Rhiannon A. Kroeger and Pamela J. Smock
Cohabitation, or living together in an intimate sexual relationship outside of marriage, has become a normative life event for most population subgroups in the United States as well as in many other countries (Smock, 2000; Lesthaeghe, 2011). While marriage remains desirable and important, with large numbers of adults expressing support for marriage and intentions to marry, recent research demonstrates that cohabitation has unmistakably altered marriage and childbearing processes (Manning, Longmore, and Giordano, 2007; Musick, 2007; Guzzo, 2009). In the United States, the majority of young adults perceive cohabitation as an acceptable arrangement, and by age 25, nearly half have spent some time in a cohabiting relationship (Payne, 2011). Cohabitation has also become common among older adults. Among adults ages 50 and over, the number of cohabitors increased by more than 1.5 million between 2000 and 2010 (Brown, Lee, and Bulanda, 2006; Brown, Bulanda, and Lee, 2012). The incidence of cohabiting partners with children is increasingly widespread, too: nearly 60% of nonmarital births are now to cohabiting parents (Lichter, 2012).
This chapter reviews and evaluates knowledge on cohabitation, with an emphasis on recent research (see Smock, 2000 for a review of earlier research). Our review is not comprehensive. Given the vast amount of research on cohabitation, there are some areas we cover minimally or not at all. As demographers and sociologists, we primarily draw on studies from those disciplines. We also do not address same-sex cohabitation, which is treated elsewhere in this volume (see Biblarz, Chapter 6, this volume). While we incorporate research on cohabitation in other regions of the world, our emphasis is on the United States.
Our chapter is organized into three general sections. First, we present basic cohabitation trends and key patterns in the United States, followed by a review of cohabitation patterns in other countries. Next, we identify and summarize the findings of four questions of importance in the extant research on cohabitation. Finally, we draw on recent research to articulate several implications that we believe are important for understanding cohabitation.
There have been tremendous increases in the prevalence of cohabitation over the past several decades. One way to illustrate such increases is through counts of the number of couples living together. Estimates indicate that approximately 7.6 million heterosexual unmarried couples were living together in the United States as of 2011. This is double the number in 2000, nearly 5 times the number in 1980, and over 17 times the number in 1960 (U.S. Census Bureau, 2011). A clear way to understand why family scholars often term the increase in cohabitation a “revolution” is via data that tell us whether individuals have ever experienced a cohabiting relationship. Approximately 58% of women ages 19–44 have ever cohabited based on data from 2006 to 2008, compared to 45% in 1995 and 33% in 1987. Among 30–34-year-old women, the trend is quite striking: in 2006–2008 roughly 70% had ever lived in a cohabiting relationship (Manning, 2010). Among very young adults, cohabitation is now the most common type of first union, such that nearly half have ever cohabited by age 25 (Kennedy and Bumpass, 2008; Payne, 2011).
Notwithstanding common perceptions that the so-called cohabitation revolution only involves young adults, cohabitation is also common among older adults and the increases have been steep over the last few decades. Estimates from a 1987 survey suggest that 22% of 40–44-year-old women had ever cohabited; data from 2006 to 2008 indicate that nearly 60% had ever done so (Manning, 2010). Moreover, among adults ages 50 and over, the number of cohabitors rose from 1.2 million in 2000 to 2.75 million in 2010 (Brown, Lee, and Bulanda, 2006; Brown, Bulanda, and Lee, 2012). Some research suggests that older generations are “learning” cohabitation from their children. For example, one study found that among parents who were single at their child’s 18th birthday, subsequent cohabitation experiences among their adult children increased the likelihood that parents themselves would cohabit rather than marry (McClain, 2011b).
Cohabitation is increasingly prominent in children’s lives, either by virtue of being born to cohabiting parents or by having a parent, typically the mother, enter into a cohabiting relationship. In fact, cohabiting households are now as likely to include children as married households, with about 40% of cohabiting households including children (Krivickas and Payne, 2010; Cohen, 2011; Lofquist et al., 2012). Up to one half of children born in the early 1990s are expected to live in a cohabiting-parent household at some point in their childhood (Bumpass and Lu, 2000; Kennedy and Bumpass, 2008). The percentage of all US births that were to cohabiting parents increased from 6% in the 1980s to 11% in the 1990s to 20% in 2006–2008 (Bumpass and Lu, 2000; Lichter, 2012). In 2001, 52% of all nonmarital births were to cohabiting parents (Manlove et al., 2010). By 2006–2008, nearly 60% of nonmarital births were to cohabiting parents (Lichter, 2012). Thus, the majority of nonmarital childbearing today occurs not to single women but to couples residing together.
