Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Feb 1.
Published in final edited form as: J Fam Issues. 2021 Oct 28;44(2):538–559. doi: 10.1177/0192513x211054466

A Cohort Comparison of Midlife Marital Quality: A Quarter Century of Change

Matthew R Wright 1, Susan L Brown 2, Wendy D Manning 2
PMCID: PMC9850426  NIHMSID: NIHMS1791575  PMID: 36683860

Abstract

Marital quality has been declining among recent cohorts, but whether this pattern characterizes middle-aged and older married adults is largely unknown. The doubling of the divorce rate among persons over the age of 50 years foretells poorer quality marriages for today’s midlife adults than a generation ago. Combining data on married individuals aged 50–65 years from the 1987–88 National Survey of Families and Households (NSFH) and the 2013 Families and Relationships Study, we conduct a cohort comparison of five dimensions of midlife marital quality. Today’s older adults report more marital disagreement and instability as well as less fairness and interaction with their spouses than their counterparts did a generation ago. The two cohorts report comparable levels of marital happiness. Consistent with the upward trend in divorce during the second half of life, the quality of midlife marriages appears to have declined over the past quarter century.

Keywords: marriage, marital quality, gray divorce, intimate relationships, dyadic relationship/quality/satisfaction

Marital quality has declined in recent decades, at least for younger adults. Compared with a generation prior, couples in the late 1990s and early 2000s reported less marital happiness and interaction as well as more marital conflict (Amato, Johnson, Booth, & Rogers, 2003; Glenn, 1998; Rogers & Amato, 1997). These findings come from data that are now dated and, to the best of our knowledge, no contemporary studies have examined cross-cohort change in marital quality, nor have they examined trends in marital quality during the second half of life. Thus, the dynamics marking marital quality in middle and later life are unclear.

This is a notable omission given the recent doubling of the divorce rate among persons over the age of 50 years, termed the gray divorce rate. Older couples who report poorer marital quality are more likely to experience gray divorce (Lin, Brown, Wright, & Hammersmith, 2018). The rise in gray divorce is tied to the shifting meaning of marriage, which is now more individualized as it entails high standards for success, including self-fulfillment and personal happiness (Cherlin, 2004), a high bar for couples to achieve and sustain, which may depress marital quality among today’s midlife married adults.

The negative association between marital quality and divorce in later life (Lin et al., 2018) coupled with the increase in the gray divorce rate suggests that midlife marital quality has likely declined over time, but this remains an empirical question. Prior research on trends in marital quality over time provides little insight either because older adults were excluded or differences by age group were not assessed (Amato et al., 2003; Glenn, 1991; Rogers & Amato, 1997, 2000).

Using data from the 1987–1988 National Survey of Families and Households (NSFH) and the 2013 Families and Relationships Survey (FRS), we assess the extent to which five dimensions of marital quality have changed between two cohorts of midlife adults aged 50 to 65 years spanning a 25-year period. We examine descriptive change in both positive and negative facets of marital quality. Moreover, we address how factors associated with midlife marital quality have shifted over time and whether the relationships between these correlates and marital quality have changed between the two cohorts. Finally, our study investigates potential gender differences in midlife marital quality changes. The findings from this study extend prior research by providing new insight into the quality of midlife marriages. Understanding marital quality among middle-aged and older adults is especially important because marital dissolution in later life now occurs more often through divorce, and both marital quality and divorce have implications for the well-being of the middle-aged population (Carr, Freedman, Cornman, & Schwarz, 2014; Umberson, Williams, Powers, Liu, & Needham, 2006; Zhang, Liu, & Yu, 2016).

Background

Although the overall divorce rate in the United States has remained stable over the past 20 years (Kennedy & Ruggles, 2014), this pattern obscures variation in the divorce rate trend by age. After accounting for population aging, the divorce rate has actually increased over the past few decades. This conclusion aligns with research showing a rising divorce rate among middle-aged and older adults (Brown & Lin, 2012; Kennedy & Ruggles, 2014; Stepler, 2017). In 1990, the divorce rate among married persons aged 50 years and older was just 4.9 per 1000, but by 2010, the rate had climbed to 10 per 1000. Even if the divorce rate does not change, the number of persons experiencing gray divorce is predicted to rise by one-third by 2030 due to the size of the middle-aged and older adult population increasing (Brown & Lin, 2012). This new research on divorce in middle and later life foregrounds the importance of investigating potential shifts in midlife marital quality in recent decades.

Moreover, marital quality plays a key role in the health and well-being of married older adults. It is well-documented that married adults tend to enjoy better health than the unmarried (Umberson & Thomeer, 2020). However, research suggests that relationship quality may be more important for health than marital status (Miller, Hollist, Olsen, & Law, 2013) and that the health benefits of marriage depend on marital quality. Previous work has shown that higher levels of marital quality may amplify the health benefits of marriage (Kiecolt-Glaser & Newton, 2001), whereas poorer quality marriages may offer no benefits or even contribute to declining health (Umberson et al., 2006; Zhang et al., 2016). The association between marital quality and well-being appears to be especially pronounced for older adults (Liu & Waite, 2014; Umberson et al., 2006).

Research also has reported that negative dimensions of marital quality may be more consequential for health than positive aspects (Miller et al., 2013; Umberson & Thomeer, 2020). Indeed, prior work has indicated that positive and negative components of relationship quality are not the inverse of each other, rather they are separate dimensions (Johnson, White, Edwards, & Booth, 1986). These findings suggest that positive and negative dimensions of marital quality should be assessed independently. Thus, we examine five dimensions of marital quality separately.

