Socioeconomic Status and Intimate Relationships

Abstract

The ways that couples form and manage their intimate relationships at higher and lower levels of socioeconomic status (SES) have been diverging steadily over the past several decades. At higher levels, couples postpone marriage and child-birth to invest in education and careers, but eventually marry at high rates and have relatively low risk for divorce. At lower levels, couples are more likely to cohabit and give birth prior to marriage, and less likely to marry at all. This review examines how SES comes to be associated with the formation, development, and dissolution of intimate relationships. Overall, research has highlighted how a couples’ socioeconomic context facilitates some choices and constrains others, resulting in different capacities for relationship maintenance and different adaptive mating strategies for more and less advantaged couples. A generalizable relationship science requires research that acknowledges these differences and one that recruits, describes, and attends to socioeconomic diversity across couples.

1. INTRODUCTION

Monogamous, emotionally intense, pair-bonds between adults, i.e., intimate relationships, have been observed in nearly every society on earth and in every historical period, so much so that they have been called “a human universal” (Jankowiak & Fischer, 1992, p. 154). Across regions and eras, the goals that couples seek within intimate relationships are generally similar as well. Throughout the world (Fletcher, Simpson, Campbell, & Overall, 2015; Neto et al., 2000), partners’ ideals for their intimate relationships can be characterized by a combination of passion, closeness, and commitment, although the relative importance of these three elements varies somewhat across cultures (Keller, 2012).

Yet, whereas couples share common goals for their intimate relationships, couples vary widely in the support available for pursuing these goals. Couples with higher incomes, more education, and more stable employment often have flexibility and opportunities for interaction that less affluent couples lack. Couples with saving accounts and property have security that couples living paycheck-to-paycheck and renting their homes do not. Together, variables like income, education, employment, and assets have been treated as proxies for socioeconomic status (SES), a broader construct referring to an individual’s overall social standing and access to resources (Baker, 2014). Although SES varies continuously across individuals, discussions of the implications of SES often draw a broad distinction between individuals who completed college and individuals without a college degree (e.g., Carbone & Cahn, 2014).

In the United States, a country with one of the highest rates of income inequality in the world (Brandolini & Smeeding, 2006; Organisation for Economic Cooperation and Development, 2015), the ways that couples form and manage their intimate relationships on either side of this distinction have been diverging steadily over the past several decades. Data from the census and long-running surveys like the National Survey of Family Growth (NSFG) indicate that women with a college degree are between 12% and 17% more likely to marry than women who did not graduate from college, and that this gap has been widening over time (Lundberg, Pollak, & Stearns, 2016; Rosenfeld & Roesler, 2019). Similar data suggest that college educated women who do marry are up to 40% less likely to divorce than married women who did not complete college (Copen, Daniels, Vespa, & Mosher, 2012; Lundberg et al., 2016), and this gap has also been widening over time (Cohen, 2019). Other indicators of SES have comparable associations with transitions into and out of marriage: lower household income (Bramlett & Mosher, 2002), higher levels of perceived financial strain (Gudmunson, Beutler, Israelsen, McCoy, & Hill, 2007), and lower levels of assets and home-ownership (Coulter & Thomas, 2019) are all associated with lower rates of marriage and higher rates of divorce. The consistency of the findings is striking, and these patterns are not confined to the United States. The influence of SES on marriage and divorce patterns is the same in Denmark, Germany, Sweden, Finland, and other countries where it has been studied (Esping‐Anderson, 2016; Jalovaara, 2002). Although the influence of SES has been documented most regularly with respect to legally recognized and socially defined forms of intimacy like marriage and divorce, SES has been shown to affect rates of entry into and exit from other forms of intimate partnerships (e.g., cohabitation) as well (e.g., Sassler, Addo, & Lichter, 2012).

Family scholars in sociology and demography have been acknowledging and grappling with these trends for decades (Cherlin, 2005; McLanahan, 2004). Psychologists, in contrast, have been mostly absent from this discussion. Despite an emerging research literature on psychological differences between individuals from more or less affluent backgrounds, research comparing intimate relationships across levels of SES remains “in its infancy” (Kraus, Piff, Mendoza-Denton, Rheinschmidt, & Keltner, 2012, p. 560). This oversight has basic and applied implications for research on couples. With respect to basic science, it suggests that relationship scientists have been neglecting to account for an important source of influence in modeling the success or failure of intimate relationships. The size of the associations between indicators of SES and marital transitions is large. For example, analyses of data from the NSFG show that the 78% of marriages of women with a college degree last at least 20 years, compared to only 41% of marriages of women with a high school diploma or less (Copen et al., 2012). In other words, women’s college education nearly doubles the 20-year survival rate of a marriage, an effect size comparable to that of variables receiving far more attention from psychologists.

With respect to applications of relationship science, the failure to account for differences between couples at different levels of SES has already had expensive consequences. Since launching the Healthy Marriage Initiative in 2001, the federal government has spent nearly a billion dollars offering low-income couples relationship skills classes and programs that were developed for middle-class couples (Manning, Brown, Payne, & Wu, 2014). Despite this substantial investment, these programs have had negligible effects on the couples whose lives they were meant to improve (Lundquist et al., 2014; Wood, Moore, Clarkwest, & Killewald, 2014), in part because the design of the programs was not informed by empirical research on the challenges that low-income couples actually face (Johnson, 2012). If future efforts to assist disadvantaged couples are to be more effective, policy makers will need to develop policies grounded in an explicit understanding of how differences in SES affect intimacy.

The goal of this article to assemble and review research and theory that bears on this question. Toward that end, most of the work described here will address marriage, as a legally and socially defined form of intimate relationship that has received extensive research attention. Most of the reviewed research also addresses relationships between men and women. Despite these two limits in scope, there is good reason to expect that many of the findings discussed in this article will generalize to other types of committed, romantic couples, and research on those couples is included whenever it is available.

2. THEORETICAL PERSPECTIVES

It is not immediately obvious how a couple’s placement within a broad socioeconomic context comes to affect partners’ private aspirations, thoughts, and behaviors. This section reviews three theoretical approaches offering alternative perspectives on this issue.