There have also been increases in the number of cohabiting unions that individuals experience. Correspondingly, family scholars have begun to differentiate between single-instance cohabitation and what has been termed “serial” cohabitation. Generally, researchers consider individuals to be serial cohabitors if they have cohabited two or more times. The average number of cohabiting unions increased from 0.44 for women and 0.55 for men in the 1970s to 1.15 for women and 1.36 for men in the early 2000s (Zeng et al., 2012). In 1995, approximately 9% of women ages 15–44 had cohabited with more than one partner. By 2002, this percentage rose to 12% (Lichter, Turner, and Sassler, 2010). Among individuals who have cohabited at least once, between 15% and 25% are estimated to be serial cohabitors (Lichter and Qian, 2008; Cohen and Manning, 2010; Lichter, Turner, and Sassler, 2010; Manning and Cohen, 2012). Serial cohabitation is more prevalent among the young and serial cohabitors enter their first cohabitation 1–2 years earlier than their counterparts who have cohabited only once (Lichter and Qian, 2008; Cohen and Manning, 2010; Lichter, Turner, and Sassler, 2010).
Researchers have for some time now been studying the connections between cohabitation and marriage, often driven by the premise that it is not possible to understand the implications of cohabitation without also considering marriage. In addition, the public, and therefore many researchers, have been motivated by whether cohabitation ought to be perceived as a “threat” to marriage. Thus, a central question has been to what extent cohabitation serves as a precursor to marriage rather than a short-lived dating relationship or a long-term alternative to marriage. In the United States, more than 65% of first marriages now begin as cohabitations (Manning, 2010; Copen et al., 2012). This is in sharp contrast to marriages occurring in earlier years. For instance, only about 10% of marriages formed in the 1960s and early 1970s were preceded by cohabitation (Bumpass and Sweet, 1989; Bumpass and Lu, 2000; Manning, 2010).
A somewhat different way to approach the issue about what cohabitation means for the centrality of marriage is to examine the “outcomes” of cohabiting relationships. Some cohabiting couples go on to marry and some do not. In the absence of wedding bells, cohabiting couples either break up or stay together for any number of years. Regardless of whether cohabitating unions end in marriage or dissolution, they are generally short-lived, with most ending within 2 years (Raley and Bumpass, 2003). About half of individuals in their first cohabitations and 16% of those in higher-order cohabitations go on to marry their partners (Lichter and Qian, 2008). Of those remaining, the majority of couples break up within a few years, and only a small percentage continues to cohabit 5 years or longer.
Cohabitation after divorce and prior to, or instead of, remarriage has also become more common. In the 1970s, roughly 24% of women cohabited after divorce, compared to about 46% in the early 2000s (Zeng et al., 2012). Correspondingly, scholars have noted that the likelihood of remarriage after divorce has declined, with cohabitation becoming a much more common route to coupledom.
In some sense, given the growing prominence of cohabitation, the question has become “who does not cohabit?” We do know that those with strong religious convictions are less likely to cohabit, as well as individuals living in communities with high levels of religious homogamy (Eggebeen and Dew, 2009; Brown, Bulanda, and Lee, 2012; Gault-Sherman and Draper, 2012).
As with many family patterns, cohabitation patterns are influenced by economic well-being, or what many researchers term “social class.” For instance, those who are more advantaged economically are less likely to cohabit than their less advantaged counterparts. Nearly three-quarters (73%) of individuals without a high school degree have ever cohabited compared to less than half (47%) of those with a college degree or more. Congruently, research indicates that serial cohabitation is more prevalent among those at the lower end of the educational spectrum (Lichter and Qian, 2008; Cohen and Manning, 2010; Lichter, Turner, and Sassler, 2010). Further, while many factors influence the likelihood of cohabitation leading to marriage, economic well-being seems to play an important role, with couples remaining reluctant to marry until they have attained economic stability (Smock, Manning, and Porter, 2005). Indeed, cohabitations are less likely to transition to marriage among the economically disadvantaged compared to their more advantaged counterparts (Lichter, Qian, and Mellott, 2006).
