Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Studies dating back to Emile Durkheim’s (1897) Suicide demonstrate that social relationships ­provide emotional, social, and economic supports that enhance physical and emotional well-being throughout the life course (House et al. 1988). Over the past three decades, however, researchers have discovered that social relationships are not universally protective for late-life well-being; rather, the protective effects of social ties vary based on the structure, nature, and quality of the relationship. Methodological advances also have enabled researchers to ascertain whether the association between relationships and health reflects social causation (i.e., direct benefits of social relationships), or selection (i.e., characteristics of those people who form and maintain relationships over the life course). Social gerontologists no longer ask, “Do social relationships affect the well-being of older adults?” Rather, they now ask, “Why, how, when, and for whom do social relations affect the health of older adults?”

In this chapter, we first discuss recent innovations in the conceptualization and measurement of older adults’ relationships, including their objective and subjective properties, structures, and functions. We then provide an overview of the characteristics and health implications of four types of relationships: marriage and romantic partnerships (and the dissolution thereof through widowhood and divorce); intergenerational relationships (i.e., parent–child and grandparent–grandchild); friendships; and lack of satisfying relationships (i.e., loneliness). We then discuss the consequences of late-life relationships for health and well-being., We conclude by highlighting the research topics and methods that hold great promise for future exploration, and the implications of recent research for social and health policy.

Conceptualizing and Measuring Social Relations in Late Life

General Properties of Social Relationships

Social relationships are a multifaceted concept, encompassing a variety of structures, functions, and qualities. They may be based on legal ties, blood relations, coresidential status, or simply a fondness for and a desire to affiliate with one another. Relationships also vary with respect to their voluntariness, permanence, and duration. Voluntariness refers to whether one chooses to enter a relationship; permanence indicates whether one is able to terminate a relationship; and duration describes the amount of time that the relationship has existed.

Family ties, particularly relationships with blood relatives such as parents, siblings, and children, are involuntary and are based on a powerful sense of obligation, thus they cannot be abandoned easily (Litwak 1985). By contrast, friendships and romantic relationships are formed voluntarily in contemporary western societies; most individuals choose to befriend or marry persons with whom they are compatible. Social ties with friends and even romantic partners today may be terminated without violating important norms regarding obligation and commitment, although adults may incur some emotional or financial costs from severing these ties (Litwak 1985). For older adults, sibling relationships generally have the longest duration of any relationship; they are formed at the birth of the younger sibling and typically persist until one’s death. Marriages, parent–child relations, and friendships, by contrast, typically begin later in the life course – when one marries, gives birth, or forms a new friendship.

Sociologists and psychologists have developed a range of measures to further capture important quantitative and qualitative aspects of social relationships. Early studies typically focused on simple quantitative characteristics such as the number of persons in one’s social network, to capture one’s social embeddedness and integration (Antonucci 1990). In the past decade, social networks researchers have developed innovative new measures to further capture quantitative aspects of one’s relationships, including their density (i.e., number of people in an individual’s network who know one another) and multiplexity (i.e., number of different types of interactions exchanged within a single dyadic relationship) (Smith and Christakis 2008).

One of the most influential research advances, however, has been the widespread recognition that qualitative (or subjective) aspects of relationships are more important to one’s well-being than are simple counts of significant others. For instance, measures of perceived support are stronger predictors of well-being than the number of persons in one’s network available to provide help. Similarly, subjective appraisals of positive and negative relationship characteristics are powerful predictors of older adults’ well-being (Rook 1998). Positive aspects include feeling loved, cared for, and understood, whereas negative aspects include demands, criticism, and conflict. A recent innovation is the conceptualization and operationalization of “ambivalence,” which refers to having both positive and negative sentiments toward a single person or relationship (Lüscher and Pillemer 1998). For example, a spouse’s well-intended desire to provide advice may be appreciated, yet also perceived as critical. Ambivalence is a common feature of late-life relationships, especially among spouses and parents and their adult children.

Functions of Social Relationships in Later Life

Social relationships also vary widely with respect to the functions they serve in older adults’ lives. Older adults often rely on members of their social network for emotional and instrumental (i.e., practical or financial) support, yet they may expect and desire different types of support based on the nature of a particular relationship. Individuals usually expect both instrumental and emotional support from family members, but emotional support only from friends. Two influential models have been developed over the past three decades to characterize older adults’ support preferences. The hierarchical compensatory model proposes that older people have a rank-ordered preference for receiving support from others (Cantor 1979). Most will turn first to family members and turn to nonfamily only when kin are unavailable. Cantor (1979) further specifies that people prefer to receive support from their spouse, followed by children, other relatives, friends, and professionals or formal organizations. Empirical studies generally support the hierarchical model; older adults are more likely to both prefer and receive assistance from a spouse, followed by children, other relatives, and friends.

