The effect of social relations with children on the education–health link in men and women aged 40 and over

https://doi.org/10.1016/S0277-9536(02)00099-0Get rights and content

Abstract

Accumulated evidence demonstrates a strong relationship between socioeconomic status (SES) and health. Our examination of this relationship focuses on education, an established indicator of SES, and tests whether social relations, particularly with children, mediate and/or moderate the education–health link for middle-aged and older parents. The data are drawn from a regionally representative sample of adults (aged 40–93) in the Detroit area, USA. All analyses are stratified by gender (N=males: 330; females: 468). A series of multiple regression analyses were performed to test whether social relations mediate the association between education and health. Although analyses revealed no mediation effect, both men and women with less education were found to have smaller social networks. Women with more education confided less in their children than women with less education did. A series of hierarchical regression analyses were performed to test whether social relations variables moderate the relationship between education and health. Separate analyses by gender indicated that men, but not women, with less education who had larger networks and who perceived emotional, financial and sick care support to be available from a child had lower scores on a health problems index. Findings indicate that the health of lower-educated men in the presence of key social supports parallels the advantaged health status of men with higher levels of education. These findings suggest that social relationships may be a protective factor for the health of men in the lower socioeconomic strata.

Introduction

Gerontologists are renewing their interest in the association between socioeconomic status (SES) and the well being of the elderly. Higher SES generally translates into a greater number of resources available for successful aging. As the aging process advances, each individual accumulates the health benefits and detriments associated with a given socioeconomic status. Researchers have documented those at each point on the socioeconomic spectrum have better health (and lower mortality rates) than those below them, and worse health than those above them (Marmot et al., 1998). Among the elderly, those who enjoy higher SES tend to fare better both psychologically and physically than their lower SES counterparts (Kubzansky, Berkman, Glass, & Seeman, 1998).

Although these SES-related health differentials operate over the life course, they have particular salience in middle and old age. Many health problems begin to manifest themselves during midlife (Marks, 1996), and they do so differentially by socioeconomic status (Haan, Kaplan, & Syme, 1989). In many cases, SES determines the point in life at which an individual will begin to experience health problems, with those of lower SES experiencing the onset of health problems earlier—i.e., often in midlife—than do their higher-SES counterparts (House et al. (1990), House et al. (1994)). Among older adults education has been demonstrated to be a key factor in successful aging, due to its association with cognitive and psychological functioning, health practices, and biological indicators of health risk (Albert et al., 1995; Kubzansky et al., 1998).

Studying the relationship between SES and health among midlife and older individuals, and in particular identifying causal mechanisms or protective factors in this group, may increase opportunities for primary and secondary prevention of disease and illness and may also contribute to a healthier older adulthood. Although current research examines a number of individual-level (e.g., health behaviors; Lantz et al., 1998) and environmental-level factors (e.g., income inequality; Lynch et al., 1998) that may underlie or shape the SES-health association, one of these factors, interpersonal social relations, has received only limited attention. Conceptualized as a resource (Ensel & Lin, 1991; Martin, Grunendahl, & Martin, 2001), social relationships have been identified as one of the many psychosocial mechanisms that may mediate or moderate SES-related health disparities (Anderson & Armstead, 1995; Turner & Marino, 1994; Vaillant, Meyer, Mukamal, & Soldz, 1998; Williams, 1990). However, little research specifically examines the link that social relations, particularly relations with adult children, may have to the association between SES and health (Haan et al., 1989; Kaplan & Lynch, 1997). We focus on the relationship between middle-aged and older parents and their adult children because this relationship is uniquely long term and is the most common source of intergenerational support exchanges (Pillemer & Suitor, 1998).

The current study focuses on two possible forms this association might take. The “mediation hypothesis” holds that the link between SES (as measured by education) and health is mediated, in part, by individuals’ perception of social support from an adult child, and the size of their social network. In other words, individuals with less education might possess fewer social resources, lack of which might account for poorer health. Alternatively, the “moderating hypothesis” posits that social integration and perceived support from important others attenuates or “buffers” the impact of low education on health problems.

A significant amount of research has examined the link between social relations, and health and well-being (Berkman, 1985; Cohen & Syme, 1985; House, Landis, & Umberson, 1988; Pearlin, 1999). This link was first demonstrated in early empirical work exploring the impact of social integration in connection with stress and the incidence of suicide (Durkheim, 1951) and later with mortality both across adulthood (Berkman & Syme, 1979; House, Robbins, & Metzner, 1982) and among the elderly (Blazer, 1982).

