In this section, we discuss the various ways in which cohabitation experiences differ by socioeconomic status. These differences are seen primarily across three avenues: ‘who,’ ‘when,’ and ‘to what end?’ Cohabitation at some point during one’s life is common today across SES, but differences by SES remain in the share who have ever cohabited and the speed at which people enter into a cohabiting union. The most noticeable differences are in the outcomes of cohabitation: continued cohabitation, transition to marriage, and dissolution of the union (breaking up).
The majority (62%) of adults, regardless of SES, now report having lived with a partner outside of marriage at some point in their lives,2 but this was not always the case. Since the 1980s, the share of women who have ever cohabited has grown across all educational attainment groups but still varies by education (see Figure 1). For example, in the late 1980s, 43 percent of women with less than a high school education had ever cohabited, compared to 31 percent of those with a college degree. Nearly three decades later, 68 percent of women with less than a high school education have cohabited, compared to 59 percent of women with a college degree.2
When do Americans cohabit?
Generally, adults with less than a high school education begin cohabiting at younger ages than those with a college degree. For example, women with less than a high school education who formed cohabiting relationships from 2006 to 2010 reported first living with a cohabiting partner at a median age of 19, whereas college-educated women began cohabiting at a median age of 25.17
The likelihood that partners will move in together tends to be highest relatively soon (within six months) after their relationship begins, with the likelihood declining after six months.18 However, this pattern also varies by socioeconomic status. For instance, during the first year of a romantic relationship, women with college-educated mothers are less likely to begin cohabiting in that year than their counterparts whose mothers do not have a college degree.18
How do cohabiting relationships end?
Cohabiting relationships can end in different ways: They can transition to marriage, or they can separate. These outcomes are tied to the varied roles of cohabitation in the broader scope of relationships. For some individuals, cohabitation is an alternative to dating; for others, it is a stage in the marriage process (or precursor to marriage); and for others still it represents an alternative to marriage.3 As a result of this variation, some couples are engaged to be married when they begin living together, while others begin living together more casually—“sliding” into shared living—and with less deliberate thought to the future of their unions.19 Although cohabitors have become less likely to be engaged or to hold definite plans to marry when they start living together, a substantial minority of those entering a first cohabiting union—more than 40 percent—do report having these plans when they started cohabiting.2,20,21,22
In general, while today’s cohabiting relationships remain intact a bit longer than those formed in the early 1980s, the majority end in separation, typically in under two years.23,24,25 Individuals who were engaged or who had plans to marry when they started living with a partner are more likely to marry than their counterparts without such plans.19 However, research finds that women’s transition from cohabitation to marriage occurs most often for those who are highly educated.18,25 For example, among women who cohabited at some point from 2005 to 2010, the probability of transitioning from cohabitation to marriage in a given year was 24 percent for those with a college degree and less than 15 percent for those with lower levels of education.26 Consistent with these findings, higher incomes among women under age 30 are associated with a reduced risk of cohabitation dissolution.24,27 Educational attainment is similarly linked to the stability of cohabiting unions. Women with at least some college education have significantly lower chances of dissolving their cohabiting unions and experiencing multiple union dissolutions early in their lives, relative to those women who did not attend college. 24,27,28
The variance by socioeconomic status in individuals’ motivations for cohabitation seems to explain much of the variation in the link between socioeconomic status and cohabitation timing and outcomes. For men and women with below-average incomes, the decision to cohabit is sometimes driven by economic constraints, such as housing costs, a lack of well-paying employment opportunities, a lack of savings, burdensome student loan debt (especially among those who did not complete a degree), or little financial support from parents.29,30,31 Men and women with higher incomes and educational attainment often report cohabiting in response to other factors—such as the end of a lease or the potential for a better commute to their workplace—as well as explicit desires to take a more serious step forward in their relationship. Having moderate or high incomes appears to allow these couples more control over when they start their cohabiting relationships, instead of feeling the need to move in together out of economic necessity.30 Moreover, to the extent that a quicker progression into living together might indicate less formulated plans for a relationship’s future, variation in the timing of cohabitation across SES may explain differentials in the stability of cohabiting unions.4
This section will cover socioeconomic differences concerning marriage. As with cohabitation, these differences can be briefly described in terms of ‘who,’ ‘when,’ and ‘to what end.’ Additionally, the ‘why’—or the reasons that people cite for marrying or not getting married—is often tied to socioeconomic conditions and considerations. Each of these questions can be answered in part by looking at economic constraints and pathways to marriage. We begin by discussing which individuals, by SES, decide to get married; we then examine timing of, attitudes toward, and constraints on marriages, and we end with a discussion of divorce and remarriage.