Marital Quality Change

Marital quality is negatively associated with gray divorce (Lin et al., 2018). Therefore, the doubling of the gray divorce rate suggests that marital quality may have plummeted over the past few decades among midlife adults (Lin et al., 2018). This remains unclear, however, as research on changes in marital quality has been sparse in recent years and relatively limited with regard to middle-aged and older adults. Studies typically have indicated that the quality of marriages in more recent samples and cohorts is lower than that of previous samples and cohorts. Between 1973 and 1988, the percentage of married persons who reported being in very happy marriages dropped (Glenn, 1991). Similarly, Glenn (1998) found marital quality declined across several 10-year marriage cohorts. Likewise, levels of marital interaction were lower and marital conflict and problems were higher in more recent marriage cohorts compared to earlier marriage cohorts (Rogers & Amato, 1997, 2000). More recently, a study investigating marital quality change between 1980 and 2000 showed stability in marital happiness and divorce proneness (Amato et al., 2003), whereas marital interaction declined over the same time span. Overall, research indicates that marital quality is lower among more recent cohorts than earlier cohorts.

These prior studies on change in marital quality at best provide limited insights on marital quality among midlife and older adults. The data examined in previous work are now somewhat old, as the most recently used data across the studies were collected in 2000. Moreover, several of the studies did not consider midlife or older adults (Amato et al., 2003; Rogers & Amato, 1997, 2000). Rather, their samples only included those aged 55 years and younger. Although Glenn (1991) retained older adults, his analysis did not assess whether there were differences in patterns of marital quality change by age group and his study preceded the onset of the gray divorce revolution.

The few published studies on marital quality among middle-aged and older married persons do not cover change from prior to the gray divorce revolution to today. As with other research, the data utilized by Glenn (1998) are now dated. Drawing on the 1973–1994 General Social Survey, Glenn indicated that marital quality was fairly stable in middle and later life. A recent study covering a narrow time span (4 years) showed within couple declines in marital quality from 2009 to 2013 among married persons in which both spouses were at least 50 years old and at least one spouse was 60 years or older (Stokes, 2017). The upshot is that it remains unclear whether marital quality among older adults has declined over the same time period that the gray divorce rate has risen.

Demographic Change and Marital Quality

A key tenet of the life course perspective is that the historical time in which people live has important implications for their lives (Elder, 1994). The two midlife (aged 50–65 years) cohorts in our study are distinct, as respondents in the 1987–1988 NSFH were born between 1922 and 1937 and those in the 2013 FRS were born from 1948–1963. The sociohistorical contexts for these two cohorts are also unique, which may contribute to lower marital quality among the FRS cohort than the NSFH cohort. For instance, the NSFH cohort came of age and married during an era of traditional gender roles. Men were often breadwinners and women were socialized as nurturers with primary responsibility for the success of their marriages (Boerner, Jopp, Carr, Sosinsky, & Kim, 2014; Sullivan, 2006). Conversely, individuals from the FRS cohort came of age more recently, during a period marked by comparatively greater emphasis on fairness and equity in gender roles, such as household labor and paid employment. Increasingly, expectations for marital success rely less on gendered roles as prevailing expectations for both women and men are to be both economic contributors and nurturers focused on marital success (Boerner et al., 2014; Pruchno, 2012; Sullivan, 2006). In other words, the growing norm of gender equality within marriage may place greater strain on marriages among the FRS cohort, as a failure to realize this equality norm could lead to lower marital quality, especially among women.

More broadly, the institution of marriage has undergone a great deal of change in the past several decades. The meaning of marriage has shifted from companionate to individualistic, characterized by a greater focus on self-fulfillment and personal growth rather than role performance (e.g., being a good husband or wife) (Cherlin, 2004; Wu & Schimmele, 2007). With more emphasis in marriages being placed on individualism, expectations have become much higher and potentially more difficult to meet. If one’s marriage falls short of these standards, marital quality is likely to be lower. Thus, with the meaning of marriage having changed over time, marital quality may be lower among today’s married midlife adults than 25 years ago.

Along with shifts in the meaning of marriage, changes in the demographic composition of the older population also may have contributed to variation in marital quality over time. One such change is the proportion of older adults in a higher order marriage. Despite the remarriage rate remaining steady among older adults over the past two decades (Brown & Lin, 2013), an increasing share is in a remarriage with approximately 30% of married older persons today in a higher order marriage, compared to just 19% in 1980 (Brown & Lin, 2012; Brown & Wright, 2017). The gray divorce rate is 2.5 times higher among remarried older adults than those in a first marriage (Brown & Lin, 2012).

Family scholars often argue that remarriage is an incomplete institution (Cherlin, 1978), lacking the norms that typically guide behavior for those in a first marriage. The absence of shared norms may then lead to greater tension and stress within a remarriage. This is particularly the case because remarriages often involve stepfamilies, which present challenges for couples (Uhlenberg & Myers, 1981). Thus, remarriage may be associated with lower levels of marital quality, relative to first marriages. Previous research on marital quality differences by marriage order, however, has yielded mixed results. Some studies show that remarried older adults report lower marital happiness than those in their first marriage (Kaufman & Taniguchi, 2006), whereas other research indicates no differences between the first married and remarried (Bulanda, 2011; Cooney, Proulx, & Snyder-Rivas, 2016) in later life. To the extent that marriage order is linked to quality, midlife marital quality may have declined over the past 25 years in part because a greater proportion of the married population is in a higher order marriage.

Gender Differences in Marital Quality Change

Women and men often experience marriage quite differently, with men potentially reaping greater benefits from being married than women (Bernard, 1972). Likewise, women and men frequently have different evaluations of their marriages. Men report higher marital quality and satisfaction, on average, than women (Amato et al., 2003; Bulanda, 2011; Cooney et al., 2016; Rogers & Amato, 1997, 2000; Stokes, 2017). Yet, it is unclear if and how cohort variation in midlife marital quality differs by gender. Previous research examining gender differences in marital quality shifts over time indicate that marital quality declines for both women and men, but that rates of decline do not significantly differ by gender (Stokes, 2017; Umberson, Williams, Powers, Chen, & Campbell, 2005; VanLaningham, Johnson, & Amato, 2001). These studies, however, used panel data to investigate change within married couples.

