Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jun 1.
Published in final edited form as: J Marriage Fam. 2016 Sep 27;79(3):833–849. doi: 10.1111/jomf.12375

Psychological Well-Being among Older Adults: The Role of Partnership Status

Matthew R Wright 1, Susan L Brown 2
PMCID: PMC5469370  NIHMSID: NIHMS811938  PMID: 28626245

Abstract

Today’s older adults are increasingly unmarried. Some are in cohabiting unions, others are dating, and many remain unpartnered. Unmarried older adults are at risk of poorer well-being than the married, but it is unclear whether older cohabitors fare worse than or similar to their married counterparts. Nor have well-being differences among cohabitors, daters, and unpartnered persons been considered. Conceptualizing marital status as a continuum of social attachment, data from Waves I and II of the National Social Life, Health, and Aging Project are used to examine how older married, cohabiting, dating, and unpartnered individuals differ across multiple indicators of psychological well-being. Among men, cohabitors appear to fare similarly to the married, and better than daters and the unpartnered. In contrast, there are few differences in psychological well-being by partnership status for women.

Keywords: aging, cohabitation, dating, marriage, mental health


Today’s older adults (i.e., those aged 50 years and older) are increasingly unmarried (Cooney & Dunne, 2001; Sassler, 2010) but they are not necessarily unpartnered (Calasanti & Kiecolt, 2007). Some are in cohabiting unions or dating relationships. Between 2000 and 2013, older adult cohabitation increased from 1.2 million to over 3.3 million individuals (Brown, Lee, & Bulanda, 2006; U.S. Census Bureau, 2014). Older adult dating relationships are even more common than cohabitation (Brown & Shinohara, 2013). The shrinking share of older adults who are married coupled with the increasing divorce rate among those over age 50 (Brown & Lin, 2012; Kennedy & Ruggles, 2014) and the rising number of older adults in the population means that we can anticipate a growing number of unmarried older adults in the coming years (Cooney & Dunne, 2001). Thus, it is likely that dating and cohabitation among older adults may become more prevalent in the future, particularly as the baby boomers age into later life.

Unmarried older adults are at risk of poorer health and well-being than are the married. Even during midlife, the level of disability is about twice as high among unmarrieds as marrieds (Lin & Brown, 2012). However, it is unclear whether cohabiting or dating mitigate the risk of lower well-being among unmarrieds. Prior studies have not considered whether well-being differs among cohabiting, dating, and unpartnered older adults. Nor is it clear whether cohabitation affords older adults protective benefits similar to marriage.

We use longitudinal data from the National Social Life, Health, and Aging Project (NSHAP; see http://www.norc.org/Research/Projects/Pages/national-social-life-health-and-aging-project.aspx) to examine how partnership status is related to psychological well-being among older adults. Theories of partnership status as a continuum of social attachment (Ross, 1995) and commitment (Kamp Dush & Amato, 2005) would lead us to expect that older married individuals enjoy the highest levels of well-being, followed by cohabitors, daters, and lastly unpartnereds. Alternatively, the well-being of older marrieds and cohabitors may be comparable given mounting evidence that later life cohabitation operates as a substitute for marriage (Brown, Bulanda, & Lee, 2012; King & Scott, 2005; Vespa, 2012). Our study breaks new ground by taking a longitudinal approach, distinguishing among cohabitors, daters, and unpartnered persons, and examining multiple measures of psychological well-being (depressive symptoms, perceived stress, and loneliness). Understanding the linkages between partnership status and well-being among older adults is especially important considering the changing patterns of marriage, cohabitation, and dating in later life.

BACKGROUND

The retreat from marriage is evident for those in the second half of life. For middle-aged adults, the share who are unmarried has risen by 50% since 1980 (Kreider & Ellis, 2011). Among those over age 65, roughly 30% of men and 60% of women were unmarried in 2008 (Manning & Brown, 2011). Increasingly, unmarried older adults are forming partnerships outside of marriage (De Jong Gierveld, 2004). Cohabitation has been on the rise among adults ages 50 and over, increasing threefold since 2000. Now nearly one in four cohabiting unions involves a partner over age 50 (Brown et al., 2006; U.S. Census Bureau, 2014; Vespa, 2012). Indeed, older adults are as likely to transition into a cohabiting union as a marriage (Brown et al., 2012). A greater share of unmarried older adults is in a dating than cohabiting relationship. Approximately 5% of older adults in the U.S. are in a dating relationship, which is 14% of the unmarried older adult population. By comparison, only 8% of older unmarrieds are cohabiting. Dating in older adulthood is much more common among men than women. Over one-quarter of older unmarried men are in a dating relationship, versus only 7% of older unmarried women (Brown & Shinohara, 2013). These trends underscore the importance of examining how partnership status, namely marriage, cohabitation, dating, or singlehood, is linked to well-being.

It is well established that married adults typically fare better in terms of health and well-being than the unmarried (Carr & Springer, 2010). Married adults in their 50s tend to experience lower levels of psychological distress, including depression and anxiety, than do the unmarried (Marks, 1996; Waite, 1995; Waite & Gallagher, 2000). But prior research typically lumps together all unmarrieds, failing to distinguish among cohabitors, daters, and unpartnereds (Lin & Brown, 2012; Liu, 2009). This is a key limitation given the growth in unmarried older adults (Lin & Brown, 2012). A few studies (Brown, Bulanda, & Lee, 2005; Hansen, Moum, & Shapiro, 2007) examine the well-being of older cohabitors but compare them with married individuals, not other types of unmarrieds. In short, we lack a good understanding of whether and how partnership status among older unmarrieds is linked to their well-being. Moreover, it is not clear whether cohabitors fare worse than or similar to their married counterparts.

Theoretical Framework: Partnership Status and Well-Being

Partnership status has been conceptualized as a continuum of social attachment (Ross, 1995). Social attachment emerges through social ties, which include emotional, social, and economic support. Ross (1995) argued that married people have the highest levels of social attachment, followed in order by those who are cohabiting with a partner, those with a nonresident partner (i.e., daters), and the unpartnered. Kamp Dush and Amato (2005) extended Ross’s (1995) work by conceptualizing marital status as a continuum of commitment. In their study, commitment was framed in terms of relationship length and salience of the relationship to the identity of the individual. Marriages last longer on average than cohabitations, and cohabitations often last longer than dating. Similarly, marriage may be more salient than cohabitation for the identity of an individual, and cohabitation might be more salient than dating. Thus, marriage involves the most commitment, followed by cohabitation, and dating, echoing Ross’s (1995) conclusion.

Those who experience greater levels of social attachment and commitment are expected to enjoy higher levels of well-being. This may be due, in part, to higher levels of emotional support for those with greater levels of attachment and commitment (Ross, 1995). That is, partnerships provide a form of social support. Of course, individuals can also receive economic support from friends and family or religious participation. Likewise, economic resources vary across partnership statuses and may also influence psychological well-being. Economic disadvantage is stressful and thus may diminish psychological well-being. It is also linked to partnership status (Pearlin & Johnson, 1977). For example, married older adults have higher household incomes than do cohabitors, whereas cohabitors tend to have greater household incomes than the unpartnered (Brown et al., 2006). Daters have greater assets than the unpartnered (Brown & Shinohara, 2013), but research has not compared older daters with marrieds or cohabitors on economic characteristics. Thus, partnership status is indicative of a continuum of social attachment in part because it captures both social and economic resources. Using data from 1990 with respondents ages 18 to 90, Ross (1995) found the expected patterns for cohabitors, daters, and unpartnered individuals, but no difference in depressive symptoms between cohabitors and marrieds.