Cohabitation patterns also vary by race/ethnicity. Generally, there is little difference across non-Hispanic whites, non-Hispanic blacks, and Hispanics in the likelihood of having ever cohabitated. Statistics on cohabitation rates among Asian-Americans have generally been more difficult to obtain because there are few data sets with sufficient sample sizes to provide solid estimates. However, recent studies utilizing various national data sources suggest that, at least through young adulthood, fewer Asians report having ever cohabited than do their white or black counterparts (Cheng and Landale, 2011; Zeng et al., 2012).
The most striking detail regarding racial/ethnic differences is that fewer cohabitations transition to marriage among non-Hispanic blacks or Hispanics compared to whites (Guzzo, 2009; Zeng et al., 2012). However, it is important to note two distinctions between blacks, whites, and Hispanics. The first is that cohabitating unions among black women are less stable than those among white women (e.g., more likely to break up and less likely to result in marriage). The second is that while relationship stability is similar among whites and Hispanics, Hispanics are more likely to remain in cohabiting unions versus marry than are whites. Scholars generally cite greater cultural acceptance of unmarried cohabitation across various parts of Latin America as being a source of this difference (Landale, Schoen, and Daniels, 2010). Racial/ethnic differences in the number of cohabitations that transition to marriage may in part be related to social class. With whites being more economically advantaged than blacks or Hispanics, one study finds that racial/ethnic differences in women’s expectations of marrying their cohabiting partners is narrowed considerably once the male partners’ income and earnings are taken into account (Manning and Smock, 2002).
Far from a US-bound phenomenon, cohabitation is prominent across much of the world. While an exhaustive review of cohabitation research in all countries is beyond the scope of this chapter, here we summarize some notable trends in cohabitation around the globe. An important point underlying these trends is that the prevalence of cohabitation in any given country is both motivated by and reflective of national context. In part, scholars have observed that cohabitation is more common in industrialized nations (Lesthaeghe, 2011). Indeed, increases in cohabitation have been observed in conjunction with industrialization in multiple “Western” countries, including Canada, Australia, and many European countries. But the importance of national context is not merely a consequence of industrialization. Indeed, the prevalence of cohabitation is reflective of cultural factors, religious traditions, and politics, and is influenced immensely by social networks (Nazio and Blossfeld, 2003; Heuveline and Timberlake, 2004; Nazio, 2008; Lesthaeghe, 2011).
This complex array of factors influencing cohabitation is apparent across Europe, where cohabitation prevalence varies widely by region. For instance, in Northern Europe, cohabitation is extremely prevalent (Soons and Kalmijn, 2009; Kasearu and Kutsar, 2011). In Sweden, over 60% of 18–45-year-old adults had ever cohabited, and the typical path to marriage is a sequence involving cohabitation, childbearing, and then (maybe) marriage (Kasearu and Kutsar, 2011). In Norway, 75% of the young adult population selects cohabitation (rather than marriage) as their first union and nearly all (90%) of married couples cohabit prior to marriage (Aarskaug Wiik, 2011). Cohabitation is increasingly prevalent in countries such as France, the United Kingdom, Germany, Austria, and Belgium, where between 30% and 50% of the adult population has ever cohabited (Kiernan, 2004;Soons and Kalmijn, 2009; Kasearu and Kutsar, 2011; Ochiai, 2011).
In Central/Eastern Europe, cohabitation levels are more modest. In Russia, recent data suggest that about 20% of the adult population has ever cohabited (Soons and Kalmijn, 2009). Although recent estimates suggest that less than 20% of adults ages 18–45 in Poland have ever cohabited, there has been a striking rise since the early 1990s (Mynarska and Matysiak, 2010; Kasearu and Kutsar, 2011). The spread of cohabitation has been slowest in Southern Europe, although again cohabitation is in upward swing (Munoz-Perez and Recano-Valverde, 2011; Impicciatore and Billari, 2012). In Italy, data from 2008 indicate that less than 11% of the adult population in Italy has ever cohabited (Soons and Kalmijn, 2009). Proportions of individuals with cohabitation experience are not much higher for Portugal (13.3%), Greece (11.4%), or Spain (17.0%).
