Abstract
Objectives
Our study compares two types of later life marital dissolution that occur after age 50—divorce and widowhood—and their associations with repartnership status (i.e., remarried, cohabiting, or unpartnered).
Method
We used data from the Health and Retirement Study to provide a portrait of later life divorce and widowhood for women and men. Next, we tested whether marital dissolution type is related to women’s and men’s repartnered status, distinguishing among remarrieds, cohabitors, and unpartnereds, net of key sociodemographic indicators.
Results
Divorcees are more often repartnered through either remarriage or cohabitation than are widoweds. This gap persists among women net of an array of sociodemographic factors. For men, the differential is reduced to nonsignificance with the inclusion of these factors.
Discussion
Later life marital dissolution increasingly occurs through divorce rather than widowhood, and divorce is more often followed by repartnership. The results from this study suggest that gerontological research should not solely focus on widowhood but also should pay attention to divorce and repartnering during later life.
Keywords: Cohabitation, Divorce, Marriage, Remarriage, Widowhood
Marital dissolution commonly occurs in later life through spousal death. However, a growing share of marital dissolutions after age 50 are voluntary. The gray divorce rate, which describes divorces that occur after age 50, has more than doubled since 1990. Nowadays, one in four people getting divorced in the United States is older than 50 years (Brown & Lin, 2012). The rise in gray divorce arguably portends an increase in later life marital dissolution, which can occur through either death of a spouse or divorce. It also raises important questions about how individuals who experience later life marital dissolution via divorce compare with those who experience dissolution via widowhood, a topic that has received little attention in the literature (Kitson, Babri, Roach, & Placidi, 1989).
Basic comparisons by current marital status are inadequate for two reasons. First, they ignore duration in the current status, confounding long-term divorcees with gray divorcees, for example. Long-term divorcees may differ from recent divorcees in terms of postdivorce adjustment and well-being. Moreover, long-term divorcees are a select group who have remained divorced and not repartnered. Thus, we focus on later life marital dissolution. Second, comparisons by current marital status disregard the differential propensity of divorcees and widow(er)s to repartner. Repartnering is more common following divorce than widowhood, underscoring the importance of accounting for the type of marital dissolution (Brown, Bulanda, & Lee, 2012; Vespa, 2012), but this conclusion rests on research that has lumped together all singles, regardless of the timing of dissolution. Current repartnership status—whether remarried, cohabiting, or unpartnered—is partially contingent on marital dissolution type. Thus, to understand how the experiences of gray divorcees and widow(er)s compare, we must account for repartnership status. Marital dissolution and subsequent repartnership are intertwined.
Our goal is to provide a national portrait of individuals who experience marital dissolution after age 50, comparing patterns for those who are divorced versus widowed, and taking into account subsequent repartnership status (i.e., remarried, cohabiting, or unpartnered). This approach acknowledges the differential selection out of marital dissolution and into repartnership, which is necessary to ensure appropriate, rigorous comparisons between the divorced and widowed in later life (Kitson et al., 1989). Using data from the Health and Retirement Study (HRS), we examine how key sociodemographic and well-being indicators are associated with one’s repartnered status, distinguishing among remarrieds, cohabitors, and unpartnereds, by marital dissolution type (i.e., widowhood vs divorce). Dating and living apart together (LAT) relationships are not included because the HRS only measures coresidential partnerships. We pay particular attention to gender differences as women are especially likely to eschew repartnership. They not only confront an increasingly unbalanced sex ratio as they age but also they are less interested in forming a new union (Talbott, 1998; Watson & Stelle, 2011). Unlike prior studies on marital status in later life that have tended to rely on contemporaneous measures, our approach is more dynamic, bridging both later life marital dissolution and subsequent repartnership status to provide fresh insights on how they are related to demographic, economic, and health factors. As later life marital dissolution more often occurs through divorce, it is important to understand whether and how it differs from widowhood (Kitson et al., 1989).
Background
The marital status composition of older adults (i.e., individuals older than 50) is shifting as fewer are either married or widowed than a generation ago (Lin & Brown, 2012; Manning & Brown, 2011). This trend reflects the transformation of American family life that has occurred across the life course. Of particular note for older adults is the changing pathways to later life marital dissolution. In 2010, more than 643,000 individuals experienced divorce after age 50 (Brown & Lin, 2012). By comparison, more than 1.2 million married individuals older than 50 years became widowed. In 1980, just 84,000 people experienced divorce after age 50 whereas about 1 million transitioned to widowhood after age 50 (authors’ calculations). As later life marital dissolution less often occurs through widowhood and increasingly results from divorce, it is vital that researchers address whether and how older adults differ according to which type of marital dissolution they experience.
Marital Dissolution
Gray divorce has accelerated in recent decades, despite the stability of the overall U.S. divorce rate during this period (Brown & Lin, 2012; Kennedy & Ruggles, 2014). This rise reflects shifting cultural meanings of marriage, which is now more individualized. Americans expect to achieve self-fulfillment through marriage (Cherlin, 2009). Today’s older adults enjoy longer life expectancies which decrease their willingness to remain in subpar marriages (Bair, 2007). Additionally, most wives are no longer economically dependent on their husbands, and thus for couples who are dissatisfied in their marriage, divorce is now a viable option (Bair, 2007). Since 1990, the divorce rate more than doubled for individuals in middle age (50–64 years) and tripled among older adults (65+ years) (Brown & Lin, 2012). Gray divorce is more common among middle-aged than older adults, non-Whites than Whites, and those with a high school compared with a college degree. Marriage order and marital duration are closely associated with gray divorce. The gray divorce rate is 2.5 times higher for those in remarriages versus first marriages and is highest among those with the shortest marriages (Brown & Lin, 2012).
