Families in Later Life: A Decade in Review

Abstract

Later-life families encompass the legal, biological, romantic, and kin-like relationships of persons ages 65 and older. Research on older families has flourished over the past decade, as population aging has intensified concerns regarding the capacities of families to care for older adults and the adequacy of public pension systems to provide an acceptable standard of living. Shifting patterns of family formation over the past half-century have created a context in which contemporary older adults’ family lives differ markedly from earlier generations. Decreasing numbers of adults are growing old with their first and only spouse, with rising numbers divorcing, remarrying, forming non-marital romantic partnerships, or living single by choice. Remarriage and the formation of stepfamilies pose challenges and opportunities as older adults negotiate complex decisions such as inheritance and caregiving. Family relationships are consequential for older adults’ well-being, operating through both biological and psychosocial mechanisms. We synthesize research from the past decade, revealing how innovations in data and methods have refined our understanding of late-life families against a backdrop of demographic change. We show how contemporary research refines classic theoretical frameworks and tests emerging conceptual models. We organize the article around two main types of family relationships: (1) marriage and romantic partnerships and (2) intergenerational relationships. We discuss how family caregiving occurs within these relationships, and offer three promising avenues for future research: ethnic minority and immigrant families; older adults without close kin (“elder orphans”); and the potentials of rapidly evolving technologies for intergenerational relationships and caregiving.

Research on families in later life has flourished during the past decade as population aging has intensified concerns regarding the capacities of families to care for older adults and the adequacy of public pension systems to provide an acceptable standard of living (Angel & Settersten, 2015; Blieszner & Bedford, 2012). Demographic shifts have created a context in which aging families today differ markedly from earlier generations. Decreasing numbers of adults are growing old with their first and only spouse, with rising numbers divorcing, remarrying, forming nonmarital romantic unions, or living single by choice (Brown & Wright, 2017). Remarriage and the creation of stepfamilies pose new challenges as older adults negotiate complex decisions such as inheritance and caregiving (Waite & Xu, 2015). Macrosocial trends, including diminishing economic opportunities for young people, have transformed intergenerational ties, with some grandparents acting as provider rather than recipient of care (Doley, Bell, Watt, & Simpson, 2015). These trends have structured the ways that care and support are exchanged within and across older adults’ legal, biological, romantic, and kin-like relationships.

Many factors shape roles and relationships within later-life families, including sociodemographic characteristics such as race and gender, life course experiences such as childhood conditions, relational factors such as one’s social networks, and macro-level factors such as public supports for and cultural norms regarding families (Carr, 2019a). Recent studies have explicated both biological and psychosocial mechanisms through which family relationships and intrafamilial exchanges protect or undermine older adults’ well-being (e.g., Donoho, Crimmins, & Seeman, 2013). Innovations during the past decade in the collection and analysis of biomarker, dyadic, family-level, and social network data have enabled researchers to examine new aspects of perennial questions, such as the implications of marital dissolution and family caregiving for older adults’ well-being and timely new topics such as older adults’ provision of spousal care in same-sex versus heterosexual marriages (Umberson, Donnelly, & Pollitt, 2018). Harmonized cross-national data resources, encompassing nations in Africa, Asia, Europe, Latin America, and North America, permit documentation of how policy and cultural contexts shape outcomes such as intergenerational exchanges and older adults’ well-being (Lee et al., 2018). These data advances have provided researchers the means to reassess and extend long-standing theoretical frameworks, such as family systems (Broderick, 1993), socioemotional selectivity (Carstensen, 1992), and stress process (Pearlin, Schieman, Fazio, & Meersman, 2005) models, and to evaluate emerging frameworks such as the gender-as-relational model (Springer, Hankivsky, & Bates, 2012). (See Table 1 for exemplar data sets used to study later-life families.)

In this article, we synthesize research published in the past decade and show how innovations in data, measures, and methods have advanced and refined our understanding of late-life families against a backdrop of demographic change. We organize the article around the following two main types of family relationships: (a) marriage and romantic partnerships and (b) intergenerational relationships. We show how caregiving occurs within these relationships and conclude by proposing three promising avenues for future research: immigrant and ethnic minority aging families, “elder orphans” growing old without close family ties, and the potentials of technology to meet older adults’ caregiving and interpersonal needs. A comprehensive discussion of late-life families is beyond the scope of a single article, so we are selective in coverage. (Related topics in this issue include immigrant families [Van Hook & Glick, 2020], work–family adaptations [Perry-Jenkins, & Gerstel, 2020], income security [Cooper & Pugh, 2020], elder abuse [Hardesty & Ogolsky, 2020], and lesbian, gay, bisexual, transgender and queer or questioning [LGBTQ] families [Reczek, 2020].)

Marriage and Romantic Partnerships

Marriage

Marriage is the most common romantic partnership among U.S. adults aged 65 years and older, with higher rates among men, Whites, and persons aged 65 to 74 relative to women, Blacks, and adults aged 85 years and older. In 2017, 70% of men but only 46% of women aged 65 years and older were currently married; this gap reflects men’s lower life expectancy, women’s greater likelihood of becoming widowed, and men’s greater tendency to marry slightly younger partners and remarry following divorce or widowhood (Brown & Wright, 2017). Older Blacks are less likely than Whites to be married, reflecting lower rates of marriage and higher rates of marital dissolution at every life course stage (Sassler & Lichter, 2020; Sweeney & Raley, 2020). The proportion of older adults who are married declines with age, as rates of widowhood increase. Among those aged 85 years and older, 60% of men but only 17% of women are married (U.S. Census Bureau, 2017).