While the spread of cohabitation has only recently been documented across many Western nations, cohabitation has long existed alongside formal marriage in Latin America, especially among rural and indigenous populations (Gonzalbo Aizpuru, 1998; Castro Martin, 2002; Fussell and Palloni, 2004). Indeed, while cohabitation has gone from negligible to normal over the past several decades in the United States, Canada, Australia, and much of Europe, statistics indicate its persistently high prevalence in Latin America since at least the 1950s (Fussell and Palloni, 2004). Current statistics suggest that more than half of all unions among men and women ages 25–34 in Columbia, Cuba, Panama, Peru, and Venezuela are cohabiting unions (Esteve, Lesthaeghe, and López-Gay, 2012). At the same time, cohabitation’s enduring presence in Latin America does not dismiss Latin American countries from the changes in family structure occurring in other parts of the world. For instance, scholars are increasingly examining whether cohabitation patterns in Latin America are becoming more like those observed in industrialized nations like the United States, especially for those in more recent cohorts and for those living in urban areas (Heaton and Forste, 2007).
The importance of national context is especially evident in many Asian countries, where cultural taboos against nonmarital sexual unions have traditionally kept cohabitation prevalence low. While increases in cohabitation have been observed for some industrialized Asian countries, the overall prevalence of cohabitation in these countries remains remarkably low. For instance, in Japan, approximately 15% of adults report having ever been in a cohabiting relationship (Tsuya, 2006; Raymo, Iwasawa, and Bumpass, 2009). In Taiwan, the percentage of women ever having cohabited increased from 11% in the late 1990s to nearly 20% in 2004 (Lesthaeghe, 2011). One exception to the low prevalence of cohabitation across Asia is the Philippines, where cohabitation has been historically important for family formation (Xenos and Kabamalan, 2007). At present, some estimates suggest that approximately 60% of young adults living in the Philippines have ever cohabited (Williams, Kabamalan, and Ogena, 2007).
Family patterns continue to be in flux, even as we write these words. Thus, many scholars are kept busy tracking the changing contours of cohabitation: How long do cohabiting unions last? What percentage of cohabitations break up? What is the percentage of children born to cohabiting couples? How do these patterns vary by subgroups? These are exactly the type of issues we summarize in the previous section. In this section, we identify four research questions as being current and central to the field. The first concerns measurement, the second concerns children, the third concerns comparisons between cohabitation and marriage, and the fourth concerns the relationship between premarital cohabitation and marital instability.
Data and measures available for measuring cohabitation have improved dramatically in recent years. Until the 1990s, many national data sources, such as the US census, did not even directly measure cohabitation. Prior to direct measurement of cohabitation, scholars inferred national cohabitation rates from household composition data (Glick and Spanier, 1980). Such inferences often led to underestimated cohabitation rates, especially among households with children (Casper and Cohen, 2000).
National surveys began to gather more detailed information about individual experiences with cohabitation in the 1980s. The landmark survey in this regard is the 1987–1988 National Survey of Families and Households (NSFH) (Sweet, Bumpass, and Call, 1988). It obtained complete cohabitation histories from a sample of women and men of all ages and spawned a surge in research on cohabitation. Many national surveys now collect cohabitation histories for each individual interviewed, detailing not only how many cohabitations a person has experienced but also intricate details for each cohabitation reported, including start and end dates, indicators of relationship quality and conflict, and information on partner characteristics such as age, race/ethnicity, gender, etc. Such a potentially vast amount of data for each cohabiting union and the characteristics of individuals, especially in surveys that follow individuals over time, is a treasure trove for family scholars.