To the best of our knowledge, no published studies address how divorcees compare with those who experience widowhood after age 50. There is an extensive literature on the disadvantages associated with widowhood, yet the comparison group usually is the continuously married (Han, Cichy, Small, & Almeida, 2014; Hughes & Waite, 2008; Hungerford, 2001; Lee & Carr, 2007; Lee & DeMaris, 2007; Sasson & Umberson, 2014; Williams, 2004). Spousal loss is related to declines in physical and mental health, particularly for men (Hughes & Waite, 2008; Lee & Carr, 2007; Lee & DeMaris, 2007; Williams, 2004). Widowhood also is associated with a reduction in one’s economic standard of living, especially among women (Hungerford, 2001). This narrow approach is consistent with the implicit assumption that has guided much of gerontological research on marital status and well-being equating marital dissolution with widowhood and ignoring divorce. Historically, this was a safe assumption, but new research on later life marital transitions cannot disregard marital dissolution that occurs through divorce. Moreover, it is important to distinguish recent gray divorces from divorces that occurred earlier in the life course because those who experienced divorce a long time ago may have since adapted to the transition. For this reason, time since dissolution (whether through divorce or widowhood) is a relevant factor.
Marital dissolution is a stressful life event, and thus whether it occurs through spousal death or divorce may be of negligible consequence (Kitson et al., 1989). However, it is possible that adaptation following marital dissolution is partially a function of the type of marital dissolution one experienced. On the one hand, divorce may have more deleterious effects than widowhood because widow(er)s have unique institutional protections such as Social Security and spousal retirement benefits that may buffer the disadvantage associated with marital loss (Lin, Brown, & Hammersmith, in press). On the other hand, divorce may have more benign effects than widowhood because the transition to divorce often involves a calculated decision-making process, at least for the initiator (Bair, 2007). Whereas widowhood is largely unpredictable and certainly involuntary, divorce may operate as a turning point (cf. Bair, 2007) that signals a fresh start and thus could be less harmful (particularly for the initiator). For these reasons, the type of marital dissolution—divorce or widowhood—may be differentially associated with repartnership status.
Repartnership After Marital Dissolution
Divorce results in two single people who might form new unions, whereas widowhood leaves just one person available to repartner. Thus, as gray divorce continues to climb and widowhood occurs less often, repartnering after marital dissolution is likely to become more common among older adults. This trend will be fueled further by the fact that age is negatively associated with both divorce and repartnering whereas widowhood rises with age. Marital dissolution is increasingly occurring among younger older adults (aged 50–64 years) and they are the group most likely to form a new union (Brown & Lin, 2012; de Jong Gierveld, 2004).
Only a few studies address the roles of the timing or type of marital dissolution, both of which may influence repartnership status. Repartnership following widowhood in the Canadian context was negatively associated with age at widowhood (Wu, Schimmele, and Ouellet, 2015). Repartnership after marital dissolution in the Netherlands was similarly likely for those who experienced dissolution through either widowhood or divorce, net of the timing of the dissolution (de Jong Gierveld, 2004). Those whose marriage dissolved after age 55 were much less likely to have formed a new relationship, especially a remarriage (de Jong Gierveld, 2004). Still, patterns of cohabitation and marriage are distinctive in the United States, and thus the results from these studies do not necessarily characterize the patterns and experiences of U.S. older adults.
In the United States, coresidential repartnership occurs through either cohabitation or remarriage. Cohabitation among older adults has grown rapidly just in the past decade or so, with the number of older cohabitors more than tripling since 2000 (Brown et al., 2012). Cohabitation offers many of the benefits of marriage and at the same time allows individuals to maintain financial autonomy that would not be possible within the legal confines of marriage. Cohabiting couples enjoy a close, intimate coresidential relationship akin to marriage, but are able to retain economic independence, allowing them to continue to receive a former spouse’s pension or Social Security benefits, for instance, and ensuring their assets can be bequeathed to their offspring (Chevan, 1996; Hatch, 1995).
Demographic Characteristics, Economic Resources, and Health
Both life course and demographic mechanisms are associated with later life repartnership status, reflecting the constraints and opportunities characterizing this life course stage (Brown et al., 2012; Bulcroft, Bulcroft, Hatch, & Borgatta, 1989; Vespa, 2012). A recent demographic portrait of older cohabitors, remarrieds, and unpartnereds offers some insights into how these groups compare in terms of demographic characteristics, economic resources, and health indicators (Brown, Lee, & Bulanda, 2006). However, the portrait is also limited in that it lumps together longstanding and recent repartnerships; the timing of the marital dissolution that preceded repartnership is not accounted for (nor was duration in the current state). Rather, the study simply describes older cohabitors, remarrieds, and unpartnereds, arguably heterogeneous groups. Older adult cohabitors are disproportionately divorced; widoweds are relatively unlikely to be cohabiting and more likely to be remarried. Demographic characteristics, including age, race, and children, are related to repartnership status. Cohabitors tend to be youngest, followed by remarrieds, and lastly unpartnereds. Older cohabitors are less often White compared with both remarrieds and unpartnereds. Older remarrieds more often have children than either cohabitors or unpartnereds (Brown et al., 2006).
The role of economic resources differs by gender. Cohabiting and remarried men appear similar, and both groups are advantaged economically compared with unpartnered men. For women, cohabitation does not afford the same level of economic benefits as remarriage. Cohabiting women are disadvantaged compared with their remarried counterparts, who are more likely to own their home and to have health insurance. In fact, cohabiting women fare worse than unpartnereds on some dimensions. For example, they are less likely to have health insurance yet more likely to work full time (Brown et al., 2006). Finally, good health is positively associated with remarriage (Vespa, 2012). Whether these patterns hold for older adults whose marital dissolution occurs after age 50 and whether they differ according to marital dissolution type are empirical questions that the present study addresses.
The Present Study
Our goal is to construct a detailed portrait of those who experience marital dissolution after age 50 through divorce or widowhood. We explicitly compare those who follow three pathways: remarriage, cohabitation, and singlehood. This approach allows us to document the composition of the population that experiences later life marital dissolution in a more nuanced and dynamic fashion by differentiating among those who went on to form a new union versus remained unpartnered. A basic comparison of those who are currently divorced or widowed ignores their counterparts who have formed a new union. Those who have repartnered following marital dissolution are likely to differ from those who remain single, and marital dissolution type is expected to shape the pathway to repartnership, which should be more common among the divorced than the widowed.