Yet there is scant empirical evidence to support this framework for today’s older adults. Prior research has not considered the entire spectrum of the partnership continuum that we investigate here. Older adults with a resident partner (i.e., married or cohabiting) or who are dating report lower levels of depressive symptoms, loneliness, and social isolation, on average, than do the unpartnered (Cornwell & Waite, 2009; De Jong Gierveld, 2002; Peters & Liefbroer, 1997). Similarly, older adults who live with a spouse or partner report lower levels of loneliness compared to those living without a spouse or partner (Greenfield and Russell, 2011). But by combining married and cohabiting adults, these studies have obscured any potential differences between the two groups (Cornwell & Waite, 2009; Greenfield & Russell, 2011). Moreover, these studies largely ignored the potential role of dating relationships in the well-being of unmarrieds (De Jong Gierveld, 2002; Greenfield & Russell, 2011). The data used by Ross (1995) and Peters and Liefbroer (1997) are now old, having been collected in the early 1990s. Thus, whether partnership status operates as a continuum of social attachment in its relationship with psychological well-being for older adults remains unclear.

In particular, the psychological well-being advantage for marrieds compared with cohabitors as predicted by the partnership as a continuum of social attachment and commitment frameworks may not hold for older adults. On the one hand, marriage could be linked to higher levels of well-being for older adults. Married individuals enjoy greater social support and tend to have more social ties than do cohabitors (Brown et al., 2006; Waite & Lehrer, 2003). Cohabitors are less likely than both the continuously married and remarried to have friends or relatives in their neighborhood, and they also report lower levels of religiosity (Brown et al., 2006). Moreover, economic differences between older cohabitors and marrieds may contribute to variation in psychological well-being. Compared to the continuously married and remarried, the profile of cohabitors is one of disadvantage. Older cohabitors have lower incomes than both groups of marrieds and are also less likely than both the continuously married and remarried to own their homes (Brown et al., 2006).

On the other hand, the psychological well-being of older cohabitors may not be appreciably worse than that of older marrieds. Several scholars have argued that cohabitation in older adulthood serves as a long-term substitute for marriage (Brown et al., 2012; King & Scott, 2005). Older cohabitors experience similar union durations and levels of relationship quality as older remarrieds, signaling comparable commitment and social attachment (Hansen et al., 2007; King & Scott, 2005). If older cohabitors have similar levels of commitment and social attachment as the married, they may also derive many of the same benefits from their relationships as do the married. Thus, we might anticipate that cohabitors have similar levels of psychological well-being as the married.

Prior research on the well-being of older cohabitors versus marrieds is limited. One study found that cohabitors reported higher levels of depressive symptoms than the married, although this difference obtained only among men (Brown et al., 2005). Older married men enjoyed particularly low levels of depressive symptoms, suggesting men benefit more than women from marriage, at least in terms of psychological well-being. Cohabiting men and women as well as married women reported comparable levels of depressive symptoms.

The Role of Gender

Gender plays a central role in both older adult partnership status and psychological well-being. Older men are less often unmarried than older women. Among unmarrieds, older women are especially likely to be unpartnered whereas men are disproportionately cohabiting or dating. For example, older men are over three times as likely as older women to be dating (Brown & Shinohara, 2013). Similarly, women are more likely than men to report distress, anxiety symptoms, loneliness, and depressive symptoms (Cairney & Krause, 2005; Mehta, Simonsick, Penninx, Schulz, Rubin, Satterfield, & Yaffe, 2003; Pinquart & Sorensen, 2001; Thoits, 1986).

Marital status also appears to have a larger association with well-being for men than for women (Gove, Hughes, & Style, 1983). This may reflect the negligible gains women experience from marriage. Persistent gender inequality within marriage may diminish the gains from marriage enjoyed by wives. Jessie Bernard (1972) argued that men have power over women in marriage and wives often fill roles, such as housework and caregiving, within the family that are less rewarding than those performed by husbands (economic provision). Even though marriages are often more egalitarian these days, nonetheless marriage remains a gendered institution that is marked by distinct expectations for husbands and wives (Hochschild & Machung, 2012; Sayer, 2005). For older adults, the gendered division of labor could be magnified as wives continue to perform housework and caregiving activities but husbands are often retired, having relinquished their economic provider role. This asymmetry may lead women to be more likely to experience poorer psychological well-being than men within marriage (Bernard, 1972; Gove & Tudor, 1973). In contrast, unmarried men appear to fare worse than unmarried women in terms of psychological well-being during midlife (Marks, 1996; Peters & Liefbroer, 1997), in part because men tend to have fewer other sources of social support than women do (Carr, 2004). This gender gap in well-being does not appear to hold for older cohabitors though, presumably because the partner provides the attachment, much like a spouse would, but without the gendered constraints of marriage (Brown et al., 2005). Overall, marital status seems to be more consequential for men than for women, and thus we anticipate greater variation in the association between partnership status and psychological well-being among men than women.

The Current Study

The current study uses nationally representative, longitudinal data on older adults to examine how partnership status is related to psychological well-being. Symptoms of depression, perceived stress, and loneliness are not uncommon in later life (Dykstra, van Tilburg, & de Jong Gierveld, 2005; Mehta et al., 2003; Mirowsky & Ross, 1992; Pinquart, 2003; Yang, 2007). We examine the utility of conceptualizing partnership status as a continuum of social attachment and commitment by comparing the psychological well-being of marrieds, cohabitors, daters, and unpartnereds. Social attachment and commitment are positively related to well-being such that individuals who live with a partner tend to be less depressed and have higher levels of subjective well-being than those who are dating or unpartnered, and daters typically fare better than the unpartnered (Kamp Dush & Amato, 2005; Peters & Liefbroer, 1997; Ross, 1995).

Thus, we anticipate cohabitors have fewer depressive symptoms, less perceived stress, and less loneliness than daters and unpartnereds. In turn, daters should report lower levels of these outcomes than the unpartnered. We also test competing hypotheses with regards to cohabitors and marrieds. On the one hand, marrieds may have higher levels of psychological well-being because they have more resources and social ties than cohabitors (Brown et al., 2006). On the other hand, there could be no differences, as cohabitation may serve as a substitute for marriage among older adults (Brown et al., 2012; King & Scott, 2005).

Demographic, economic, social relationship, and physical health indicators are associated with partnership status and psychological well-being. Moreover, some of the relationship between partnership status and well-being may be a function of social support and economic resources, according to the partnership status as a continuum of social attachment perspective.

Demographic characteristics include race and age. Blacks are disproportionately unmarried (Cherlin, 2010) and more likely to date at older ages than are Whites (Brown & Shinohara, 2013). Some studies note that non-Whites report higher levels of depressive symptoms than Whites (Mirowsky & Ross, 1989), while others find that Blacks report less psychological distress than Whites (Bratter & Eschbach, 2005). Partnership status influences health more at older ages (Williams & Umberson, 2004). The proportion cohabiting declines with age (Brown et al., 2006; Chevan, 1996), and daters are younger, on average, than non-daters (Brown & Shinohara, 2013).

Economic disadvantage is linked to lower psychological well-being (Pearlin & Johnson, 1977). Cohabitors are more likely to have private health insurance, higher household incomes and to be employed full time than are the unpartnered, but are less likely to be insured and have lower incomes than the married (Brown et al., 2006). Daters tend to be better educated and have more assets than the unpartnered (Brown & Shinohara, 2013). Accounting for these factors may diminish the well-being advantage experienced by daters versus unpartnereds.