The functional specificity model counters that supportive relationships are best matched according to each partner’s needs and resources (Litwak 1985). For example, a married woman may turn to her best friend for emotional support or household help, if she feels that her husband is ill equipped to do so. Studies consistently show that emotional support from friends is more strongly associated with emotional well-being than comparable support from family, suggesting that support is most effective when provided by persons who are most qualified to do so.

Both the hierarchical compensatory and functional specificity models were developed to characterize relationships at one point in time, however. The “convoy model,” by contrast, emphasizes that individuals maintain convoys of significant others who provide both instrumental and emotional support over the life course (Kahn and Antonucci 1980). The desirability, value, and types of ­support needed from a member of the convoy may change over the life course. The number of persons included in one’s convoy, as well as one’s level of closeness to convoy members, shifts over the life course. The model also emphasizes reciprocity and “support banks,” where deposits are made early in the life course in anticipation of future needs or “withdrawals.” This is one mechanism through which support to older family members becomes obligatory; adult children may feel obliged to “give back” to the parents and grandparents who supported them earlier in the life course.

All three models imply that older adults desire support from significant others. However, several theorists acknowledge that individual-level and cultural factors may affect the comfort with which one solicits and accepts help. Cohler (1983) has argued that self-reliance and autonomy are core values of capitalist societies, and they inhibit expression of dependency among older adults in the United States. For example, studies repeatedly show that older adults – even those with serious physical health limitations – do not want to reside with their adult children but prefer to remain independent and reside in their own homes: that is, they want “intimacy at a distance.” Feeling dependent on significant others, especially for health-related caregiving, may be highly distressing to frail older adults and even hasten their desire for death. In sum, older adults’ social relationships are not a monolithic entity; rather, they differ with respect to their number, structure, subjective qualities, and functions. As we shall see next, each of these aspects of relationships carries important implications for the well-being of older adults.

Marriage and Other Romantic Relationships

Concepts and Patterns

Heterosexual marriage is the most common type of romantic relationship among older Americans. In 2008, 73.2% of men and 42.9% of women aged 65 and older were married. A larger proportion of women (41.9%) than men (13.8%) are currently widowed, reflecting men’s higher mortality risk and greater tendency to remarry. Only 10.3% of older men and women are currently divorced. Divorce was relatively rare and stigmatized during the young adult years of current cohorts of older adults, and most of those who did divorce ultimately remarried. Only 4.1% have never married (U.S. Bureau of the Census 2008).

As of February 2011, same-sex marriage is legally performed in five U.S. states and is recognized, though not legally performed, in three states. Exact numbers of legally married gay and lesbian couples are difficult to determine, because many couples report being “married” despite having no U.S. marriage license and because the legal status of unions may shift as new state legislation is implemented. Of those gay men and women who say they are married, 21.0 and 16.5% are 65 or older, respectively (O’Connell and Lofquist 2009). By contrast, 13% of all Americans are aged 65 and older, thus older adults are overrepresented among gay and lesbian partners who self-identify as married.

Many older adults, like younger persons, are in committed nonmarital romantic relationships. In 2000, heterosexual cohabitation was the relationship of choice for 4% of unmarried persons over aged 50 in the United States (Brown et al. 2006). According to estimates from the U.S. Census, same-sex cohabitation is an uncommon status for older Americans. Only 5.5% of gay male and 4.5% of lesbian cohabiting couples are aged 65 or older, although this may reflect the way that gay couples are measured in the Census. Reports are based on one’s own gender, one’s report of the cohabitant’s gender, and the householder’s report of the relationship between the two (O’Connell and Lofquist 2009). By contrast, studies based on nonrandom samples indicate that as many as ­one-quarter to two-thirds of coupled gay and lesbian persons aged 50 or older live with their partner (Grossman et al. 2000).

Some older couples in committed romantic relationships do not coreside – an arrangement called living apart together. In the United States, 7% of women and 6% of men report that they “live apart together” (Strohm et al. 2008). Older adults who choose this arrangement tend to own their own homes, and neither partner wants to move and combine their possessions nor create inheritance complications for their children. Older women also may not want to take on the homemaking and caregiving responsibilities that often accompany coresidence.

Researchers know relatively little about dating in later life. Although the American Association of Retired Persons commissioned a national study of dating in 2003, their sample of more than 3,000 persons focused on those aged 40–69 only. Carr (2004) estimated that 18 months after becoming widowed, only one-fifth of men and one-tenth of women in their 60s or 70s residing in the Detroit area were dating – although nearly 40% of men and 15% of women said that they were interested in dating. In the next decade, we suspect that much more will be learned about dating. Aging Baby Boomers, many of whom divorced and re-entered the dating market in later life, may feel more comfortable and less stigmatized in discussing their nonmarital, perhaps even casual, sexual, and romantic relationships.