The effects of marriage on older adults’ well-being are generally consistent with stress process models, which delineate the pathways through which stress affects mental health and the factors moderating these linkages, such as coping resources (Pearlin et al., 2005). Instrumental and emotional support from one’s spouse can directly enhance well-being and buffer the impact of late-life stressors such as functional impairment. Conversely, marital strain can undermine well-being directly and intensify the harmful impacts of chronic and acute stressors (Carr, Cornman, & Freedman, 2017). During the past decade, researchers have made new discoveries regarding the contexts in which and pathways through which marriage affects older adults’ well-being. First, dyadic studies reveal the importance of relationship context, comprising both partners’ experiences, traits, and perceptions. Second, researchers are documenting how sexual relationships contribute to late-life well-being, identifying nuanced aspects of sexuality that extend beyond simplistic measures such as frequency. Finally, biomarker data reveal the physiological pathways through which intimate partnerships affect health and well-being.

Dyadic Approaches to Understanding Late-Life Marriage.

Research on late-life marriage has long relied on data provided by one spouse only, capturing a one-sided and static appraisal of the relationship (Boerner, Jopp, Carr, Sosinsky, & Kim, 2014). Dyadic data allow explorations of the relational context, including the perceptions, (re)actions, and characteristics of both partners. As such, they provide an ideal site for evaluating new theoretical frameworks such as the gender-as-relational model, which proposes that gender roles are not predetermined by early socialization but instead are “dynamic and situational” and are constructed and negotiated within relationships (Springer et al., 2012, p. 1661). This model challenges the dated argument of “his” and “hers” marriage, which holds that marriage is uniformly more protective for men than women due to the gendered nature of macrosocial structures (Bernard, 1972). Rather, the gender-as-relational model recognizes that gendered experiences also are constructed via microlevel exchanges.

Consistent with the gender-as-relational model, dyadic analyses reveal the distinctive ways husbands and wives are affected by stressors common in later-life marriages. Studies of older heterosexual couples find that husbands’ health problems including depression and disability are linked with wives’ reduced emotional well-being, yet parallel patterns are not detected for men (Carr, Cornman, & Freedman, 2016; Iveniuk, Waite, Laumann, McClintock, & Tiedt, 2014). This asymmetry may reflect the fact that wives feel a greater responsibility for and play a more intensive role in caring for ailing spouses than do husbands. However, comparative studies find that this partner inequality in spousal caregiving is more pronounced in heterosexual couples than same-sex couples (Umberson et al., 2018). The burden of spousal care borne by women is more intense for those married to men than women, challenging the conflation of “male” and “female” with gendered caregiving roles, even among a cohort of older adults who had been socialized to comply with those role expectations (Salari & Zhang, 2006). These studies challenge the notion that “his” marriage uniformly confers greater health benefits than “hers” and reveal how older husbands’ and wives’ experiences are conditional on the marital context, including partners’ health, caregiving arrangements, and the gender composition of the dyad.

Sexuality in Later-Life Partnerships.

Research on late-life sexuality has expanded during the past decade, shedding new light on an overlooked mechanism linking marital status to older adults’ well-being (Waite, Iveniuk, Laumann, & McClintock, 2017). Sexual activity is positively associated with physical and mental health (Wong & Waite, 2015), and currently married older adults are more likely than their unpartnered counterparts to have been sexually active in the past year. Older men tend to report higher levels of sexual interest than their female counterparts, although sexual interest for both declines with age (Das, Laumann, & Waite, 2012). Dyadic analyses demonstrate the importance of taking a gender-relational approach, where gender differences in sexual interest, behaviors, and satisfaction—once presumed to be largely biological in nature—are shaped by the marital context. For example, an analysis of the National Social Life, Health and Aging Project (NSHAP) data found that older husbands (but not wives) whose homes were rated by interviewers as messy or disorderly reported diminished interest in sex (Schafer, Upenieks, & Iveniuk, 2017). The investigators attributed this waning sexual interest to the husband’s perception that their wife violated the culturally prescribed gender role of homemaker.

Advances in social networks methods have facilitated novel explorations of how older adults’ friends and confidantes bear on their sexual health. One analysis of NSHAP social network data found that older married men were at greater risk of erectile dysfunction when their wives had more frequent contact with their husband’s confidants than did the husband (Cornwell & Laumann, 2011). The authors suggested that the wives’ centrality in the husbands’ social networks outside the marriage may undermine his feelings of autonomy, masculinity, and privacy, which may bear on his sexual performance. This work demonstrates that the nature of spouses’ sexual relations and its effects on older adults’ well-being vary based on the relationship context, a key feature in both stress process and gender-as-relational models.

Biological Responses to Romantic Partnerships.

Biological measures of health (referred to as biomarkers) enable researchers to document how support and strain within marriage “get under one’s skin” to affect older adults’ health. These data are increasingly collected as a part of nationally representative surveys in the United States and abroad, such as the English Longitudinal Study of Aging, Health and Retirement Study (HRS), Midlife in the United States (MIDUS), and NSHAP. They have contributed to a burst of research linking marriage and romantic partnerships with physiological indicators of immune, cardiovascular, nervous, musculoskeletal, and circulatory system health (Umberson & Thomeer, 2020). Decades of social science research have documented the behavioral and psychosocial pathways linking spousal ties to health, emphasizing the social support and control functions of those ties (Carr & Springer, 2010). Biomarkers have facilitated the specification of physiologic responses linking relationships to disease and well-being among older adults (Hobcraft, 2009).

Analyses of biomarker data provide further evidence that marriage is not uniformly more health-enhancing or depleting for older men relative to women (Bernard, 1972) but, rather, the ways that gender is performed and negotiated within the dyad bear on both partners’ health (Springer et al., 2012). Analyses of MIDUS data find that marital quality and support are linked with superior immune system functioning (interleukin-6 and C-reactive protein; Donoho et al., 2013) and bone health (bone mineral density) for women only (Miller-Martinez et al., 2014). These results suggest that gender differences are conditional on marital context; older women’s health appears to be protected in the context of highly supportive marriages, whereas men may receive health-enhancing benefits regardless of the union’s emotional quality (Boerner et al., 2014).