Social relationships differ by partnership status and are related to psychological well-being. They are also one of the main mechanisms through which partnership status is expected to operate on well-being. Cohabitors fare worse in terms of social relationships, such as having friends or relatives in their neighborhood and religiosity, than do either the married or the unpartnered (Brown et al., 2006), whereas daters have more social connections than the unpartnered (Brown & Shinohara, 2013). How cohabitors and daters compare is unclear. We tap social relationships using a social support scale. Cohabitors have lower religiosity than either the married or unpartnered (De Jong Gierveld, 2004), and religiosity is negatively associated with depressive symptoms (Brown et al., 2005). Living children may provide support to their parents, contributing to higher levels of psychological well-being. Conversely, children could add stress to their parents, contributing to lower levels of psychological well-being.

Little work has considered physical health among cohabitors, but daters report better self-rated health than the unpartnered, on average (Brown & Shinohara, 2013). Limitations in Activities of Daily Living (ADLs) are associated with additional stress in a relationship, as well as more depressive symptoms, at least among married individuals (Booth & Johnson, 1994).

We estimate models separately for men and women with the expectation that marital status is more influential for the well-being of men than women (Bernard, 1972; Gove & Tudor, 1973; Gove et al., 1983). Married men are expected to have better psychological well-being than married women (Bernard, 1972; Brown et al., 2005; Gove & Tudor, 1973), whereas unmarried men tend to have lower levels of psychological well-being in midlife than do unmarried women (Marks, 1996; Peters & Liefbroer, 1997), although there may be no difference between cohabiting men and women (Brown et al., 2005).

METHOD

Data for this study came from Waves I and II of the National Social Life, Health, and Aging Project (NSHAP). Wave I of the NSHAP data were collected by the National Opinion Research Center and the University of Chicago using a nationally representative sample of 3,005 people between the ages of 57 and 85 years in 2005–2006. The sample was designed by and carried out through the household sampling process of the Health and Retirement Study (HRS). Of the households identified by the HRS, 4,400 individuals were selected, limited to one per household. Of those selected, 92% were eligible for the NSHAP study, and the response rate for the interviews was 75.5% (O’Muircheartaigh, Eckman, & Smith, 2009). The NSHAP data were collected in three ways: an in-home interview, a self-administered questionnaire, and biomeasures. A wide range of topics were covered in the study, including sexual behavior, physical and mental well-being, social networks, and social and intimate relationships. Wave II of the NSHAP data were collected in 2010–2011 from nearly 3,400 respondents, and include Wave I respondents, individuals who did not participate in Wave I although they were invited, and cohabiting partners and spouses of respondents (O’Muircheartaigh, English, Pedlow, and Kwok, 2014). The 2,261 respondents who were in both Waves I and II were utilized in the current study. The Wave II response rate for Wave I respondents was 89% (O’Muircheartaigh et al., 2014). These data were ideal for this study for several reasons. First, in the NSHAP marital status question, cohabitors were directly measured as a category separate from other unmarried individuals. Second, respondents in the NSHAP who were neither married nor cohabiting were asked if they had a romantic, intimate, or sexual partner, allowing us to identify daters.

To be included in the analytic sample, respondents had to have valid data for both Waves I and II (n= 2,261) and not be missing on partnership status (n = 2,258). For the analysis of depressive symptoms, two cases were eliminated because of missing information on the dependent variable. The final sample size for the analysis of depressive symptoms was 2,256, of which 1,176 were women and 1,080 were men. For perceived stress, 420 respondents were eliminated due to missing information on the dependent variable (n = 1,838). Of the 1,838 respondents, 956 were women and 882 were men. Finally, 355 respondents were missing on the loneliness variable (n = 1,903). The sample includes 991 women and 912 men. To account for missing information on the other study variables, mode imputation was employed for categorical variables and mean imputation for the continuous variables. The dependent variables were measured at Wave 2, and all of the other study variables were measured at Wave 1.

Measures

Dependent variables

Three mental health outcomes were considered in this study. The first dependent variable was depressive symptoms experienced in the past week, measured at Wave II. NSHAP included eleven items, both positive and negative, from the Center for Epidemiologic Studies Depression Scale (CESD). The positive items included “was happy” and “enjoyed life.” The negative items included “felt depressed,” “felt everything was an effort,” “sleep was restless,” “felt lonely,” “people were unfriendly,” “felt sad,” “could not get going,” “did not feel like eating,” and “felt people disliked me.” Response options for all of the items included rarely or none of the time, some of the time, occasionally, and most of the time. Following Payne, Hedberg, Kozloski, Dale, and McClintock (2014), who have validated psychological well-being measures using the NSHAP data, we coded the items with rarely or none of the time = 0, some of the time = 1, occasionally and most of the time = 2. The two positive items were reverse coded so that a higher score reflected poorer psychological well-being. To create the scale, we summed the nonmissing items for respondents with at least seven valid responses. We then found the average of the valid items and multiplied it by 11 to put all respondents on the same scale. Finally, we followed the recommendation of Payne et al. (2014) who suggest dichotomizing the measure to identify “frequent” depressive symptoms (Payne et al., pp. S103). Those who score nine or higher were coded as 1 and 0 otherwise.

The second dependent variable was perceived stress at Wave II. We used four items noted by Payne et al. (2014) to create our measure of perceived stress. Respondents were asked how often in the past week they felt the following: “I was unable to control important things in my life,” “I felt difficulties were piling up so high I could not overcome them,” “I felt confident about my ability to handle personal problems,” and “I felt that things were going my way.” Response options for all four items were “rarely or none of the time,” “some of the time,” “occasionally,” and “most of the time.” Following Payne et al. (2014), the negative items were coded as rarely or none of the time = 0, some of the time = 1, occasionally = 2, and most of the time = 3. The positive items were coded in the reverse order. To construct perceived stress, we summed the nonmissing items for respondents with valid responses on at least three of the measures. Next, we took the average of the valid items and multiplied by 4 to put respondents on the same scale. Following Payne et al. (2014, p. S109), we created a dichotomous variable to tap respondents who perceived stress more than rarely. Those who scored one or higher were coded as 1 and 0 otherwise.

The final dependent variable was loneliness at Wave II. Loneliness is constructed using three items, including “how often do you feel that you lack companionship?,” “how often do you feel left out?,” and “how often do you feel isolated from others?” Response options were “never,” “hardly ever,” “some of the time,” and “often.” Following Payne et al. (2014), we coded never and hardly ever = 0, some of the time = 1, and often = 2 for each item. To construct the score for loneliness, we summed the nonmissing items for respondents with valid responses on at least two measures. Next, we took the average of the valid items and multiplied by 3 to put all respondents on the same scale.

We treat the three psychological well-being constructs as separate measures. Their correlations are fairly low (0.20 for depressive symptoms and perceived stress, 0.31 for depressive symptoms and loneliness, and 0.25 for perceived stress and loneliness) and prior research has established the distinctiveness of these three constructs (Payne et al., 2014).

Focal independent variable

Partnership status was a series of binary variables coded as: married, cohabiting (reference), dating, and unpartnered, with 1 = yes and 0 = no. Cohabitation was a direct question, as respondents were given the response option of living with a partner when asked their marital status. Daters were those who are neither married nor cohabiting and have a romantic, intimate, or sexual partner. Cohabitors were the reference category because our hypotheses compare cohabiting persons to their married and unmarried counterparts.

Demographic characteristics

Gender was measured as a dichotomous variable with 1 = man and 0 = woman. Race was also a dichotomous variable with 1 = White and 0 = Nonwhite. Age was measured as a continuous variable in years.

Economic resources

Education was measured as a dichotomous variable, with some college or more coded as 1 and a high school degree or less coded as 0 (modest sample size precludes further specification). Employment status reflects respondents who are currently employed versus those who are not, with 1 = employed and 0 = not employed. Private health insurance was coded as 1 if the respondent has such insurance and 0 otherwise. Assets were measured as the logged dollar value of the respondent’s household assets after accounting for debts.