We investigate how remarried, cohabiting, and unpartnered individuals compare across several core domains that prior research has shown are related to union status in later life, including demographic characteristics (age, race-ethnicity, and presence of children), economic resources (education, employment, home ownership, and assets), and health (health insurance and self-rated health). Importantly, we account for key features of the marital pathway, namely dissolution type (i.e., divorce vs widowhood), the time since dissolution, and whether the dissolved marriage was a first or remarriage.
Our analyses assess whether marital dissolution type is related to repartnership status. We anticipate that divorce after age 50 is more likely to be followed by repartnership than is widowhood. Other characteristics of the marital dissolution pathway are also relevant. Time since marital dissolution should be positively related to repartnership. Marital dissolution of a remarriage (vs a first marriage) is hypothesized to be positively associated with repartnership. These associations should hold net of the demographic characteristics, economic resources, and health indicators that are related to repartnership status in later life (Brown et al., 2006).
All analyses are conducted separately for women and men because gender plays a pivotal role in shaping marital pathways in later life. Older women are disproportionately widowed, reflecting the tendency for women to marry men older than themselves and women’s longer life expectancy compared with men. At the same time, women are much less likely than men to be repartnered, not only because they face an unbalanced sex ratio that advantages men but also because many older women do not want to form a new union. They are often reluctant to assume the caregiving responsibilities that would be entailed by repartnership (Talbott, 1998). We anticipate that variation by marital dissolution type and repartnership status is more pronounced for women than men.
Method
Data came from the HRS, a longitudinal study of a nationally representative, continuous cohort of individuals born before 1960 and their spouses or partners. At the time of initial interview, respondents were aged 50 or older. The HRS covers a range of topics, including economics and retirement, health, and family life. Complete marital histories along with detailed information about cohabiting unions during the study period are available, making the data ideal for the purposes of this study. The HRS began interviewing in 1992 with a cohort of individuals born in 1931–1941, and reinterviews have been conducted every other year. In 1998, a cohort born before 1924 from the study of HRS’s Asset and Health Dynamics among the Oldest Old and a new cohort born in 1924–1930 (Children of the Depression Era) were combined with the HRS. Additionally, to ensure that the HRS remains representative of those older than 50 years, the sample is replenished by adding a new, younger cohort to the study every 6 years. The response rates for the baseline interviews of various cohorts hover around 70%–82% and roughly 90% or higher for follow-up interviews. For this study, sampling weights were used to adjust for the unequal probability of selection (for Blacks, Hispanics, and respondents living in Florida), nonresponse, and sample attrition (Ofstedal, Weir, Chen, & Wagner, 2011).
In total, the HRS sample consists of 38,006 respondents. We restricted the analytic sample to respondents interviewed in 2010 (n = 22,033), who had experienced a marital dissolution (divorce or widowhood) at age 50 or older (n = 4,918), who reported their marital status in 2010 (n = 4,917), and whose current marital status was consistent with their marital history, resulting in 4,886 respondents. Of these respondents, 1,465 had experienced a divorce and 3,421 had experienced widowhood after age 50.
Measures
Repartnership status differentiated among respondents who were currently remarried, cohabiting, or unpartnered (reference category) at the 2010 interview.
Marital dissolution pathway
Marital dissolution type captured the first marital dissolution after age 50: divorced or widowed (reference category). Years since marital dissolution was measured by subtracting the year in which the dissolution occurred from 2010. Dissolved remarriage was coded 1 if the dissolved marriage was a remarriage and 0 if it was a first marriage.
Demographic characteristics
Age was classified into one of four categories: 50–64 years (reference category), 65–74 years, 75–84 years, and 85 and older. For the multivariate analyses, age was coded in years and centered (at age 50). Race-ethnicity was composed of four categories: White (reference category), Black, Hispanic, and other race. Parenthood status was measured by any resident children (reference category), nonresident children only, and childless.
Economic resources
Education was an ordinal variable: less than high school, high school diploma (reference category), some college, and college or more. Employment distinguished among those who were working full-time (35+ hours per week) (reference category), part-time (less than 35 hours per week), and not in the labor force. Home ownership was coded 1 if respondent owned a home and 0 otherwise. Assets were gauged by five categories: in debt, $0–50,000 (reference category), $50,001–100,000, $100,001–250,000, and $250,001 or more.
Health
Health insurance was coded 1 for respondents who reported that they had private or public insurance and 0 for those who reported no health insurance. Self-rated health ranged from 1 (poor health) to 5 (excellent health).
Analytic Strategy
We began by establishing the prevalence of divorce versus widowhood separately for women and men across four age groups (50–64, 65–74, 75–84, and 85+) to document how these two types of marital dissolution unfold during middle-age and older adulthood. Next, we provided a descriptive profile of older women and men by marital dissolution type. This allowed us to establish repartnership patterns following later life marital dissolution. For women and men, we estimated the shares who were currently remarried, cohabiting, and unpartnered separately for those who experienced divorce versus widowhood after age 50. We also examined whether women’s and men’s demographic characteristics, economic resources, and health differ by marital dissolution type. Finally, we estimated multinomial logistic regression models of the competing risks of being remarried or cohabiting versus unpartnered (reference category) separately for women and men to test whether repartnership status varies by marital dissolution type, net of other factors (demographic characteristics, economic resources, and health) associated with union status in later life. All variables are entered simultaneously. Gender differences were tested using fully interactive models (i.e., all variables are interacted with gender simultaneously), and the significant gender differences in coefficients are denoted in the regression table for men. Our models only offer correlational evidence and should not be construed as causal. A multiple imputation procedure was used to handle missing cases such that the missing value for a single variable was imputed as a function of other covariates in the analysis (Acock, 2005). To preserve the randomness of the imputed variables, the study results were based on 10 random, multiple-imputed replicates. All analyses were conducted in Stata using svy commands to adjust for the complex sampling design of the HRS.