Prior scholarship on changing marital quality across cohorts also has yielded somewhat mixed results. As previously noted, these findings regarding gender differences are limited because they pertain only to adults aged 55 years and younger. The rise in marital discord between 1964 and 1980 was larger for wives than husbands, though the increases for women and men did not achieve significance (Rogers & Amato, 2000). A more recent study found that the difference in marital happiness between women and men was smaller in 2000 than in 1980 (Amato et al., 2003). Consistent with the shrinking gender happiness gap found by Amato et al. (2003), shifting demographic patterns in later life unions suggest that marital quality may differ less by gender today than 25 years ago. Men are more likely to repartner in later life following a marital dissolution than are women (Brown, Lin, Hammersmith, & Wright, 2018; Schimmele & Wu, 2016), and across the life course, remarriage is more common among men than women (Brown & Lin, 2013; Raley & Sweeney, 2020). With remarriage being more common among men than women, men may have experienced more decline in marital quality than women. Overall, little is known about how changes in marital quality across cohorts of midlife and older adults vary by gender.

The Current Study

The current investigation is designed to examine how marital quality has changed among two cohorts of middle-aged adults spanning a 25-year period. Prior research has consistently shown that marital quality tends to decline across cohorts (Amato et al., 2003; Glenn, 1991, 1998; Rogers & Amato, 1997, 2000), although this work is limited in that middle-aged and older adults are either excluded or the data analyzed are dated. The doubling of the gray divorce rate over the past two decades, coupled with the negative association between marital quality and dissolution (Lin et al., 2018) suggest that marital quality among the midlife population also may have declined, but research to date has not examined whether this is the case.

We anticipate that marital quality has declined for several reasons. First, the context of marriage has changed in recent decades. Marriage today is characterized by individualism, with an emphasis on personal growth and self-fulfillment (Cherlin, 2004; Wu & Schimmele, 2007). Given the higher expectations for marriages, marital quality is likely to have declined between cohorts if spouses are increasingly unable to meet each other’s desires. Moreover, demographic changes in marriages may have contributed to decline in marital quality over time. A greater proportion of older adults today are in a higher order marriage, and some research suggests that quality tends to be lower in remarriages than first marriages (Kaufman & Taniguchi, 2006). We consider whether the association between remarriage and marital quality differs between cohorts.

Gender is a key factor in reports of marital quality among older adults. Men typically report higher marital quality than women (Bulanda, 2011; Cooney et al., 2016; Stokes, 2017). Research among adults younger than 55 years found that changes in marital quality across cohorts is similar for women and men or that the gap between them has decreased over time (Amato et al., 2003; Rogers & Amato, 2000). Demographic patterns in later life unions suggest that cohort variation in marital quality may differ for men and women. Remarriage is more common for men than women (Brown & Lin, 2013; Raley & Sweeney, 2020). Thus, we might anticipate a greater decline in marital quality among men than women between cohorts. We test gender interactions to determine whether the association between cohort and marital quality varies for women and men.

Our analysis includes controls for several demographic and economic characteristics associated with marital quality. Key demographic factors include age, gender, race, whether the respondent is in a remarriage, marital duration, and resident children. Age is positively associated with marital quality (Umberson et al., 2005). Research has consistently shown that men report higher levels of marital quality than women (Amato et al., 2003; Bulanda, 2011; Cooney et al., 2016; Rogers & Amato, 1997, 2000; Stokes, 2017). Blacks tend to have lower marital quality than Whites and Hispanics (Bulanda, 2011; Bulanda & Brown, 2007), whereas marital quality is similar between Whites and Hispanics (Bulanda & Brown, 2007). Studies on remarriage are mixed, with some suggesting that remarrieds have lower marital quality than first marrieds (Kaufman & Taniguchi, 2006) and others finding no differences by marriage order (Bulanda, 2011; Cooney et al., 2016). Marital quality typically decreases as marriage duration increases (Umberson et al., 2005; VanLaningham et al., 2001). We include a control for children in the household. The presence of young children is negatively associated with happiness and interaction (Bulanda, 2011). We also account for several economic factors that are related to marital quality in later life. Findings regarding education are mixed. One study reported that higher levels of education are linked to less happiness and interaction (Bulanda, 2011). Another study found that greater education was negatively associated with marital satisfaction for men, but positively related among women (Kaufman & Taniguchi, 2006). Still other research indicates that education is positively associated with marital satisfaction for men (Kulik, 2002). Income is a significant predictor of marital quality for older men, though it is generally not correlated with quality among women (Bulanda, 2011; Kaufman & Taniguchi, 2006). Employment is positively associated with marital quality (Bulanda, 2011; Moen, Kim, & Hofmeister, 2001) and negatively with gray divorce (Brown & Lin, 2012).

Method

This study uses data from two nationally representative surveys to perform a cohort comparison of midlife marital quality over a 25-year time period. The early cohort (the 1987 cohort) was derived from Wave I of the NSFH, collected in 1987–1988 from a multistage probability sample of 13,017 respondents. The NSFH included individuals aged 19 years and older, with oversamples of minorities, recently married couples, cohabiting couples, and single-parent and stepfamily households (Sweet, Bumpass, & Call, 1988). Data were collected through an in-person interview with a randomly chosen adult in the household selected to be the primary respondent. The response rate in the first wave of the NSFH was 74%.

The late cohort (the 2013 cohort) was derived from the FRS, which was devised and sponsored by the National Center for Family and Marriage Research (NCFMR) at Bowling Green State University. Consisting of 7517 adults aged 18–65 years, the FRS was collected in 2013 by GfK group (formerly Knowledge Networks) through an online panel called KnowledgePanel, which was established using probability sampling with random digit dialing and address-based sampling. The sample frame covered both the online and offline population, as GfK provided technology and internet access as necessary. The completion rate for the FRS was 58.8%. These data are appropriate for the current study because the FRS was broadly modeled on the first wave of the NSFH. Thus, both datasets include the same items measuring marital quality, allowing for the examination of change over the past 25 years for the two midlife cohorts.

We examined married adults aged 50 to 65 years with valid data on each of the marital quality measures. There were 2263 adults aged 50–65 years in the NSFH. We excluded 901 respondents who were unmarried or whose marital status could not be determined. Nine respondents who married prior to age 15 years were also eliminated, as were 229 individuals who were missing on one or more of the marital quality measures. This produced a final sample size of 1124 from the NSFH for our early (1987) cohort.