There have also recently been significant improvements in cohabitation measurement in existing surveys. The Current Population Survey (CPS), an ongoing survey that is part of the US Bureau of the Census, made major changes in 2007. The structure of the questioning prior to 2007 was such that only cohabitations involving the “reference” person were counted, effectively excluding all cohabitations involving members of the household other than the reference person. Qualitative research had shown that some cohabiting couples reside with others, including roommates, parents, or other relatives, and that this situation was more likely for less economically advantaged couples (Manning and Smock, 1995). In addition, qualitative research indicated that the term used by the CPS (“unmarried partner”) did not resonate or make sense to many people (Manning and Smock, 2005). The CPS thus changed their term to “boyfriend/girlfriend/partner.” There were also improvements that made it possible to identify the relationship of children in the household to their parent(s). These changes have been important. In 2008, 5% of all cohabiting unions in the CPS were between two household members that did not include the reference person. Moreover, changing the term “unmarried partner” to “boyfriend/girlfriend/partner” captured 20% more cohabitations (Kennedy and Fitch, 2012).
In addition to the NSFH and CPS, today various nationally representative survey data sources exist that measure cohabitation. The proliferation of such data has allowed family scholars to examine links between cohabitation and a seemingly endless supply of individual-level and relationship-level outcomes. Two data sources that include intricate cohabitation histories are the National Longitudinal Survey of Youth and the National Longitudinal Study of Adolescent Health (Bureau of Labor Statistics, 2002, 2005; Harris, 2009).
At the same time, several factors remain a challenge for accurate measurement of cohabitation. First, cohabitations are often reported inconsistently, depending upon who is asked. For example, among parents and children in cohabiting stepfamily households, only one-third of children whose mothers described their family as a cohabiting stepfamily household did the same. The remainder did not “count” their mothers’ cohabiting partner in defining their families (Brown and Manning, 2009). Second, individuals in couples sometimes provide differing responses about whether or not they are cohabiting (Waller and McLanahan, 2005). Third, many individuals are unable to specify an exact start date for their cohabitations, because many cohabiting unions form gradually and often unintentionally over a period of time (Manning and Smock, 2005). A fourth issue is the timing of data collection. For instance, when retrospective cohabitation histories are collected, the accuracy and consistency of the information gathered decreases as the length of time between the interview date and when the cohabitation occurred increases (Hayford and Morgan, 2008). In other words, there is a recall problem. An example is that individuals who had a child together are more likely to describe their relationship status at the time of the birth as “cohabiting” when asked directly after childbirth than when asked 5 years later (Knab and McLanahan, 2007).
Just as numerous studies emerged in the 1980s and 1990s examining the ramification of divorce for children, family scholars are intensively studying the implications of cohabitation for children. In general, research suggests that cohabitation negatively affects child well-being across multiple outcomes. Growing up in a cohabiting household compared to a married household is linked to poorer educational outcomes, more behavioral problems, and poorer health and well-being among children (Willetts and Maroules, 2004; Raley, Frisco, and Wildsmith, 2005; Brown, Lee, and Bulanda, 2006; Apel and Kaukinen, 2008; Bulanda and Manning, 2008; Schmeer, 2011).
More specifically, parental cohabitation is associated with earlier sexual initiation and teen pregnancy among adolescent girls (Bulanda and Manning, 2008). In turn, involvement in nonromantic sexual relationships during adolescence increases the likelihood of involvement in cohabitation during early adulthood, and cohabitations in early adulthood tend to be unstable (Raley, Crissey, and Muller, 2007; Schoen, Landale, and Daniels, 2007). Upon the dissolution of their own cohabiting unions, mothers experience declines in mental health, physical health, and income (Avellar and Smock, 2005; Meadows, McLanahan, and Brooks-Gunn, 2008). These patterns suggest a cycle such that family instability in one generation leads to family instability in the next generation, with negative effects on well-being for both parents and children. Because cohabitation is more prevalent among those who have fewer socioeconomic resources as children, this cycle has implications for the social transmission of well-being from one generation to the next.
