The Relationship Context of Premarital Serial Cohabitation

Abstract

Even though serial cohabitation is on the rise, it has not been integrated into recent family research. We analyze the National Survey of Family Growth (NSFG) cycle 6 to explore the relationship context of serial cohabitation for women throughout emerging adulthood (N=3,397). We provide a relationship context for serial cohabitation by examining the age at first cohabitation, duration of cohabiting unions, marital expectations and transitions, as well as premarital sexual histories. Furthermore, we examine the change in these relationship indicators across women’s birth cohorts. We find that serial cohabitors’ co-residential unions are about the same duration as single-instance cohabiting unions. Serial cohabitors start cohabiting younger, report lower marital expectations than single-instance cohabitors and a smaller proportion marry before age 30. Women who have more premarital sex partners have significantly greater odds of serial cohabiting. These findings indicate that women face increasingly complex relationship trajectories during emerging adulthood.

The median age of first marriage continues to increase. For women it is 26 years and for men, 28 (U.S Bureau of the Census 2009). Yet, this does not mean that young (“emerging”) adults are remaining single or living alone until marriage. Indeed, the courtship process now includes cohabitation as the modal pathway to marriage, a process that often begins with dating (or sexual relationships), transitions into cohabitation, and culminates with marriage (Cherlin 2009). As the age at first marriage rises, the opportunities to experience greater numbers of premarital sexual partnerships throughout emerging adulthood have also grown.

Prior research on cohabitation often focuses on one key cohabiting union and overlooks the advent of serial cohabitation, defined in this study as two or more cohabiting relationships. The central goal of this study is to describe the relationship context of serial cohabitation. We accomplish this by examining the age at first cohabitation, duration of co-residential unions, cohabitors’ marital expectations and transitions, as well as women’s premarital sexual partnerships. Given the increase in serial cohabitation over the late 1990s and early 2000s (Lichter, Turner and Sassler 2010), we consider the change in the relationship context across three birth cohorts of emerging adult women. We also pay special attention to women’s premarital sexual histories by including their number of premarital sex partners as a main predictor of serial cohabitation. In line with Nock (1995), this work recognizes the importance of premarital cohabitation and serial cohabitation’s burgeoning role in the marriage process during emerging adulthood. Our findings have implications for understanding a group of cohabitors who are most at risk of marital dissolution once they transition into marriage (e.g., Teachman 2003; Lichter and Qian 2008).

1.1 Emerging Adulthood: A Developmental Perspective

Using a developmental perspective, we explore the meaning of serial cohabitation throughout emerging adulthood. Because the age at first marriage is at an historical highpoint in the United States, there is increased time during late adolescence and early adulthood for a wide variety of premarital sexual and relationship experiences. The contemporary courtship process is now likely to include multiple sexual, dating and cohabiting relationships before entering marriage (Cherlin 2009). The term “emerging adulthood” has been coined to encapsulate the notion of this complex and less rigid pathway into adulthood, which results from shifts in the nature of educational opportunities, leaving the parental home, career development, early parenthood, and delayed marriage (Arnett 2004). Emerging adulthood is a time of both adolescent and adult commitments and responsibilities, and is characterized by instability, self-focus and exploration (Arnett 2004; Erikson 1968).

One important aspect of emerging adulthood is social learning through romantic unions to understand what types of relationships work best. The relationship quest in emerging adulthood shifts from the more adolescent question of “Who would I enjoy being with now?” to “Who am I?” and “Who do I want to be with through life?” (Arnett 2004). There are opportunities for exploring many relationships as there is little normative pressure to settle down or for these relationships to operate inside the purview of close parental attention. Although the age range defining emerging adulthood varies across prior studies, it typically ranges from late adolescence through the twenties (Arnett 2004). Thus, in this paper, we examine women’s cohabitation experiences between the ages of 16–30. This investigation furthers the extant literature on cohabitation by examining the fluid nature of union formation throughout emerging adulthood. We do not simply limit our analysis to one premarital cohabiting or sexual relationship during a time when many women are starting and ending several relationships. Our goal is thus to contextualize women’s multiple cohabitation transitions within their premarital sexual histories.