The association between social relations and physical health is complex, however. First, the umbrella term “social relations” includes multiple dimensions, often referred to as network structure, social support, and support quality; each of these may influence health in unique ways (Antonucci, 1994; Lynch, 1998). Second, the potential of social relationships to offset the negative effects of stress on health has been widely discussed and demonstrated over the past several decades (see Pearlin, 1999 for a review). Social relations may also have a beneficial “main” effect on health, by “reducing feelings of isolation and alienation and by restraining deviant and injurious behavior” (Rook, 1994) or by providing support in managing health problems (Thoits, 1995). Social relations have been found to have an association with the prevalence, incidence and prognosis of cardiovascular disease in populations that include elderly members (see Berkman, 1985; Seeman, 1996). Seeman (1996) notes the positive impact of pre-stroke social integration on post-stroke recovery in an elderly sample. Yet social relations can also represent a source of conflict and disappointment, negatively influencing health (Antonucci, 2001; Thoits, 1995). Among the elderly, negative social interactions may have a stronger effect on health than positive support does (Rook, 1984). Thus, although social support and social integration are associated with better health in a general sense, the association is likely to be multifaceted and to vary with the specific indicators of constructs that are used in a given study. In the current study, by focusing on one important source of support, adult children, we seek to contribute to the literature by avoiding generalizations that aggregate across important differences in type and source of social relations.

Empirical evidence points to the likelihood that the three phenomena of primary interest in this study and their interrelationships—education, health and family relations—may be critically shaped by gender. There are significant gender differences in SES throughout the life course, with men for the most part acquiring higher education levels than women (Anderson & Armstead, 1995; Barer, 1994; Longino, 1988). Additionally, health status of adults also varies by gender. In general, chronic illness is more prevalent among women, but men die an average of 7 years earlier than women (Barer, 1994; Verbrugge, 1985).

Research suggests that social relations and support are experienced differently by men and women. We are just beginning to document and understand these differences (Antonucci, 1994; Marks, 1996; Turner & Marino, 1994). For instance, women are known to engage in numerous and more intimate relationships and are known to often be the “kin keepers” of the family (Turner & Troll, 1994). Women are also more likely than men to be primary caregivers in their social networks (Bengston, Rosenthal, & Burton, 1996), and women perceive a greater amount of social support to be available than men do (Antonucci & Akiyama, 1987; Ross & Mirowsky, 1989; Turner & Marino, 1994). Other research has suggested that men may benefit more from social relations than women do because they are less likely to personalize troubled aspects of a relationship but are equally or more likely to benefit from the positive aspects (Antonucci, Akiyama, & Lansford, 1998; House et al., 1988).

The analyses presented here, separated by gender, are designed to move the field forward by documenting these complex interrelationships. In this paper, we do not seek to reaffirm the well-established gender differences in social relations, education and health for this cohort of middle-aged and older people. Rather we seek to explore whether differences in social relations differentially impact the education–health link for men in contrast to women. Further, although some research on the family social relations of middle-aged and older people has been conducted, too often it has focused on either the spousal relationship specifically or social relationships in the aggregate. In the present study, we focus on a particularly unexplored area: social relations with adult children.

According to the mediating hypothesis to be tested, weaker social relations represent one of the causal pathways between low levels of education and increased health problems. This hypothesis is suggested by prior theoretical and empirical work that links the availability of social capital (on both an individual and aggregate level) to SES and to health outcomes. As Ross and Wu (1995) note, psychosocial resources are not randomly distributed throughout society. Instead, like stressors, hardships, beliefs, and behaviors, the distribution of psychosocial resources is socially patterned. Accordingly, SES may affect the social capital that is represented by social networks, supports and exchanges. This proposition makes intuitive sense; for instance, it is easy to imagine that the economic hardships characteristic of the lower rungs of the SES ladder adversely affect the nature and extent of an individual's social relationships. Empirical evidence supports this notion (e.g., Roschelle, 1997). Research has also demonstrated an association between lack of social capital on an aggregate level and health indicators. For example, Kawachi, Kennedy, Lochner, and Prothrow-Stith (1997) found that social capital on the US state level, as measured by levels of social trust and civic participation among state residents, was negatively associated with state mortality rates, and that the relationship between income inequality and mortality at this level was substantially mediated through social capital.