2.1. The Psychological Orientation Account

Just as Attachment Theory proposes that specific early experiences with our primary caregivers give rise to characteristic expectations of intimacy throughout our lives (Bowlby, 1979), the psychological orientation account of SES suggests that early experiences with social ranking and unequal distribution of resources produce characteristic patterns of thought among people from different socioeconomic backgrounds (Kohn, 1969). As a result of shared cultural norms, common experiences in more or less powerful social roles, and habits developed as adaptive responses to different circumstances, this view describes individuals at varying levels of SES with stable, general, but distinct ways of approaching and interacting with other people (Kraus et al., 2012). In higher-SES communities, where access to resources and opportunities can be taken for granted, individuals should develop habits of self-focus and expectations of control, neglecting situational constraints on their choices and behaviors. In lower-SES communities, where limited access to resources and opportunities do constrain choices and behaviors, individuals should develop a heightened sensitivity to environmental cues and to patterns of interdependence with other people in particular. This account thus positions the unique interpersonal approaches of higher- and lower-SES individuals as a mediator of the effects of SES on relationships.

An emerging empirical literature provides support for the general idea that people from lower-SES backgrounds are indeed more “collectivist” than people from higher-SES backgrounds. For example, compared to their more affluent peers, individuals raised in lower-SES communities have been found to be more accurate at judging the emotions of strangers (Kraus, Cote, & Keltner, 2010), more motivated to seek out community (as opposed to personal wealth) when confronted by uncertainty (Piff, Stancato, Martinez, Kraus, & Keltner, 2012), and more prone to experiencing other-related emotions like compassion, love, and awe than self-related emotions like contentment and pride (Piff & Moskowitz, 2017). Analyses of large, representative data sets indicates that lower-income individuals do spend more time socializing and less time alone than higher-income individuals (Bianchi & Vohs, 2016), consistent with the description of lower-income individuals as more connected to their communities.

To date, research informed by the psychological orientation perspective has not examined SES differences in intimate relationships directly. Yet the distinction between collectivist and individualist orientations across levels of SES echoes a similar distinction that has been applied to intimate relationships extensively, that between communal and exchange orientations (Clark & Mills, 1993). To the extent that the collectivist orientation ascribed to lower-SES individuals overlaps with communal orientations toward relationships, then lower-SES individuals should demonstrate the hallmarks of that disposition, including attentiveness to the needs of the partner. To the extent that the self-reliance ascribed to higher-SES individuals overlaps with exchange orientations toward relationships, then higher-SES individuals should demonstrate the markers of that disposition, including a tendency to keep track of benefits and costs within the relationship (Clark, Mills, & Powell, 1986). Endorsing a communal orientation, as opposed to an exchange orientation, has consistently been associated with more stable, rewarding relationships (Clark & Mills, 2012), so this perspective predicts that the relationships among individuals from lower-SES backgrounds should, on average, be more successful than relationships among individuals from higher-SES backgrounds (Kraus et al., 2012).

There are two concerns with this account. First, cross-cultural research has begun to question the idea that collectivist orientations necessarily improve interpersonal relationships. Individuals with an orientation that prioritizes connection and relationships may be especially likely to experience conflict when those relationships disappoint, and indeed people living in collectivist West Africa are significantly more likely to report having enemies than people living in individualist North America (Adams, 2005). Similarly, individuals living in collectivist regions of China report higher levels of suspicion toward in-group members than individuals living in less collectivist regions and individuals living in America (Liu, Morris, Talhelm, & Yang, 2019). In other words, the fact that people raised in lower-SES families may be more attentive to the needs of strangers does not imply that they will be similarly attentive towards their intimate partners, and the fact that people raised in higher-SES families are less connected to strangers does not imply that they will feel similarly disconnected from their intimate partners. The second, and more significant, concern is that there are no data to support the prediction that a lower-SES background is associated with more successful relationships. On the contrary, as already noted, virtually all available evidence strongly supports the opposite pattern, suggesting the need to look further for an adequate account.

2.2. The Stress and Resources Account

Whereas the psychological orientation account suggests that economic contexts operate on couples indirectly through their effects on partners’ interpersonal styles, alternative accounts emphasize the direct effects of the environment on how couples develop and maintain their relationships. Stemming from his work with military families after World War II, Ruben Hill’s seminal Crisis Theory (1949) was the first to acknowledge that stress external to a couple can bring partners closer together or push them apart, depending on the resources available for coping with that stress. Contemporary versions of this perspective identify two specific routes through which challenging circumstances affect intimacy (Karney & Neff, 2013; Neff & Karney, 2017). First, conditions outside the couple directly shape the content of partners’ interactions with each other. Couples confronted with financial strain or health issues, for example, must devote time and energy toward negotiating their response to those problems; couples free from such issues have more time and energy to spend on activities that nurture intimacy, like having sex and pursuing shared interests (Bodenmann, Ledermann, & Bradbury, 2007). Second, conditions outside the couple can enhance or inhibit the quality of partners’ interactions with each other. Under stress, the difficult work of empathy and understanding becomes even more difficult (Conger & Conger, 2008; Neff & Karney, 2004). As a result, couples facing imminent stress find the tasks of resolving problems and providing each other with support more challenging than comparable couples that are not under stress (Bodenmann et al., 2015).

From this perspective, differences between intimate relationships at different levels of SES would be expected to the extent that life is more challenging for people living in lower-SES communities than for people living in higher-SES communities. Indeed, those challenges are manifold and well-documented. For example, SES is powerfully associated with health outcomes, such that individuals in lower-income communities experience higher rates of cancers and chronic illnesses like diabetes, hypertension, and cardiovascular disease, more illnesses that limit activities of daily living, higher rates of infant mortality, and lower life expectancy than individuals in higher-income communities (Braveman, Cubbin, Egerter, Williams, & Pamuk, 2010). Individuals with lower SES have higher rates of unemployment, and when they are employed they are more likely to be working non-standard hours in jobs that lack flexibility, paid sick leave, and health benefits (Enchautegui, 2013; Presser, 2000). In any given 12-month period, individuals with lower-SES are more likely than individuals with higher-SES to experience acute, uncontrollable stressful events that demand a response, such as the death of a loved one, health crises, and income instability (McLeod & Kessler, 1990).

Magnifying the impact of these demands is a relative lack of material, interpersonal, and intrapersonal resources among individuals at lower levels of SES. By definition, lower SES is associated with fewer assets and less savings, limiting the direct coping responses available to individuals at lower levels of SES relative to those with high SES. Research reveals social disadvantages as well: compared to their working-class and more affluent peers, the poor report greater social isolation, fewer ties to potential sources of social support (e.g., family and friends), and more frequent experiences of interpersonal conflict (Mickelson & Kubzansky, 2003; Stephens, Cameron, & Townsend, 2014). Perhaps as a consequence, individuals with lower SES experience lower levels of subjective well-being (Luhmann, Murdoch, & Hawkley, 2014), more frequent experiences of sadness (Kushlev, Dunn, & Lucas, 2015), and heightened emotional reactivity to daily events (Gallo & Matthews, 2003), all of which can interfere with the effective stress management.