Other novel biological measures, including sleep actigraphy data, reveal the specific pathways through which marital dynamics affect late-life health. Selcuk, Stanton, Slatcher, and Ong (2017) found that perceived partner responsiveness improved both self-reported and actigraphy measures of sleep, yet these effects operated through the reduced levels of depression and anxiety enjoyed by those with more supportive partners. Taken together, studies incorporating biomarkers and complex measures of marital relations shed light on the physiological mechanisms through which marital relations enhance or undermine older adults’ well-being.

Marital Transitions

For many older adults, especially women, later life is spent outside a marital relationship. Among persons aged 65 years and older in 2016, only 13% of men were divorced or separated and 12% were widowed. Comparable proportions for women were 16 and 34, respectively. Among those aged 85 years and older, only 7% of men and 8% of women were divorced or separated, although women were more than twice as likely as men to be widowed (72% vs. 35%; U.S. Census Bureau, 2017). Widowhood historically was the primary pathway out of marriage, yet in recent decades rates of “gray” divorce have increased dramatically and now account for one third of all later-life dissolutions (Brown & Wright, 2017). Marital dissolutions have been described as among the most distressing life transitions (Holmes & Rahe, 1967), yet contemporary studies reveal how older adults adapt and even thrive following dissolution by seeking new romantic partnerships (Brown, Lin, Hammersmith, & Wright, 2019), engaging in productive activities such as volunteering (Carr, Kail, Matz-Costa, & Shavit, 2017), and strengthening ties with adult children (Carr & Boerner, 2013a). This work adds a new dimension to classic frameworks such as socioemotional selectivity theory (Carstensen, 1992), which describes how older adults contract, maintain, or expand their social networks and interactions to satisfy their personal needs in the face of age-related losses.

Widowhood.

Longitudinal analyses show that older bereaved spouses are vulnerable to mental health symptoms including depression, anxiety, grief, and loneliness, yet most return to preloss levels of emotional well-being within 2 years of the death (Sasson & Umberson, 2014). (For an exception, see research on complicated grief [Shear et al., 2011].) Recognizing that adaptation is the norm rather than the exception, recent research has moved beyond mental health outcomes and instead has explored how older widow(er)s experience growth and re-engage interpersonally and socially with attention to the ways that relationship histories condition these adaptations.

Analyses of the Changing Lives of Older Couples (CLOC) data reveal how widow(er)s’ relationships with their adult children change following loss and how these changes are contingent on both the nature of the late marital relationship and the presence of new romantic partnerships. Widow(er)s who were highly dependent on their late spouse reported fewer negative interactions with their adult children following the death, suggesting that children may play a compensatory role as their parent adapts to widowhood (Carr & Boerner, 2013a). New romantic partnerships also bear on older widowers’ relationships with adult children in complex ways. Widowers who are dating report declines in the quality of their relationship with children, especially if the relationship was already strained prior to the death. However, men who had close ties with their children prior to loss grew even closer once they started dating (Carr & Boerner, 2013b). This work is consistent with family systems theories, which underscore the interdependence of family ties. In particular, these analyses reveal how widowed persons’ adaptations are affected not only by ties within the nuclear family but also ties with late spouses and romantic partners who may become future spouses.

Older bereaved spouses also reap benefits from extending their ties beyond children and romantic partners. Friendships, especially those that are easily accessible, increase widow(er)s’ emotional well-being (de Vries, Utz, Caserta, & Lund, 2014). Community engagement, especially volunteering, provides older widow(er)s a regular routine, a sense of purpose, and a network of peers with whom to socialize. Longitudinal analyses of HRS data show that volunteering a minimum of 2 hours a week diminishes older widow(er)’s loneliness symptoms (Carr, Kail, et al., 2017). Older widow(er)s’ selective engagement with kin and nonkin relationships may partially compensate for the loss of their spouse and may meet short-term and longer term emotional needs, consistent with a core theme of socioemotional selectivity theory (Carstensen, 1992).

Gray Divorce.

Only 15% of persons aged 65 years and older today are divorced, yet this proportion is projected to rise as the large Baby Boomer cohort reaches old age (Sweeney & Raley, 2020). Rates of gray divorce, or marital dissolutions among persons aged 50 years and older, more than doubled between 1990 and 2010 (Lin, Brown, Wright, & Hammersmith, 2016). This steep rise reflects factors implicated in divorce at earlier ages, including greater cultural acceptance of divorce and women’s heightened economic independence. Factors specific to older adults also have contributed. With increasing life span, midlife adults may recognize that they have many years ahead and are choosing not to spend them in an unsatisfying marriage. Others wait until their children have left home to divorce. Risk factors for gray divorce are similar to those identified for divorce earlier in life; rates are higher among younger rather than older adults, Blacks rather than Whites, and high school versus college graduates (Lin et al., 2016). Remarriages and shorter lived marriages are more likely than first and longer duration marriages to dissolve (Brown & Lin, 2012).

Few population-based studies have documented the implications of gray divorce for older adults’ well-being; rather, most studies document the long-term effects of dissolutions that happened years earlier. However, qualitative research describes how older adults adapt to gray divorces. Crowley (2019) carried out 80 in-depth interviews with recently divorced older adults and found that common concerns included loneliness and financial worries, whereas positive consequences included higher levels of overall happiness, a feeling of “liberation,” and enhanced independence and freedom. Consistent with studies of earlier life divorce, women experienced more dramatic drops in economic well-being, but men experienced a “social penalty.” Because men typically have weaker social networks prior to divorce, they are especially vulnerable to the loss of friendships and strained parent–child ties, which may reduce the availability or willingness of adult children to provide care or support. These results are consistent with a core theme of family systems theory—the interdependence of family members—and illustrate how the stressors of one person can spill over and restructure relationships with other family members.