Implications for Health and Well-Being

Empirical studies consistently show that married persons are healthier than their unmarried counterparts; strong effects are found for a range of outcomes including all-cause mortality (Johnson et al. 2000) and psychological distress (Johnson and Wu 2002). Social selection and causation are the dominant explanations proposed for the so-called “marriage benefit.” The social selection perspective holds that healthy and financially secure people are more likely to marry and remain married over the life course, thus accounting for the association between marital status and health. The social causation perspective counters that marriage provides social control, economic, and psychosocial benefits that directly enhance health.

Marriage is a key source of social control; married people are less likely than unmarried persons to smoke, drink excessively, and engage in risky behaviors like not wearing seat belts. Spouses, especially wives, may help their partner with health-enhancing behaviors, such as regular visits to the doctor, exercise, healthy eating, and complying with medication regimens (Schone and Weinick 1998). Partnered adults also have higher household incomes than single individuals, and economic well-being predicts good health. In general, persons in a committed relationship are more likely than the unpartnered to report that they have a confidante, and that they feel loved and supported. Although representative studies of long-term gay and lesbian relationships are rare, mounting research suggests that they provide at least some of the same benefits for older adults’ well-being as do heterosexual marriages (Grossman et al. 2000).

One of the most provocative discoveries in recent decades is the recognition that the benefits of a romantic relationship vary considerably based on legal and structural aspects of the relationships. First marriages are more protective than remarriages, legal marriage is more protective than cohabitation or dating, and heterosexual relationships are somewhat more protective than gay and lesbian relationships. For example, the “benefits” of remarriage for mental health and self-rated physical health are more modest than for first marriages (Barrett 2000). These protective effects also are short-lived, appearing only in the early stages of the remarriage transition (Blekesaune 2008). Similarly, cohabitors fare better than unpartnered persons yet worse than married persons in terms of depressive symptoms (Brown 2000), all-cause mortality (Koskinen et al. 2007), and self-rated physical health (Wu et al. 2003). A recent analysis of merged data from the General Social Survey, National Health and Social Life Survey, and the Chicago Health and Social Life Survey found that partnered gays and lesbians were similar to married persons and straight unmarried cohabitors in terms of self-rated health but fared poorer on measures of happiness (Wienke and Hill 2009).

Most scholars concur that these patterns reflect social selection more than social causation, however. For example, older heterosexual cohabitors have a greater mortality risk than their married peers, due largely to the cohabitors’ relatively lower socioeconomic status. However, some studies suggest that cohabiting relationships are qualitatively different from marriages. Marcussen (2005) found that older cohabiting men receive fewer caregiving benefits than their married peers, perhaps because older cohabiting women felt less obligation than married women to provide this time-intensive care. Moreover, gay partnerships and remarriages face distinctive stressors that persons in heterosexual first marriages are spared of, including homophobia and strains of negotiating relationships with ex-spouses (and stepchildren), respectively. Future studies should explore the distinctive stressors experienced by a range of romantic relationships, including cohabitation, same-sex unions, and higher order marriages, to help pinpoint precisely why their health benefits are not comparable to those reaped in marriage.

Across all types of romantic relationships, however, health benefits are contingent upon relationship quality. Survey-based studies show that negative marital interactions increase one’s risk of poor physical and emotional health (Liu and Umberson 2008). A pathbreaking development over the past decade has been the identification of the physiological pathways through which relationships “get under our skin,” with particular attention to cardiovascular, endocrine, immune, metabolic, and sympathetic nervous systems (Ryff and Singer 2001).

Experimental studies typically induce either conflict or closeness among couples in laboratory settings, and then gauge physiological responses. Relationship conflict can impair immune response, slow wound healing, heighten susceptibility to infectious agents, and increase cardiovascular reactivity – all factors that may increase mortality risk among older adults (Robles and Kiecolt-Glaser 2003). Experiments focusing on positive interactions show that inducing physical contact and closeness under a stressful condition may decrease blood pressure and heart rate, and increase oxytocin, a hormone that weakens the impact of stress (Grewen et al. 2005). Negative interactions typically have a more powerful impact on health than positive interactions, and effects are stronger for women than men (Robles and Kiecolt-Glaser 2003). In sum, while long-term committed relationships provide emotional and health-enhancing supports, these benefits vary widely based on the legal status and quality of the relationship, and the personal characteristics of those who enter into (and remain in) such unions.