To the extent that couples with lower SES confront these disproportionate challenges and disadvantages, the stress and resources perspective predicts that the complex work of intimacy should likewise be more challenging for these couples, relative to couples with higher SES.

2.3. The Mating Strategies Account

In contrast to the previous two accounts, both developed by psychologists, the Mating Strategies account derives from economics and sociology, where analyses of the interplay between social structures and family structures have a long tradition (e.g., Becker, 1973). Central to this perspective is the idea that mating strategies, i.e., the values that guide decisions about long-term intimate relationships, are profoundly influenced by individuals’ implicit understanding of their immediate socioeconomic context.

One element of that context that has received particular attention is the gender ratio among the pool of eligible mates. In their classic study Too Many Women?: The Sex Ratio Question, Guttentag and Secord (1983) drew upon historical and demographic data to support what was then a radical hypothesis: “that the number of opposite-sex partners potentially available to men or women has profound effects on sexual behaviors and sexual mores, on patterns of marriage and divorce, childrearing conditions and practices, family stability, and certain structural aspects of society itself” (p. 9). When men outnumber women, men must compete to gain access to the most desirable mates, and women consequently can be more selective. Under such conditions, men are more open to long-term commitments (in order to ensure access to relatively scarce partners), more invested in their own careers (in order to appear more attractive to potential mates), and more likely to meet women’s demands for fidelity (Angrist, 2002). Cultures characterized by high male to female ratios therefore tend to have high rates of marriage, low rates of divorce, and low rates of non-marital fertility. When women outnumber men, in contrast, men become more selective and women have to compete for partners. Under these conditions, men can afford to make commitments lightly, as women with few alternatives either accept infidelity as the cost of access to a desirable partner or withdraw from long-term relationships with men altogether. Cultures characterized by low male to female ratios therefore tend to have low rates of marriage, high rates of divorce, and high rates of non-marital fertility.

Economists Carbone and Cahn (2014) argue that gender ratios are crucial for understanding differences in mating strategies across levels of SES, in part because income inequality affects men differently than it affects women. Men’s employment and earning tends to vary more than women’s; there are more men than women at the very top of the payscale (e.g., corporate CEOs remain disproportionately male) and also more men at the very bottom (Wilkinson & Pickett, 2011). In light of the powerful tendency for men and women to seek out and find partners within their own level of education and income (Schwartz & Mare, 2005), this means that the ratio of eligible men to eligible women in the pool of potential mates varies according to SES. Among the upper levels, a greater number of men compete for the smaller number of women who manage to manage to transcend institutional sexism and achieve relative parity in accomplishment. Among the lower levels, a greater number of women grow mistrustful of the smaller number of men who are employed and earning enough to support a family.

This perspective therefore predicts that higher- and lower-SES men and women will pursue distinct mating strategies, as rational adaptations to their different socioeconomic contexts. Among the upper class, these strategies should motivate deep investment in education and careers, followed by strong and lasting romantic commitments to like-minded partners, as a foundation for having children and grooming them for future success. Among the lower class, these strategies should motivate avoidance and even mistrust of long-term romantic commitments, and an acceptance of child-bearing outside the context of marriage.

2.4. Applying the Theories

Empirical work on SES and intimacy has drawn upon the general accounts reviewed here to derive two sorts of predictions about differences between couples at different levels of SES. First, the perspectives highlight specific aspects of intimate relationships that are likely to vary across levels of SES, i.e., main effects of SES. The prediction from the psychological orientation account that partners from higher levels of SES will depend on their interpersonal connections less than partners from lower levels of SES is such a main effect, as is the prediction from the stress and resources account that couples at lower levels of SES will experience more stress and have more difficulty managing that stress than couples at higher levels of SES.

The second sort of prediction is the idea that the same variables may affect intimate relationships differently at different levels of SES, i.e., an interaction. For example, one implication of the stress and resources account is that coping strategies that are effective for couples at higher levels of SES may be ineffective or even counter-productive at lower levels of SES, such that SES may moderate the implications of particular coping strategies for relationship outcomes.

Both sorts of predictions have been examined with respect to understanding how SES is related to: a) how couples at different levels of SES meet and form intimate relationships, b) how those relationships are maintained over time, and c) when and why those relationships end. The next sections consider each of these outcomes in turn.

3. SES AND RELATIONSHIP FORMATION

In 1950, roughly 85% of women without a college education were married, compared to 70% of college graduates. Over the next 60 years, this difference flipped, such that rates of marriage for college-educated women have been climbing even as rates of marriage for women who do not attend college have dropped to less than 60% (Lundberg & Pollak, 2015; Musick, Brand, & Davis, 2012). Yet, with regard to understanding how committed intimate relationships begin across levels of SES, marriage rates represent the mere tip of the iceberg. Research has identified substantial differences between couples at different levels of SES at every stage of the development of intimate relationships.

3.1. SES and Pathways to Commitment

In considering how romantic relationships begin, Bozon and Heran (1989) distinguished between three contexts in which partners might meet for the first time. “Public” contexts are spaces that anyone can enter, like a park or a bar or a bookstore. “Select” contexts are spaces like a university or a place of employment that can only be entered by individuals who have met a specific set of criteria. “Private” contexts are spaces restricted by a particular social network, like a friendship or family group. These distinctions matter because the pool of mates is generally more diverse in public spaces, even as the availability of social support and approval is generally lower for relationships formed in such spaces, as partners are less likely to share backgrounds and social ties. In contrast, private and select spaces, by restricting access to individuals who share a class and a background, are likely to promote homogamy and thereby higher levels of social approval and support for couples. Studies drawing from survey data and in-depth interviews show that couples at lower levels of SES are more likely to meet each other in public spaces, whereas couples at higher levels of SES are more likely to meet each other in private or select spaces (Lampard, 2007; Sassler & Miller, 2014a). In other words, disadvantaged individuals are more likely to meet partners in contexts that exacerbate their disadvantages, whereas advantaged individuals are more likely to meet partners in contexts that consolidate their advantages.

Once they have met, the timing of major transitions within the relationship varies across levels of SES as well. In some respects, relationship formation is accelerated among couples at lower levels of SES. For example, more disadvantaged couples report having sex sooner after first meeting, and, once they have had sex, are more likely to transition to cohabitation and do so sooner than more advantaged couples (Sassler, Addo, & Hartmann, 2010; Sassler, Michelmore, & Holland, 2016). That is where the accelerated development ends, however. Among cohabiting couples, those without a college education are less likely to be engaged and substantially less likely to marry than those who completed college, and a more rapid progress from sexual involvement to living together makes subsequent marriage even less likely (Sassler, Michelmore, & Qian, 2018; Sassler & Miller, 2011).