Remarriage.

Analyses of HRS data estimate that 20% of women and 25% of men repartner following later-life dissolutions, with women more likely to remarry and men more likely to cohabit (Brown et al., 2019). Remarriage can mitigate the financial and psychosocial strains associated with late-life marital dissolution. For example, remarriage may enrich the lives of divorced and widowed persons by diminishing their symptoms of depression and loneliness (Carr & Springer, 2010). Persons in first versus higher-order marriages generally report comparable marital quality (Cooney, Proulx, & Snyder-Rivas, 2016).

Nonmarital Unions

Remarriage is not the only pathway to a new partnership. Older adults are increasingly living with a romantic partner without legally formalizing the union (cohabitation), maintaining an exclusive romantic relationship yet living separately in their own homes (living apart together), or dating without the promises of coresidence or exclusivity.

Cohabitation.

The number of older cohabitors has increased dramatically during the past 2 decades. Of persons aged 50 years and older, 14% are in a cohabiting union, with numbers increasing from 950,000 in 2000 to more than 4 million in 2016 (Brown & Wright, 2017). Cohabitation is more common than remarriage among divorced and widowed adults aged 50 years and older, especially men (Brown et al., 2019). This increase reflects a greater acceptance of nonmarital sexual relations and rising rates of gray divorce, as cohabitations are more common following divorce than widowhood. Most older cohabitors report the same sense of relationship commitment and satisfaction as their remarried peers, and their unions are long lasting, with a mean duration of 10 years. Most end due to death rather than a break-up (Brown, Bulanda, & Lee, 2012).

The benefits of cohabitation are generally comparable to marriage, with some indications that it may be a preferred relationship type among some older adults. A recent study found older male cohabitors reporting even better psychological well-being than their married peers (Brown & Wright, 2017). Unlike remarriage, cohabitation generally allows an older adult to continue receiving their late spouse’s pension benefits. Some divorced older adults, especially women, may prefer cohabitation to marriage as it carries less rigid gendered expectations regarding household roles such as spousal caregiver (Noël-Miller, 2011). Cross-national comparisons suggest that the benefits of cohabitation are comparable to marriage in nations where it is a culturally normative and legally protected institution (Stavrova, Fetchenhauer, & Schlösser, 2012). For example, in the Netherlands, registered cohabitors have the same rights as married couples when it comes to inheritance and pension receipt following a partner’s death, which may lessen the financial anxieties associated with being single (Perelli-Harris & Gassen, 2012). Cultural norms and social policies that afford cohabitating unions similar protections, as marital unions may further heighten the desirability of cohabitation.

Living Apart Together (LAT).

Some older adults choose to live apart from their committed long-term romantic partner, with each maintaining their own home. An estimated 7% of older adults in the United States are in LAT relationships. This arrangement affords the autonomy of living in one’s own home and the flexibility to choose when to spend time with one’s partner (Connidis, Borrell, & Karlsson, 2017). LAT relationships spare older adults from the strains of relocating and combining households, the legal complexities regarding co-owning a home with their partner, and complications for their offsprings’ inheritance. As such, LATs are especially preferable among older adults with coresidential children or grandchildren (de Jong Gierveld & Merz, 2013). Cross-national analyses show that persons in LATs report lower relationship satisfaction than their married peers, with the largest gap in nations with less cultural, political, and social support for intimate partnering outside of marriage (Tai, Baxter, & Hewitt, 2014).

Dating.

About 15% of unmarried older adults are currently dating, with rates considerably higher among men than women (27% vs. 7%), reflecting the skewed sex ratio among older adults (Brown & Shinohara, 2013). Older men and women do not differ in their desire to date, with some women preferring a casual relationship to marriage as it allows them to maintain their autonomy and avoid providing care to their aging partners (Brown & Wright, 2017). Daters, when compared with nondaters, have advantages, including higher levels of education, assets, better health, and more social connections; these advantages may make them more attractive partners (Brown & Shinohara, 2013). Little is known about the impact of dating on older adults’ well-being, nor the ways that dating relationships might progress into LAT relationships, cohabitations, or marriage—an important query for future research.

Lifelong Singlehood

Lifelong singlehood is rare; slightly more than 5% of adults aged 65 years and older in 2016 had never married (U.S. Census Bureau, 2017). However, this proportion is projected to increase to as many as 11% (and nearly 25% of older Black women) by 2060, raising concerns about who will provide care and financial support for these adults growing old alone (Tamborini, 2007). Current cohorts of older single women have physical and mental health on par with their married counterparts, reflecting their tendency to seek out and create meaningful relationships and families by choice (DePaulo, 2013). However, qualitative evidence suggests that they bear the emotional labor of having to explain and account for their lifelong singlehood, especially because of the centrality of marriage, children, and grandchildren to the identities of current cohorts of older women (Band-Winterstein & Manchik-Rimon, 2014).

Never-married older men report poorer physical and mental health, fewer friendships, less social support, and an elevated mortality risk relative to their married, divorced, and widowed peers; these patterns are partially accounted for by negative social selection associated with unmarried men including fewer socioeconomic resources and poorer early life health (Koropeckyj-Cox, 2005). Lifelong singlehood is a major risk factor for older men’s and women’s economic insecurity; in 2014, never-married men and women aged 65 years and older were roughly five times as likely as their married counterparts to live beneath the federal poverty line (23% and 21% vs. 4.5%, respectively). This disparity is largely a product of single adults’ reliance on only one income during the life course, without the economies of scale and financial partnerships enjoyed by couples. It also reflects the structure of Social Security benefits, which rests solely on single workers’ own income—despite the fact that they consistently earn less than their married counterparts (Lin, Brown, & Hammersmith, 2017). This financial divide is projected to widen further as the Baby Boom (born 1946–1964) and Generation X (born 1965–1980) cohorts enter old age. For these cohorts, marriage rates are much higher for college-educated persons and Whites relative to less-educated persons and blacks, with these gaps widening over time (Schwartz & Han, 2014). For future cohorts of older adults, lifelong singlehood may be both a cause and consequence of economic disadvantage.