Results
About 34% (result not shown) of first later life marital dissolutions occur through divorce. For women, most dissolutions are the result of spousal death, but nearly one quarter (24.1%) are due to divorce. Marital dissolution has more often occurred through divorce (56.2%) among women currently between the ages of 50 and 64 years, but for women aged 65 and older marital dissolution disproportionately has occurred through spousal death (72.2%–95.4% depending on age group). For men, divorce plays a larger role. Slightly more than half (52.1%) of later life marital dissolutions among men occur through divorce rather than widowhood. For men aged 50–64 years, 78.9% of marital dissolutions have occurred through divorce. Even among 65- to 74-year-old men, the share of marital dissolutions that are due to divorce is 59% (vs 27.8% for women). And at ages 75–84 years, 31.7% of men’s marital dissolutions reflect divorce, not widowhood. There is pronounced gender asymmetry not only in widowhood but also in divorce. Given a marital dissolution, men are especially likely to experience divorce whereas women are particularly likely to become widowed (Table 1).
Table 1.
Women | Men | |||||
---|---|---|---|---|---|---|
Divorced | Widowed | n | Divorced | Widowed | n | |
Age (years) | ||||||
50–64 | 56.2 | 43.8 | 538 | 78.9 | 21.2 | 321 |
65–74 | 27.8 | 72.2 | 904 | 59.0 | 41.0 | 482 |
75–84 | 13.8 | 86.2 | 1,085 | 31.7 | 68.3 | 487 |
85+ | 4.6 | 95.4 | 789 | 11.6 | 88.4 | 280 |
Weighted % | 24.1 | 75.9 | 52.1 | 47.9 | ||
Unweighted N | 755 | 2,561 | 3,316 | 710 | 860 | 1,570 |
Descriptive Results
Repartnering following later life marital dissolution is common, particularly among those who experienced a divorce, which is consistent with our expectations and underscores the utility of differentiating by marital dissolution type. The role of marital dissolution type is more prominent for women than men, as we anticipated. For women, repartnership through either remarriage or cohabitation is about three times more common following divorce than widowhood, as shown in Table 2. Among divorced women, 15.3% are remarried, 8.6% are cohabiting, and 76.1% are unpartnered. Few widowed women are repartnered with just 3.6% remarried and 2.4% cohabiting. A whopping 94% are single. Among men, divorcees are just 50% more likely to be remarried or cohabiting than widowed men. For divorced men, 28.3% are remarried, 14.6% are cohabiting, and 57.1% remain single. Widowed men are more often single (75.7%) and less often either remarried (17.6%) or cohabiting (6.7%). As expected, a larger share of women are unpartnered than men. Among men, remarriage is twice as common as cohabitation regardless of marital dissolution type. Women are also more often remarried than cohabiting, but the difference is modest, especially among widowed women.
Table 2.
Women | Men | |||||
---|---|---|---|---|---|---|
Divorced | Widowed | Difference | Divorced | Widowed | Difference | |
Marital status in 2010 | ||||||
Remarried | 15.3a | 3.6b | *** | 28.3a | 17.6b | *** |
Cohabiting | 8.6a | 2.4b | *** | 14.6a | 6.7b | ** |
Unpartnered | 76.1a | 94.0b | *** | 57.1a | 75.7b | *** |
Marital dissolution pathway | ||||||
Years since dissolution | 10.8 | 11.0b | 10.3 | 8.3b | *** | |
Dissolved remarriage | 51.8 | 19.4 | *** | 46.8 | 18.7 | *** |
Demographic characteristics | ||||||
Age 50–64 | 47.5 | 11.7 | *** | 48.5 | 14.2 | *** |
Age 65–74 | 29.6 | 24.3 | * | 33.0 | 25.0 | * |
Age 75–84 | 18.8 | 37.1 | *** | 15.6 | 36.6 | *** |
Age 85 and older | 4.1 | 26.9 | *** | 2.9 | 24.2 | *** |
White | 77.9 | 84.9 | ** | 78.0 | 84.1 | * |
Black | 11.4 | 7.8 | * | 10.7 | 8.5 | |
Hispanic | 8.2 | 5.4 | ** | 8.6 | 5.6 | |
Other race | 2.5 | 1.9 | 2.7 | 1.8 | ||
Any resident child | 24.1a | 20.1 | 17.0a | 18.1 | ||
Nonresidential children only | 72.2a | 74.9 | 79.2a | 75.4 | ||
Childless | 3.7 | 5.0 | 3.8 | 6.5 | ||
Economic resources | ||||||
Less than high school | 16.8 | 23.9b | *** | 13.9 | 28.6b | *** |
High school | 31.6 | 41.3b | *** | 28.6 | 30.3b | |
Some college | 26.1 | 20.8 | * | 25.5 | 19.1 | * |
College or more | 25.5 | 14.0b | *** | 32.0 | 22.0b | *** |
Full-time employment | 25.7a | 6.7 | *** | 39.9a | 9.5 | *** |
Part-time employment | 7.2 | 2.2b | *** | 3.4 | 0.9b | ** |
Not in the labor force | 67.1a | 91.1 | *** | 56.7a | 89.6 | *** |
Owns home | 55.8 | 65.3 | ** | 55.5 | 67.9 | *** |
In debt | 11.6 | 3.7 | *** | 10.1 | 3.6 | ** |
$0–50,000 | 34.4 | 27.0 | *** | 28.7 | 24.1 | |
$50,001–100,000 | 10.3 | 13.2 | 11.8 | 11.3 | ||
$100,001–250,000 | 18.4 | 21.5 | 17.4 | 21.8 | ||
$250,001 or more | 25.2a | 34.6 | ** | 32.0a | 39.2 | * |
Health | ||||||
Has health insurance | 89.0 | 97.2 | *** | 88.9 | 97.9 | *** |
Self-rated health | 3.1 | 3.0 | 3.2 | 3.1 | ||
Unweighted N | 755 | 2,561 | 710 | 860 |
Notes. aDivorced women and men significantly differ at p < .05.
bWidows and widowers significantly differ at p < .05.