The FRS included 2952 adults aged 50–65 years. Of these respondents, 1206 were excluded because they were unmarried or their marital status could not be determined. Three additional individuals were eliminated because they reported having married prior to age 15 years. Finally, 111 respondents were excluded due to missing data on one or more of the marital quality measures. Our final sample size for the FRS was 1632 in the late (2013) cohort.

Measures

The question wording and response options on the surveys were generally identical across the two datasets. We note where differences existed and how we handled them.

Dependent variables.

Five dimensions of marital quality were examined in this study: disagreement, fairness, happiness, interaction, and instability. Our disagreement measure consisted of respondent reports on how often they had open disagreements about five items, including household tasks, money, spending time together, sex, and in-laws. In the NSFH, response options were “never,” “less than once a month,” “several times a month,” “about once a week,” “several times a week,” and “almost every day.” The FRS included “once a month” as an extra response category. Thus, we combined the “less than once a month” and “once a month” responses in the FRS to make the measures comparable. The scale score for disagreement was created by summing the nonmissing items for those with valid responses to at least three of the items, taking the average, and multiplying by five to place all respondents on the same scale. The Cronbach’s alpha coefficients for disagreement were 0.67 for the NSFH and 0.74 for the FRS.

Our second marital quality measure was fairness. Respondents were asked how they felt about the fairness in their relationships regarding three items: household chores, working for pay, and spending money. The response options included “very unfair to respondent,” “somewhat unfair to respondent,” “fair to both,” “somewhat unfair to spouse,” and “very unfair to spouse.” Following Brown and Booth (1996), low values signified less fairness and high values greater fairness, with each item coded as 0 = very unfair to respondent or spouse,1= somewhat unfair to respondent or spouse, and 2 = fair to both. The scale was constructed by summing the nonmissing items for respondents with valid data on at least two of the three items. Next, we took the average of the valid items and multiplied by three to put respondents on the same scale. The Cronbach’s alpha coefficients were 0.70 in the NSFH and 0.64 in the FRS.

The final three dimensions of marital quality were single item measures. Happiness was a dichotomous variable based on respondent reports to the question “Taking things all together, how would you describe your marriage?” Families and Relationships Survey respondents were asked to describe their current relationship. Response options in both surveys ranged from 0 = very unhappy to 6 = very happy. Consistent with Schoen, Astone, Rothert, Standish, and Kim (2002), we created a variable in which respondents in the top two categories were coded as 1 = happy and all other respondents were coded as 0 because relationship happiness tends to be quite skewed. Interaction was the frequency of time that the respondent spent alone with their spouse, talking, or sharing an activity in the past month. We coded the response options as 0 = never, 1= about once a month, 2 = two or three times a month, 3 = about once a week, 4 = two or three times a week, and 5 = almost every day. Finally, instability was measured by the question “During the past year, have you ever thought that your marriage (relationship) might be in trouble?” We coded it as 1 = yes and 0 = no.

Focal independent variable.

Cohort was included as a dichotomous variable, with NSFH respondents interviewed in 1987 coded as 0 and FRS respondents from 2013 coded as 1.

Demographic characteristics.

Age was a continuous measure in years. Gender was a dichotomous measure coded as 0 = man and 1 = woman. Race was a categorical variable coded as White (reference), Black, and other race. Whether the respondent had any resident children was captured as a dichotomous variable with 1 = yes and 0 = no. Respondents were considered to have children in the household if any related person(s) younger than 18 years was identified as residing with the respondent. Marital duration was measured as a continuous variable in years. Remarriage was a dichotomous variable coded as 1 = currently in a higher order marriage and 0 = in a first marriage.

Economic resources.

Education was a categorical variable with less than high school, high school (reference), some college, and college or more. Income was the household income in the year of the survey. The response options differed slightly between the two surveys. In both, household income was a continuous measure in dollars. However, respondents in the FRS who did not provide a value selected a range that included their income. The 1987 income values from the NSFH were adjusted for inflation, and are presented in 2013 dollars. To obtain consistency across data sources we created a five category measure: less than $25,000 (reference), $25,000–39,999, $40,000–74,999, $75,000–99,999, and $100,000 or greater. This follows prior research that has utilized household income using the FRS data (Brown, Manning, & Wu, 2017). Employment was a dichotomous variable indicating whether the respondent was working at a paying job, coded as 1 = yes and 0 = no.

Analytic Strategy

We first document trends in the five marital quality dimensions and the demographic and economic factors considered in the study. Next, multivariate analyses were estimated for the full sample to determine if there were significant changes, net of other factors, between cohorts across each dimension of marital quality. We estimated ordinary least squares regression (OLS) models for the continuous outcomes of disagreement, fairness, and marital interaction. Logistic regression was used for marital happiness and instability because they are dichotomous variables. Models including interactions were examined to assess the extent to which marital quality changes between cohorts vary by gender. Finally, we estimated models to test whether the association of remarriage and marital quality varied between cohorts.

Missing data were modest in both datasets, with two exceptions. In the NSFH, less than 1% of cases were missing on all but two variables. Nearly one-quarter of cases were missing on income and just over 20% were missing on marital duration. We created flags for each of these variables, with missing respondents coded as 1 and not missing 0. Less than 2% of cases in the FRS were missing on the variables included in the study. To handle missing cases, we used mean substitution for continuous variables and modal substitution for categorical variables. We present weighted descriptive statistics and unweighted multivariate regression models. Johnson and Elliott (1998) found that design effects are more influential on descriptive statistics than multivariate models. Weighted and unweighted models analyzing marital happiness in the NSFH produced similar results (Johnson & Elliott, 1998). Thus, analysis of unweighted models should not produce much, if any, bias.