1.2 Cohabitation

Nock (1995) documents the increasingly important role of cohabitation in the marriage process by examining the substantive differences between cohabiting and marital unions. His work reveals that cohabitation and marriage differ on a host of relationship quality criteria, with cohabitors reporting significantly lower relationship quality than married couples. The author is especially concerned about understanding cohabitation as a step toward marriage and the implications of premarital cohabitation on the quality and stability of marital unions. This same concern is echoed in Nock (2009) where he argues that marriage generates certain benefits that cohabitation cannot and further, that while marriage signals a number of desirable characteristics that a husband/wife may possess -- maturity, independence, productivity, and trustworthiness -- cohabitation does not signal to society that one possesses these desirable traits. Nock argues that as the rate of marriage continues to decline, the social power of such signals may increase. In keeping with Nock’s past research focus, this study addresses serial cohabitation’s role in the marriage formation process by concentrating on the serial cohabitation of emerging adult women prior to marrying for the first time and describing the meaning of serial cohabitation. To this end, this investigation focuses on women’s complex premarital relationship careers throughout emerging adulthood and discusses the implications of premarital serial cohabitation for subsequent marriage and divorce.

Most prior studies of cohabitation have focused on one cohabiting union and have typically overlooked serial cohabitation, cohabiting with two or more partners. Past work on serial cohabitation is sparse (exceptions Lichter et al. 2010; Lichter and Qian 2008; Teachman 2003; Teachman and Polonko 1990). Furthermore, past work that does focus on serial cohabitation is often limited to the experiences of ever-married women, thus excluding those women who delay or forgo marriage (exception Lichter et al. 2010). This gap in the extant literature is problematic considering that serial cohabitation is increasing. Indeed, in 2002 about 25% of women report having serial cohabitation experience (Lichter et al. 2010). Lichter et al.’s (2010) research focuses on the increase of serial cohabitation and lays important groundwork by documenting the predictors of serial cohabitation. Our investigation complements their findings by presenting the relationship context of serial cohabitation, thereby giving scholars more insight into the nature of these unions and how they relate to women’s other sexual relationships. Lichter et al. (2010) investigate serial cohabitation with great breadth by studying both never- and ever-married women, considering serial cohabitation before and after first marriage, examining women of a wide age range (19 through 44), and analyzing two different survey cycles of the NSFG (1995 and 2002). Our work extends their findings with a more detailed analysis of recent birth cohorts of women (1958–1972), whose cohabitation experiences occurred within a narrower age range (16–30), and prior to first marriage. In doing so, we are able to consider serial cohabitation within emerging adulthood and focus on how it links to the marriage process.

1.3 Premarital Sexual Partnerships

Women’s sexual histories are considered in this study, not only as a descriptive measure designed to examine relationship context, but also as the focal predictor of serial cohabitation. Sexual experiences, especially earlier onset of sexuality in adolescence, are associated with greater odds of cohabitation in adulthood (Lichter et al. 2010; Raley, Crissey and Muller 2007). Therefore, we expect to find that premarital sexual partnerships in emerging adulthood serve as an important predictor of serial cohabitation. Prior research suggests that parallel to the rise in cohabitation since the 1980s is the de-linking of sex from marriage. In 2002, less than ten percent of married women, under age 35, waited until marriage to have sex (Chandra et al., 2005). According to evidence from the 1995 NSFG, Teachman (2003) reports that about three-quarters of married women had intercourse with someone besides their husband prior to marriage. Women and men also have more sexual partners in their lifetime than in decades past. For example, the 1995 NSFG indicates that women between the ages 30–44 had an average of 3 sexual partners in their life time. According to analysis of the 2002 cycle, this average rose to 4 sexual partners (Mosher et al. 2005). Considering these prior findings, we examine the change in the number of premarital sexual partnership across three birth cohorts of women during emerging adulthood.

Although the number of sex partners a man or women will have prior to marriage has increased, researchers rarely include this measure in investigations of union formation. This is ironic given that dating and sexual relationships are direct precursors to cohabitation and marriage. Research has focused on cohabiting sexual relationships, but relatively little attention has been paid to how non-residential sexual relationships or dating relationships influence transitions into cohabitation or marriage.