Intergenerational Relationships

Similar to marriage and intimate partner relationships, intergenerational relations are critical in meeting older adults’ socioemotional and practical needs. Research has traditionally focused on parent–child relationships and emphasized the quality and closeness of these ties, the implications thereof for older adults’ well-being, and the inter- and intragenerational exchange of instrumental, social, and financial support. During the past decade, this research has advanced through the use of dyadic and within-family differences (WFD) approaches that incorporate the perspectives of both generations. Researchers also have moved beyond studying biological parent–child relationships and have expanded their foci to include increasing numbers of stepchildren and step-grandchildren, a function of divorce and remarriage among older adults and their adult children. Studies of intergenerational relationships among aging families is informed by family systems approaches (Broderick, 1993). As such, it illustrates the interdependent, interconnected, and dynamic nature of older adults’ family ties and shows how these ties are consequential for late-life well-being.

Parent–Child Relationships

Dyadic and WFD analyses show that parent–child relationship quality is at least partly in the eye of the beholder. Dyadic analyses document generational asymmetries, where children are more likely than parents to notice, report, and respond to negative exchanges such as arguments (Fingerman, Huo, & Birditt, 2020). Adult children report giving more emotional, financial, and instrumental support than their parents report receiving, whereas older parents report giving more practical support and less advice than their adult children report receiving (Kim, Zarit, Eggebeen, Birditt, & Fingerman, 2011). These data support self-enhancement theory, such that people view their own behaviors positively (perhaps excessively so) to enhance their sense of self (Krueger, 1998). Children are motivated to view themselves as dedicated caregivers to their parents, whereas parents are motivated to see themselves as a source of practical support to their children without being intrusive or overbearing with advice.

Dyadic parent–child data capture just one child’s views, yet children may have considerably different experiences even within the same family. During the past decade, researchers have made strides in collecting family-level data, from both mother and father and from or about multiple children. These data are analyzed using WFD approaches and are especially valuable for understanding the support networks of older adults (Suitor et al., 2017). Studies using these complex data have found, for instance, that older fathers report lower ambivalence toward daughters than sons, whereas mothers report less ambivalence toward sons than daughters (Pillemer, Munsch, Fuller-Rowell, Riffin, & Suitor, 2012). WFD approaches also have been instrumental for studying once-verboten topics such as parental favoritism. Researchers have found that mothers tend to show favoritism toward children with whom they hold similar values, and thus have more close-knit and less conflictual relationships with them (Suitor, Gilligan, Johnson, & Pillemer, 2013). These data also reveal who older adults select as their preferred caregiver, with most naming the child with whom they are closest emotionally.

Dyadic and family-level data have enabled researchers to address novel questions regarding the impact of parent–child relations on older adults’ well-being. Historically, research relied on coarse measures assessing a parent’s overall appraisal of their relationship with all children or the characteristics of one randomly selected child. However, WFD studies have found that older parents are “only as happy as their least happy child,” such that having one adult child with significant personal problems undermines parental well-being, yet having one child with considerable successes does not enhance well-being. However, parents with multiple successful children enjoy increased well-being (Fingerman, Cheng, Birditt, & Zarit, 2012). These results demonstrate the importance of considering multiple family members, as parental well-being may be differently affected by their ties with each particular child.

Moving Beyond Biological Children

Demographic trends during the past 5 decades, including rising rates of divorce, remarriage, and reconfigured families, mean that growing numbers of older adults are in step- or “blended” families (Sweeney & Raley, 2020). Stepfamilies are formed when one or both partners have children from previous relationships. For remarriages in which the wife is of childbearing age, the couple also may have children together, creating blended families with both biological and stepchildren. HRS data show that 40% of U.S. middle-aged and older parents have stepchildren (Lin, Brown, & Cupka, 2018), whereas 20% of grandparents have at least one step-grandchild, whether through own, spouse’s, or an adult child’s remarriage (Yahirun, Park, & Seltzer, 2018).

An important yet unresolved question is whether stepchildren and biological children differ with respect to the care they provide their (step)parents. Evidence suggests that stepchildren may not provide the same level of care and support as biological children, especially when family ties are fraught (Sherman, Webster, & Antonucci, 2013). However, one analysis of HRS data suggested that linkages between family structure and parental well-being are complex and differ by gender (Pezzin, Pollack, & Schone, 2013). Older adults with stepchildren only had shorter life spans, higher rates of institutionalization, and earlier onset of disability than those with biological-only children. However, the results were more complex for blended families. The fathers of blended families had poorer longevity than fathers of biological-only children, whereas the mothers of blended families evidenced greater longevity and later onset disability than mothers of biological-only children. Women may be better able to negotiate high-quality relationships with their own and their partner’s children, leading to a greater overall network of adult children to draw on during old age. Remarried men, by contrast, might lose some support their biological children would have provided given the well-documented negative effect of divorce on men’s social ties to children (Noël-Miller, 2013).

Grandparent–Grandchild Ties

Research on grandparent–grandchild ties has expanded in the past decade. Increases in life expectancy have resulted in a greater number of families with three or more generations alive at the same time (Margolis & Wright, 2016). Rising age at first marriage and first birth mean that people are making the transition to grandparenthood at older ages, although this transition comes earlier, as young as one’s 30s or 40s, among people from disadvantaged economic backgrounds, Blacks, and Latinos—relative to wealthier persons and Whites (Skopek & Leopold, 2017). Grandparents today are older, on average, than previous generations, yet they are playing an increasingly outsized role in caring for grandchildren.