There are sizeable differences in marriage across socioeconomic status—larger than those seen for cohabitation—which is consistent with research suggesting an economic bar to marriage that lower-income men and women struggle to reach.32 Prior to the turn of the 21st century, college-educated women were less likely to be currently married than women without a college degree, but the opposite has become true in the last 20 years.33 In 2016, 59 percent of women with a college degree or higher were currently married, compared to 27 percent of women with less than a high school education.34 Similarly, the first marriage rate (the number of first marriages per 1,000 unmarried women) was substantially higher for those with a Bachelor’s degree or Master’s degree or higher (65.6 and 78.2, respectively) than for those without a college degree (ranging from 27.3 for those with less than a high school degree to 38.8 for those with some college).35 As with current marital status, patterns of having ever married exhibit marked educational differences that have grown over time, as shown in Figure 2. Moreover, people increasingly marry someone with the same level of education as themself.36
A similar gradient exists for marriage by other indicators of socioeconomic status. Among both men and women, full-time employment, earnings, and higher incomes are positively associated with the likelihood of getting married.37,38,39 These findings indicate that economically advantaged individuals are more likely to marry than their less advantaged counterparts. For women, the importance of earnings for marriage increases with age.39 In other words, work characteristics are a stronger predictor of marriage for women in their 30s than those in their 20s. Being in a high-status occupation (one that requires high levels of education) also increases the chances that a woman will marry.40
When do men and women marry?
In the United States, men and women generally delay marriage until after they have completed their education. But while those with a college degree are married at older ages than those without, they are ultimately more likely to marry than their less-educated counterparts.41 Research shows that women with a college degree tend to marry at older ages than women with less than a college education (at median ages of 29 and 27, respectively).5
Despite shifts in marital behavior, most young men and women expect to get married at some point in their lives. From 1976 to 2017, the percentage of high school seniors who expect to marry has stayed consistently high, at 75 to 80 percent.42 Overall, men and women report a range of reasons for wanting to marry, including love, the desire to start a family, and long-term stability;43 some also cite “improving their economic circumstances” as a motivation to get married.44,45 Still, most young adults in the United States center marriage around love and commitment rather than finances.44,46 Attitudes and decisions about marriage also vary by SES. For example, expectations of marriage (i.e., one’s view of the likelihood of getting married someday) are higher among high school seniors whose parents have a college degree (79%) than among those whose parents have less than a high school degree (69%).42 Some evidence also suggests that men and women with lower levels of education tend to hold more pessimistic views about their own likelihood of marriage.43
Unmarried individuals often highlight the importance of financial resources when making decisions about potential partners and marriage.47,45 For example, men and women tend to be less willing to partner with those who have unstable jobs or low incomes.48 Interviews with cohabiting couples with low incomes reveal that many wish to marry in the future but will only do so when they feel more financially stable and can afford a nice wedding.44,49 Similarly, some research finds that, for example, individuals with jobs that have standard work schedules, fringe benefits, and union memberships are more likely to marry.50,51 In fact, financial incentives for married couples—such as tax breaks, Social Security pension funds, and insurance benefits—may encourage some couples to marry.14 These patterns are consistent with the argument that lacking adequate financial resources is a key barrier to marriage. Once these barriers are surmounted, the odds of marrying increase significantly.32,44
Research also finds socioeconomic differences in marital stability. Individuals with a college education, for example, have a lower likelihood of divorce than those with less education.52,53,54,55 In 2018, the divorce rate for first marriages (also referred to as the first divorce rate, defined as the number of divorces per 1,000 women in first marriages) was lowest among women with a Bachelor’s degree or higher (12.4), whereas the highest first divorce rate (18.9) was among those with some college experience but without a formal degree (Figure 3)55 (see footnotea for clarification).