Social ties

Religious attendance was coded as 1 if the respondent indicated that he or she attended church at least once a week and 0 otherwise. Living children was coded as 1 if the respondent has at least one living child and 0 if not. Social support was a scale constructed from four items, including how often respondents can open up to their friends about their worries, how often they can open up to their family, how often they can rely on their friends when they have a problem, and how often they can rely on their family when they have a problem. Each item ranged from 1 (hardly ever (or never) to 3 (often). The scale was the sum of the four items, with scores ranging from 4 to 12. The Cronbach’s alpha for the scale was .64.

Disabilities

Respondents reported whether they had difficulty with several activities, including “walking one block,” “walking across the room,” “eating,” “dressing,” “bathing,” “using the toilet,” “getting in and out of bed,” “driving during the day,” or “driving at night.” Those who report they had difficulty with at least one of the activities were coded as 1 for having limitations in activities of daily living (ADL), and all others were coded as 0. A dichotomous variable was preferable to a continuous measure because 40% of the sample had 0 difficulties and 70% fall between 0–2 out of a possible 27 on a summed scale, creating a highly skewed distribution.

Analytic Strategy

Descriptive statistics for all variables are presented by gender separately for marrieds, cohabitors, daters, and unpartnered individuals. Multivariate analyses were conducted using logistic regression for depressive symptoms and perceived stress because these outcomes are dichotomous. Ordinary least squares regression (OLS) was used for loneliness because loneliness is a continuous variable. Multivariate models were estimated to assess differences between partnership groups net of other factors. Finally, we examined separate models for men and women to determine if gender differences exist in the effect of partnership status on psychological well-being. Descriptive and multivariate analyses were conducted using the svy procedure in Stata to correct for the complex sampling design of the NSHAP.

RESULTS

Descriptive Results

Table 1 presents the distribution of partnership status for the total sample and by gender. In the full sample, 64.7% of respondents were married, 2.5% were cohabiting, 5.1% were dating, and 27.7% were unpartnered. Partnership status varied by gender, as men were much more likely than women to be partnered. Among women 41.1% were unpartnered. Just over half of women (52%) were married, 2.5% were cohabiting, and 3.9% were dating. Among men, over three quarters (78%) of the respondents were married, whereas just 2.4% were cohabiting, and 6.4% were dating. Only 13.2% of older men in the sample were unpartnered.

Table 1.

Partnership Status Distribution by Gender

Total Sample Women Men

% N % N % N
Married 64.7 1,434 52.5 600 78.0 834
Cohabiting 2.5 48 2.5 23 2.4 25
Dating 5.1 122 3.9 43 6.4 79
Unpartnered 27.7 652 41.1 510 13.2 142
Unweighted N 2,256 1,176 1,080

Descriptive statistics for women are shown by partnership status in Table 2. About 14.8% of cohabiting women reported frequent depressive symptoms which is largely comparable to that of married women at 17%. The shares for dating and unpartnered women were a bit higher at 27.1% and 22.6%, respectively. Most older women experience considerable stress regardless of partnership status. About 68.6% of marrieds, 72% of cohabitors, 85.4% of daters, and 70.2% of unpartnereds reported at least moderate stress. Loneliness was also common among older women, with frequency levels hovering around “some of the time” for married and cohabiting women and in between “some of the time” and “often” for dating and unpartnered women.

Table 2.

Weighted Percentages and Means by Partnership Status for Women

Married Cohabiting Dating Unpartnered
  Depressive Symptoms (%) 17.0 14.8 27.1 22.6
  Perceived Stress (%) 68.6 72.0 85.4 70.2
  Loneliness (mean) 1.0 1.0 1.8 1.5
Demographic Characteristics
  White 90.3 84.7 81.4 78.1
  Nonwhite 9.7 15.3 18.6 21.9
  Age 66.3 63.3 65.8 69.9
Economic Resources
  Some college or more 60.8 64.0 53.9 46.6
  High school or less 39.2 36.0 46.1 53.4
  Currently employed 32.0 51.8 40.9 30.2
  Not in the labor force 68.0 48.2 59.1 69.8
  Private health insurance 74.4 89.3 66.8 63.2
  No private health insurance 25.6 10.7 33.2 36.8
  Assets ($1,000s) 613,234 286,312 511,264 275,529
Social Support
  Living children 95.6 89.0 98.3 88.2
  No living children 4.4 11.0 1.7 11.8
  Attends services at least once a week 53.7 10.5 30.5 53.7
  Attends less than once a week 46.3 89.5 69.5 46.3
  Social support scale 9.8 9.8 9.4 9.6
Disabilities
  At least one ADL limitation 47.6 44.0 53.1 55.9
  No ADL limitations 52.4 56.0 46.9 44.1
  Unweighted N 600 23 43 510

Variation in demographics, economics, social support, and health by partnership status are evident in Table 2. Dating (19%) and unpartnered (22%) women were disproportionately nonwhite. On average, cohabitors were the youngest of the partnership statuses (63) and the unpartnered were the oldest (70). The married (66) and daters (66) fall in between cohabitors and unpartnereds. Cohabiting (64%) and married (61%) respondents were more likely than daters (54%) to have reported having at least some college education, whereas less than half of unpartnereds reported higher education (47%). Cohabitors were the most likely of the groups to be employed (52%), with the unpartnered being the least likely (30%). Similarly, cohabitors were the most likely to have private health insurance (89%) and the unpartnered were the least likely (63%). Finally, marrieds held the highest amount of assets on average at $613,234, and the unpartnered had the lowest ($275,529).

In terms of social support, dating (98%) older adults were the most likely to have children, followed by the married (96%). The percentages of cohabitors (89%) and unpartnereds (88%) who noted they have living children are similar. Cohabitors (11%) were by far the least likely to report attending religious services once a week or more. Marrieds (54%) and the unpartnered (54%) were the most likely, followed by daters (31%). There were few apparent differences across partnership groups in reports of social support from friends and family. Finally, cohabitors (44%) least frequently indicated that they had at least one ADL limitation, but the unpartnered (56%) were the most likely to have one or more limitation.

Table 3 reports descriptive statistics for men separately by partnership status. There was notable partnership status variation in psychological well-being among men. About 11% of married men reported frequent depressive symptoms compared with just 1% of cohabiting men. For dating men, the share was roughly 12.3% and for unpartnered men it was considerably higher at 22.7%. The relative advantage for cohabiting men also emerged for the indicator of perceived stress. Whereas 65.8% of married men were at least moderately stressed, only 43.4% of cohabiting men were. The proportions experiencing stress were higher for dating and unpartnered men at 64% and 70.9%, respectively. Loneliness was more comparable for cohabiting and married men. Each group reported feeling lonely slightly less than “some of the time,” whereas for daters and the unpartnered the range was between “some of the time” and “often.”

Table 3.