Grandparent–grandchild relationships vary across demographic subgroups, with Blacks and those of lower socioeconomic status taking on more time- and labor-intensive roles, often serving as custodial caregiver when their own children struggle to care for their offspring. Older adults with greater economic advantages tend to provide more sporadic supports such as babysitting, transportation, and recreational activities (Harrington Meyer & Kandic, 2017). Policy contexts also shape theses exchanges; cross-national analyses of 11 European nations show that grandparents are more likely to provide intensive care in nations with little availability of and weak cultural support for formal child care services (Di Gessa, Glaser, Price, Ribe, & Tinker, 2015).

Census data show that of the 65 million grandparents in the United Sates today, 10% live with at least one grandchild (Casper et al., 2016). Multi-generation households are more common among blacks and immigrants, relative to whites and U.S-born older adults (Casper, Florian, Potts, & Brandon, 2016). Of the 7 million households with coresidential grandparents and grandchildren, 60% were grandparent-headed households, and one third had no parent present at all (i.e., “skip generation”). Historical analyses confirm that grandparents have long provided support for younger generations, dating back to the 19th century, and that multigeneration households provide support to whichever generation is most in need (Ruggles, 2011). Most multigeneration households form due to needs of the child or grandchildren, rather than the older adult, consistent with a core theme of family systems theory, that the experiences of one generation trigger reactions and responses from older and younger generations.

For most older adults, the grandparent–grandchild relationship is a source of joy and optimism, with participants in a recent qualitative study describing it as more enjoyable than raising their own children (Mansson, 2016). The benefits of grandparenthood are generally greater for maternal grandparents than paternal grandparents in part because adult daughters play an active role in maintaining and fostering multigenerational ties (Fingerman, 2004). Grandparents, in turn, provide benefits to their grandchildren and other family members. Financial resources from grandparents, irrespective of parental resources, are associated with the educational outcomes of grandchildren, suggesting that grandparent resources can serve either as a substitute for or supplement to parental resources (Deindl & Tieben, 2017). Investments from grandparents also may reap long-term benefits; longitudinal analyses show that grandparents who regularly assisted with grandchild care received more instrumental and emotional support from their children 13 years later (Geurts, Poortman, & Van Tilburg, 2012).

However, when grandparents do more than occasional babysitting or recreational outings with their grandchildren, they are more likely to report burden and poorer outcomes for themselves and the grandchild (Arpino & Bordone, 2014; Baker & Mutchler, 2010). For coresidential grandparents, the strains can be immense. They are often distressed by the underlying problems that rendered their children unable to care for their offspring, such as addiction, imprisonment, financial insecurity, or premature death. Because the custodial role frequently is entered into by those already facing physical or financial challenges, the added physical and financial demands of raising a child are particularly daunting, consistent with stress accumulation perspectives (Hayslip, Fruhauf, & Dolbin-MacNab, 2017). Roughly 21% of custodial grandparents live beneath the poverty line, 25% have a disability, and 40% have provided this care for more than 5 years (Ellis & Simmons, 2014).

Family Caregiving

Romantic partnerships and intergenerational social ties are critical sources of support in later life, with the demands on family caregivers intensifying as rising numbers of older adults survive into their 80s and 90s and experience the physical and cognitive declines that often accompany extreme old age (Hebert, Weuve, Scherr, & Evans, 2013). Caregiving remains a gendered task, with wives and daughters providing the most care, although siblings, sons, husbands, and grandchildren also commonly care for aging relatives.

One of the most significant issues facing contemporary society is the shortage of family caregivers, a challenge that will only escalate as the large Baby Boom cohort reaches old age (Redfoot, Feinberg, & Houser, 2013). This shortage is attributable primarily to population trends including declining fertility rates during the past 5 decades, such that older adults have fewer children on whom they may rely. Economic and geographic mobility mean that adult children may live far from their parents, and dramatic changes in women’s roles mean that daughters and sisters—who were once presumed to be a readily available source of care—may be unavailable to provide care or must do so while juggling caregiving with the demands of paid employment (National Academies of Sciences, Engineering, & Medicine, 2016).

Despite the ubiquity of family caregiving, it is neither an easy nor intuitive task. Family caregivers are typically unpaid and perform increasingly technical aspects of care, such as administering medications, giving injections, and cleaning wounds—often without adequate support or training (Polenick, Leggett, & Kales, 2017). A perennial concern of both researchers and practitioners is caregiver strain, or the detrimental physical, emotional, and financial impacts of caring for a family member. (For a review of the positive impacts of caregiving, see Cho, Ory, & Stevens [2016].) Caregiving is a chronic stressor that undermines one’s mental health by overwhelming one’s coping resources and erodes one’s physical health through gradual “wear and tear” on the immune system (National Academies of Sciences, Engineering, & Medicine, 2016). Consistent with stress process models, whether and how caregiving affects late-life well-being varies based on the nature of the care, including its duration, intensity, and tasks performed as well as the quality of one’s relationship with the care recipient both currently and earlier in the life course (National Academies of Sciences, Engineering, & Medicine, 2016). Longitudinal analyses found that caregiver burden was worse among those caring for a parent who had been abusive or neglectful roughly 50 years earlier, when the caregiver was an adolescent (Kong & Moorman, 2016). Conversely, family caregivers who had earlier engaged in reciprocal family exchanges with kin showed less distress when providing care to kin in later life (Leopold, 2012).