Results

Descriptive Results

Table 1 displays means and distributions for the study variables by cohort. The differences between cohorts were statistically significant in four of the marital quality measures. Marital quality averages are lower today across four of the five dimensions than they were 25 years ago. The overall levels of disagreement were quite low. In the 1987 cohort, the mean score on the disagreement scale was 2.4 (out of 25) and increased to 3.5 in 2013. The average on the fairness scale was relatively high, at 5.5 (out of 6) in 1987. Respondents in the 2013 cohort reported less fairness in their marriages, though the level was still quite high, with an average score of 5.1. Despite higher disagreement and a decline in fairness for the 2013 cohort, the share of respondents who reported being happy slightly increased from 78.3% to 79.1%. Levels of marital interaction declined between cohorts. The average in the 1987 cohort was 4.4 (out of 5) and was 4.1 in 2013. Just over one-tenth (12.5%) of respondents in 1987 thought their marriage could be in trouble. By 2013, the proportion had increased to nearly one in five (17.3%). Overall, marital quality among midlife married individuals tended to be relatively high, though across four of the five measures, marital quality declined between cohorts. Marital happiness, however, increased only slightly between cohorts.

Table 1.

Weighted Descriptive Statistics by Cohort.

Cohort
Variable 1987 2013
Marital quality (means)
 Disagreement 2.4 3.5***
 Fairness 5.5 5.1***
 Happiness (%) 78.3 79.1
 Interaction 4.4 4.1***
 Instability (%) 12.5 17.3**
Demographic characteristics
 Age 57.3 57.1
Gender (%)
 Woman 49.3 50.3***
 Man 50.7 49.7***
Race/ethnicity (%)
 White 87.7 75.1
 Black 6.7 10.0***
 Other race 5.6 14.9***
Children living in household (%)
 Resident child 17.0 18.0
Marriage characteristics
 Marital duration (mean) 30.8 26.1***
 Remarriage (%) 19.9 36.6***
Education (%)
 Less than high school 24.8 9.5***
 High school 44.0 38.1***
 Some college 13.8 25.5***
 College 17.4 26.9***
Household income (%)
 Less than $25,000 8.3 7.4**
 $25,000–$39,999 6.4 11.5**
 $40,000–$74,999 21.1 25.4***
 $75,000–$99,999 12.8 20.6***
 $100,000 or more 51.4 35.1***
Employment (%)
 Employed 59.0 64.4***
Unweighted N 1124 1632
Open in a new tab
*

p < .05,

**

p < .01,

***

p < .001.

Descriptive statistics for the demographic and economic variables are also presented by cohort in Table 1. The average age was approximately 57 years for both cohorts. A slightly greater proportion of the 1987 cohort was men (51%) than women (49%), whereas the 2013 cohort was evenly split. The racial/ethnic composition of married respondents significantly differed across time. Nearly nine in ten (88%) respondents in 1987 were White, compared to three-fourths (75%) in 2013. The proportion of Blacks increased from 7% to 10%, and the percentage of respondents of other races almost tripled from 6% to 15%. The increase in the other race category is likely due to the rise in the Hispanic population over time. There was a small uptick in the percentage of middle-aged adults with at least one resident child (17%–18%), though the difference was not statistically significant. As expected, there were significant changes in marital duration and the proportion remarried between cohorts. The average marital duration declined by nearly 5 years. In 1987, the mean duration was almost 31 years, compared to just 26 years in 2013. The decrease in marital duration is likely due, in part, to the increasing proportion of remarried individuals. One-fifth (20%) of respondents in the 1987 cohort were in a higher order marriage. Over one-third (37%) of married persons in the 2013 cohort were in a remarriage.

Turning to economic characteristics, married individuals in 2013 reported higher education than those in 1987. For both cohorts, having a high school degree was the modal category. Not quite half (44%) of respondents in 1987 had a high school degree, and this declined to 38% in 2013. Almost one-fourth (25%) of respondents in 1987 had less than a high school degree, compared to fewer than 10% in 2013. Both the some college and college degree categories increased over time. In 1987, about 14% had some college and this rose to 26% in 2013. Similarly, 17% had a college degree in 1987 and the figure increased to 27% in 2013. The proportions of respondents in the lowest (less than $25,000) and highest ($100,000 or more) household income brackets both significantly declined between 1987 and 2013, whereas each of the three middle categories increased over time. Over half (51%) of respondents fell into the highest bracket in 1987, but only 35% did in 2013. Finally, just under two-thirds (64%) were still employed in 2013, a significantly higher percentage than in 1987 (59%).

Multivariate Results

The results of multivariate models predicting each marital quality indicator are presented in Table 2. As shown in Model 1, cohort was positively related to disagreement. Married midlife adults in 2013 reported higher levels of disagreement than those in 1987, net of demographic and economic factors. Notably, women reported more disagreement than men. Neither marital duration nor being in a remarriage was associated with disagreement. Likewise, Whites did not significantly differ from Blacks or those in other racial/ethnic groups. Age was negatively associated with disagreement. On the other hand, those with resident children had higher levels of disagreement than those without children in the household. Overall, disagreement among midlife adults has increased between cohorts.

Table 2.

Regression Models Predicting Marital Quality (n = 2756).

Variable Model 1
 Model 2
 Model 3
 Model 4
 Model 5
Disagreement
 Fairness
 Marital happiness
 Marital interaction
 Instability
b b OR b OR
Cohort
 1987 (ref)
 2013 1.02*** −0.43*** 1.13 −0.39*** 1.43**
Demographic characteristics
 Age −0.04* 0.02*** 1.01 0.02* 0.96**
Gender
 Man (ref)
 Woman 0.29* −0.21*** 0.70*** −0.03 1.39**
Race/ethnicity
 White (ref)
 Black 0.24 −0.01 0.78 −0.02 0.73
 Other race 0.32 −0.00 1.28 −0.06 0.82
Children living in household
 Resident child 1.02*** −0.25*** 0.71** −0.83*** 1.27
Marriage characteristics
 Marital duration −0.01 −0.00 1.01 0.00 0.99
 Marital duration flag 0.29 −0.18* 0.79 −0.23* 1.27
 Remarriage 0.23 −0.05 0.99 0.03 1.57**
Education
 Less than high school −0.10 −0.07 1.06 −0.17* 0.79
 High school (ref)
 Some college 0.01 −0.01 0.84 −0.01 1.05
 College 0.24 −0.06 0.91 0.05 1.35*
Household income
 Less than $25,000 (ref)
 $25,000–$39,999 −0.10 0.05 0.81 −0.09 1.01
 $40,000–$74,999 −0.42 0.06 1.18 0.10 0.76
 $75,000–$99,999 −0.41 −0.02 0.98 0.04 0.88
 $100,000 or more −0.43 1.05 0.05 0.78
 Income flag −0.02 −0.17 0.88 −0.10 1.25
Employment
 Employed 0.19 0.03 0.95 −0.08 1.07
Constant 4.14*** 4.49*** 2.80 3.71*** 1.61
Open in a new tab
*

p < .05,

**

p < .01,

***

p < .001.

