies have investigated the roles of childhood circumstances, education, income, and wealth in those differences. Although achieved SES factors do contribute to differing health status, they have not been found to explain the health differences by race–ethnicity. If SES factors do not explain these differences, what does? In addition to potential physiological differences by virtue of ancestral geography, other factors that may underlie those differences include background status factors such as parental education, birth weight, other childhood health and economic circumstances; differential educational quality; effects of accumulated daily hassles and/or discrimination; lower income and wealth; and interactions of all of these factors (Hayward, Crimmons, Miles, & Yu, 2000; National Research Council, 2003; Whitfield, Weidner, Clark, & Anderson, 2002; Williams & Wilson, 2001). Ultimately, race and ethnic differences in health probably result from patterns of institutional racial and ethnic discrimination that produce differential social pathways contributing to different health outcomes.
Educational Levels
Perhaps the most common indicator in research on “social status and health” is the individual's level of schooling. The salutary relationship between levels of schooling (or education) and health is one of the strongest and best documented linkages in this area (e.g., Adler, Boyce, Chesney, Folkman, & Syme, 1993; Adler et al., 1994; Kuh & Davey-Smith, 1997; Ross & Mirowsky, 1999; Ross & Wu, 1995, 1996). Summarizing their research on educational levels, Mirowsky and Ross (2003) argue that amount of schooling creates most of the relationship between achieved social statuses and health, and that it is primarily through the increased sense of personal control that results from higher levels of schooling that greater health outcomes are achieved. Yet schooling is not independent of social background. Education is highly dependent on parental SES, with some 50 percent of the variance in education being attributable to social background variables (see Alwin, McCammon, & Wray, 2004).
Socioeconomic Differences
Dozens of studies also show strong relationships between income, wealth, and occupational status and health outcomes: The higher the SES, the better the health (e.g., Ettner, 1996). Although some researchers (e.g., see Smith, 1998, 1999) have argued that the direction may be reversed (e.g., health produces SES), most research concludes that SES (causally subsequent to schooling) enhances health outcomes through safer and less stressful workplaces, greater access to information and health insurance, and social networks. Wealth may be more important than levels of schooling in explaining differences for African Americans (Smith & Kington, 1997). Regardless of the possible feedback effects of poor health on SES, there is substantial evidence that indicates that both micro- and macro-level socioeconomic inequalities have consequences for health (e.g., Adler, Marmot, McEwen, & Stewart, 1999; Kawachi, Kennedy, & Wilkinson, 1999; Lantz, Lynch, House, Lepkowski, Mero, Musick, & Williams, 2001; Marmot & Wilkinson, 1999; Preston & Taubman, 1994; Rogers, 1995).
Childhood Experiences and Family Background
Studies of social status and health have focused primarily on adult life circumstances, especially SES, and indicators of health or wellness. It is often assumed either (a) that childhood family social conditions do not affect adult health, or (b) that if they do, their effects are entirely transmitted by current social statuses (see Hayward & Gorman, 2004). Yet it is known that unequal investment in children's lives from before birth is likely to have consequences for health in adulthood (Conley & Bennett, 2000; Preston & Haines, 1991), and there is growing evidence that childhood adversity linked to social inequalities in childhood has important consequences for health later in life (see Barker, 1998; Blackwell, Hayward, & Crimmins, 2001; Brunner, Davey-Smith, Marmot, Canner, Beksinska, & O'Brien, 1996; Costa, 1999; Elo & Preston, 1992; Gunnell, Frankel, Nanchahal, Braddon, & Davey-Smith, 1996; Hayward & Gorman, 2004; Kuh & Ben-Shlomo, 1997; Kuh, Power, Blane, & Bartley, 1997; Kuh & Wadsworth, 1993; Lundberg, 1993, 1997; Lynch, Kaplan, & Salonen, 1997; Power & Peckham, 1990; Preston et al., 1998; Schwartz, Friedman, Tucker, Tomlinson-Keasey, Wingard, & Criqui, 1995; Wadsworth, 1997; Winkelby, Jatulis, Frank, & Fortmann, 1992).