Several researchers who focus on elderly populations have examined the association between social relations and SES. Weaker social support has been found among the elderly in lower-SES groups when compared to those in higher-SES groups, and is related to greater levels of stress among the lower class elderly (Murrell & Norris, 1991). In a national probability sample of elderly individuals, Krause and Borawski-Clark (1995) found that older adults of higher SES had more contact with friends, provided more support to others and expressed greater satisfaction with the support they received. Another study, however, found that lower levels of education among individuals aged 70–79 were associated with larger social networks and less negative support (Kubzansky et al., 1998).

Several general-population studies have examined social relations and their link to education level. Ross and Wu (1995) found, in a national probability sample of adults aged 20–64, that the level of social support increased with education level. Marmot et al. (1998) report that in a nationally representative sample of adults in the United States the lowest educational group was disproportionately represented in the highest quintile of family/friend strain and lowest quintile of family/friend support. There is some limited evidence, therefore, to suggest that weaker social relations may be characteristic of individuals with lower education compared to their more educated counterparts, lending plausibility to the idea that social relations are one mediator of the education–health link.

According to the moderating hypothesis to be tested, the quantity and quality of social relations are not necessarily linked to education in a systematic way; however, the presence of these resources among individuals with low education mitigates the negative effects of this status on health. This hypothesis borrows from the notion of “buffering” that has been a central part of research on the stress process (Pearlin, Menaghan, Lieberman, & Mullan, 1981; Wheaton, 1985), and follows other empirical research that has conceptualized low education as a chronic stressor (Wu & Rudkin, 2000). The notion of low education as a stressor is supported by research showing that low education is significantly associated with exposure to certain undesirable life events and that individuals with less education may be more vulnerable to the deleterious effects of stressors (McLeod & Kessler, 1990; also see Adler et al., 1994).

The buffering hypothesis, as used here, suggests that social relationships have a more important effect on health in the presence of the chronic stressors associated with low education than they do in the absence of these stressors. For example, social support may act as a buffer against such stressors as economic strain by increasing feelings of mastery or enhancing coping skills (Russell & Cutrona, 1991; Williams, 1990), whereas these personal resources may not be as consequential among higher-educated individuals. Similarly, supportive social relationships may promote positive health behaviors and facilitate access to medical care among elderly individuals of low education (Berkman, 1985); whereas these relationships may not play such a key role among those with higher levels of education, who have better access to health-related resources. Vaillant et al. (1998) note that social support may play its strongest role in the absence of other social buffers. These observations all point to the likelihood that social support may play an especially important role among people under the chronic “stressor” of low education, and therefore that their primary role in the SES–health association is as a buffer.

The present study seeks to investigate the impact of social relations on the education–health link among middle-aged and older adults. Men and women aged 40 and over are included because of the well-known finding that life expectancy is shorter and chronic health problems emerge earlier among those in the lower SES strata (Marmot et al., 1998). Specifically, our analysis explored

  • (1)

    whether perceived support from child and network size mediate the association between education and health problems for men and women, and

  • (2)

    whether these aspects of social relations moderate the relationship between education level and health problems for men and women.

Section snippets

Description of data set

The Survey of Social Relations was collected through the Survey Research Center at the University of Michigan, and focused on the dynamics of social relations, stress and mental health across the life span. The data were collected in 1992–93. The sample was drawn from a stratified probability sample of 1702 people ranging in age from 8 to 93 in the greater Detroit metropolitan area. People aged 60 and above were over sampled. The overall response rate for the study was 72%. The sample for the

Results

Table 1 presents the distribution of the sample in terms of gender, marital status, and education level. Approximately 25% of both men and women in the sample did not complete high school. Seventy-one percent (n=59) of males who did not complete high school were married or living with a partner, whereas 23% (n=28) of females who did not complete high school were married or living with a partner. The health problem index score for both men and women with less than a high school education was

Discussion

The results of our analysis suggests that (1) certain aspects of social relations vary with years of education; (2) the association between education and health is not mediated by social relations; (3) there are aspects of social relations that moderate the association between education and health; and (4) that patterns of (1) and (3) occur differentially for men and women. In the following paragraphs the implications of these findings are considered.

Acknowledgements

The authors gratefully acknowledge Hiroko Akiyama for comments made on earlier versions of this manuscript, Halimah Hassan for her advice and support in the analyses and Lisa Byrd for assistance in preparing this manuscript. This study was supported by NIA grant AG13490-02, NIMH grant 5R01MH46549-02 and NIA grant T32-AG0017.