The jury is still out, however, about whether it is cohabitation per se that “causes” poorer outcomes for children or whether the issue is really a matter of “selection.” That is, children experiencing parental cohabitation are disadvantaged in numerous ways compared to children born and raised by married parents. Economic status or social class is one of those ways. That is, some of the difference in the well-being of children in cohabiting compared to married couple households is explained by differences in the economic situation of the two types of families (Manning and Lamb, 2003; Brown, 2004; Artis, 2007). Consider the following: 47% of children in cohabiting households are living in poverty, compared to 11% of children in married couple households (Williams, 2012a). There is also a stark difference in the percentage of children covered by private health insurance. In cohabiting households, the percentage is 34% compared to 74% for children in married-couple households (Williams, 2012b).
The negative linkage between parental cohabitation and child well-being is also partly explained by family instability. Scholars have known for some time that family structure instability (e.g., divorce) typically does not enhance child well-being. Children born to cohabiting parents are five times more likely to see their parents separate than are children born to married parents (Osborne, Manning, and Smock, 2007). Family instability is associated with worse educational attainment and more behavioral problems during childhood (Cavanagh and Huston, 2006; Cavanagh, Schiller, and Riegle-Crumb, 2006). Moreover, studies suggest that the impact of family instability may endure well into adulthood (Amato et al., 2008; Ryan et al., 2009).
At the same time, there are some nuances that should not be overlooked when making generalizations about cohabitation not being good for children. Relationship quality between cohabiting parents is important and positively related to parental engagement with the child (Carlson et al., 2011). This, in turn, is associated with greater relationship stability among cohabiting parents (Cross Barnet, Cherlin, and Burton, 2011; McClain, 2011a). Further, should cohabiting parents dissolve their unions, high-quality relationships between parents significantly diminish negative effects of union instability on child well-being (Fomby and Osborne, 2010).
Certainly, the influence of cohabitation on child well-being is subject to national context. For example, in countries such as Sweden, where cultural norms and national policy treat cohabitation as more akin to marriage, cohabitation does not adversely affect child well-being (Kalenkoski, Ribar, and Stratton, 2005; Ono and Yeilding, 2009). In this sense, any detrimental effects of cohabitation on child well-being observed in the United States might be caused by government policy that favors marriage. Indeed, prominent US scholars have noted that more marriage-neutral policies may be necessary to foster the well-being of all children, regardless of family structure (Cherlin, 2004).
Much of what we know about cohabiting unions is drawn from comparisons to marital unions. Scholars have been doing this for quite some time. In fact, in Smock’s 2000 review chapter, she notes that scholars often “attempt to gauge the meaning of cohabitation by comparing it to marriage” (p. 12). This is understandable given that in order to know what something “is,” contrasts with similar categories (in this case living arrangements) can be tremendously useful. In the previous section, we explained that a number of researchers are examining how cohabitation may differ from marriage in terms of implications for child well-being. In this section, we demonstrate that comparisons of cohabitation to marriage characterize a chunk of research much larger than that focused exclusively on effects for children.
Conceptually, cohabitation has been referred to as an “incomplete institution” that lacks the stability found in marital unions (Nock, 1995, 2005), and much research has characterized cohabiting unions as having less commitment and lower relationship quality compared to marital unions (Treas and Giesen, 2000; Stanley, Whitton, and Markman, 2004; Marcussen, 2005;Hamplova and Le Bourdais, 2009; Yabiku and Gager, 2009). In fact, some studies suggest that adults in cohabiting unions experience more depressive symptoms compared to their married counterparts, in part because of such differences in relationship commitment and quality (Brown, 2000; Lamb, Lee, and DeMaris, 2003; Marcussen, 2005).
At the same time, some emerging research contradicts the idea that marriage is inherently “better” than cohabitation, and even suggests that cohabitation may be more beneficial than marriage for some aspects of well-being. Musick and Bumpass (2012) find that marriage is not necessarily more beneficial than cohabitation for physical or psychological well-being. Other studies report no differences in levels of relationship conflict or satisfaction when comparing married individuals and long-term cohabiting couples, with long term defined as 4 years (Willetts, 2006). In addition, a recent study shows that most married individuals who cohabited first are indistinguishable from those who did not in terms of trajectories of wealth accumulation during marriage (Vespa and Painter, 2011). Notably, too, cohabitation seems to be especially stable among older adults ages 50 and over. Older cohabitors report significantly higher levels of relationship quality and satisfaction than younger cohabitors, although they are less likely to have plans to marry their partners (King and Scott, 2005). In addition, older adults report similar levels of relationship quality and commitment across various indicators, regardless of whether they are married or cohabiting (Brown and Kawamura, 2010; Noel-Miller, 2011).