The Loss of Romantic Relationships: Divorce and Widowhood

Later life is marked by the loss of important social relationships, including the deaths of friends, siblings, parents, and spouses. Relationship loss via divorce is very rare among current cohorts of older adults; the vast majority of divorced older adults dissolved their marriages in young or middle adulthood, and many subsequently remarried. Studies using cross-sectional and administrative data show that currently divorced older adults have an elevated risk of all-cause, cardiovascular disease, cancer, and suicide mortality relative to married persons (Johnson et al. 2000). However, studies based on multiwave data showed that much of this gap is due to social selection, particularly the disadvantageous health and personality traits of those who divorce and do not ultimately remarry (Sbarra and Nietert 2009). As later life divorce becomes more normative among members of the Baby Boom and subsequent cohorts, however, researchers will need to delve more fully into the distinctive ways that late-life marital dissolution affects health and well-being.

Widowhood, by contrast, is a distressing and health-compromising transition faced overwhelmingly by older adults; two-thirds of the 2 million deaths occurring in the United States each year befall persons aged 65 and older (Federal Interagency Forum on Aging-Related Statistics 2009). Older bereaved spouses lose their primary source of emotional, instrumental, and financial support along with the disruption of daily routines and practices. While early stress theories suggested that widowhood was universally and intensely distressing, emerging evidence shows that the physical and emotional consequences of widowhood vary widely, based on one’s gender and the nature of one’s late relationship.

Men are more likely than women to experience physical health declines, increased disability, and heightened risk of mortality after their wives die. While popular lore claims that widowers may “die of a broken heart,” research shows that it is the loss of a helpmate and caretaker that is really the culprit. Wives typically monitor their husbands’ diets, remind them to take their daily medications, and urge them to give up vices like smoking and drinking (Umberson et al. 1992). Widowers are more likely than married men to die of accidents, alcohol-related deaths, lung cancer, and chronic ischemic heart disease during the first 6 months after their loss, but not from other causes that are less closely linked to health behaviors (Martikainen and Valkonen 1996). Even worse for older men is that their wives often are their primary (or only) source of social support and integration; when a man loses his wife, he also loses an important connection to his social networks.

By contrast, women’s more emotionally intimate social relations over the life course are an important resource as they adjust to spousal loss. Older widows typically receive more instrumental and emotional support from their children than do widowers, given mothers’ closer relationships with their children throughout the life course. Women also are more likely to have larger and more varied friendship networks than men, and these friendships provide an important source of support to women as they cope with their loss (Ha 2008).

The well-being of older widow(er)s also is linked to the emotional climate of the late marriage. Early writings, based on the psychoanalytic tradition, proposed that bereaved persons with the most troubled marriages would suffer heightened and pathological grief (Parkes and Weiss 1983). This perspective held that persons who had conflicted marriages would find it hard to let go of their spouses, yet also feel angry at the deceased for abandoning them. However, longitudinal studies that track married persons over time through the widowhood transition find that the loss of high-quality marriages is most distressing. Wheaton (1990) found that the emotional consequences of role loss are contingent upon one’s “role history”; widowhood is distressing when the marital relationship had been satisfying. In sum, while classic bereavement theories proposed that a romantic partner’s death is universally distressing, more recent work suggests that reactions to loss are contingent upon precisely what was lost: the loss of a beloved helpmate has a more profound impact than the death of a difficult or distant spouse.

Life-Long Singlehood

Never married persons have been nearly absent from social relations research. This absence reflects the fact that only 4% of persons aged 65 and older in the United States have never married (U.S. Census Bureau 2008). As such, researchers using sample surveys often do not have enough cases to study the health and well-being of the never married. Mortality is one of the few outcomes studied, because mortality and marital status data are available on vital registries and large administrative data sets such as the National Longitudinal Mortality Study (Johnson et al. 2000). However, these sources include only basic demographic measures, so researchers cannot explore the psychosocial pathways through which singlehood affects health.

The few studies exploring other indicators of single older adults’ well-being reveal quite a positive picture. Older never married women enjoy mental health (Pudrovska et al. 2006) and physical health (Cwikel et al. 2006) equal to their married peers, and superior to their formerly married peers. These patterns partly reflect selection, where older cohorts of never married women are better ­educated than their married and formerly married peers, and have richer economic resources than their divorced or widowed peers. Moreover, never married women have adjusted to their status over time; they have chosen relationships that offer socioemotional support (Pudrovska et al. 2006), and they rely on formal services such as meal preparation services to help manage their instrumental needs (Cwikel et al. 2006). In sum, while research on the health of never married persons is sparse, the evidence generally reveals that singlehood may carry health advantages for women, yet these advantages typically reflect social selection factors including greater educational and economic resources of never married women in current cohorts of older adults.