Weighted Percentages and Means by Partnership Status for Men

Married Cohabiting Dating Unpartnered
  Depressive Symptoms (%) 11.0 1.0 12.3 22.7
  Perceived Stress (%) 65.8 43.4 64.0 70.9
  Loneliness (mean) 0.9 0.7 1.3 1.9
Demographic Characteristics
  White 87.2 81.2 75.4 80.0
  Nonwhite 12.8 18.8 24.6 20.0
  Age 66.0 65.8 68.1 68.8
Economic Resources
  Some college or more 63.9 49.7 63.2 53.3
  High school or less 36.1 50.3 36.8 46.7
  Currently employed 49.7 56.7 44.2 28.9
  Not in the labor force 50.3 43.3 55.8 71.1
  Private health insurance 77.9 75.5 71.0 58.0
  No private health insurance 22.1 24.5 29.0 42.0
  Assets ($1,000s) 587,628 324,489 783,456 340,455
Social Support
  Living children 96.7 85.7 78.4 81.3
  No living children 3.3 14.3 21.6 18.7
  Attends services at least once a week 45.2 10.5 20.4 29.2
  Attends less than once a week 54.8 89.5 79.6 70.8
  Social support scale 8.7 8.2 9.2 8.7
Disabilities
  At least one ADL limitation 36.8 48.4 29.0 38.2
  No ADL limitations 63.2 61.6 71.0 61.8
  Unweighted N 834 25 79 142

The percentages of older men who were cohabiting (19%), dating (25%), and unpartnered (20%) were disproportionately nonwhite. The unpartnered (69) were the oldest, on average, whereas cohabitors and the married were the youngest (66). The married (64%) were the most likely to have some college or greater education. In contrast, cohabitors (50%) were the least likely. But, cohabiting (57%) men were the most likely to be employed, and the unpartnered (29%) were the least likely. The married (78%) had the highest percentage with private health insurance, while the unpartnered (58%) had the lowest. Daters ($783,456) had the highest assets, whereas cohabitors had the fewest ($340,455).

There was variation in social support across partnership groups among men. Married (97%) men were the most likely to have children, and daters (78%) were the least likely. Marrieds (45%) were also the most likely to attend religious services weekly, and cohabitors (11%) were the least likely to do so. Men who were dating (9.2) reported the highest level of social support from friends and family, while cohabitors (8.2) noted the least social support. Finally, the highest percentage reporting at least one ADL limitation was among cohabitors (48%), whereas daters (29%) had the lowest percentage.

Multivariate Results

We estimated models separately for men and women. The results for women are reported in Table 4 and those for men are reported in Table 5. Model 1 of Table 4 presents a zero order model predicting frequent depressive symptoms for women. As shown in Model 1, no significant differences emerged between cohabiting and married, dating, and unpartnered women, indicating the likelihoods of reporting frequent depressive symptoms were comparable. Dating women did not significantly differ from the unpartnered on depressive symptoms (result not shown).

Table 4.

Regression Models Predicting Depressive Symptoms, Perceived Stress, and Loneliness for Women

Depressive Symptoms Perceived Stress Loneliness

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

OR SE b SE OR SE OR SE b SE b SE

Partnership Status
  Married 1.18 0.68 1.14 0.64 0.85 0.44 0.84 0.10 0.03 0.26 0.18 0.26
  Cohabiting (ref)
  Dating 2.15 1.39 1.80 1.13 2.27 1.50 1.98a 0.13 0.79 0.44 0.77 0.47
  Unpartnered 1.68 0.96 1.31 0.72 0.91 0.51 0.68a 0.10 0.53 0.29 0.59 0.30
Demographic Characteristics
  Race (1 = White) 1.27 0.24 0.71 0.04 0.04 0.13
  Age (in years) 0.99 0.02 1.02 0.01 −0.01 0.01
Economic Resources
  Education (1 = Some col +) 0.60** 0.10 0.81 0.04 −0.02 0.12
  Employment (1 = employed) 0.71 0.16 0.76 0.04 −0.07 0.15
  Private health insurance 0.85 0.21 0.97 0.05 −0.08 0.11
  Assets (logged) 0.95 0.04 0.87* 0.02 −0.05 0.03
Social Support
  Living children 0.87 0.29 0.96 0.07 −0.07 0.19
  Weekly attendance or more 0.97 0.17 0.87 0.05 −0.25* 0.11
  Social support scale 0.80*** 0.04 0.86** 0.01 −0.15*** 0.03
Disabilities
  ADLs 2.00** 0.38 1.19 0.04 0.28* 0.10
  Constant 0.17** 0.10 5.26*** 9.51 2.57+ 1.34 42.91** 52.14 0.98*** 0.25 3.57*** 0.77
  Unweighted N 1,176 1,176 956 956 991 991

Note: Analyses are weighted to account for complex sampling design.

+

p = .10,

*

p < .05,

**

p < .01,

***

p < .001.

a

daters and unpartnereds significantly different

Table 5.

Regression Models Predicting Depressive Symptoms, Perceived Stress, and Loneliness for Men

Depressive Symptoms Perceived Stress Loneliness

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

OR SE OR SE OR SE OR SE b SE b SE

Partnership Status
  Married 12.51* 12.02 18.42** 18.57 2.51 1.44 3.29 1.99 0.20 0.26 0.41 0.25
  Cohabiting (ref)
  Dating 14.13* 15.61 16.52* 19.77 2.31 1.55 2.94 2.12 0.57 0.31 0.66* 0.31
  Unpartnered 29.69*** 28.92 28.61** 29.06 3.17+ 1.96 3.48+ 2.16 1.13** 0.37 1.13** 0.36
Demographic Characteristics
  Race (1 = White) 0.68 0.18 0.98 0.21 −0.15 0.12
  Age (in years) 1.02 0.02 1.01 0.01 0.01 0.01
Economic Resources
  Education (1 = Some col +) 0.59* 0.14 0.69* 0.12 −0.15 0.13
  Employment (1 = employed) 0.65+ 0.16 1.10 0.26 −0.23 0.14
  Private health insurance 1.05 0.25 0.95 0.16 0.11 0.16
  Assets (logged) 0.94 0.06 0.92 0.05 −0.04 0.03
Social Support
  Living Children 0.52 0.24 0.66 0.21 −0.61* 0.25
  Weekly attendance or more 0.71 0.17 1.00 0.19 −0.08 0.13
  Social support scale 0.82*** 0.04 0.91+ 0.05 −0.07 0.04
Disabilities
  ADLs 1.29 0.30 1.45* 0.25 0.24* 0.10
  Constant 0.01*** 0.01 0.09 0.16 0.77 0.42 2.86 3.61 0.74** 0.25 2.32** 0.72
  Unweighted N 1,080 1,080 882 882 912 912

Note: Analyses are weighted to account for complex sampling design.

+

p = .10,

*

p < .05,

**

p < .01,

***

p < .001.

Results for the full model for women are reported in Model 2 of Table 4. Similar to the zero order model, there were no differences between cohabiting, married, dating, and unpartnered women. Daters did not significantly differ from the unpartnered on depressive symptoms (result not shown). Those with at least some college education were less likely to report frequent depressive symptoms than those with less education. Respondents who felt they received more social support had lower odds of reporting frequent depressive symptoms. Finally, having at least one ADL limitation was associated with greater odds of frequent depressive symptoms. In short, these findings indicate that partnership status was largely unrelated to depressive symptoms for older women.

Models 3 and 4 of Table 4 display the results for regressions predicting perceived stress among women. The zero order model shows that cohabiting women did not significantly differ from their married, dating, or unpartnered counterparts. Dating women did not differ from unpartnered women, with the unpartnered slightly less likely (p < .10) to report perceiving stress more than rarely (result not shown). Similarly, in the full model, reported in Model 4 of Table 4, there were no significant differences in perceived stress by partnership status. Daters had higher odds of perceiving stress more than rarely than unpartnered women. Those with greater assets were less likely to perceive stress more than rarely. Women with higher levels of social support were also less likely to perceive stress more than rarely, relative to those reporting less support. Consistent with depressive symptoms, these results suggest that partnership status was largely unrelated to perceived stress among older women.

Models 5 and 6 of Table 4 display the results of the analysis of loneliness among women. The zero order model in Model 5 shows no significant differences in loneliness by partnership status. The coefficients for daters and the unpartnered were marginally significant (p < .10), suggesting that cohabiting women may be slightly less lonely than dating and single women. Cohabiting and married women reported similar levels of loneliness. Dating women did not significantly differ from the unpartnered on loneliness (result not shown). The addition of controls changed the results in the full model. Cohabiting and dating women were no longer significantly different on loneliness, while cohabiting women were marginally less lonely than unpartnered women. Cohabiting and married women did not differ on loneliness. Daters did not significantly differ from the unpartnered on loneliness (result not shown). Attending religious services at least weekly and social support were negatively associated with loneliness, whereas having at least one ADL limitation was positively related to loneliness.