One arena that has been studied is how premarital sexual relationships within and outside of cohabitation influence the stability of marriages. Teachman (2003) uses the 1995 NSFG to examine women who were first married between 1970 and 1995. He finds that women who had no premarital sex or only had premarital sex with their future spouse shared similar odds of marital dissolution. Women who had sex with someone besides their husband had a higher risk of marital dissolution. He further investigates how all sexual unions (cohabitation and other sexual partnerships) influence marital instability and finds that women who only cohabited with their spouse or only had premarital sex with their spouse had similar odds of marital dissolution as women who did not cohabit. Women who serially cohabited and/or had premarital sex with someone besides their husband had higher odds of marital dissolution than women who never cohabited. Teachman’s findings suggest that both sexual history and cohabitation history influence marital stability. To better understand romantic relationship dynamics today, scholars should include sexual histories as well as cohabitation histories. Thus, in our work, we examine the interplay between the women’s number of premarital sex partners and cohabitation patterns, as well as determine whether and how women’s premarital sexual histories are linked to serial cohabitation.

1.4 Relationship Context

Prior work suggests that cohabiting unions are typically short-lived (averaging about two years in duration as cohabitations) (Kennedy and Bumpass 2008). Indeed, only about one third of women’s first cohabitations last as long as 3 years before dissolution or marriage (Goodwin, Mosher and Chandra 2010). However, no prior work has compared the duration of serial cohabitors’ co-residential unions to single-instance cohabitation. This is problematic considering that union duration is one dimension of assessing a couples’ commitment to their relationship. Since, by definition, serial cohabitors have more than one co-residential union, we expect that the duration of their cohabiting unions may be shorter than their single-instance cohabiting counterparts.

Prior research on marriage expectations among cohabitors (Manning and Smock 2002; Brown 2000a) does not acknowledge the role of serial cohabitation. However, previous research indicates that when serial cohabitors do marry, they may be at greater risk of subsequent marital instability than single-instance cohabitors (DeMaris and MacDonald 1993; Teachman and Polonko 1990; Teachman 2003; Lichter and Qian 2008). In other words, people who cohabit with two or more partners have a higher likelihood of experiencing a divorce than those who only cohabit with one person before marriage. In fact, serial cohabitors have lower quality and less stable cohabiting relationships than single-instance cohabitors (Stets 1993; Brown 2000b). Brown (2000b) also reports higher levels of depression among serial cohabitors than single-instance cohabitors. These findings suggest that serial cohabitors may be more willing to end relationships that they do not find emotionally satisfying, and that they may also have certain sociodemographic characteristics associated with higher rates of marital dissolution. Thus, we expect that serial cohabitors will report lower marital expectations than their single-instance cohabiting counterparts.

Prior work on the marriage transitions of cohabitors has considered marital history, but most studies have not examined how cohabitation history influences marriage entry (Manning and Smock 2002; Sassler and Schoen 1999). Analysis of women who generally first married in the 1980s and early 1990s indicates that serial cohabitors are less likely to marry than single-instance cohabitors (Lichter and Qian 2008). Indeed serial cohabitors, even when controlling for education and economic characteristics, remain less likely to marry than their single-instance cohabiting counterparts (Lichter and Qian 2008). Thus, using recent data, we expect that emerging adult women with more cohabitation experience (i.e. serial cohabitors), will be less likely to have married by age 30 than single-instance cohabitors.

1.5 Control Variables

Prior work indicates that serial cohabitation has increased (Lichter et al. 2010), thus we expect that women born in more recent birth cohorts will have higher odds of serial cohabiting than women born in older birth cohorts. Furthermore, several sociodemographic characteristics have been found to be associated with union formation: specifically the odds of ever cohabiting (e.g Kamp Dush, Cohan and Amato 2003; Kennedy and Bumpass 2008), as well as serial cohabitation (Lichter and Qian 2008; Lichter et al. 2010). “At risk” populations, such as women with low educational attainment or whose mothers had low educational attainment, are over-represented among serial cohabitors (Lichter and Qian 2008; Lichter et al. 2010). Women whose parents divorced are more likely to ever cohabit (Kamp Dush et al. 2003) and women who did not grow up in a two-parent household are also more likely to serially cohabit (Lichter et al. 2010). Blacks tend to serially cohabit more often than whites and Hispanics and foreign-born women are underrepresented among serial cohabitors (Lichter et al. 2010). In light of these trends, we expect to find that respondents and mothers who have not earned a high school degree and women who did not grow up in a two parent household are more likely to serial cohabit. We also expect that blacks will be more likely to serially cohabit than their white and Hispanic counterparts and foreign-born Hispanics may be less likely to serially cohabit than whites.