Interestingly, there does not seem to be a direct link between income and divorce.56,57 Rather than low incomes leading to stress and hardship that destabilize marriages, higher incomes and more economic resources appear to act as barriers to divorce—in other words, divorce is more costly for those with more resources (debt, however, does increase the risk of divorce).58 The roles of wives’ earnings and employment in divorce are nuanced, as well.59 The large-scale increase of women in the labor force since the 1970s has lessened women’s economic dependence on their husbands and has altered the gendered division of labor (such as housework and child care). As women’s employment has become more common, its association with divorce has diminished. The most recent research suggests that women’s employment increases the risk of divorce but only among women who report low marital satisfaction.60 Similarly, although research initially suggested that the risk of divorce grew when wives’ share of the couple’s relative income increased, more recent research suggests this is no longer the case.56,61 On the other hand, despite shifts in gender roles, there is still evidence that men’s employment matters: Husbands’ lack of employment increases the risk of divorce.56
About one quarter of all marriages formed in recent years are remarriages, rather than first marriages, for at least one partner in the couple.62 Some research finds a link between SES and the likelihood that a previously married man will eventually remarry, but this link is not as strong as the link between SES and the rate of first marriages.38 This work also finds that men with a college degree are 40 percent more likely to remarry than those without a high school education, and that employed men are much more likely to remarry than unemployed men. Income, however, is not associated with the odds that men will be remarried.38 For women, there is no evidence that socioeconomic status is associated with the odds of remarriage.38,63
Research findings show extensive differences across socioeconomic status when it comes to marriage, providing answers to the questions: Who marries? Who does not? Why do these outcomes occur? If a couple does marry, when do they do it? Do they stay married? As with cohabitation, for both men and women, the decisions of whether and when to marry are driven by economic constraints. People seem to get married when they have the resources to pursue the formal union, treating it as a capstone linked to SES attainment. People also seem better able to maintain marriages when they have fewer economic constraints. Discussions concerning the ‘who’, ‘when,’ and ‘why’ of marriage in the United States should pay close attention to income, education, and material resources.
Implications for Research and Practice
Implications for research
Socioeconomic differences in the formation and dissolution of romantic unions are sizeable, and in many cases have widened in recent years. This brief has highlighted many of the ways that SES is linked to the formation and stability of cohabiting and marital unions, but more work is needed. Much of the research we have covered, for instance, documents variation but uses sources with relatively little contextual information (such as the U.S. Census Bureau’s American Community Survey); therefore, this work cannot tell us much about the underlying processes that contribute to variation. Our findings would be strongly supplemented by more qualitative work that deliberately probes and explores the obstacles and experiences that different socioeconomic groups face in the context of romantic unions, and that discusses the underlying reasons for these differences. Additionally, widening inequality and the various economic shocks of the past decade (including uneven recovery from the Great Recession and the COVID-19 pandemic) may accelerate the growing divergence in family formation between those with a college degree and stable jobs, and those without. More work is also needed to investigate how other indicators of individual socioeconomic status (e.g., debt or home ownership, employment characteristics, and related factors such as health insurance)—and how these indicators may interact with aggregate socioeconomic factors (e.g., unemployment rates, housing costs, public policies)—may shape individuals’ and couples’ decisions around forming and staying in cohabiting and marital relationships.
This brief did not cover non-coresidential relationships because few data sources collect such information. This means we know little about dating relationships or other relationship types, such as “living apart together” relationships in which individuals are in committed relationships but do not live together.64,65 Data on non-coresidential relationships would provide valuable insights into relationship processes and behaviors and how they might vary by SES.
Socioeconomic status is just one factor by which family and relationship trends, behaviors, and outcomes vary: Cohabitation and marriage are stratified along other important dimensions, including race and/or ethnicity, nativity, sexual orientation and gender, and religion. Importantly, focusing on any single avenue of stratification has become more carefully scrutinized due to the growing recognition of intersectionality—the idea that people have multiple and overlapping identities and experiences, and face different forms and degrees of oppression and opportunity. Extensive literature documents racial and/or ethnic differences in union formation and stability (see Raley, Sweeney, & Wondra  for a review of this literature),66 yet much of this work fails to account for important heterogeneity by considering, for instance, that the experiences of immigrant Black men and women differ than their native-born counterparts, or that the umbrella category of Asian obscures variation among those of Japanese, Chinese, Pakistani, and other ancestries.67 New research suggests that, although part of racial and/or ethnic differences in marriage are due to differences in the distribution of SES, Black men and women have experienced larger declines in marriage than their White counterparts within the same SES level.68 Moreover, the limited evidence examining both same- and different-gender couples suggests important avenues of variation in relationship formation across sexual orientation. For example, same-gender couples are more likely to be interracial than different-gender couples.69 Thus, the ways in which SES shapes the trends described in this brief may look different within and across other dimensions of stratification.