Parent/Child Relationships and Grandparent/Grandchild Relationships

Thus far, we have focused solely on relationships between romantic partners; however, intergenerational relations also are a critical source of support (and strain). We now discuss characteristics and health consequences of parent–child and grandparent–grandchild relations.

Patterns and Concepts

Approximately 85% of American adults aged 65 and older are parents. Childless older adults are a heterogeneous group, including those who are child-free by choice, and those who desired children yet did not ultimately become parents. In 2010, 80% of women aged 60–64 will have at least one grandchild, but they will have fewer grandchildren than same-age women in prior cohorts. Due to increases in longevity, children will have more living grandparents than in the past: In 2020, nearly half of 10-year-olds will have four living grandparents, and four-fifths of 30-year-olds will have at least one living grandparent (Uhlenberg 2005).

Implications for Health and Well-Being

Nearly two-thirds of parents aged 50 and older rate their relationships with all of their adult children as excellent, and relationships tend to become even better as parents and children age (Birditt et al. 2009). However, even the best filial relationships can experience tensions, which may threaten older adults’ well-being. Older adults who rate even one of their relationships with an adult child to be anything less than excellent report less happiness and more depressive symptoms (Ward 2008). Older parents may experience ambivalent feelings in their relationships with adult children when children face major problems, such as an illness, divorce, a legal problem, or substance abuse.

Such major problems in the life of an adult child also may affect older adults’ health and well-being indirectly – via taking on the role of caregiving or custodial grandparent. Grandparents who are raising their grandchildren are a topic of heightened concern in the United States today. More than 1.5 million children (2.0% of all children) now live with at least one grandparent and no parent, and this figure underestimates the number of children whose primary noncoresidential caregiver is a grandparent (U.S. Census Bureau 2008). Grandparents step into a parenting role, or are asked to step in by a family member or the state, when their children (i.e., parents of the grandchildren) experience serious problems such as physical or mental illness, substance abuse, incarceration, or homelessness. Love and obligation may motivate grandparents to keep their grandchildren in the family and out of the foster care system.

Serving as primary caregiver to a grandchild is highly stressful: Grandparents worry about their children’s problems and the financial demands of raising grandchildren – many of whom have emotional or developmental difficulties – all while navigating their own aging process. The early stages of caregiving are especially difficult. New custodial grandparents may forgo their own preventive behaviors, and experience increased depressive symptoms and poor physical health (Hughes et al. 2007). Despite the strains imposed by coresidence with their grandchildren, grandparents provide an indispensable safety net. Moreover, for many older adults, grandchildren provide an important source of happiness and purpose.

Friendships in Later Life

Older adults are more than romantic partners or caregivers to the younger generations; most maintain vibrant friendships with peers. Research on late-life friendship has flourished over the past three decades. Friendship is distinct from family relationships in that it is voluntary, nonobligatory, and typically based on the exchange of emotional rather than instrumental support (Blieszner and Roberto 2004). Friends are usually of similar ages, whereas family relationships often are cross-generational and thus may be marked by an imbalance of power. Because of their voluntary nature, friendships require more agency and motivation than familial relationships; they must be sought out, cultivated, and maintained – or in the case of unsatisfying relationships – dissolved. Healthy friendships also tend to have an equitable “give and take.” Reciprocity, or being able to give back support equivalent to what was received, is more important to the quality and sustenance of friendships than kin relations (Rook 1987). Because friendships are less institutionalized than family relationships, however, norms for affiliation may be unclear, creating potential for discord and misunderstanding (Blieszner and Roberto 2004).

Friendships carry more rewards than costs, including companionship, shared leisure, emotional support, social integration, and informational assistance. Friendship is a more powerful predictor of older adults’ psychological well-being and life satisfaction than family relationships. Persons with rich friendship networks tend to have better physical health than those with more tenuous ties, yet this relationship is mutually influential; those with good health also are best equipped to maintain friendships (Rawlins 2004).

Like family relationships, friendships differ in the levels of instrumental and expressive support they provide. Rawlins (2004) has observed that friendships have “two general modes”: confidantes and companions. The former are based on intimate conversation, caring, strong emotional attachment, commitment, and loyalty. The latter, by contrast, involve socializing, social interactions with groups of individuals, limited emotional attachment, and reciprocity.

Late-life friendships vary by gender and social class. Older women’s friendships involve more intimate self-disclosure, whereas men’s friendships often are based on shared activities. As a result, women tend to have closer same-sex friendships than men do, and this is particularly true for married individuals. Most married women have close confidants beyond their spouse, but men typically do not (Antonucci 1990). Middle class older adults tend to have more friends than their working class peers (Phillipson 1997), perhaps due to their greater financial resources and greater likelihood of living geographically far from their kin.