Family caregiving takes a particular toll on women because their episodes tend to be more time and labor intensive. Women spend an average of 6.1 years—nearly 10% of their adult lives—providing care to others, whereas men spend just 4.1 years or 7% of their adult lives (National Academies of Sciences, Engineering, and Medicine, 2016). This gap partly reflects women’s greater tendency to provide spousal care, which is particularly time intensive given its coresidential nature. Women also are more likely to perform tasks that must be done regularly and around the clock such as personal care, whereas men are more likely to provide assistance with occasional tasks such as yard work. However, a recent comparison of same- and different-sex marriages found that gender disparities in elder care were limited to heterosexual marriages; wives bore the brunt of caring for parents and in-laws, receiving sparse support from their husbands (Reczek & Umberson, 2016). In contrast, men and women in same-sex couples were more cooperative and equitable in providing emotional support and time to their parents and in-laws.

Although caregiving often is presented as an example of how families respond to the needs of its oldest members, mounting research demonstrates that older adults are increasingly the provider rather than the recipient of such care. Older adults play an active role in caring for grandchildren, especially when they coreside or have custody. Older adults also are a major source of care to their peers and even their elders; persons aged 75 years and older are one of the most rapidly growing groups of caregivers in the United States, numbering 3 million. Half are caring for a spouse, whereas others are caring for their siblings, friends, neighbors, or even a superannuated parent. The typical 75-year-old or older caregiver has been in the role for 6 years, spending an average of 34 hours per week on care—10 hours more per week than younger caregivers (National Alliance for Caregiving, 2015). Older caregivers are a vulnerable population as they often must manage co-occurring stressors including their own health concerns, yet have relatively few sources of support; the friends and siblings that they might otherwise turn to also are very old and may be too ill to pitch in. The distinct needs and adaptations of older caregivers is a critical issue that warrants attention from researchers and policy makers in the next decade (Wolff, Spillman, Freedman, & Kasper, 2016).

Conclusion and Future Directions

The lives of older families in the 21st century are a product of demographic changes that have unfolded during the past half century, including rising age at first birth and first marriage, the rise (and eventual plateau) of divorce rates and consequently reconfigured families, and increasing life expectancy. Interpersonal dynamics within families also have evolved, reflecting sweeping cultural and technological changes, including a blurring of gendered boundaries within and outside the family, greater acceptance of non-marital and same-sex romantic partnerships, and advances in medical technologies that have enabled older adults to live more active, engaged lives than their predecessors. Yet this research also underscores the centrality of the family in providing care and support to growing numbers of persons who are chronically ill and disabled, a function of unprecedented advances in life expectancy.

Against this backdrop, contemporary research on later-life families has flourished and its scope widened due in part to the expansion of data resources, refinement of analytic and theoretical approaches, and heightened public interest in how families are sustaining the health and well-being of their older members. The research highlighted in this article demonstrates the importance of incorporating multiple actors’ perspectives, whether spousal dyads or parents and children, in studies of relationship quality and dynamics. This attention to within-couple and intergenerational dynamics has enabled researchers to reevaluate and challenge the assumptions of theoretical perspectives including family systems, socioemotional selectivity, stress process, and gender-as-relational models. Contemporary research also exploits rich self-reported and biomarker indicators of health, enabling researchers to identify pathways linking social relationships to late-life well-being. This research extends beyond the social control and support functions of social ties and sheds light on the ways that sexual relationships and physiological responses to relationship-related stress affect later-life health and well-being.

We conclude by suggesting three areas of research that may be especially important in the next decade, as the population aged 65 years and older grows increasingly diverse and as families continue to redefine their roles and responsibilities to support their older members. First, research on ethnic minority and immigrant older adults has expanded somewhat during the past decade, yet this represents just the tip of the iceberg. By 2060, non-Hispanic Whites are projected to make up roughly half of the older population, whereas Blacks will account for 12%, Hispanics 22%, Asians 9%, and persons who identify otherwise 3%. Migration patterns have shifted so that nonnative born older adults in future cohorts may have experiences that differ from their predecessors. More than half of non-U.S.-born older adults today migrated here prior to 1970, mostly from Europe and Canada, whereas more recent immigrants have come from Latin America, Asia, and to a lesser extent, Africa (Abdul-Malak & Wang, 2016). Immigrant older adults may face challenges including linguistic isolation, attitudes and expectations regarding family caregiving that conflict with those of their children and grandchildren, and obstacles to maintaining social ties with kin who remained in their country of origin, especially in periods of restrictive immigration policy (Bryceson, 2019).

Second, a critical population for future research is persons aging without families; rising numbers of older adults in the United States and worldwide do not have a spouse or romantic partner, have outlived both parents, and are either childless or have children who live far away (Carr, 2019b). An estimated one in five older adults are “elder orphans” or at risk of becoming one in the coming decades (Carney, Fujiwara, Emmert, Liberman, & Paris, 2016). Isolated older adults are especially vulnerable to age-related challenges including loneliness, abuse and neglect, and lack of a readily available pool of caregivers when faced with illness. Future research will need to explore the creative adaptations employed by those growing old without family, with attention to the ways that public and community supports might be deployed to meet their basic needs.

Finally, researchers have just scratched the surface in exploring how evolving information and communication technologies, robotics, and artificial intelligence technologies might enhance (or undermine) the well-being of older adults and their families. The current and projected caregiver shortage has intensified the need to develop and evaluate such innovations. “Smart home” technologies assist physically remote caregivers monitor their loved ones’ safety and contact emergency service providers as needed. Telemedicine applications can help caregivers to monitor older adults’ symptoms and deliver health services efficiently (Czaja, 2016). However, the effectiveness of and older adults’ receptiveness to these technologies is unclear, with some evidence suggesting that they cannot provide the emotional and personal care that older adults desire (Lehoux & Grimard, 2018). Assistive technologies may be necessities for persons aging alone who lack proximate sources of support (Khosravi, Rezvani, & Wiewiora, 2016), and information and communication technologies such as video chat platforms are especially important for older immigrants whose kin live overseas or rural older adults whose families are geographically distant.