Model 2 of Table 2 displays the results for fairness, which were consistent with disagreement. Cohort was negatively and significantly associated with fairness indicating that married persons in 2013 noted lower levels of fairness than their counterparts in 1987. As with disagreement, women reported less fairness than men. Marital duration and being in a remarriage were not associated with fairness. Whites did not differ from the other racial/ethnic groups on fairness. Across the other control variables, age was positively related to fairness and having at least one resident child was linked to lower levels of fairness among midlife married individuals. Respondents missing on marital duration noted less fairness in their marriages than those who were not missing. In sum, these findings regarding fairness suggest that marital quality has declined between cohorts.

There was no significant difference in marital happiness by cohort in Model 3. In other words, it appears that marital happiness remained stable between cohorts. Consistent with the prior indicators, women noted lower quality than men, as marital happiness was higher among men than women. Marital duration and remarriage were not significant predictors of marital happiness. White respondents did not differ from other racial/ethnic groups on marital happiness. The presence of children in the household was negatively associated with marital happiness. In short, our examination of marital happiness indicates no difference by cohort in this indicator of marital quality.

Turning to marital interaction, the results of the OLS regression model are presented in Model 4. As with the disagreement and fairness indicators, cohort was significantly related to marital interaction. Marital interaction declined between cohorts and was lower among married midlife adults in 2013 than in 1987. There were no differences between women and men on marital interaction, nor did Whites differ from Blacks or other race respondents. Marital duration and remarriage were not significant. A few of the other control variables were predictors of marital interaction. As age increased, levels of marital interaction did as well. Having a resident child was associated with lower levels of interaction. Respondents with less than a high school education reported less marital interaction than those with a high school degree. Those with missing information on marital duration reported less marital interaction than those who were not missing. Again, these results suggest that marital quality has diminished between cohorts.

Finally, Model five contains the results for a logistic regression model predicting whether respondents thought their relationship was in trouble over the past year. Similar to the results for disagreement, fairness, and marital interaction, the findings regarding instability indicated that the quality of midlife adults’ marriages has declined between cohorts. The odds of having thought the marriage was in trouble during the past year were 1.4 times higher in 2013 than in 1987. Women were more likely than men to report marital instability. Similarly, remarried individuals had higher odds of being in an unstable marriage than those in their first marriage. There were no differences by marital duration. Likewise, Whites did not significantly differ from Blacks or other race respondents. Age was associated with a lower likelihood of believing one’s marriage was in trouble over the past year. Having a resident child increased the odds of marital instability, though the effect was only marginally (p < .10) significant.

Additional models were estimated to examine if the association of remarriage and marital quality varied by cohort (not shown). None of the remarriage and cohort interactions were significant, indicating that the relationship between remarriage and marital quality did not differ between cohorts. To determine if there were gender differences in the association between cohort and marital quality, we tested gender and cohort interactions (not shown). Only one of the interactions was significant. Men experienced a steeper decline in marital interaction than women between 1987 and 2013. Consistent with our expectations, it appears that marital interaction diminished more between cohorts among men than women.

Discussion

An extensive literature documents a decline in marital quality across cohorts, yet middle-aged and older adults have been largely overlooked in prior research. The gray divorce revolution, marked by a doubling of the divorce rate among married adults aged 50 years and older, signals that marital quality could be declining during the second half of life. Indeed, marital quality is negatively associated with the risk of gray divorce (Lin et al., 2018). Combining data on married adults aged 50–65 years from the 1987–88 NSFH and the 2013 FRS, we tested whether marital quality has decreased between these two cohorts spanning a 25-year period.

Our study shows that marital quality has dropped among midlife adults. Across four out of five dimensions of marital quality, today’s midlife marriages are characterized by lower quality than midlife marriages were a quarter century ago. Midlife adults report more marital disagreement, less fairness, and less interaction with their spouses these days. They are also more likely to report that their marriage may be in trouble than were their counterparts in the late 1980s. Although it appears that marital happiness has remained unchanged between cohorts, the preponderance of the evidence indicates that today’s midlife married adults are in poorer quality marriages than their counterparts were a generation ago.

These marital quality differentials hold net of controls for demographic and economic factors. We anticipated that the quality of today’s midlife marriages may be lower than in the past in part because a rising share of these marriages are remarriages, not first marriages. Our results indicate that net of other factors, the quality of remarriages and first marriages do not appreciably differ, although a key exception is that marital instability is higher in remarriages than first marriages. Thus, the higher remarriage levels do not appear to be the reason for the decline in marital quality. We also investigated whether the association between remarriage and marital quality differed between cohorts. None of the interactions were significant, suggesting that the relationship does not vary by cohort.

The lower levels of marital quality among the more recent cohort of older adults may be related to changes in expectations for marriage, as well as equality norms within contemporary marriages. Relative to a generation ago, today’s marriages are more often characterized by greater focus on self-fulfillment and personal growth (Cherlin, 2004; Wu & Schimmele, 2007). Consequently, expectations for marriage have become higher and may be more difficult to meet and sustain. Likewise, greater emphasis on fairness and gender equality in more contemporary marriages could also lead to higher expectations for marital success that are harder to meet. We were not able to explicitly assess the extent to which these higher expectations are related to marital quality, an important task for future research.

Given the sharp rise in the gray divorce rate over the past few decades, we might have expected greater cohort differences in marital quality than we uncovered. That the differences were not larger could be due to selection into divorce. Along with the increase in gray divorce, older adults today are more accepting of divorce compared to older birth cohorts (Brown & Wright, 2019). Couples with the poorest quality marriages in the more recent cohort presumably would have been more likely to dissolve their marriages rather than remain in an unhappy relationship, leaving those with relatively higher quality marriages remaining in the cohort.