Exaggerating these patterns is the very different association between relationship status and childbearing across higher and lower levels of SES. As has been widely documented, over the past few decades age at first birth has been rising steadily for women who complete college (McLanahan, 2004). The more affluent tend to delay child-rearing until careers and stable relationships are in place; for most college graduates, child rearing begins after marriage (Lundberg et al., 2016). Accordingly, when they cohabit, these couples are more rigorous about using contraception, and use more effective methods (Sassler & Miller, 2014b). In contrast, less educated women are on average now nearly ten years younger when they experience their first birth, and that birth is roughly five times more likely to take place outside the context of marriage (Lundberg & Pollak, 2015).

In sum, it is increasingly the case that couples at higher and lower levels of SES take divergent paths in forming intimate relationships (Zhang & Ang, 2020). For those who complete college, escalating commitments take time but proceed steadily from sex to cohabitation to marriage and finally to first parenthood. For those without a college education, sex happens quickly after meeting, cohabitation soon after that, and first parenthood is mostly likely to precede marriage, if marriage happens at all.

3.2. Relationship Values and Expectations Across Levels of SES

How to understand the stark differences in relationship formation across higher and lower levels of SES? An assertion of some social critics, and the premise of federal investment in programs that promote the benefits of marriage (Small, Harding, & Lamont, 2010), is that couples in less advantaged communities simply do not value marriage as much as more affluent couples, and so would benefit from exposure to traditional family values. Research that has directly compared the aspirations of higher- and lower-income groups finds no evidence to support these assertions. Indeed, recipients of public assistance and low-income, moderate-income, and high-income respondents are all equally likely to agree with statements like “A happy healthy marriage is one of the most important things in life” and “People who have children together ought to be married” (Trail & Karney, 2012). In several respects, lower-income respondents endorse more traditional family values than their higher-income counterparts: they are significantly more favorable toward the idea that parents should remain married for the sake of their children and more likely to agree that divorce reflects badly on spouses as people. Consistent with the psychological dispositions account, those from less advantaged contexts appear to idealize strong, stable relationships as much, if not more, than people from more advantaged contexts. Nor are there measurable SES differences in the ideal structure of those relationships. As the two-income family has become normative (Raley, Mattingly, & Bianchi, 2006), couples at all levels of SES increasingly desire that both partners contribute equally to housework (Carlson, Miller, Sassler, & Hanson, 2016) and childcare (Carlson, Hanson, & Fitzroy, 2016) and are happier with their relationships to the extent that these standards are met (Carlson, Miller, & Sassler, 2018).

What distinguishes lower-income from higher-income couples are not values and standards but expectations. When middle-income or more affluent couples imagine a desired future, the desired two-income family appears to be well within reach, and for most college educated people, it is. Higher-income women who delay marriage to attend college can be confident that another college-educated partner will be available and willing to marry them after they graduate (Musick et al., 2012). Working-class and poor couples aspire toward the same ideal future, but “forces beyond the control of the individual tear up the pathways that make realization of those aspirations possible” (Carbone & Cahn, 2014, p. 32). Unlike their college-educated peers, women who do not attend college cannot be confident of finding a partner who earns a living wage at all (Gerson, 2010). With the availability of their desired relationships less certain, poorer couples are accordingly more skeptical about the future of their current ones (Gibson-Davis, Edin, & McLanahan, 2005; Miller, Sassler, & Kusi-Appouh, 2011).

3.3. Mating Strategies Across Levels of SES

Realistic assessments of the economic and mating prospects available at different levels of SES give rise to distinct mating strategies among richer and poorer segments of the population. At higher levels of SES, delaying marriage and child-birth increases the chances of finding a suitable mate, because, as the sociologist Valerie Oppenheimer (1988) argued in a classic analysis, the likely achievements of a college graduate become clearer with age. Consider the possible career paths that an ambitious and talented high school graduate may take. This plucky individual may continue on a path to success or may falter, may remain focused or may develop self-destructive habits, may choose a lucrative career or become an academic. The more time that passes, the greater the reduction in uncertainty about how this person’s life will turn out. At higher levels of income, where relatively low male-to-female gender ratios allow women to be selective, and where women’s own earning potential allow a degree of independence, the adaptive choice is to delay making a long-term commitment to a romantic partner until some of these unknowns become known.

The behavior of college-educated couples is consistent with this idea. When such couples cohabit, they do so with confidence that they will eventually marry, but meanwhile they scrupulously avoid pregnancy and invest in their careers (Miller et al., 2011). On average, this strategy pays off. The longer that individuals delay marriage, the more likely they are to marry someone with a similar level of education (Schwartz & Mare, 2005). Those who wait until they have maximized their own earning potential are increasingly likely to marry other top-earners (Greenwood, Guner, Kocharkov, & Santos, 2014; Schwartz, 2010), setting the stage to pass those resources on to the next generation (Cornelson & Siow, 2016).

Like their high SES peers, those at lower levels of SES would also like to find partners capable of matching or exceeding their own contributions to the household income (Gerson, 2010). In poorer communities, however, a combination of high rates of incarceration and high rates of unemployment leave relatively few men meeting this criterion (Lichter, McLaughlin, Kephart, & Landry, 1992) and these trends have hit African American men especially hard (Banks, 2011). Lower-income women, whose employment is less affected by boom and bust cycles and who have easier access to public assistance as the likely caregiver of any children, thus face a significant undersupply of men that they would consider reasonable marriage material, and no realistic optimism that this situation will change with time. Given that men who work less outside the home generally do not compensate by contributing more within the home (Schneider, 2011), low-income women reasonably question whether making a long term commitment to a partner who cannot earn a living is in their best interests (Edin & Reed, 2005) and whether such a union is likely to last (Burton, Cherlin, Winn, Estacion, & Holder-Taylor, 2009). Low-income men likewise have reasons to resist long-term commitments. Those with employment and earnings adequate for supporting a family know they are in short supply, and thus can play the field with impunity. Those who lack steady employment or sufficient income know they are unable to meet their partner’s expectations, and may resist committing to avoid facing their disappointment (Harknett & McLanahan, 2004).

Awareness of their options in the mating market affects how individuals at different levels of SES approach intimacy (Carbone & Cahn, 2014). Contexts where men’s economic prospects are uncertain inhibit couples from forming long-term commitments. Thus, when unemployment rates rise, marriage rates fall (Harknett & Kuperberg, 2011). Indeed, the decline in men’s economic prospects and the rise in male incarceration account for a significant portion of the overall decline in marriage rates over the past 45 years (Schneider, Harknett, & Stimpson, 2018). In contrast, partners with more education, and therefore more reason to be confident about their economic futures, are more likely to foresee making tangible investments in their current relationship, like opening joint bank accounts (Emery & Le, 2014).