Turning now to our models for men, the results for men’s depressive symptoms, shown in Models 1 and 2 of Table 5, were largely in line with our hypotheses. As expected, cohabiting men were less likely to report frequent depressive symptoms than either daters or unpartnereds. However, contrary to our expectations, cohabiting men actually had lower odds of frequent depressive symptoms than the married. Dating men did not differ from the unpartnered on depressive symptoms (result not shown). Model 2 shows the results for men once control variables were included. The pattern of results was consistent with the zero order model, as cohabiting men were less likely to exhibit frequent depressive symptoms than the married, daters, and the unpartnered. The difference between the unpartnered and daters was not significant in the full model (result not shown). Men with at least some college education were less likely to note frequent depressive symptoms. The employed had marginally lower odds of frequent depressive symptoms, compared to the unemployed. Respondents who reported higher levels of social support were less likely to report frequent depressive symptoms than those receiving less support. In sum, partnership status was associated with depressive symptoms among men.

Model 3 of Table 5 reports the results of the zero order model for perceived stress. No significant differences emerged between cohabiting men and the married or daters. However, relative to cohabiting men, unpartnered men exhibited a marginally higher likelihood of perceiving stress more than rarely. Daters did not differ from the unpartnered on perceived stress (result not shown). The results were similar in the full model, shown in Model 4 of Table 5. The difference between cohabitors and the married was marginally significant, with cohabiting men having lower odds of perceiving stress more than rarely. However, the odds ratio for the unpartnered was also marginally significant, with cohabiting men less likely to have reported moderate or greater stress than unpartnered men. Cohabitors and daters did not differ. Dating men did not differ from the unpartnered (result not shown). Education was negatively associated with perceiving stress more than rarely, and those who received higher levels of social support were marginally less likely to perceive stress more than rarely, compared to those who reported lower levels of support. Having at least one ADL limitation was positively associated with perceived stress.

The results for loneliness were in line with our expectations, as differences by partnership status emerge for men. Model 5 of Table 5 indicates that cohabiting men reported lower levels of loneliness than do unpartnered men. The difference between daters and cohabitors was marginally significant (p < .10), with dating men noting greater levels of loneliness than cohabiting men. No significant difference between cohabiting and married men was evident. Daters reported less loneliness than the unpartnered, though the difference was only marginally significant (p < .10) (result not shown). Model 6 of Table 5 presents the results after the inclusion of covariates. As expected, both dating men and unpartnered men reported higher levels of loneliness than cohabiting men. Cohabiting and married men did not significantly differ. Dating men did not significantly differ from the unpartnered on loneliness, net of other factors (result not shown). Among the covariates, employment was marginally negatively associated with loneliness. Having living children was negatively associated with loneliness, while noting an ADL was positively related to loneliness. In short, cohabiting men were less lonely than their dating and unpartnered counterparts, but were similar to the married.

Supplemental tests (results not shown but available upon request) were conducted to determine whether gender differences in the associations between partnership status and well-being were statistically significant. We estimated full models for women and men combined that included interaction terms for gender and marital status. Significant interaction effects emerged for both depressive symptoms and perceived stress. Among marrieds, men were more likely than women to report frequent depressive symptoms (OR = 14.96, p = 0.03). Similarly, among daters (OR = 10.68, p = .09) and unpartnereds (OR = 24.85, p = 0.01), men were more likely to report frequent depressive symptoms than women. Married men were more likely to perceive stress more than rarely than married women (OR = 3.76, p = 0.09) and unpartnered men were also more likely to indicate perceiving stress more than rarely than unpartnered women (OR = 4.68, p = 0.05). For loneliness, there were no significant gender differentials. These patterns signaled that marital status sometimes operated differently for women and men, underscoring our rationale for conducting separate analyses by gender.

Discussion

A declining share of U.S. adults is married, and this trend is evident among older adults. Cohabitation is growing rapidly among those ages 50 and older and dating is even more common than cohabitation among older unmarrieds. Our study addressed whether cohabiting and dating partnerships are related to better psychological well-being among the unmarried, as well as whether cohabitors and the married enjoy comparable levels of well-being. Drawing on Ross’s (1995) theory of marital status as a continuum of social attachment, and Kamp Dush and Amato’s (2005) continuum of commitment, we used Waves I and II of the NSHAP data to investigate whether the psychological well-being of older adults followed the expected continuum, with married individuals reporting the highest well-being, followed by cohabitors, daters, and lastly unpartnered persons. We also considered whether cohabitors and marrieds reported comparable well-being given that prior research suggests cohabitation in later life serves as an alternative to marriage (Brown et al., 2012; King & Scott, 2005). Our approach included multiple indicators of psychological well-being, including depressive symptoms, stress, and loneliness, that have been validated in prior research using the NSHAP data (Payne et al., 2014).

As we expected, partnership status was more salient for men than women. Among women, no partnership differences in depressive symptoms and perceived stress were observed. Cohabiting women were less lonely than the unpartnered, but were similar to dating and married women. In short, there appear to be few protective benefits of partnership status for older women, at least for psychological well-being. These findings for women are contrary to our expectations based on the marital status as a continuum of social attachment perspective. Partnered women may gain fewer benefits than men from their unions due to gendered spousal roles in older adulthood that often involve caregiving (Bernard, 1972). Wives tend to be the caregivers and husbands care recipients. Caregiving is stressful and can be detrimental to well-being (Pinquart & Sorensen, 2006). Additionally, prior work has shown that women’s self-rated health was not diminished by transitions to either widowhood or divorce (Williams & Umberson, 2004), which aligns with our finding that unpartnered and partnered women fared similarly in terms of psychological well-being.

Among men, stark differences in psychological well-being emerged by partnership status. Cohabitors were less likely to report frequent depressive symptoms and to perceive stress more than rarely compared with either daters or the unpartnered. Similarly, cohabitors reported lower levels of loneliness, on average, than either daters or the unpartnreds. These patterns aligned with our expectations. But, there were few differences between dating men and their unpartnered counterparts, which is not consistent with the continuum perspective. For unmarried men, co-residence is the key factor. Indeed, our study uncovered a cohabitation advantage over marriage for men. Cohabiting men less often reported frequent depressive symptoms and were somewhat less likely to perceive stress more than rarely compared with the married, which was contrary to our expectations. This pattern was not consistent with the partnership status as a continuum of social attachment perspective which posited marrieds would fare better than cohabitors. Nor did it support our alternative hypothesis that cohabitation operates as a substitute for marriage in later life and thus the well-being of the two groups should be comparable. In fact, these results are at odds with an earlier study showing that married men report fewer depressive symptoms than their cohabiting counterparts (Brown et al., 2005). This anomalous result could be an artifact of the changing composition of older cohabiting men as the earlier study relied on data from the late 1990s. Or, it simply could be due to the small number of male cohabitors in the sample. Alternatively, mortality selection may play a role, as cohabiting men have higher rates of mortality than married men and psychological distress is positively associated with morality (Liu & Reczek, 2012). Surviving older cohabiting men therefore should be especially robust which may translate into higher levels of well-being. Regardless, our findings merit attention in future research to determine whether they are replicable with other data on today’s older cohabitors. It also would be worthwhile to examine how the two groups compare on other dimensions of well-being.