References (62)

  • N.E Adler et al.

    Socioeconomic status and healthThe challenge of the gradient

    American Psychologist

    (1994)
  • M.S Albert et al.

    Predictors of cognitive change in older personsMacArthur studies of successful aging

    Psychology and Aging

    (1995)
  • N.B Anderson et al.

    Toward understanding the association of socioeconomic status and healthA new challenge for the biopsychosocial approach

    Psychosomatic Medicine

    (1995)
  • T.C Antonucci

    Social support networksHierarchical mapping technique

    Generations

    (1986)
  • T.C Antonucci

    A life-span view of women's social relations

  • T.C Antonucci

    Social relationsan examination of social networks, social support, and sense of control

  • T.C Antonucci et al.

    An examination of sex differences in social support among older men and women

    Sex Roles

    (1987)
  • T.C Antonucci et al.

    Negative effects of close social relations

    Family Relations

    (1998)
  • T.C Antonucci et al.

    Dynamics of social relationships in midlife. In handbook of midlife development

    (2001)
  • B.M Barer

    Men and women aging differently

    International Journal of Aging and Human Development

    (1994)
  • R.M Baron et al.

    Moderator-mediator variable distinction in social psychological researchConceptual, strategic, and statistical considerations

    Journal of Personality and Social Psychology

    (1986)
  • V Bengston et al.

    Paradoxes of families and aging

  • L.F Berkman

    The relationships of social networks and social support to morbidity and mortality

  • L.F Berkman et al.

    Social networks, host resistance, and mortalityA nine-year follow-up study of Alameda county residents

    American Journal of Epidemiology

    (1979)
  • D.G Blazer

    Social support and mortality in an elderly community population

    American Journal of Epidemiology

    (1982)
  • S Cohen et al.

    Social support and health

    (1985)
  • E Durkheim

    Suicide

    (1951)
  • W.M Ensel et al.

    The life stress paradigm and psychological distress

    Journal of Health and Social Behavior

    (1991)
  • M.N Haan et al.

    Socioeconomic status and healthOld observations and new thoughts

  • J.S House et al.

    The association of social relationships and activities with mortalityProspective evidence from the Tecumseh community health study

    American Journal of Epidemiology

    (1982)
  • J.S House et al.

    Social relationships and health

    Science

    (1988)
  • J.S House et al.

    Age, socioeconomic status, and health

    Millbank Quarterly

    (1990)
  • J.S House et al.

    The social stratification of aging and health

    Journal of Health and Social Behavior

    (1994)
  • E.L Idler et al.

    The meanings of self-ratings of health

    Research on Aging

    (1999)
  • G.A Kaplan et al.

    Special report: Socioeconomic factors and cardiovascular diseaseA review of the literature

    Circulation

    (1993)
  • .A Kaplan G et al.

    EditorialWhither studies on the socioeconomic foundations of population health?

    American Journal of Public Health

    (1997)
  • I Kawachi et al.

    Social capital, income inequality, and mortality

    American Journal of Public Health

    (1997)
  • N Krause et al.

    Social class differences in social support among older adults

    The Gerontologist

    (1995)
  • L.D Kubzansky et al.

    Is educational attainment associated with shared determinants of health in the elderly? Findings from the MacArthur studies of successful aging

    Psychosomatic Medicine

    (1998)
  • P.M Lantz et al.

    Socioeconomic factors, health behaviors, and mortalityResults from a nationally representative prospective study of US adults

    Journal of the American Medical Association

    (1998)
  • P Liberatos et al.

    The measurement of social class in epidemiology

    Epidemiological Reviews

    (1988)
  • Cited by (61)

    • The onset of falls and its effects on perceived social exclusion and loneliness. Evidence from a nationally representative longitudinal study

      2022, Archives of Gerontology and Geriatrics
      Citation Excerpt :

      They found that older and less educated persons with ‘more functional limitations had fewer friends’ (Ha, 2019). Other studies reported that individuals with a low educational level had smaller social networks (Antonucci, 2003; Broese van Groenou and van Tilburg, 2003; Ajrouch, 2005). Furthermore, Ross and Wu found that a lower educational level was associated with less social support (Ross and Wu, 1995; Marmot, 1998; Ross and Zhang, 2008).

    • Social Relationships and Adaptation in Later Life

      2022, Comprehensive Clinical Psychology, Second Edition
    View all citing articles on Scopus
    View full text