While this study provides new insights into marital quality trends, it has some limitations. First, it is cross-sectional, so we are only capturing married individuals at one point in time. Ideally, we would have longitudinal data with which to compare how different cohorts perceive their marital quality over time. But such data are not available. We rely on the best available data to address whether and how marital quality is shifting during the second half of life. Second, we acknowledge that having couple-level data would be ideal. Although the NSFH included spouses, the FRS did not, precluding couple-level analyses. Couple-level data are especially important for assessments of gender and relationship quality. Third, our analyses are limited to marriages. Thus, we are not reflecting on changes in relationship quality and this may be important as there are substantial increases in cohabitation among older adults (Brown & Wright, 2017). Finally, it is possible that survey questions regarding marital quality could be interpreted differently by respondents across the two cohorts, but we were not able to assess any potential variation in interpretations between the groups.

Our study uncovers a pattern of midlife marital quality decline, which mirrors what has been unfolding among younger adults. This trend toward poorer marital quality among midlife adults also provides new insights on the gray divorce phenomenon. Midlife adults are less satisfied in their marriages these days than their counterparts were 25 years ago, and this dissatisfaction can lead to gray divorce (Lin et al., 2018). The erosion of midlife marital quality not only raises the specter of gray divorce, but also prompts concern about the potential ramifications for individual health and well-being. The health benefits of marriage are largely confined to those in high quality marriages. Individuals in lower quality marriages often exhibit no appreciable health advantage relative to their unmarried counterparts and sometimes fare worse (Umberson et al., 2006; Williams, 2003; Zhang et al., 2016). Thus, future research should address how marital quality and stability are linked to the health and well-being of midlife adults.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: 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 (P2CHD050959).