*p < .05. **p < .01. ***p < .001.
For women and men alike, the characteristics of the marital dissolution pathway differ for those who experienced divorce versus widowhood after age 50. The number of years since the dissolution occurred averages about 10.3 for divorced men versus 8.3 for widowers. The time since dissolution is longer for widows than widowers. For both women and men, dissolving a remarriage (vs a first marriage) was more than twice as common among divorcees as widoweds. About 51.8% of divorced women versus 19.4% of widowed women dissolved a remarriage. Among men, 46.8% of divorcees versus 18.7% of widowers dissolved a remarriage.
Women’s and men’s demographic characteristics differ by marital dissolution type. The age distribution of the divorced skews considerably younger than the widowed, as expected. For example, among divorced women, nearly half (47.5%) are in the 50–64 age range versus just 11.7% of widowed women. More than one quarter (26.9%) of widowed women are aged 85 or older whereas just 4.1% of divorced women are in this oldest age group. The pattern among men is similar with 48.5% of divorced men and just 14.2% of widowed men between the ages of 50 and 64 years. A mere 2.9% of divorced men are aged 85 or older versus 24.2% of widowed men. Among women, widoweds are disproportionately White compared with divorcees (84.9% vs 77.9%). Divorced women are more often Black (11.4%) or Hispanic (8.2%) than are widoweds (7.8% and 5.4%, respectively). Likewise, widowers (84.1%) are disproportionately White compared with men who are divorced (78%). A coresident child is more common among divorced women (24.1%) than divorced men (17.0%).
Marital dissolution type is closely linked to women’s and men’s economic characteristics. The divorced are more educated, on average, than widoweds. Among women, 25.5% of divorcees have a college degree versus just 14.0% of widows. Similarly, 32.0% of divorced men versus 22.0% of widowers completed college. Divorcees are more often working full time whereas widoweds are especially likely to be out of the labor force. About 25.7% of divorced women work full time compared with just 6.7% of widowed women. For men, the figures are 39.9% and 9.5%, respectively. Home ownership is much higher among widoweds (65.3% for women and 67.9% for men) than divorcees (55.8% for women and 55.5% for men), and this holds for women and men alike. Divorcees are about twice as likely to be in debt as widoweds, regardless of gender. For women, 11.6% of divorcees versus 3.7% of widoweds are in debt. For men, the shares are 10.1% and 3.6%, respectively. Indeed, widoweds report greater assets than divorcees. Nearly all widoweds (97.2% of women and 97.9% of men) report having health insurance, whereas coverage among divorcees stands between 89.0% (women) and 88.9% (men). There is no significant difference in self-rated health across marital dissolution type or gender.
Multivariate Results for Women
The multinomial logistic regression models are shown in Table 3 for women and Table 4 for men. We discuss each in turn. The characteristics of the marital dissolution pathway are linked to repartnership status net of other factors. As expected, women who experienced a divorce are more likely to be repartnered than their counterparts who became widowed. The risk of being remarried is 2.7 times higher, and the risk of cohabitation is twice as high for women who were divorced versus widowed. Divorced women are similarly likely to be remarried or cohabiting. Time since marital dissolution is positively associated with both remarriage and cohabitation among women. The greater the number of years since dissolution, the more likely women are to be remarried versus cohabiting. Women who experienced the dissolution of a remarriage have a higher risk of being remarried but are no more likely to be cohabiting than the unpartnered. Some demographic characteristics are related to repartnership status. Younger women are more likely to be either remarried or cohabiting than older women. Repartnership status does not vary by race-ethnicity for women. An empty nest, signaled by having only nonresidential children, is positively associated with both remarriage and cohabitation. Women with an empty nest are significantly more likely to be cohabiting than remarried. Childless women are more likely to stay single than to be remarried. The childless also are much less likely to be remarried than cohabiting. Economic resources are largely unrelated to women’s repartnership status, although women who are not in the labor force have a risk of being remarried that is three times higher than women who are full-time employed. Finally, having health insurance and being in good health are both positively related to remarriage among women but unrelated to cohabitation. Women with health insurance have four times the risk of being remarried as cohabiting.
Table 3.
Remarried vs unpartnered | Cohabiting vs unpartnered | Remarried vs cohabiting | |
---|---|---|---|
Marital dissolution pathway | |||
Divorced (vs widowed) | 2.70*** | 2.07* | 1.31 |
Years since dissolution | 1.10*** | 1.05** | 1.05* |
Dissolved remarriage | 1.88** | 1.20 | 1.56 |
Demographic characteristics | |||
Age (centered) | 0.90*** | 0.90*** | 0.99 |
White (ref) | |||
Black | 0.68 | 0.95 | 0.72 |
Hispanic | 0.67 | 0.47 | 1.43 |
Other race | 1.04 | 0.55 | 1.89 |
Any resident child (ref) | |||
Nonresidential children only | 1.70+ | 6.98*** | 0.24* |
Childless | 0.13* | 2.73 | 0.05** |
Economic resources | |||
Less than high school | 0.91 | 0.89 | 1.06 |
High school (ref) | |||
Some college | 0.77 | 1.07 | 0.72 |
College or more | 1.31 | 1.21 | 1.09 |
Full-time employment (ref) | |||
Part-time employment | 1.19 | 1.03 | 1.15 |
Not in the labor force | 3.02* | 1.38 | 2.19 |
Owns home | 1.44 | 0.82 | 1.74 |
In debt | 1.56 | 0.56 | 2.80 |
$0–50,000 (ref) | |||
$50,001–100,000 | 2.25* | 0.63 | 3.57* |
$100,001–250,000 | 1.75+ | 0.52 | 3.35+ |
$250,001 or more | 2.03+ | 1.33 | 1.53 |
Health | |||
Has health insurance | 3.35* | 0.84 | 4.01* |
Self-rated health | 1.42*** | 1.14 | 1.25 |
Constant | 0.001*** | 0.02*** | 0.07* |
Note. + p < .10. *p < .05. **p < .01. ***p < .001.