Recent evidence suggests that family SES background and pre-adult intellectual resources have significant effects on health outcomes at midlife, but that they may be entirely mediated by adult achievements (Alwin, Wray, & McCammon, 2003). Although adult achievements—education, occupation, and economic resources—all have significant, independent effects on health outcomes at midlife, these proximate SES factors mediate the effects of family SES background and intellective factors. For purposes of prediction, prior SES and intellective factors can be ignored, although the independent contributions of some proximate SES factors (e.g., education) are overestimated by doing so. These results underscore the importance of assessing total effects rather than direct effects in assessing the influences of SES factors across the life span.
Older Age
The literature is quite consistent in the finding that the effects of inequalities on health persist across the life span. Indeed, from the point of view of the cumulative advantage/disadvantage hypothesis, the relationship between many types of social status and health increases in strength (Dannefer, 1987, 1988; O'Rand, 1996). Despite this persistence and growing strength of the relationship between indicators of social status and health with increasing age, the level of the association is generally found to decline precipitously in older age (e.g., House, Lepkowski, Kinney, Mero, Kessler, & Herzog, 1994; Marmot & Shipley, 1996; Robert & House, 1996). Rather than being evidence against the hypothesis of the effects of social status on health, the leveling of its effect in older age is probably just the opposite—namely, evidence of increasing social inequalities in health that systematically reduce estimates of statistical relationship in the population (see Dannefer, 2003). In addition, another common explanation of this finding is mortality selection, in that at older ages the surviving population is substantially more homogenous in many measures of social status, although there is some evidence that contradicts this interpretation (Beckett, 2000). However, research on aging, social status, and health has yet to provide a definitive explanation of the seemingly contradictory findings of increasing inequality in both status and health and their decreasing levels of association.
Future Trends And Research Needs
In conclusion, we would like to briefly suggest the following areas for future work. First, there is a continued need for theoretical development. Specifically, there is a need to develop a broader theoretical linkage between social conditions and health in understanding (a) the institutional- and societal-level opportunities and constraints that provide barriers to health care and health promotion (e.g. access to jobs, health insurance, and medical care) and (b) the contextual factors that help mediate the effects of social factors and health (e.g., the role of social isolation, social support, and intergenerational social capital) (see Hagestad & Dannefer, 2001). Existing studies indicate that the more traditional health-related risk factors (e.g. smoking, drinking, and exercise) account for little of the health and mortality inequalities (Haan, Kaplan, & Camacho, 1987; House, Kessler, Herzog, Mero, Kinney, & Breslow, 1992; Lantz, House, Lepkowski, Williams, Mero, & Chen, 1998; Winkelby, Fortmann, & Barrett, 1990; Wray & Alwin, this issue). Instead, other risk factors—including psychosocial factors such as social support, chronic and acute stressors, and self-efficacy—may account for more of the social differentials in health (House et al., 1994). Further, explorations of the concept of social capital, including participation in religious services (e.g., Idler, 1994; Idler, Musick, Ellison, George, Krause, Ory, Pargament, Powell, Underwood & Williams, 2003; Macinko & Starfield, 2001; Veenstra, 2000) and indicators of human agency (Mirowsky & Ross, 2003; Williams, 1990), are important avenues for future research.
Second, there is a need for improvement in research design and methods of analysis, including the continued development of longitudinal research designs and, within this framework, the development of innovative research strategies that obtain more valid assessments of the processes involved in the development of health disparities. This includes (a) the improved measurement of health, specifically measurements relying less on self-reports and more on clinical assessments in the framework of longitudinal designs; (b) the development of a multidimensional approach to the measurement of health that includes all aspects of health and functioning, including mental health and cognitive functioning; and (c) the application of better analysis tools that link exogenous social factors to trajectories of growth/decline in health outcomes, e.g., latent growth curve approaches (see Alwin, Hofer, & McCammon, in press).
Finally, there is a need for greater attention to intervention strategies that provide a translational component to basic research and stimulate the use of research findings by health professionals and medical care organizations. Even if they do not reduce inequalities in status, such strategies may nonetheless reduce their impact on individual differences in health outcomes.
Acknowledgments
This research was supported by a grant from the National Institute on Aging (R01 AG15437-04) “Socioeconomic Status, Resources and Health” to Duane F. Alwin, PI, and Linda A. Wray, Co-PI.
This article was presented at the conference “Health Inequalities Across the Life Course” held June 5–7, 2004 at The Pennsylvania State University, University Park.
We gratefully acknowledge the assistance of Pauline Mitchell and Ryan McCammon.
- The Gerontological Society of America