We view this research as important and would add that conclusions about whether marriage is “better” or “worse” across a number of domains may change as cohabitation becomes more normative (e.g., institutionalized) in society and marriage perhaps less so. Indeed, one demographer argues that marriage is becoming “deinstitutionalized” with roles and responsibilities of spouses becoming less clear over time (Cherlin, 2004).
For many years numerous studies found that premarital cohabitation raised the odds of divorce and this became the conventional wisdom: That cohabitation decreases marital stability (Smock, 2000). That is, a couple that cohabits before marriage is at greater risk of eventual divorce.
The reasons? It was posed that either cohabitation changed people in ways leading them to be more likely to end a marriage or that those more prone to divorce were more likely to cohabit in the first place. While not mutually exclusive, the first explanation was that there is something about the experience of cohabitation that raises the odds of divorce (e.g., people learn that relationships can be temporary and that there are alternatives to marriage). The latter is what is termed the “selection” explanation: People who cohabit prior to marriage differ from those who do not across a multitude of factors (e.g., values, attitudes, relationship skills), and these differences increase the likelihood of marital instability.
This conventional wisdom has been overturned with the release of studies using more recent data. One study finds that the positive relationship between premarital cohabitation and marital instability has diminished for more recent birth and marriage cohorts and that cohabitation is not selective of individuals with higher risks of divorce. There is even some evidence that cohabitation may help stabilize remarriages (Reinhold, 2010). Another study finds that for those marrying in the late 1990s and beyond, premarital cohabitation is not associated with marital instability for either women or men (Manning and Cohen, 2012; see also Copen et al., 2012). The declining impact of premarital cohabitation on marital stability observed in the United States is likely related to its increasingly normative presence. Indeed, research focusing on other regions of the world finds that as cohabitation becomes more normative within a country, the effects of premarital cohabitation on subsequent divorce risk decline considerably (Liefbroer and Dourleijn, 2006). This is consistent with research in both the United States and across Europe, demonstrating that peers’ direct experiences with cohabitation have influences on individuals’ attitudes toward cohabitation and decisions about whether to cohabit themselves (Nazio, 2008; Manning, Cohen, and Smock, 2011).
The past several decades have seen sweeping changes in the family. Cohabitation has now become a normative stage in the life course for most population subgroups: young adults, older adults, the well-off, and the working class. In closing, we make five observations about current knowledge on cohabitation.
The first is that while cohabitation has become a normative life event for nearly all population subgroups, we wish to underscore that its role varies considerably by social class. For economically advantaged adults, cohabitation is more of a step toward marriage; they are substantially more likely to see their cohabitations “end” in marriage versus dissolution. In contrast, individuals in the most disadvantaged groups are more likely to see their cohabitations end, rather than serve as a step to marriage, and are more likely to enter a number of cohabiting unions throughout the life course. This instability has ramifications for their children as well; children born to cohabiting parents are much more likely to see their marriages end than those born to married parents (Manning, Smock, and Majumdar, 2004).
A second observation is that the complex and dynamic nature of cohabitation and marriage – of family life – has important implications for how scholars craft their research questions. As noted earlier, much research attempts to understand cohabitation by comparing it to marriage. Going forward, it is clear that this is not necessarily the most useful way to study cohabitation, and families more broadly. If we continue to build our knowledge of cohabitation based on how it compares to marriage, we will miss some crucial insights. We encourage more research that is acknowledging other forms of coupledom such as Living Apart Together and even Living Together Apart; these may well provide considerable traction for understanding what may, or may not be, unique about cohabitation (Regnier-Loilier, Beaujouan, and Villeneuve-Gokalp, 2009; Cross Barnet, Cherlin, and Burton, 2011).