Table 4.
Remarried vs unpartnered | Cohabiting vs unpartnered | Remarried vs cohabiting | |
---|---|---|---|
Marital dissolution pathway | |||
Divorced (vs widowed) | 1.19b | 1.34 | 0.89 |
Years since dissolution | 1.13*** | 1.05* | 1.08** |
Dissolved remarriage | 1.02a | 1.26 | 0.81 |
Demographic characteristics | |||
Age (centered) | 0.97*b | 0.96*b | 1.01 |
White (ref) | |||
Black | 0.98 | 0.70 | 1.40 |
Hispanic | 2.16* | 0.51 | 4.28* |
Other race | 3.12+ | 2.77 | 1.12 |
Any resident child (ref) | |||
Nonresidential children only | 1.03 | 3.76*** | 0.27** |
Childless | 0.19** | 1.02 | 0.18* |
Economic resources | |||
Less than high school | 0.88 | 2.11*a | 0.42* |
High school (ref) | |||
Some college | 0.99 | 0.91 | 1.09 |
College or more | 1.58* | 1.40 | 1.13 |
Full-time employment (ref) | |||
Part-time employment | 0.83 | 1.33 | 0.63 |
Not in the labor force | 0.66+b | 0.50+a | 1.31 |
Owns home | 3.42***b | 1.20 | 2.86** |
In debt | 1.29 | 1.58 | 0.82 |
$0–50,000 (ref) | |||
$50,001–100,000 | 0.79a | 0.68 | 1.16 |
$100,001–250,000 | 0.86 | 0.83 | 1.04 |
$250,001 or more | 0.96 | 1.03 | 0.93 |
Health | |||
Has health insurance | 1.15a | 0.55 | 2.08 |
Self-rated health | 1.13 | 1.14 | 0.99 |
Constant | 0.05*** | 0.08* | 0.58 |
Multivariate Results for Men
Turning now to men, the marital dissolution pathway is comparatively unimportant. Repartnership status does not differ by marital dissolution type—men are similarly likely to be remarried or cohabiting regardless of whether they experienced a divorce or widowhood. Time since marital dissolution is positively associated with being repartnered among men, and the more time that has elapsed the more likely they are to be remarried versus cohabiting. Age is negatively associated with being remarried or cohabiting for men. Hispanic men are more likely to be remarried than either unpartnered or cohabiting. Men with an empty nest are more likely to be cohabiting than either unpartnered or remarried. Childless men are particularly unlikely to be remarried. Their risk of cohabiting is greater than that of being remarried. Economic resources are linked to men’s repartnership status. College educated men have a higher risk of being remarried. Men not in the labor force are less likely to be either remarried or cohabiting. Home ownership is associated with a greater risk of being remarried relative to either unpartnered or cohabiting. Men’s assets are not significantly associated with repartnership status, however. Nor are men’s health indicators related to their repartnership status.
The fully interactive model reveals that some of the correlates are differentially associated with repartnership status for women versus men, as indicated by the superscripts in Table 4. The greater propensity of divorced versus widowed individuals to be remarried is more pronounced among women than that among men. Similarly, the positive association between having dissolved a remarriage (vs first marriage) and being remarried is larger for women than that for men. The negative association between age and being either remarried or in a cohabiting union is greater for women than that for men. The positive relationship between having health insurance and being remarried is larger for women than that for men, but the positive relationship between home ownership and being remarried is larger for men than that for women. Some correlates operate in different directions by gender. Low education (i.e., less than high school) is positively associated with being in a cohabiting union for men and negatively related to cohabitation among women. Women who are not in the labor force are more likely to be either remarried or cohabiting whereas men are less likely to be repartnered. Moderate earnings ($50,001–100,000) are associated with increased odds of being repartnered for women and decreased odds of repartnership for men. These gender differentials underscore the value of estimating separate models for women and men.
In supplemental analyses (not shown, available upon request), we examined whether the associations between the demographic, economic, and health factors and repartnership status differed by marital dissolution type, but we found that in general these factors operate similarly regardless of whether the older adult’s marriage ended through death of a spouse or divorce. The role of marital dissolution type was similar regardless of age; there was no significant interactions between these two variables in their effects on repartnership status. Also, marital dissolution operated similarly regardless of marriage order.
Discussion
The acceleration in divorce coupled with declining widowhood among older adults is reshaping later life marital dissolution patterns. Moreover, these shifts foretell growth in repartnership as divorcees are more likely to be remarried or cohabiting than are widoweds. This differential is magnified by the fact that divorcees tend to be younger, on average, than the widowed and age is negatively related to repartnership. By explicitly comparing later life divorce and widowhood, we move beyond prior work that either has primarily focused on widowhood vis a vis continuous marriage or has failed to distinguish between individuals who experienced marital dissolution earlier versus later in the life course. In short, we fill a notable gap in the literature on union transitions and repartnerships in later life (Sassler, 2010).
Our analysis of the Health and Retirement Study relied on the detailed union histories available in the data set to construct a national portrait of women and men following later life marital dissolution. Rather than simply examining current marital status, we developed a more dynamic indicator of repartnership status to compare remarrieds, cohabitors, and unpartnereds following a divorce versus widowhood after age 50. Additionally, we considered how key demographic, economic, and health factors were related to repartnership status, net of the characteristics of the marital dissolution pathway (i.e., marital dissolution type, time since dissolution, and whether the dissolved marriage was the first or higher order).