The family lives of older adults have undergone dramatic transitions in the 21st century and will continue to evolve in response to demographic, technological, and cultural changes. Researchers have made critical strides in documenting the pace, sources, and consequences of these changes, yet public policies that meet the needs of older adults and their families lag woefully behind (Carr, 2019a). For Baby Boomers and the cohorts that follow, divorce rather than widowhood is projected to be the main path out of marriage, yet Social Security benefits are still structured to penalize divorced persons (especially those who had short-term marriages) relative to their widowed counterparts (Lin, Brown, et al., 2017). Growing numbers of older adults with chronic or terminal illness lack a residential caregiver and may be ineligible for the home-based services offered by some hospice providers (Aldridge Carlson, Barry, Cherlin, McCorkle, & Bradley, 2012). Most older adults are cared for by family members, yet only a small fraction of working-age caregivers are eligible for an unpaid family leave based on the conditions of the Family and Medical Leave Act, and only 12% of private-sector employees receive paid family leave (Department of Labor, 2015). The United States stands out as the only wealthy nation to lack a coordinated system of providing paid caregiving leave (Osterman, 2017). These concerns will only intensify in the coming decades, and policy makers will require rigorous research to guide evidence-based practices that sustain the well-being of older adults and their families.

Appendix

Selected Data Resources for Studying Later-Life Families in the United States

Data set	Study summary	Respondents	Example publications	Website
Changing Lives of Older Couples (CLOC)	Prospective study of spousal bereavement in later life in the United States. Interview preloss and at 6-, 18-, and 48-month follow-ups	1,545 married persons at baseline (including 423 spousal dyads); 250 of whom become bereaved	Carr & Boerner, 2013a, 2013b	https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/3370
Disability and Use of Time (DUST)	Daily diary and survey study of married couples aged 60+ years in the United States	DUST sampled more than 500 older couples in the 2009 PSID. Followed up in 2013	Carr et al., 2016, 2017	https://psidonline.isr.umich.edu/Guide/Brochures/PSID-Aging.pdf
Family Exchanges Study (FES)	Survey of three generations of U.S. family members in 2008 and 2013	633 families, with nearly 2000 family members	Fingerman et al., 2012; Suitor et al., 2017	https://sites.utexas.edu/adultfamilyproject/research/family-exchanges-study/
Health and Relationships Project	Survey of same- and opposite-sex married couples aged 35 to 65 years in Massachusetts in 2014–2015	More than 800 individuals (~400 couples) include male same-sex, female same-sex, and different-sex marriages	Reczek & Umberson, 2016; Umberson et al.,2018	https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/37404
Health and Retirement Study (HRS)a	Multiwave study of older adults in the U.S. started in 1992, with attention to finances and health	Persons born 1931–1941 and their spouses. Subsequent cohorts replenished every 2 years at age 51	Brown et al., 2019; Lin et al., 2018	http://hrsonline.isr.umich.edu/
Longitudinal Study of Generations (LSOG)	Multiwave survey of four generations, focused on values and affiliation	300 four-generation families in southern California, starting in 1971 with multiple follow-ups	Suitor et al., 2017	http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/22100
Longitudinal Study of Generations (LSOG)	Multiwave survey of four generations, focused on values and affiliation	300 four-generation families in southern California, starting in 1971 with multiple follow-ups	Suitor et al., 2017	http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/22100
Midlife Development in the United States (MIDUS)	Psychosocial and biological factors in health, among U.S. adults born 1920–1975 interviewed at three waves since 1995	More than 3,000 adults, 950 of their siblings, and nearly 1,000 twin pairs	Donoho et al., 2013; Miller-Martinez et al.,2014	http://midus.wisc.edu/
National Health and Aging Trends Survey (NHATS)	Sample of U.S. adults ages 65+, interviewed in 201 land 2013; sample replenished in 2015	More than 8,500 older adults; with caregivers interviewed in companion National Survey of Caregivers (NSOC)	Wolff et al., 2016	https://www.nhats.org/
National Social Life, Health, and Aging Project (NSHAP)	A three-wave survey of adults born 1920–1947 and their spouses, focused on biosocial factors in health and aging	More than 3,000 interviewed at 2011 baseline, with spouse/partner interviews at subsequent waves in 2013 and 2015	Cornwell & Laumann, 2011; Schafer et al., 2017	http://www.norc.org/Research/Projects/Pages/national-social-life-health-and-aging-project.aspx
Panel Study of Income Dynamics (PSID)	The original 1968 sample included 18,000 individuals in 5,000 families. All children of original sample tracked over time	Nearly 70,000 people have participated in the PSID, and as many four generations are represented	Yahirun et al., 2018	http://psidonline.isr.umich.edu/
Wisconsin Longitudinal Study (WLS)	Multiwave study tracking a random one-third sample of all high school seniors in Wisconsin in 1957	10,317 high school graduates, a randomly selected sibling, and spouse. Data linkage to high school friends	Kong & Moorman, 2016	http://www.ssc.wisc.edu/wlsresearch/
Within Family Differences Study (WFD)	Mixed-methods study of 566 U.S. mothers aged 65–75 years and their adult children interviewed in 2001–2003, and 2008–2011	Baseline interviews with 566 mothers and 774 adult children and 420 mothers and 835 children at follow-up	Suitor et al., 2017	https://web.ics.purdue.edu/~jsuitor/within-family-differences-study/

^*Note: Parallel studies conducted in China (China Health and Retirement Survey), Costa Rica (Costa Rican Longitudinal and Health Aging Study), England (English Longitudinal Study of Aging), India (Longitudinal Aging Study in India), Ireland (Irish Longitudinal on Ageing), Japan (Japanese Study of Aging and Retirement), Korea (Korean Longitudinal Study of Aging), Mexico (Mexican Health and Aging Study), and 20 European nations and Israel (Survey of Health, Ageing and Retirement in Europe) are available at https://g2aging.org/