An important discovery over the past three decades is that friendships change in quantity, quality, and importance over the life course, particularly as one faces late-life transitions including widowhood and the onset of health problems, and as one’s needs change (Litwak 1985). Laura Carstensen’s influential socioemotional selectivity theory proposes that older adults selectively choose to maintain relatively fewer, but higher quality, relationships as they age and experience declines in health. Casual, less rewarding ties may lapse while only the most meaningful relationships are maintained (Carstensen et al. 1999). Empirical evidence confirms that as adults age they evidence a reduction in the number of friends reported (Phillipson 1997). This decrease in the number of one’s friendships may not entirely reflect preferences, however. As older adults retire, their ties to workplace friends weaken, while physical health declines, caregiving demands, and functional limitations may impede one’s ability to maintain social ties. Some scholars speculate that online relationships may become increasingly important to older adults, especially those with serious mobility limitations, as they enable social interaction with acquaintances and “weak tie” relationships (Wright 2000).

An important avenue that researchers are only starting to explore is the identification of specific tasks for which friends are “substitutable” when family ties are not available. For example, a recent study of older adults’ choices for health care proxy designations revealed that nearly all (85–90%) married persons and parents named a spouse or child to make end-of-life decisions for them. Among unmarried childless persons, by contrast, nearly one-quarter appointed a friend to play this important role (Carr and Khodyakov 2007). Other studies reveal that widows and widowers are less likely to seek out new romantic partnerships if their social and emotional needs are met by supportive friends (Carr 2004). Among future cohorts of older adults, for whom divorce rates are higher and fertility rates lower than for current cohorts, friendships may be an essential source of instrumental and emotional support.

Absent and Unfulfilling Social Relationships: Loneliness and Social Isolation

Old age historically has been considered a time of social isolation. Disengagement theory proposed that it was beneficial for both the older adult and society if the elder were to gradually withdraw from his or her social roles and relationships (Cumming and Henry 1961). Similarly, classic role theories held that the loss of the work role for men (via retirement) and loss of the wife role for women (via widowhood) would leave older adults socially isolated and despondent (see Biddle 1986, for review). More recent research counters, however, that while loneliness and social isolation are problematic, they are neither inevitable nor universal features of aging.

Over the past three decades, researchers have recognized that loneliness is not triggered by a quantitative lack of relationships, but by a lack of satisfaction with the number or quality of one’s relationships. Contemporary researchers have identified two statistically and conceptually distinct subtypes of loneliness: emotional loneliness refers to the absence of an intimate confidante, while social loneliness refers to the absence of a broader social network. The two types are highly correlated; widowed persons, those living alone, or those living far away from their friends and families consistently report higher levels of both types of loneliness than persons who are more socially integrated (de Jong Gierveld and Havens 2004).

The mere presence of proximate relationships does not ward off loneliness, however. An estimated 25% of older married persons report emotional and social loneliness; this pattern is particularly common among persons whose spouses are ill, who have a dissatisfying (or nonexistent) sexual relationship, and who have infrequent or conflicted conversations (De Jong et al. 2009). As de Jong Gierveld and Havens (2004) noted, loneliness depends on one’s “standards as to what constitutes an optimal network of relationships.”

Despite its subjective nature, loneliness is a serious problem for many older adults; it is linked to sleep problems, poor cardiovascular health, and elevated blood pressure, each of which carries long-term consequences for mortality risk (Cacioppo et al. 2002). Loneliness also may be a particularly acute social problem for older adults in future cohorts. Smaller families and an increased prevalence of divorce and childlessness among future cohorts of older adults may create a context where one maintains objectively fewer relationships (see Chap. 13), thus triggering social ­loneliness. More importantly, however, some have argued that current cohorts of midlife adults have ­unrealistically high expectations for what their social relationships should provide (e.g., one’s partner should be their “soulmate”); if these lofty expectations go unfulfilled, then older adults may report higher levels of emotional loneliness, as well. In sum, it is not how many relationships one maintains that matters for one’s late-life well-being, but the extent to which those relationships are deemed personally satisfying and fulfilling.

Looking Forward to the Next Three Decades

Over the past three decades, scholars have made major advances in conceptualizing and measuring the multifaceted nature of older adults’ social relationships, and documenting the implications of these relationships for health and well-being. We anticipate that five topics are ripe for further exploration in the coming decades: sexuality; population-based studies of elder abuse; racial and ethnic differences in late-life social relationships; social relations of Baby Boomers; and the use of research methods that incorporate multiple reporters and multiple sources of health and relationship data.