A third observation concerns what is missing from our review. We did not fully address scholarship on the reasons for the increase in cohabitation in the United States or elsewhere, beyond noting the importance of national context for the prevalence of cohabitation in any given country. At present, many scholars examining cohabitation in the United States are engaged in qualitative studies that focus on motives to cohabit among individuals or couples. These studies find that some young adults perceive living together as a chance to evaluate their potential success as marital partners, while others have no intentions to someday marry their partners (Edin, Kefalas, and Reed, 2004; Manning and Smock, 2005; Huang et al., 2011). For the latter, reasons for cohabiting often include finances, convenience, and housing needs (Sassler, 2004). This is consistent with research indicating that individuals from more disadvantaged backgrounds frequently cohabit due to financial necessity, while holding marriage as a union that can only be entered once one is established financially (Cherlin, 2004; Cherlin and Fomby, 2004; Gibson-Davis, Edin, and McLanahan, 2005; Smock, Manning, and Porter, 2005). Such research is consistent with the pattern of working-class couples moving from dating to cohabitation more quickly than middle-class couples while at the same time being less likely to be engaged than their more advantaged counterparts (Sassler and Miller, 2011a).
Our fourth observation stems from a review on cohabitation a dozen years ago undertaken by the second author (Smock, 2000). In that review, Smock noted that while early research and popular opinion was that cohabitation was more gender egalitarian than marriage, there was other research to suggest that this distinction was overstated. She concluded that cohabitation is gendered in much the same way that marriage continues to be, focusing on studies showing an unequal division of housework in both cohabiting and marital unions and that men’s economic characteristics appear to have more influence on whether cohabiting couples marry (Smock and Manning, 1997; Gupta, 1999). As she wrote, “That these findings emerge even among cohabitors, a subgroup less traditional in terms of attitudes toward gender roles, suggests the intractability of a division of labor that assigns breadwinning to men and homemaking to women” (Smock, 2000, p. 15).
Has research suggested a different conclusion more than a decade later? In large part, research conducted over the past dozen years has reinforced Smock’s conclusions. For instance, research using recent data suggests that women experience more severe economic consequences when exiting cohabiting relationships than do their male counterparts (Avellar and Smock, 2005). Further, qualitative research is suggestive of gendered views and dynamics of cohabitation. One study indicates that women perceive cohabitation more as a step in the marriage process than do men. In addition, women appear to be more concerned that cohabitation may delay marriage and reduce chances of a marriage proposal, whereas men express views that the downside of cohabitation is that it constrains their freedom (Huang et al., 2011). Another study finds that women in cohabiting relationships often feel that the man is in control of the progression of the relationship (Sassler and Miller, 2011b). At the same time, some recent studies suggest that although cohabiting women do more housework than cohabiting men, both decisions about housework and the division of housework are more egalitarian among cohabiting couples compared to married couples, with cohabiting women spending between 5 and 7 fewer hours on housework per week than their married counterparts (Batalova and Cohen, 2002; Baxter, 2005; Cunningham, 2005; Davis, Greenstein, and Gerteisen Marks, 2007).
Our fifth and final observation is that cohabitation has brought challenges to the study and our understanding of families. What we mean is that families today are characterized by considerable flux, complexity, and ambiguity, and probably more so for the less advantaged in our society. As we noted earlier in this chapter, a substantial proportion of adolescents and their mothers do not even agree about who is in their family. Families are increasingly likely to be characterized by part-time residents; these may include half-siblings or children of cohabiting partners. And even when a cohabiting relationship ends, ties between members who once lived together may remain, even between a child and his or her mother’s former cohabiting partner, to give just one example. Thus, what constitutes a “family” is changing and is increasingly changing over the lifetimes of members of families themselves. Surveys are trying to grapple with how to measure families, acknowledging this complexity and that families increasingly stretch over two or more households. This is a challenge for survey research, which has tended to sample households and many surveys have been based on the assumption that all family members live in the same household. It is likely to be an ongoing challenge. Arguably, cohabitation (and marriage and families) will continue to change as they have throughout history.
Rhiannon A. Kroeger acknowledges support by an F32 NICHD Ruth L. Kirschstein National Research Service Award (F32 HD072616), as well as an NICHD center grant to the Population Research Center at The University of Texas at Austin (R24 HD042849). The second author acknowledges support by an NICHD center grant to the Population Studies Center at the University of Michigan (R24 HD041028).