Yet the same considerations have very different effects on approaches to parenthood. Because they can be more secure about their futures, the more affluent have much to gain from delaying pregnancy. The longer college educated women delay, the greater their own career attainment and the more likely they are to find a high quality mate to help them raise their children (Hymowitz, Carroll, Wilcox, & Kaye, 2013). With less certainty about their prospects of finding an economically viable partner, however, women without a college education have fewer reasons to delay pregnancy. On the contrary, for poor mothers, parenthood provides a socially valued role that the available career opportunities may not (Edin & Kefalas, 2005). In communities where there are few models of successful marriage but many models of successful single parenting, childbearing outside marriage carries little stigma (Cherlin, Cross-Barnet, Burton, & Garrett-Peters, 2008). It is not surprising, then, that when confronted with cues to their own mortality individuals from poor backgrounds express a desire to have children sooner, even at the expense of their own education, whereas individuals from wealthier backgrounds express a desire to postpone parenthood to invest more in their careers (Griskevicius, Delton, Robertson, & Tybur, 2011). When considering the timing of forming a family, individuals at different levels of SES adopt mating strategies suited to their accurate understanding of their options.

4. SES AND RELATIONSHIP MAINTENANCE

Once intimate relationships form and two people try to sustain their romantic connection over time, couples at all levels of SES confront some similar challenges. For example, couples with high, medium, and low household incomes are all equally likely to report struggles with being a parent/having children, communication, sex, household chores, each other’s parents, and spending time together (Trail & Karney, 2012). Some problems, regardless of context, appear to be endemic to the task of raising a family and negotiating differences with another person. Other problems, however, do vary across socioeconomic strata. The same data indicate that, compared to more affluent couples, lower-income couples are more likely to report relationship difficulties arising from outside the relationship itself, like problems with money, substance abuse, infidelity, and conflicts over friends. The additional challenges that couples at lower levels of SES face have broad implications for their ability to maintain their relationships.

4.1. SES Moderates the Effects of Other Stressors

Informed by the stress and resources account, research on couples has often examined the effects of specific stressors and resources one at a time, as if each might affect couples independently. Yet the privileges associated with higher SES and the corresponding challenges associated with lower SES generally cluster together (Kraemer, Stice, Kazdin, Offord, & Kupfer, 2001). If the effects of different sources of influence on couples are additive, i.e., they do not interact with each other, then examining each independently is warranted. If different demands and supports do interact, however, then the way any particular challenge affects couples may depend on couples’ level of SES and the associated constellation of additional challenges that couples confront.

One possibility is that couples at lower levels of SES become more resilient, adapting to each stressor more effectively because they have had to adapt more frequently (e.g., Neff & Broady, 2011; Schetter & Dolbier, 2011). The result would be a buffering effect: any given stressor may have a weaker impact on less advantaged couples for whom other stressors are already likely to be present. An alternative possibility is that couples at lower levels of SES will have more difficulty responding to any given stressor because their coping resources are likely to be taxed by other simultaneous demands (e.g., Buck & Neff, 2012). The result would be an exacerbation effect: any given stressor may have a stronger negative impact on couples who are less advantaged overall.

Research evaluating these alternatives drew upon survey data from Florida, California, New York, and Texas (the four most populous states in the U.S.) that included oversamples of respondents at lower levels of income and respondents receiving public assistance (Rauer, Karney, Garvan, & Hou, 2008). Respondents provided data on seven established correlates of lower relationship satisfaction, all of which are also known to be more prevalent at lower levels of SES: lack of a high school diploma, psychological distress, substance abuse, financial strain, acute stress, social isolation, and intimate partner violence. Consistent with prior research, each risk factor was associated with lower relationship satisfaction. In addition, associations between specific risks and relationship satisfaction were significantly moderated by the accumulation of other risk factors, and the direction of the interaction supported an exacerbation effect. Specific risk factors were more strongly associated with lower relationship satisfaction when they occurred in the context of other risks; the more additional risks, the stronger the association.

To the extent that SES is associated with multiple sources of risk, the way any given stressor affects a relationship is likely to depend on the couple’s level of SES. The same events and circumstances that may be irritating to couples at with greater resources may be catastrophic for couples lacking those resources. Research that acknowledges the effects of SES on intimate relationships has been far more likely to examine main effects than interactions, but studies that have examined SES as a moderator support this idea. For example, analyses of data from the General Social Survey have compared associations between employment uncertainty (“Thinking about the next 12 months, how likely do you think it is that you will lose your job or be laid off?”) and work-family conflict (“How often do the demands of your job interfere with your family life?”) at different levels of household income (Fan, Lam, & Moen, 2019). Employment insecurity was associated with greater work-family conflict for the lowest income men, but not for higher-earning men. Similar research has found that, across multiple indices of SES, stressful life events and mental health problems are more strongly associated with relationship satisfaction for poorer individuals than for more affluent individuals (Maisel & Karney, 2012).

In sum, lower levels of SES do not merely confront couples with greater challenges; a context of disadvantage makes those challenges harder on the relationship. Higher levels of SES do not merely protect couples from external stress; a context of advantage makes the challenges that couples do confront easier to manage.

4.2. SES and Dyadic Processes

When partners are depleted by demands outside the relationship, the work of maintaining that relationship (e.g., negotiating differences, accommodating each other’s minor transgressions, allocating chores) becomes more difficult (Buck & Neff, 2012; Neff & Karney, 2004). For example, couples who are perfectly capable of communicating effectively during periods of calm find themselves less able to do so when placed under stress (Bodenmann et al., 2015). To the extent that it accounts for a broad array of more or less stressful circumstances, SES should therefore exert a main effect on dyadic processes, such that effective techniques of relationship maintenance are easier for more advantaged couples and harder for less advantaged couples.