Sexuality and Romantic Relationships in Later Life

Even at the turn of the 21st century, most studies of older adults’ romantic relationships focus on the protective effects of instrumental and emotional support, and fail to consider another critically important component: sexuality. In the future, we expect that scholars will explore more fully the role of sexuality in older adults’ relationships – regardless of whether the sexual relationship occurs within the context of a legal or coresidential relationship. Of particular interest is how sexual aspects of older adults’ relationships are protective for health and well-being, and how aging-related changes in physical functioning may affect the quality and nature of sexual relations among long-term partners. Waite et al. (2009) suggest that older adults with high-quality sexual and intimate relations will have better trajectories of physical and mental health than those whose relationships function less well (or who lack such relationships). The recently collected National Social Life, Health and Aging Project (NSHAP), a nationally representative sample of community-dwelling individuals aged 57–85, provides in-depth measures of sexual behavior, practices, and health and will be an invaluable resource as social gerontologists further investigate both heterosexual and homosexual older adults’ sexual relationships.

The Dark Side of Relationships: Elder Abuse

In the past three decades, researchers have documented that even high-quality relationships may be strained or ambivalent. Yet relatively little population-based research focuses on extreme negativity in older adults’ relationship. Elder abuse and neglect comprise intentional physical, emotional, financial, or sexual abuse, as well as failure by a caregiver to meet an older adult’s basic needs (Bonnie and Wallace 2005). Elder abuse is a serious though understudied social problem; data historically have come from small, nonrepresentative samples; the criminal justice system; or agency or caregiver reports.

In 2004, however, the NSHAP obtained reports of respondents’ recent experiences of mistreatment, and for those who reported mistreatment, their relationship to the perpetrator. NSHAP also obtained detailed demographic, health, cognitive functioning, and social relationships data (Laumann et al. 2008). These data reveal that family members are the most common perpetrators of elder abuse; verbal and physical abuse are most often committed by romantic partners and children, whereas financial abuse is most frequently perpetuated by children and siblings. In the next decade, sociologists will have the resources to fully explore the risk factors for and consequences of multiple types of elder mistreatment, with particular attention to the ways that family relations may elevate or buffer against the risk of multiple types of late-life abuse.

Racial and Ethnic Differences in Late-Life Social Relationships

Most research on late-life social relationships focuses on white Americans. This pattern partly reflects the fact that older blacks and Latinos are underrepresented in large-scale sample surveys, given their elevated rates of mortality and morbidity. However, documenting the nature and consequences of social relationships for ethnic minorities is an important inquiry. Some scholars have proposed that blacks’ low rates of marriage relative to whites contribute to blacks’ elevated risk of mortality and morbidity (Kaplan and Kronick 2006). However, this argument rests on the assumption that marriage benefits blacks’ and whites’ health similarly.

This assumption requires interrogation. Studies reveal that blacks report poorer marital quality and more marital conflict than whites, and that the economic gains to marriage are less for blacks than whites given black men’s disadvantaged economic prospects (Broman 1993). Given these patterns, marriage may be less protective to blacks compared to whites. Further, given the very high levels of intergenerational integration and support in Latino and Asian families, it is plausible that the benefits of marriage vis a vis other social relationships are weaker than they are for whites. Identifying the distinctive contributions of marriage, parenthood, friendships, and sibling relationships to the health of ethnic minorities will be a valuable line of inquiry in the future decades. In 2006, whites, blacks, Latinos, and Asians accounted for 81, 9, 6, and 3% of the U.S. population, respectively. By 2050, these proportions will be 61, 12, 18, and 8%, respectively (Federal Interagency Forum on Aging-Related Statistics 2009).

As the demographics of older Americans shift, social gerontologists should identify the distinctive correlates of health and well-being for all ethnic and racial groups. Small-scale qualitative studies may be particularly useful in revealing the ways that the distinctive cultural views and practices of ethnic families affect late-life health and well-being. For example, Confucian values including filial piety affect the ways adult children monitor the health of aging parents among Chinese Americans (Park and Chesla 2007). Gendered cultural views such as machismo (i.e., men’s adherence to traditionally masculine, high-risk behaviors) and marianismo (i.e., women’s self-sacrifice for spouse and children) in Latino families affect both family relations and health practices – which may have important implications for later-life health (Cianelli et al. 2008). We are optimistic that future research blending qualitative and quantitative research may better illuminate the ways that cultural context shapes the relationship between family and health among ethnic and racial subgroups.

Social Relationships of Baby Boomers

Much of what scholars know about older adults’ social relationships is based on the experiences of current cohorts of older adults, those men and women born in the early 20th century who were socialized into rigid gender-typed social roles during their formative years and who often ­maintained a traditional division of labor in their marriages, where men were primary breadwinners, and women were primary caretakers for their husbands and children. It is not surprising, then, that the studies reported in this chapter reveal that women have closer ties to friends and children than their husbands, that women are an important source of health control to their husbands, and that men typically fare worse than women when their spouse dies.