From our study, we can conclude that older cohabiting men are no worse off than their married counterparts, and in fact they appear to appraise their psychological well-being more positively than do married men. This finding does not negate the possibility that cohabitation serves as an alternative to marriage in later life (Brown et al., 2006; King & Scott, 2005). Indeed, cohabitation may offer some unique benefits versus marriage, at least for men’s psychological well-being. Whereas cohabitation among young adults is marked by poorer outcomes, including lower levels of psychological well-being (Brown, 2000; Lamb, Lee, & DeMaris, 2003), the two types of unions are more comparable for older adults.

Our findings call into question the marital status as a continuum of social attachment perspective, which does not hold at all for older women and is only partially supported for older men. The meaning and purposes of partnerships are arguably unique in later life. Older adults seek out partners for different reasons than do younger adults. For example, older adults place greater emphasis on companionship whereas younger adults focus on love (Bulcroft & Bulcroft, 1991). Most older adults in cohabiting and dating partnerships have been married previously and are often uninterested in remarriage, and this is especially true for women (McWilliams & Barrett, 2014). Future research would benefit from the development of theory that is specifically geared toward older adult partnerships and recognizes the unique experiences of older women and men (Calasanti & Kiecolt, 2007; Cooney & Dunne, 2001).

Indeed, our study underscores the importance of examining the linkage between partnership status and well-being separately by gender. The experiences of women and men are distinctive. Consistent with prior work (Gove et al., 1983), these findings suggest that union status is more influential on the mental health of men than women in later life. We demonstrate that this conclusion applies across partnership statuses. The psychological well-being gains from partnership are negligible for women, but considerable for men.

There are some limitations to our study. For example, our sample size is modest, particularly among cohabitors and daters, which limits the statistical power of our analyses. However, the proportions of cohabitors and daters in our sample are consistent with population estimates for older adults (Brown et al., 2006; Brown & Shinohara, 2013). Moreover, we found notable differences in psychological well-being by partnership status, as well as significant differences between men and women, despite our modest sample size. A related limitation is that many of the covariates were measured as binary variables due to our modest sample size. It is preferable to have more detailed measures when possible. A small share of respondents may have experienced a partnership status transition between waves. This typically occurred through widowhood among the married but our findings remain robust. Whether these married to widowed individuals are included or excluded in the analytic sample, the results remain the same. For other partnership statuses, there are too few transitions to model and cell sizes become untenable. Finally, although our study can distinguish between cohabitors and daters, we are not able to explicitly capture unmarried older adults who are in Living Apart Together (LAT) relationships, which are arguably distinctive from dating in that they are more committed and may endure longer (and, by extension, provide greater mental health benefits) (Duncan & Phillips, 2011).

Overall, our study sheds new light on the role of partnership status among older adults by addressing the potential benefits of having an unmarried partner. Whereas prior research on older adult well-being has largely focused on marriage and widowhood, we document variation in psychological well-being among unmarried persons, particularly for men. Our findings demonstrate the need to consider the benefits of nonmarital unions for older adults, especially as the unmarried cohabiting and dating populations continue to grow in the future. Our study provides new evidence that cohabitation and marriage may offer similar benefits in later life by uncovering the comparable levels of psychological well-being enjoyed by the two groups.

Acknowledgments

The authors thank I-Fen Lin for her helpful comments on earlier versions of this paper. This research was supported in part 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 (R24HD050959).

Contributor Information

Matthew R. Wright, Department of Sociology, 222 Williams Hall, Bowling Green State University, Bowling Green, OH 43403-0222, mrwrigh@bgsu.edu, Phone: 419-372-2294

Susan L. Brown, Department of Sociology, 239 Williams Hall, Bowling Green State University, Bowling Green, OH 43403-0222, brownsl@bgsu.edu, Phone: 419-372-9521