Research guided by the Family Stress Model has developed this idea, proposing that economic pressures on couples and families directly increase partners’ emotional distress, which in turn exacerbates conflicts within the relationship (Conger, Rueter, & Elder, 1999). Observational research on the marital interactions of couples ranging in SES consistently supports this perspective, showing that spouses experiencing difficulty paying bills exhibit more hostility (e.g., anger and contempt) and less warmth (e.g., praise and affection) than spouses whose financial obligations are being met (Conger et al., 1990; Masarik et al., 2016). More objective indicators of SES reveal similar effects. For example, research on African American couples drew upon census data to show that couples living in economically disadvantaged neighborhoods expressed less warmth toward each other during a videotaped problem-solving interaction than couples living in more affluent neighborhoods (Cutrona et al., 2003). Even more broadly, the same study found that, controlling for other correlates, couples living in rural areas in Georgia (as opposed to mid-sized cities in Iowa) also expressed less warmth during their interactions, presumably because life is more challenging for African American couples in the South than in the mid-West. Couples’ economic contexts account for the quality of their communication over and above the effects of personal history, depressive symptoms, and relationship satisfaction (Williamson, Karney, & Bradbury, 2013), likely because lower-income couples face more severe and intractable problems (Jackson et al., 2016). A disagreement about where to spend the holidays is often easier to resolve than a disagreement about which bills to leave unpaid, as heated as both conversations can be.

When their relationships need help that partners cannot provide themselves, more affluent couples are significantly more likely than poorer couples to seek out and receive interventions (Halford, O’Donnell, Lizzio, & Wilson, 2006), mostly because couples at lower levels of SES fear the cost of such services and are uncertain about where to find providers (Williamson, Karney, & Bradbury, 2018). By the time that they do seek therapy for relationship issues, lower-income couples are more distressed and their problems are more severe than higher-income couples seeking the same interventions (Williamson, Nguyen, Rothman, & Doss, 2019). As a consequence, relationship enhancement programs and therapies that have proven to be effective for higher-income couples may not be as effective for lower-income couples (e.g., Lundquist et al., 2014).

An additional obstacle to intervening with lower-income couples is that most empirically validated couples therapies (e.g., cognitive behavioral approaches; Bradbury & Bodenmann, in press) were developed and designed for use with middle-and upper-income couples. Efforts to disseminate such interventions to disadvantaged couples assume that the same dyadic processes are likely to be adaptive across levels of SES, even if enacting those behaviors may be harder for couples under stress. An alternative possibility is that socioeconomic contexts can moderate the sorts of behaviors that best promote intimacy, such that behaviors that are adaptive for couples at higher levels of SES may be maladaptive for couples at lower levels, and vice versa.

One study that has explored this possibility examined the implications of the demand/withdraw pattern. One of the most frequently studied communication patterns in relationship science, the demand/withdraw pattern occurs when one partner seeks a change in the relationship and the other partner, preferring the status quo, withdraws to avoid making that change (Christensen & Heavey, 1990). Among the relatively affluent couples who represent the majority of the participants in research on intimate relationships, the demand/withdraw pattern is a sign that partners are not invested in each other’s happiness, as the withdrawing partner is refusing to make a change that presumably could be made. Thus, the demand/withdraw pattern is reliably associated with lower relationship satisfaction among affluent couples (Christensen, Eldridge, Catta-Preta, Lim, & Santagata, 2006). But what if the requested change were impossible? Within lower levels of SES, many of the changes that one partner could request of the other (e.g., get a better job, contribute more to household income, participate more in child-rearing) may be outside of the control of the other partner. In such circumstances, withdrawal might be an adaptive response. In fact, analyses of data from two studies that included diverse couples from a wide range of SES converge to support this idea, finding that the demand/withdraw pattern is negatively associated with relationship outcomes among upper-income couples, but positively associated with relationship outcomes among lower-income couples (Ross, Karney, Nguyen, & Bradbury, 2019). Although these findings await further replication, they suggest that adapting to lower levels of SES may not only make some relationship maintenance processes more difficult, it may also alter the way those processes function within the relationship. A task for future research is to explore other behaviors and processes that may serve distinct functions for more or less advantaged couples.

4.3. SES and Relationship Satisfaction

As a result of their increased vulnerability to acute and chronic stress (McLeod & Kessler, 1990), and the impact of those challenges on couples’ ability to communicate effectively, stress and resource perspectives (e.g., Conger & Conger, 2008; Karney & Neff, 2013) predict that, on average, couples at lower levels of SES should experience less satisfying intimate relationships than couples at higher levels. Yet this association has not been documented often, for several reasons. First, large-scale studies of couples and family structures (e.g., the National Survey of Family Growth) often neglect to assess relationship quality. Second, studies that do include assessments of relationship satisfaction often sample from a restricted range of SES. College student samples, for example, are by definition attending and likely to complete college, limiting the variability in SES available to analyze (Kerkmann, Lee, Lown, & Allgood, 2000). Third, researchers that do collect data on community samples often treat measures of SES as sources of extraneous variance to be statistically controlled, rather than measures of interest in their own right.

Instead, psychological researchers are more likely to examine measures of subjective financial strain (e.g., “During the past year, how much difficulty have you had paying your bills?”). Such measures are reliably associated with relationship satisfaction, such that couples who perceive more financial difficulties generally report more conflicts and lower satisfaction with their relationships (Conger et al., 1990; Masarik et al., 2016). The problem with such measures, however, is that self-reported financial strain is an imperfect proxy for a couples’ status within a broader social structure. Couples at lower levels of SES, by virtue of being poor, are certainly more likely than wealthier couples to experience financial strain, but some poor couples adapt well to their limited resources (e.g., Conger et al., 1999), and some wealthier couples may still have occasional difficulties paying their bills.

More objective indices of SES, like education and income, have also been associated with relationship satisfaction, but without the benefit of shared method variance these associations tend to be smaller (Voydanoff, 1990). Education, perhaps because it does not fluctuate much over time within adult populations, has been the more reliable predictor. Analyses of data from multiple large-scale, representative surveys in the U.S. and Germany find that couples with more education report more love for each other and more satisfaction with their relationships than couples with less education (Hardie, Geist, & Lucas, 2014; Hardie & Lucas, 2010; Maisel & Karney, 2012). Household income, in contrast, was not significantly associated with partners’ evaluations of the relationship in any of those studies, although a positive correlation between income and relationship satisfaction has been observed in studies from other countries (e.g., Zainah, Nasir, Hashim, & Yusof, 2012). Analyses of data from the Taiwan Panel Study of Family Dynamics suggest that the association between income and relationship satisfaction may be curvilinear, such that working class couples report greater relationship satisfaction than the poorest couples, but higher levels of income offer no additional advantage (Tao, 2005). Such an effect is consistent with the view that income improves well-being only to the extent that it allows people to meet their basic needs (Diener & Biswas-Diener, 2002), and with other data showing differences between the poorest couples and working-class couples, but not between working-class couples and more advantaged couples (Dakin & Wampler, 2008; Williamson & Lavner, 2019).