As expected, a substantial share of later life marital dissolutions occurs through divorce, not through widowhood. For older women, about one quarter of dissolutions are the result of divorce. In contrast, a majority of the dissolutions that occur among older men are due to divorce, not due to death of a spouse. These patterns underscore that gerontological research should focus not only on widowhood but also on divorce as a nontrivial share of marital dissolutions in later life is voluntary.
Although the majority of older adults who experienced later life marital dissolution remain single, marital dissolution type sets the stage for repartnership status. Divorced women are about twice as likely as widowed women to be in a repartnership, whether remarriage or cohabitation. This gap is not accounted for by sociodemographic factors and persists with the inclusion of key correlates of repartnership status, including age, socioeconomic status, and health. This finding aligns with the notion that many older women are reluctant to repartner because they do not want to be a caregiver again (Talbott, 1998; Watson & Stelle, 2011). Most widowed women probably served as caregivers to their spouses whereas most divorced women did not. Avoiding a reprise of the caregiving role figures prominently in widowed women’s narratives about their disinterest in repartnership (McWilliams & Barrett, 2014).
Among men, divorcees are more often repartnered than are widowers although the differential by marital dissolution type is less stark for men than women. In fact, the differential is reduced to nonsignificance in the multivariate model for men, indicating divorced men and widowers are similarly likely to be repartnered. This makes sense considering that caregiving is not a salient role for most husbands. Rather, older single men are eager to be repartnered to ensure they are cared for (McWilliams & Barrett, 2014).
Overall, these trends reinforce the utility of distinguishing by marital dissolution type. For women, this delineation is important because repartnership status differs for those who are divorced versus widowed. For men, the widowed and divorced appear more similar but an exclusive focus on widowers misses over half of older men who experience marital dissolution because the majority of men’s dissolutions occur through divorce.
Moreover, our study illustrates that repartnership following later life dissolution is arguably a distinct process that may differ from how long-term singles navigate repartnership. We show that repartnership following marital dissolution after age 50 more often occurs through remarriage rather than cohabitation. This finding is not consistent though with those of prior studies on later life union formation among all singles (including those who have been single for decades) that uncover no apparent preference for marriage as similar shares of singles form marital and cohabiting unions (Brown et al., 2012).
The factors associated with repartnership among older adults who experienced later life marital dissolution align with traditional, gendered mate selection processes in which women exchange youth and vitality for men’s economic provision. For women, youthfulness (i.e., age), an empty nest, good health, and not working are positively associated with being remarried. Remarriage may reduce economic need among women and thus be related to not working. Alternatively, women who are not working may be particularly motivated to repartner. For men, economic resources are tied to their remarriage status; men’s college degree and home ownership are associated with greater risks of being remarried. And men who are not working are less likely to be either remarried or cohabiting. Health is not related to men’s remarriage or cohabitation status.
Relatively few factors are expressly related to cohabitation although a notable exception is having an empty nest which is positively associated with being in a cohabiting union relative to either being remarried or unpartnered. This pattern is consistent with that identified in a recent qualitative study on LAT relationships among older adults showing that couples tend to live separately in LAT relationships until their children are no longer in the household and then the couple can coreside (de Jong Gierveld & Merz, 2013).
There are some limitations to the current study. Although we go beyond prior work by bridging marital dissolution and subsequent repartnership status, we capture repartnered status at a single point in time. This approach permits us to compare individuals across repartnership status groups (remarried, cohabiting, and unpartnered) net of marital dissolution type, but we acknowledge that some respondents may have formed and dissolved other unions since the marital dissolution occurred. This was the case for fewer than 7% of the sample and inclusion of this indicator variable in the multivariate analyses did not alter the findings. Some remarried respondents may have premaritally cohabited (this was the case for just 9% of remarrieds). Also, the demographic, social, and health factors are measured at the same time point as repartnership status, and thus our results should not be interpreted as causal. Instead, our goal was to provide a rich description of those who experienced either divorce or widowhood after age 50, explicitly modeling the possibility that some individuals will have repartnered through remarriage or cohabitation whereas others remain single. We acknowledge that some older adults experience union dissolution through the death of a cohabiting partner. As cohabitation becomes more popular in later life, future research should consider this type of dissolution. Finally, we do not examine non-coresidential unions such as LAT or dating relationships because they are not measurable in the HRS data. A nontrivial fraction of our single respondents are likely to be in a dating relationship (Brown & Shinohara, 2013). Non-coresidential partnerships are a viable alternative to remarriage or cohabitation and are worthy of attention in future research.
The standard juxtaposition of marriage versus widowhood in gerontological research on marital status ignores the growing fraction of later life marital dissolutions that occur through divorce. In fact, older men whose marriages dissolve experience divorce about as often as widowhood. For older women, widowhood still predominates, but about one quarter of dissolutions are due to divorce. Marital dissolution type is related to repartnership status with divorcees considerably more likely to be either remarried or cohabiting compared with widoweds, especially among women. The sociodemographic portraits of these groups are distinctive, too. In short, our study provides new evidence about the importance of conceptualizing later life marital dissolution more broadly to encompass the distinctive pathways that unfold following divorce versus widowhood.
Funding
The research for this article was supported by a grant to S. L. Brown and I.-F. Lin from the National Institute on Aging (R15AG047588). Additional support was provided by the Center for Family and Demographic Research, Bowling Green State University, which has core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R24HD0509059).
Acknowledgments
Author contributions: S. L. Brown and I.-F. Lin conceived the study, planned the analyses, and wrote the article. A. M. Hammersmith and M. R. Wright prepared the data file and performed the analyses.