References

  1. Bernard J. The future of marriage. New Haven, CT: Yale University Press; 1972. [Google Scholar]
  2. Booth A, Johnson DR. Declining health and marital quality. Journal of Marriage and the Family. 1994;56:218–233. [Google Scholar]
  3. Bratter JL, Eschbach K. Race/ethnic differences in nonspecific psychological distress: Evidence from the national health interview survey. Social Science Quarterly. 2005;86:620–644. [Google Scholar]
  4. Brown SL. The effect of union type on psychological well-being: Depression among cohabitors versus marrieds. Journal of Health and Social Behavior. 2000;41:241–255. [PubMed] [Google Scholar]
  5. Brown SL, Bulanda JR, Lee GR. The significance of nonmarital cohabitation: marital status and mental health benefits among middle-aged and older adults. Journal of Gerontology: Series B: Psychological and Social Sciences. 2005;60:21–29. doi: 10.1093/geronb/60.1.s21. [DOI] [PubMed] [Google Scholar]
  6. Brown SL, Lee GR, Bulanda JR. Cohabitation among older adults: A national portrait. Journal of Gerontology: Series B: Psychological and Social Sciences. 2006;61:71–79. doi: 10.1093/geronb/61.2.s71. [DOI] [PubMed] [Google Scholar]
  7. Brown SL, Bulanda JR, Lee GR. Transitions into and out of cohabitation in later life. Journal of Marriage and Family. 2012;74:774–793. doi: 10.1111/j.1741-3737.2012.00994.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brown SL, Lin IF. The gray divorce revolution: Rising divorce among middle-aged and older adults, 1990–2010. Journals of Geronotology: Series B: Psychological and Social Sciences. 2012;67:731–741. doi: 10.1093/geronb/gbs089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brown SL, Shinohara SK. Dating relationships in older adulthood: A national portrait. Journal of Marriage and Family. 2013;75:1194–1202. doi: 10.1111/jomf.12065. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bulcroft RA, Bulcroft KA. The nature and functions of dating in later life. Research on Aging. 1991;13:244–260. [Google Scholar]
  11. Cairney J, Krause N. The social distribution of psychological distress and depression in older adults. Journal of Aging and Health. 2005;17:807–835. doi: 10.1177/0898264305280985. [DOI] [PubMed] [Google Scholar]
  12. Calasanti T, Kiecolt KJ. Diversity among late life couples. Generations: Journal of the American Society on Aging. 2007;31:10–17. [Google Scholar]
  13. Carr D. The desire to date and remarry among older widows and widowers. Journal of Marriage and Family. 2004;66:1051–1068. [Google Scholar]
  14. Carr D, Springer KW. Advances in families and health research in the 21st century. Journal of Marriage and Family. 2010;72:743–761. [Google Scholar]
  15. Cherlin AJ. Demographic trends in the United States: A review of research in the 2000s. Journal of Marriage and Family. 2010;72:403–419. doi: 10.1111/j.1741-3737.2010.00710.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chevan A. As cheaply as one: Cohabitation in the older population. Journal of Marriage and the Family. 1996;58:656–667. [Google Scholar]
  17. Cooney TM, Dunne K. Intimate relationships in later life: Current realities, future prospects. Journal of Family Issues. 2001;22:838–858. [Google Scholar]
  18. Cornwell Erin York, Linda J, Waite Social disconnectedness, perceived isolation, and health among older adults. Journal of Health and Social Behavior. 2009;50:31–48. doi: 10.1177/002214650905000103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. De Jong Gierveld J. The dilemma of repartnering: Considerations of older men and women entering new intimate relationships in later life. Ageing International. 2002;27:61–78. [Google Scholar]
  20. De Jong Gierveld J. Remarriage, unmarried cohabitation, living apart together: Partner relationships following bereavement or divorce. Journal of Marriage and Family. 2004;66:236–243. [Google Scholar]
  21. Duncan S, Phillips M. People who Live Apart Together (LATs): New family form or just a stage? International Review of Sociology. 2011;21:513–532. [Google Scholar]
  22. Dykstra PA, van Tilburg TG, de Jong Gierveld J. Changes in older adult loneliness: Results from a seven-year longitudinal study. Research on Aging. 2005;27:725–747. [Google Scholar]
  23. Gove WR, Tudor JF. Adult sex roles and mental illness. American Journal of Sociology. 1973;78:812–835. doi: 10.1086/225404. [DOI] [PubMed] [Google Scholar]
  24. Gove WR, Hughes M, Style CB. Does marriage have positive effects on the psychological well-being of the individual? Journal of Health and Social Behavior. 1983;24:122–131. [PubMed] [Google Scholar]
  25. Greenfield Emily A, David Russell Identifying living arrangements that heighten risk for loneliness in later life: Evidence from the U.S. National Social Life, Health, and Aging Project. Journal of Applied Gerontology. 2011;30(4):524–534. [Google Scholar]
  26. Hansen T, Moum T, Shapiro A. Relational and individual well-being among cohabitors and married individuals in midlife: Recent trends from Norway. Journal of Family Issues. 2007;28:910–933. [Google Scholar]
  27. Hochschild A, Machung A. The second shift: Working families and the revolution at home. New York: Penguin; 2012. [Google Scholar]
  28. Kamp Dush CM, Amato PR. Consequences of relationship status and quality for subjective well-being. Journal of Social and Personal Relationships. 2005;22:607–627. [Google Scholar]
  29. Kennedy S, Ruggles S. Breaking up is hard to count: The rise of divorce in the United States, 1980–2010. Demography. 2014;51:587–598. doi: 10.1007/s13524-013-0270-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. King V, Scott ME. A comparison of cohabiting relationships among older and younger adults. Journal of Marriage and Family. 2005;67:271–285. [Google Scholar]
  31. Kreider RM, Ellis R. Current Population Reports, P70–125. Washington, DC: U.S. Census Bureau; 2011. Number, timing, and duration of marriages and divorces: 2009. [Google Scholar]
  32. Lamb KA, Lee GR, DeMaris A. Union formation and depression: Selection and relationship effects. Journal of Marriage and Family. 2003;65:953–962. [Google Scholar]
  33. Lin IF, Brown SL. Unmarried boomers confront old age: A national portrait. The Gerontologist. 2012;52:153–165. doi: 10.1093/geront/gnr141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Liu H. Till death do us part: Marital status and U.S. mortality trends, 1986–2000. Journal of Marriage and Family. 2009;71:1158–1173. [Google Scholar]
  35. Liu H, Reczek C. Cohabitation and U.S. adult mortality: An examination by gender and race. Journal of Marriage and Family. 2012;74:794–811. [Google Scholar]
  36. Manning WD, Brown SL. The demography of unions among older Americans, 1980-Present: A family change approach. In: Settersten RA, Angel JL, editors. Handbook of Sociology of Aging. New York: Springer; 2011. pp. 193–210. [Google Scholar]
  37. Marks NF. Flying solo at midlife: Gender, marital status, and psychological well-being. Journal of Marriage and the Family. 1996;58:917–932. [Google Scholar]
  38. McWilliams S, Barrett AE. Online dating in middle and later life gendered expectations and experiences. Journal of Family Issues. 2014;35:411–436. [Google Scholar]
  39. Mehta KM, Simonsick EM, Penninx BWJH, Schulz R, Rubin SM, Satterfield S, Yaffe K. Prevalence and correlates of anxiety symptoms in well-functioning Older adults: Findings from the Health Aging and Body Composition Study. Journal of the American Geriatrics Society. 2003;51:499–504. doi: 10.1046/j.1532-5415.2003.51158.x. [DOI] [PubMed] [Google Scholar]
  40. Mirowsky J, Ross CE. Social causes of psychological distress. New York: Aldine de Gruyter; 1989. [Google Scholar]
  41. Mirowsky J, Ross CE. Age and depression. Journal of Health and Social Behavior. 1992;33:187–205. [PubMed] [Google Scholar]
  42. O’Muircheartaigh C, Eckman S, Smith S. Statistical design and estimation for the national social life, health, and aging project. Journals of Gerontology: Series B: Psychological and Social Sciences. 2009;64:i12–i19. doi: 10.1093/geronb/gbp045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. O’Muircheartaigh C, English N, Pedlow S, Kwok P. Sample design, sample augmentation, and estimation for wave 2 of the NSHAP. Journals of Gerontology: Series B: Psychological and Social Sciences. 2014;69:s15–s26. doi: 10.1093/geronb/gbu053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Payne C, Hedberg E, Kozloski M, Dale W, McClintock MK. Using and interpreting mental health measures in the national social life, health, and aging project. Journals of Gerontology: Series B: Psychological and Social Sciences. 2014;69:s99–s116. doi: 10.1093/geronb/gbu100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pearlin LI, Johnson JS. Marital status, life-strains, and depression. American Sociological Review. 1977;42:704–715. [PubMed] [Google Scholar]
  46. Peters A, Liefbroer AC. Beyond marital status: Partner history and well-being in old age. Journal of Marriage and the Family. 1997;59:687–699. [Google Scholar]
  47. Pinquart M, Sorensen S. Influences on loneliness in older adults: A meta-analysis. Basic and Applied Social Psychology. 2001;23:245–266. [Google Scholar]
  48. Pinquart M. Loneliness in married, widowed, divorced, and never-married older adults. Journal of Social and Personal Relationships. 2003;20:31–53. [Google Scholar]
  49. Pinquart M, Sorensen S. Gender differences in caregiver stressors, social resources, and health: An updated meta-analysis. Journals of Gerontology: Series B: Psychological and Social Sciences. 2006;61:P33–P45. doi: 10.1093/geronb/61.1.p33. [DOI] [PubMed] [Google Scholar]
  50. Ross CE. Reconceptualizing marital status as a continuum of social attachment. Journal of Marriage and the Family. 1995;57:129–140. [Google Scholar]
  51. Sassler S. Partnering across the life course: Sex, relationships, and mate selection. Journal of Marriage and Family. 2010;72:557–575. doi: 10.1111/j.1741-3737.2010.00718.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Sayer LC. Gender, time and inequality: Trends in women's and men's paid work, unpaid work and free time. Social Forces. 2005;84:285–303. [Google Scholar]
  53. Thoits PA. Multiple identities: Examining gender and marital status differences in Distress. American Sociological Review. 1986;51:259–272. [Google Scholar]
  54. Vespa J. Union formation in later life: Economic determinants of cohabitation and remarriage among older adults. Demography. 2012;49:1103–1125. doi: 10.1007/s13524-012-0102-3. [DOI] [PubMed] [Google Scholar]
  55. Waite LJ. Does marriage matter? Demography. 1995;32:483–507. [PubMed] [Google Scholar]
  56. Waite LJ, Gallagher M. The case for marriage: Why married people are happier, healthier, and better off financially. New York, NY: Doubleday; 2000. [Google Scholar]
  57. Waite LJ, Lehrer EL. The benefits from marriage and religion in the United States: A comparative analysis. Population and Development Review. 2003;29:255–275. doi: 10.1111/j.1728-4457.2003.00255.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Williams K, Umberson D. Marital status, marital transitions, and health: A gendered life course perspective. Journal of Health and Social Behavior. 2004;45:81–98. doi: 10.1177/002214650404500106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. U.S. Census Bureau. Table UC3. Opposite sex unmarried couples by presence of biological children/1 under 18, and age, earnings, education, and race and Hispanic origin/2 of both partners: 2010. Current Population Survey, 2010 Annual Social and Economic Supplement. 2014 [Google Scholar]
  60. Yang Y. Is old age depressing? Growth trajectories and cohort variations in late-life depression. Journal of Health and Social Behavior. 2007;48:16–32. doi: 10.1177/002214650704800102. [DOI] [PubMed] [Google Scholar]

RESOURCES