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

  1. Amato PR, Johnson DR, Booth A, & Rogers SJ (2003). Continuity and change in marital quality between 1980 and 2000. Journal of Marriage and Family, 65(1), 1–22. 10.1111/j.1741-3737.2003.00001.x [DOI] [Google Scholar]
  2. Bernard J (1972). The future of marriage New Haven, CT: Yale University Press. [Google Scholar]
  3. Boerner K, Jopp DS, Carr D, Sosinski L, & Kim SK (2014). “His” and “her” marriage? The role of positive and negative marital characteristics in global marital satisfaction among older adults. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 69(4), 579–589. 10.1093/geronb/gbu032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brown SL, & Booth A (1996). Cohabitation versus marriage: A comparison of relationship quality. Journal of Marriage and the Family, 58(3), 668–678. 10.2307/353727 [DOI] [Google Scholar]
  5. Brown SL, & Lin I-F (2012). The gray divorce revolution: Rising divorce among middle-aged and older adults, 1990–2010. Journals of Gerontology: Series B: Psychological and Social Sciences, 67(6), 731–741. 10.1093/geronb/gbs089 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brown SL, & Lin I-F (2013). Age Variation in the Remarriage Rate, 1990–2011. (FP-13–17). National Center for Family & Marriage Research Retrieved from. http://ncfmr.bgsu.edu/pdf/family_profiles/file134878.pdf.
  7. Brown SL, Lin I-F, Hammersmith AM, & Wright MR (2018). Later life marital dissolution and repartnership status: A national portrait. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 73(6), 1032–1042. 10.1093/geronb/gbw051 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brown SL, Manning WD, & Wu H (2017). Relationship quality in midlife: A comparison of dating, LAT, cohabitation, and marriage [DOI] [PMC free article] [PubMed]
  9. Brown SL, & Wright MR (2017). Marriage, cohabitation, and divorce in later life. Innovation in Aging 1(2):1–11. 10.1093/geroni/igx015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Brown SL, & Wright MR (2019). Divorce attitudes among older adults: Two decades of change. Journal of Family Issues, 40(8), 1018–1037. 10.1177/0192513X19832936 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bulanda JR (2011). Gender, marital power, and marital quality in later life. Journal of Women & Aging, 23(1), 3–22. 10.1080/08952841.2011.540481 [DOI] [PubMed] [Google Scholar]
  12. Bulanda JR, & Brown SL (2007). Race-ethnic differences in marital quality and divorce. Social Science Research, 36(3), 945–967. 10.1016/j.ssresearch.2006.04.001. [DOI] [Google Scholar]
  13. Carr D, Freedman VA, Cornman JC, & Schwarz N (2014). Happy marriage, happy life marital quality and subjective well-being in later life. Journal of Marriage and Family, 76(5), 930–948. 10.1111/jomf.12133. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cherlin A (1978). Remarriage as an incomplete institution. American Journal of Sociology, 84(3), 634–650. 10.1086/226830 [DOI] [Google Scholar]
  15. Cherlin AJ (2004). The deinstitutionalization of American marriage. Journal of Marriage and Family, 66(4), 848–861. 10.1111/j.0022-2445.2004.00058.x [DOI] [Google Scholar]
  16. Cooney TM, Proulx CM, & Snyder-Rivas LA (2016). A profile of later life marriages: Comparisons by gender and marriage order. In Divorce, separation, and remarriage: The transformation of family (pp. 1–37). Bingley, England: Emerald Group Publishing Limited. [Google Scholar]
  17. Elder GH Jr. (1994). Time, human agency, and social change: Perspectives on the life course. Social Psychology Quarterly, 57(1), 4–15. 10.2307/2786971 [DOI] [Google Scholar]
  18. Glenn ND (1991). The recent trend in marital success in the United States. Journal of Marriage and Family, 53(2), 261–270. 10.2307/352897. [DOI] [Google Scholar]
  19. Glenn ND (1998). The course of marital success and failure in five American 10-year marriage cohorts. Journal of Marriage and Family, 60(3), 569–576. 10.2307/353529 [DOI] [Google Scholar]
  20. Johnson DR, & Elliott LA (1998). Sampling design effects: Do they affect the analyses of data from the National survey of families and households? Journal of Marriage and the Family, 60(4), 993–1001. [Google Scholar]
  21. Johnson DR, White LK, Edwards JN, & Booth A (1986). Dimensions of marital quality: Towards methodological and conceptual refinement. Journal of Family Issues, 7(1), 31–49. 10.1177/019251386007001003 [DOI] [Google Scholar]
  22. Kaufman G, & Taniguchi H (2006). Gender and marital happiness in later life. Journal of Family Issues, 27(6), 735–757. 10.1177/0192513X05285293 [DOI] [Google Scholar]
  23. Kennedy S, & Ruggles S (2014). Breaking up is hard to count: The rise of divorce in the United States, 1980–2010. Demography, 51(2), 587–598. 10.1007/s13524-013-0270-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kiecolt-Glaser JK, & Newton TL 2001. Marriage and health: His and hers. Psychological Bulletin, 127(4), 472–503. 10.1037/0033-2909.127.4.472 [DOI] [PubMed] [Google Scholar]
  25. Kulik L (2002). Marital equality and the quality of long-term marriage in later life. Ageing & Society, 22(4), 459–481. 10.1017/S0144686X02008772 [DOI] [Google Scholar]
  26. Lin I-F, Brown SL, Wright MR, & Hammersmith AM (2018). Antecedents of gray divorce: A life course perspective. Journals of Gerontology: Social Sciences, 73(6), 1022–1031. 10.1093/geronb/gbw164 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Liu H, & Waite L (2014). Bad marriage, broken heart? Age and gender differences in the link between marital quality and cardiovascular risks among older adults. Journal of Health and Social Behavior, 55(4), 403–423. 10.1177/0022146514556893 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Miller RB, Hollist CS, Olsen J, & Law D (2013). Marital quality and health over 20 years: A growth curve analysis. Journal of Marriage and Family, 75(3), 667–680. 10.1111/jomf.12025 [DOI] [Google Scholar]
  29. Moen P, Kim JE, & Hofmeister H (2001). Couples’ work/retirement transitions, gender, and marital quality. Social Psychology Quarterly, 64(1), 55–71. 10.2307/3090150 [DOI] [Google Scholar]
  30. Pruchno R (2012). Not your mother’s old age: The baby boomers at age 65. The Gerontologist, 52(2), 149–152. 10.1093/geront/gns038 [DOI] [PubMed] [Google Scholar]
  31. Raley RK, & Sweeney MM (2020). Divorce, repartnering, and stepfamilies: A decade in review. Journal of Marriage and Family, 82(1), 81–99. 10.1111/jomf.12651 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Rogers SJ, & Amato PR (1997). Is marital quality declining? The evidence from two generations. Social Forces, 75(3), 1089–1100. 10.1093/sf/75.3.1089 [DOI] [Google Scholar]
  33. Rogers SJ, & Amato PR (2000). Have changes in gender relations affected marital quality? Social Forces, 79(2), 731–753. 10.2307/2675515 [DOI] [Google Scholar]
  34. Schimmele CM, & Wu Z (2016). Repartnering after union dissolution in later life. Journal of Marriage and Family, 78(4), 1013–1031. 10.1111/jomf.12315 [DOI] [Google Scholar]
  35. Schoen R, Astone NM, Kim YJ, Rothert K, & Standish NJ (2002). Women’s employment, marital happiness, and divorce. Social Forces, 81(2), 643–662. 10.1353/sof.2003.0019 [DOI] [Google Scholar]
  36. Stepler R (2017). Led by baby boomers, divorce rates climb for America’s 50+ population. Fact Tank: Pew Research Center http://www.pewresearch.org/fact-tank/2017/03/09/led-by-baby-boomers-divorce-rates-climb-for-americas-50-population/.
  37. Stokes JE (2017). Do “his” and “her” marriages influence one another? Contagion in personal assessments of marital quality among older spouses over a four-year span. Social Psychology Quarterly, 80(2), 204–215. 10.1177/0190272517701701 [DOI] [Google Scholar]
  38. Sullivan O (2006). Changing gender relations, changing families: Tracing the pace of change over time (Gender Lens Series) New York, NY: Rowman & Littlefield. [Google Scholar]
  39. Sweet J, Bumpass L, & Call V (1988). The design and content of the national survey of families and households NSFH Working Paper, No. 1. Madison, WI. http://www.ssc.wisc.edu/cde/nsfhwp/nsfh1.pdf. [Google Scholar]
  40. Uhlenberg P, & Myers MAP (1981). Divorce and the elderly. The Gerontologist, 21(3), 276–282. 10.1093/geront/21.3.276 [DOI] [PubMed] [Google Scholar]
  41. Umberson D, & Thomeer MB (2020). Family matters: Research on family ties and health, 2010 to 2020. Journal of Marriage and Family, 82(1), 404–419. 10.1111/jomf.12640 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Umberson D, Williams K, Powers DA, Chen MD, & Campbell AM (2005). As good as it gets? A life course perspective on marital quality. Social Forces, 84(1), 487–505. 10.1353/sof.2005.0131 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Umberson D, Williams K, Powers DA, Liu H, & Needham B (2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47(1), 1–16. 10.1177/002214650604700101 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. VanLaningham J, Johnson DR, & Amato P (2001). Marital happiness, marital duration, and the u-shaped curve: Evidence from a five-wave panel study. Social Forces, 79(4), 1313–1341. 10.1353/sof.2001.0055 [DOI] [Google Scholar]
  45. Williams K (2003). Has the future of marriage arrived: A contemporary examination of gender, marriage, and psychological well-being. Journal of Health and Social Behavior, 44(4), 470–487. 10.2307/1519794 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Wu Z, & Schimmele C (2007). Uncoupling in late life. Generations, 31(3), 41–46. [Google Scholar]
  47. Zhang Z, Liu H, & Yu Y-L (2016). Marital biography and health in middle and late life. In Bookwala J (Ed.), Couple relationships in the middle and later years: Their nature, complexity, and role in health and illness (pp. 199–218). Washington, DC: American Psychological Association. [Google Scholar]

RESOURCES