References
- Acock A. C. (2005). Working with missing values. Journal of Marriage and Family, 67, 1012–1028. doi:10.1111/j.1741-3737.2005.00191.x [Google Scholar]
- Bair D. (2007). Calling it quits: Late-life divorce and starting over. New York, NY: Random House. [Google Scholar]
- Brown S. L., Bulanda J. R., Lee G. R. (2012). Transitions into and out of cohabitation in later life. Journal of Marriage and the Family, 74, 774–793. doi:10.1111/j.1741-3737.2012.00994.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brown S. L. Lee G. R., & Bulanda J. R (2006). Cohabitation among older adults: A national portrait. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 61, S71–S79. doi:10.1093/geronb/61.2.s71 [DOI] [PubMed] [Google Scholar]
- Brown S. L., Lin I.-F. (2012). The gray divorce revolution: Rising divorce among middle-aged and older adults, 1990-2010. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 67, 731–741. doi:10.1093/geronb/gbs089 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brown S. L., & Shinohara S. K (2013). Dating relationships in older adulthood: A national portrait. Journal of Marriage and Family, 75, 1194–1202. doi:10.1111/jomf.12065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bulcroft K. Bulcroft R. Hatch L., & Borgatta E. F (1989). Antecedents and consequences of remarriage in later life. Research on Aging, 11, 82–106. doi:10.1177/0164027589111004 [DOI] [PubMed] [Google Scholar]
- Cherlin A. J. (2009). The marriage-go-round. New York, NY: Knopf. [Google Scholar]
- Chevan A. (1996). As cheaply as one: Cohabitation in the older population. Journal of Marriage and the Family, 58, 656–667. doi:10.2307/353726 [Google Scholar]
- de Jong Gierveld J. (2004). Remarriage, unmarried cohabitation, living apart together: Partner relationships following bereavement or divorce. Journal of Marriage and Family, 66, 236–243. doi:10.1111/j.0022-2445.2004.00015.x [Google Scholar]
- de Jong Gierveld J., & Merz E. M (2013). Parents’ partnership decision making after divorce or widowhood: The Role of (step)children. Journal of Marriage and Family, 75, 1098–1113. doi:10.1111/jomf.12061 [Google Scholar]
- Han E. A. Cichy K. E. Small B. J., & Almeida D. M (2014). Daily emotional and physical reactivity to stressors among widowed and married older adults. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69, 19–28. doi:10.1093/geronb/gbt035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hatch R. (1995). Aging and cohabitation. New York, NY: Garland. [Google Scholar]
- Hughes M. E., & Waite L. J (2008). Health in household context: Living arrangements and health in later life. Journal of Health and Social Behavior, 43, 1–21. doi:10.2307/3090242 [PMC free article] [PubMed] [Google Scholar]
- Hungerford T. L. (2001). The economic consequences of widowhood on elderly women in the United States and Germany. The Gerontologist, 41, 103–110. doi:10.1093/geront/41.1.103 [DOI] [PubMed] [Google Scholar]
- Kennedy S., Ruggles S. (2014). Breaking up is hard to count: The rise of divorce in the United States, 1980–2010. Demography, 51, 587–598. doi:10.1007/s13524-013-0270-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kitson G. C. Babri K. B. Roach M. J., & Placidi K. S (1989). Adjustment to widowhood and divorce: A review. Journal of Family Issues, 10, 5–32. doi:10.1177/019251389010001001 [Google Scholar]
- Lee G. R., & DeMaris A (2007). Widowhood, gender, and depression. Research on Aging, 29, 56–72. doi:10.1177/0164027506294098 [Google Scholar]
- Lee M. A., & Carr D (2007). Does the context of spousal loss affect the physical functioning of older widowed persons? A longitudinal analysis. Research on Aging, 29, 457–487. doi:10.1177/0164027507303171 [Google Scholar]
- Lin I.-F., Brown S. L. (2012). Unmarried Boomers confront old age: A national portrait. The Gerontologist, 52, 153–165. doi:10.1093/geront/gnr141 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lin I.-F. Brown S. L., & Hammersmith A. M.(in press). Marital biography, social security receipt, and poverty. Research on Aging. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Manning W. D., & Brown S. L (2011). The demography of unions among older Americans, 1980-present: A family change approach. In Settersten R. A., Jr., Angel J. L. (Eds.), Handbook of sociology of aging (pp. 193–210). New York, NY: Springer. doi:10.1007/978-1-4419-7374-0_13 [Google Scholar]
- McWilliams S., & Barrett A. E (2014). Online dating in middle and later life: Gendered expectations and experiences. Journal of Family Issues, 35, 411–436. doi:10.1177/0192513x12468437 [Google Scholar]
- Ofstedal M. B. Weir D. R. Chen K. T., & Wagner J (2011). Updates to HRS sample weights (HRS Documentation Report DR-013). Ann Arbor, MI: Survey Research Center, University of Michigan. Retrieved from http://hrsonline.isr.umich.edu/sitedocs/userg/dr-013.pdf [Google Scholar]
- Sassler S. (2010). Partnering across the life course: Sex, relationships, and mate selection. The Journal of Marriage and the Family, 72, 557–575. doi:10.1111/j.1741-3737.2010.00718.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sasson I., Umberson D. J. (2014). Widowhood and depression: New light on gender differences, selection, and psychological adjustment. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69, 135–145. doi:10.1093/geronb/gbt058 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Talbott M. M. (1998). Older widows’ attitudes towards men and remarriage. Journal of Aging Studies, 12, 429–449. doi:10.1016/S0890-4065(98)90028-7 [Google Scholar]
- Vespa J. (2012). Union formation in later life: Economic determinants of cohabitation and remarriage among older adults. Demography, 49, 1103–1125. doi:10.1007/s13524-012-0102-3 [DOI] [PubMed] [Google Scholar]
- Watson W. K., Stelle C. (2011). Dating for older women: Experiences and meanings of dating in later life. Journal of Women & Aging, 23, 263–275. doi:10.1080/08952841.2011.587732 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Williams K. (2004). The transition to widowhood and the social regulation of health: Consequences for health and health risk behavior. The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 59, S343–S349. doi:10.1093/geronb/59.6.S343 [DOI] [PubMed] [Google Scholar]
- Wu Z. Schimmele C. M., & Ouellet N (2015). Repartnering after widowhood. The Gerontologist, 70, 496–507. doi:10.1093/geronb/gbu060 [DOI] [PubMed] [Google Scholar]