SES affects more than relationship satisfaction at one time; it is also associated with how relationship satisfaction changes over time. A limitation of the research on SES and relationship satisfaction presented thus far is that it all compares couples assessed on a single occasion. Such cross-sectional analyses necessarily exclude couples who have already dissolved by the time of data collection, and these are likely to be the most vulnerable couples. To avoid this problem, longitudinal research has assessed changes in marital satisfaction across the early years of marriage in newlywed couples that vary widely in SES (Jackson, Krull, Bradbury, & Karney, 2017). Like the cross-sectional surveys, this study found that couples with higher or lower household incomes did not begin their marriages at different levels of marital satisfaction. Nor did they differ in the rate at which their marital satisfaction changed over the first four years of their marriage. What household income did predict was the variability between assessments and between spouses within a couple. For the more affluent couples, marital satisfaction was more stable across assessments, and husbands and wives within a couple tended to agree on their evaluations of the relationship. For the poorer couples, marital satisfaction was more variable across assessments, perhaps reflecting the greater turbulence of life at lower levels of SES. Moreover, poorer husbands and wives were less likely to evaluate their relationships the same way, consistent with the idea that the demands on spouses at lower levels of SES constrain their shared leisure time, and thus their ability to form a common understanding of their relationship (Presser, 2000).

5. SES AND RELATIONSHIP DISSOLUTION

There are several reasons why one might expect intimate relationships to be more enduring at lower levels of SES. First, the psychological dispositions account proposes that people from lower-income backgrounds may value connection and interdependence more than people from higher-income backgrounds (Kraus et al., 2012), and survey data from people at different levels of SES offer some support for this view (Trail & Karney, 2012). Second, given that marriage rates are lower in poorer communities than in wealthier communities (Rosenfeld & Roesler, 2019), selection into marriage may favor the most stable, committed couples in poorer communities, in contrast to wealthier communities where less committed couples are still likely to get married. Finally, couples who lack resources may remain in unsatisfying relationships because they cannot afford to bear the financial costs of separating (Harknett & Schneider, 2012).

Despite these considerations, however, intimate relationships among lower-income couples are, as noted earlier, far less stable than among higher-income couples. Whether they are married or cohabiting, more disadvantaged couples are at greater risk of dissolving and when they do dissolve they dissolve sooner, compared to more advantaged couples (Copen et al., 2012; Lundberg et al., 2016; Rosenfeld & Roesler, 2019). Together these trends suggest that, whatever factors there may be that protect relationships at lower levels of SES, they are on average overwhelmed by several other factors that greatly exacerbate their risk. For example, couples at lower levels of SES are from the outset of their relationships more likely than couples at higher levels of SES to experience challenges, including lower relationship satisfaction, premarital and extra-pair children, and financial strain, that are well-established precursors of dissolution and divorce (Karney & Bradbury, 1995). Greater fluctuation in relationship satisfaction trajectories, also more prevalent among lower-income couples, has been linked to greater risk of subsequent break-ups as well (Arriaga, 2001; Whitton, Rhoades, & Whisman, 2014).

Yet, while the relative risks and challenges among couples at different levels of SES have been well-documented, how more or less advantaged couples decide when to end their intimate relationships has received far less attention. Data on these decisions comes mostly from research that has reviewed reasons for divorce as listed on divorce applications (e.g., Kitson & Holmes, 1992), as well as longitudinal studies of couples that have been able to obtain data from one or both partners after the relationship ended (Amato & Previti, 2003). Across studies and methods, some reasons for divorce are equally likely across levels of income (e.g., infidelity), but persistent differences between couples at higher and lower levels of SES have also emerged. Among couples with higher incomes and college diplomas, the most frequent reasons for divorce include lack of communication, incompatibility, lack of love, and personality differences. In contrast, couples with lower incomes and those who did not attend college are significantly more likely to cite physical abuse, substance abuse, and failure to contribute to the household.

The difference between the more relationship-centered reasons for divorce cited by couples at higher levels of SES and the more instrumental reasons for divorce cited by couples at lower levels of SES maps onto the mating strategies account of SES and intimate relationships (Carbone & Cahn, 2014). College-educated individuals can take many things for granted: their partners are likely to be employed and relatively unlikely to be incarcerated, addicted, or abusive. With their material needs satisfied, these fortunate individuals are free to prioritize emotional connection with their romantic partners, and to end their relationships when their expectations for connection are not met. Individuals who did not complete college also value and aspire to emotional connections with their partners, but they simultaneously have more salient concerns. They cannot take for granted that their partners will participate in child-rearing, refrain from antisocial behaviors, or contribute to their household income. This is especially true for lower-income wives, 70% of whom earn more than their husbands, compared to 34% in higher-income couples (Glynn, 2012). Their sense of whether their partners are fulfilling their roles as parent and provider therefore plays a larger role in their decisions about whether or not their relationships are worth maintaining.

Differences between couples at higher and lower levels of SES do not end after the relationship does. All along the socioeconomic spectrum, divorce predicts marked declines in wealth and financial security, especially for women (Hogendoorn, Leopold, & Bol, 2019; Smock, Manning, & Gupta, 1999). Individuals who have graduated from college, however, remarry at high rates, and when they do, they generally recover economically (Ozawa & Yoon, 2002). Men without a college education are less likely to remarry (Shafer & James, 2013), and when couples at lower levels of SES do find new partners, their financial recovery is not as strong (Ozawa & Yoon, 2002). In other words, the effects of differences in SES on how couples enter and exit intimate relationships tend to reinforce differences in SES, such that poorer couples, at greater risk of dissolving, are likely to become poorer still, whereas wealthier couples, at lower risk of dissolving and higher likelihood of remarrying successfully, accumulate greater wealth to pass on to the next generation (Greenwood et al., 2014).

6. CONCLUSIONS AND RESEARCH PRIORITIES

Although individuals who have graduated from college or who expect to graduate from college comprise just under 40% of Americans between the ages of 25 and 29, and an even smaller proportion of older cohorts (U.S. Department of Education, 2019), they represent the vast majority of all participants in research on marriage and intimate relationships (Karney & Bradbury, 1995). Paying such disproportionate scrutiny to the minority of couples could be excused if intimate relationships are generally similar across levels of SES. As this review has made clear, that not the case. On the contrary, as summarized in Table 1, intimate relationships at different levels of SES differ from each other in many notable ways.