For members of the Baby Boom cohort, the 75 million persons born between 1946 and 1964, late-life relationships may be reinvented. Future cohorts of older women are more likely than their predecessors to have received college degrees, held professional occupations, and shared childrearing tasks with their husbands. Future cohorts of men, by contrast, may have more emotionally intimate friendships than prior cohorts of men who were socialized to be self-sufficient and independent. Thus, we might expect future cohorts to fare better in the face of spousal loss – given that women may have the economic resources and men the interpersonal resources to cope with bereavement. Further, given that Baby Boomers faced fewer social obstacles to divorce during their young and midlife years, we might expect that their late-life marriages will be of higher quality than ­current cohorts; older adults today may have faced social or economic pressure to remain in ­dissatisfying marriages. Given major shifts in family structure and gender role socialization over the past 30 years, we expect that scholars’ understanding of older adults’ social relationships may be transformed in coming decades.

Methodological Innovations

Sociologists overwhelmingly rely on large-scale sample survey data to document the nature and consequences of older adults’ relationships. However, this approach has a number of limitations. First, survey-based studies of social relations typically focus on a single individual’s report, rather than information from both members of a romantic dyad, or multiple members of a family or social network. Ironically, research purportedly exploring spousal dynamics or parent–child relationships typically relies on a single person’s appraisal of the relationship.

Further, research reveals that a parent may have a very different relationship with each of his or her children, yet most standard survey items rely on an overall appraisal of one’s relationship with “children” (Davey et al. 2009). It is not clear whether these measures tap one’s assessment of the child with whom one is closest, most emotionally distant, or an average across children. New analytic techniques, such as dyadic data analysis, allow researchers to use information from multiple reporters to estimate how much each person’s outcome is associated with both own and partner characteristics (Kenny et al. 2006). This approach enables researchers to explore questions such as: how do both spouses’ reports of marital conflict affect each spouse’s health and health behaviors, and to what extent does one partner’s health behaviors and practices affect the other partner’s outcomes?

Second, survey-based studies of relationships and health historically have relied on self-reports of symptoms and conditions, rather than physiological indicators that may capture short-term responses to relationship characteristics and strains. As such, scholars still do not fully understand how older adults’ social relationships “get under the skin” to affect physical and emotional health. Laboratory research, conducted primarily by psychologists, has made important advances, by measuring the physiological responses of older couples placed in either stressful or supportive settings (Robles and Kiecolt-Glaser 2003). In the past decade, a number of large representative sample surveys of older adults including the Health and Retirement Study, Midlife Development in the United States, National Health and Nutrition Examination Survey, NSHAP, and Wisconsin Longitudinal Study have supplemented their self-reported health data with extensive genetic and biological indicators, such as immune response measures. We are optimistic about the scientific discoveries that may develop in the next three decades, as interdisciplinary teams of researchers continue to ­investigate the complex ways that demographic, socioeconomic, biological, psychosocial, and genetic factors link social relationships to health and well-being among older adults.

Implications for Policy and Practice

Over the past 30 years, researchers have demonstrated that older adults maintain diverse social networks including but not limited to spouses, former spouses, children, grandchildren, siblings, and friends. Although socially integrated persons enjoy better health than those who are more isolated, the protective effects of social relations vary widely based on the quality of those relationships. These findings have important implications for social policy.

Current policies tend to privilege legal and biological ties over all other relationships. For instance, if an incapacitated older adult has not appointed a health care proxy, many states have policies that give priority to specific family members as substitute decision makers; spouses and children are typically at the top of the hierarchy. Likewise, the majority of U.S. states do not grant gay and lesbian partners the legal right to formalize their union. Public policies, and especially health care policy, should be based on an expansive definition of “family,” and allow older adults to include as their decision makers, advocates, and beneficiaries whomever they consider their closest and most meaningful ties.

Although the health-enhancing effects of high-quality social relationships cannot be understated, it is essential to recognize that at least part of these benefits reflect social selection characteristics; that is, the preexisting traits of persons who enter in and out of particular relationships. For example, researchers have found that the deleterious health effects of divorce and grandparent–grandchild coresidence and the relatively weak health protection provided by cohabitation reflect the fact that those who divorce, serve as custodial grandparents, and cohabiters tend to have fewer economic resources than their counterparts. As such, a financial safety net that provides at least minimal quality housing, food, health care, and economic security may be the most effective policy for promoting physical and emotional well-being among older adults.