Implications for practice
The trends described in this brief point to a number of circumstances—such as housing insecurity, high levels of debt, and a lack of stable employment, among others—that many individuals and couples with fewer socioeconomic resources perceive as barriers to marriage. For those who do form cohabiting unions or marry, there is evidence that couples in economically challenged circumstances are more likely to experience a break-up or divorce. HMRE programs can use this information to identify aspects of financial literacy and economic security that may contribute to adults’ motivation for forming coresidential unions, as well as the stability of those unions. Given the importance of economic resources and stability to the lives and well-being of couples and families, many HMRE programs already supplement relationship skills education with content that promotes financial stability and economic self-sufficiency.70 For example, HMRE programs may implement curricula that include information and guidance on financial management and planning, budgeting, workforce development, or employment training.71,72 However, HMRE program providers may be able to further tailor their programs with a better understanding of how SES shapes couples’ relationship formation and stability, and what factors are most relevant for cohabiting versus married couples.
In recognition of the complex connections between relationship stability and economic self-sufficiency, recent federal HMRE programming has encouraged more comprehensive programmatic approaches that support economic stability and mobility. However, additional evaluation research is needed to formally test whether these services translate into more relationship stability.73,74,75 Moreover, evaluations have found that job and career services offered through HMRE programs tend to be underused, even among those with low incomes.76,77
HMRE programs can go further to communicate the relevance of their job and career service offerings for enhancing relationship quality and stability as a way to better market those aspects of programming to current or prospective clients. Programs can also consider, however, whether such underutilization is a result of low awareness of these offerings’ relevance, or whether some low-income participants or communities may have higher-priority economic concerns affecting their relationships that would be better addressed through other services.78
HMRE programs looking to enhance their focus on economic stability could partner with outside groups and organizations (for example, workforce development and training organizations) to provide a wider range of services, along with more targeted services offered directly to program participants. They might also consider the bidirectionality of the association between some aspects of economic stability and relationship quality, and emphasize the relevance of the general relationship skills typically taught in HMRE programs (e.g., communication skills) for promoting workplace success.
This brief is based on a comprehensive review of professional and scientific journal articles, book chapters, and reports published in or after 2010, along with a small number of important publications from before 2010. The research cited in this brief uses data primarily from nationally representative surveys. Much of the research on union formation and dissolution uses data that are only available for women, especially when analyzing trends over time. Some major surveys, such as the National Survey of Family Growth (NSFG), only began including men relatively recently (2002 for the NSFG). Other surveys (namely household-based surveys on civilians in noninstitutional settings) have a sample design that excludes many young men because they are in the military, incarcerated, or otherwise not formally attached to a household. Despite this limitation, we discuss research for men in addition to women when possible.
The patterns and trends described in this brief cover a broad range of time, although the brief focuses on more recent periods when possible, depending on the available data and research. Likewise, various aspects of relationship formation discussed in the brief may not appear in each section due to a lack of research or data. This brief covers only published research studies and previously recorded data and does not include any original descriptive analyses. As such, the overview is limited to the foci and definitions used by other researchers. Finally, in this brief, the discussion is limited to different-gender relationships because there is relatively little published research on trends and patterns in same-gender relationship formation. For a deeper discussion of recent research on union formation and dissolution, we direct readers to Sassler and Lichter (2020) and Raley and Sweeney (2020).4,50
a With education, sometimes it is more about ‘why’ people do or do not complete their education than about the actual amount of education—such that the relationship between divorce rates and educational attainment is not a linear one. Also, the ‘some college’ group is distinct: Higher education is expensive in the United States; those who pursued higher education but did not complete a degree took on the economic burdens of college without the explicit economic advantage conferred through an official degree.