Table 1:

Summary of Key Similarities and Differences between Intimate Relationships at Higher and Lower levels of SES

Dimension of Assessment	Couples with a Bachelor’s degree or higher	Couples with less than a Bachelor’s degree
Aspirations	Agree that “A happy healthy marriage is one of the most important things in life”; Idealize two-parent, two-income, egalitarian families
Expectations	Confident that ideals are within reach	Skeptical of prospects
Gender ratios	More marriageable men than women	Fewer marriageable men than women
1st meeting	More likely in select or private spaces	More likely in public spaces
Early development	Delayed sex and cohabitation	Rapid sex and cohabitation
Marriage rates	Highly likely to marry	Less likely to marry
Child-birth	Typically follows marriage	Typically precedes marriage
Similarity between mates	Likely to pair with someone from the same level of SES
Relationship maintenance	More warmth, less hostility, no benefit to withdrawal	Less warmth, more hostility, withdrawal can be effective
Trajectories of satisfaction	Higher satisfaction, more stable over time	Lower satisfaction, more turbulence over time
Divorce rates	Less likely to divorce	More likely to divorce
Reasons for divorce	More emotional: lack of communication, incompatibility, lack of love	More instrumental: physical abuse, substance abuse, failure to contribute
Seeking treatment	More likely to seek out and find couples’ therapy	Less access to interventions and assistance
After the relationship	Likely to remarry and recover financially	Less likely to remarry, financial consequences are lasting

Studying the intimate relationships of people who are attending or who have completed college means studying couples who take for granted that their material needs will be met. They expect to marry someone who will be an income earner and an involved parent at the same time, and, confident that such a person will be available, they are willing to delay cohabitation, marriage, and child-birth until they find the right mate, investing in their own education and careers in the meantime. When they do commit to a long-term partner, they have resources and support that minimize the impact of external stressors, leaving them free to prioritize emotional support and connection, work through disagreements, and get therapy when they encounter problems too difficult to handle by themselves. As a consequence, their likelihood of breaking up is low, but when they do break-up, they are likely to find new partners.

Individuals who do not attend or complete college aspire to these same outcomes, but their likelihood of achieving them is far lower, and they know it. A lack of well-paying jobs translates to a lack of marriageable men, giving rise to distrust between genders and an accurate suspicion on both sides that long-term commitments may not be worth the risk. With fewer anticipated benefits to delaying parenthood, child-birth happens sooner and most frequently outside the context of marriage. Those who do find long-term partners face increased exposure to challenges from outside the relationship, like health issues, unemployment, and financial strain, raising intractable problems that must be managed, taking priority over activities that nurture intimacy. With sources of assistance for the relationship hard to find and difficult to afford, couples without a college education are at greater risk of breaking up, and less likely to form new committed relationships when they do break up. The financial damage caused by the dissolution of a long-term relationship is therefore likely to be permanent, widening the income inequality that gave rise to the differences between couples at higher and lower levels of SES in the first place.

In sum, couples at different levels of SES live in disparate worlds, each of which constrains some of their options while facilitating others. As a result, scholarly conclusions about intimacy will often fail to generalize across levels of SES, and it is well past time for relationship science to acknowledge that fact. A first step in toward reckoning with SES in relationship science would be to develop habits of describing research that make SES visible. Currently, it is acceptable for researchers to describe a sample of “couples” as if the demographic characteristics of those couples makes no difference to interpretations of their results. Reporting on household income and level of education within samples should be a minimum requirement to allow informed evaluations of how far effects within a study are likely to generalize.

A second step is to include couples from a broader range of SES in research on intimacy. Doing so is likely to require increased funding for relationship research, as studying couples outside of university settings can be costly. Even with resources to sample within communities, less educated couples, with correspondingly less exposure to the idea of participating in research, are significantly less likely to respond to solicitations to participate (Karney et al., 1995). MTurk and other on-line panels are imperfect solutions to this problem, as the population of people who participate regularly in those panels is mostly college-educated as well (Huff & Tingley, 2015; Ipeirotis, 2010). One promising strategy is to design sampling frames around specific communities, using census data to identify neighborhoods where residents are higher or lower in SES on average (Elliott et al., 2013). Whatever way this field rises to meet these challenges, rise it must, as a science that relies primarily on data from college graduates is unlikely to develop a basic understanding of intimate relationships and the role they play in adult lives. A science that relies primarily on data from college students is even less viable.

Once relationship science accepts the need to focus less narrowly on the most accessible populations, several directions stand out as priorities for future research. First, whereas this review has focused on effects associated with SES, those effects are likely to interact with culture, race, and ethnicity in ways that have yet to be examined. In the United States, race and ethnicity are strongly associated with SES, in that white people are overrepresented among higher levels of SES and underrepresented at lower levels (Williams, 1996). Yet the association is far from perfect, so it is a mistake to treat race and ethnicity as proxies for SES. To the extent that researchers acknowledge that the differences in intimate relationships at different levels of SES may be independent from differences associated with race, ethnicity, and culture, then the door is opened to explore how race, ethnicity, and culture may moderate the implications of SES for couples. Suggestive research along these lines finds that, whereas higher SES is associated with greater self-focus in the United States and other Western countries (consistent with the psychological orientation account), higher SES is associated with less self-focus and greater other-focus in Japan and other Eastern countries (Miyamoto et al., 2018). The implications of SES for mating strategies, relationship maintenance, and dissolution may also vary across cultures, e.g., cultures that vary in the centrality of family ties (Campos, Perez, & Guardino, 2014).

Acknowledging the impact of SES also points out new approaches for promoting stronger intimate relationships. One of the limitations of federal policies like the Healthy Marriage Initiative is that they focused narrowly on educational programs that either promote pro-marriage attitudes or teach specific communication skills expected to benefit couples (Manning et al., 2014), interventions that have little impact on the lower-income couples that the programs targeted (Lundquist et al., 2014; Wood et al., 2014). If many of the obstacles facing lower-income couples are related to their level of SES, then an alternative strategy would be to develop programs that address those obstacles directly (Karney, Bradbury, & Lavner, 2018). Analyses of data from the Fragile Families study, for example, suggest that improving labor markets in lower-income communities would lead to significant increases in marriage rates within those communities (Harknett & Kuperberg, 2011). Indeed, a review of five anti-poverty programs indicates that they have consistent positive effects on relationship stability for participants, even though such programs do not address relationships directly (Lavner, Karney, & Bradbury, 2015). From this perspective, all programs that work to alleviate income inequality are likely to benefit intimate relationships in important ways, leading some to suggest that a family impact statement be a routine part of evaluating new policies (Bogenschneider, 2014).

Finally, integrating SES within models of intimate relationships supports a broader point: micro-level dyadic processes are always facilitated and constrained by the macro-level contexts in which they take place. A rich understanding of how couples manage disagreements, exchange support, and have sex therefore requires recognizing the surrounding history, geography, culture, economy, and political structure that restricts some choices in the moment and makes others possible. Interdisciplinary collaborations to draw links between these levels of analysis could be the foundation of a generative future for relationship science.