2. Current Investigation

The primary goal of this paper is to examine the relationship context of serial cohabitation among women in emerging adulthood. In doing so, this study moves beyond prior research in three key ways. First, previous research has typically only included cohabitation history as one of many independent covariates to investigate its association with relationship quality and the odds of marital disruption among cohabiting couples or ever-cohabiting couples. In contrast, this investigation concentrates on the pathway to marriage by specifically examining serial cohabitation during emerging adulthood, and prior to first marriage, among never- and ever-married women. Second, this investigation does not merely revisit the correlates of serial cohabitation, but examines its very nature by documenting the relationship context in which such multiple unions occur and assessing how similar or distinct serial cohabitors are from their single-instance counterparts. Third, we include women’s number of premarital sex partners as the main predictor of serial cohabitation to help us understand how serial cohabitation fits in with the broader set of sexual experiences of emerging adult women. Past research indicates that this measure is positively correlated with women’s likelihood of marital disruption (Teachman 2003) and is most likely another significant determinant of women’s serial cohabitation. However, it typically has not been included in past research on serial cohabitation (e.g., DeMaris and MacDonald 1993; Teachman and Polonko 1990; Lichter and Qian 2008; Lichter et al. 2010). Taken together, this work will move forward our understanding of cohabiting relationships during emerging adulthood and build on Nock (1995) by exploring serial cohabitation’s growing role in the marriage process.

3. Materials and Methods

We use cycle 6 (2002) of the National Survey of Family Growth (NSFG) which is a national probability sample, representing the household population of the United States, ages 15–44. Interviews took place between mid-March of 2002 and the end of February in 2003. In-person interviews were completed with 7,643 women. The interviews were voluntary and confidential and the response rate was 80 percent for females. The NSFG deliberately over-sampled smaller groups in the population, thus these analyses are weighted. These data are appropriate for this investigation because they contain detailed marriage, cohabitation, and sex histories for women allowing us to analyze the number of women’s cohabitating unions and non-cohabiting sexual partnerships. The NSFG does not provide parallel cohabitation and sexual histories for men; therefore, due to these data limitations, we restrict analyses to women.

The analyses focus on women’s cohabitation behavior while unmarried, between the ages of 16–30 (during emerging adulthood) and are based on a sample of 3,397 women. We restrict our analyses to women who were born between 1958 and 1972 and compare the premarital cohabitation experiences of three birth cohorts of women: those who were born between 1958–1962, 1963–1967, and 1968–1972. Among the ever-married women in this sample, we count the number of cohabiting unions which occurred between the ages of 16–30, prior to date of first marriage. Among never-married women we count the number of cohabiting unions which occurred between the ages of 16–30, prior to the date of interview. Therefore, we examine only cohabitation experiences that occurred prior to first marriage. Initially we restricted our sample to ever-married and never-married women’s cohabitation experiences between the ages of 18–30; however a substantial proportion of women with cohabitation experience began cohabiting at ages 16 and 17 (10%). This sample is limited to women who provided valid data on the start dates of their cohabiting unions as well as the date of their first marriage (when appropriate), thus we eliminated 97 respondents from our initial sample.

3.1 Dependent Variable

The NSFG defines cohabitation as living together while not being married and “having a sexual relationship while sharing the same usual address.” Ever-married women were asked whether they had “lived with him [their first husband] before marriage.” Both never- and ever-married women were asked whether they had “lived with any other/a) man.” An additional question was posed to all women to determine whether they were currently cohabiting at the time of the interview. Women’s current cohabitations were also included in the estimation of the number of cohabitations before first marriage.

The dependent variable is the number of cohabiting unions that occurred between the ages of 16–30. For ever-married women, we count the number of cohabiting unions that occurred prior to date of first marriage. For never-married women, we count the number of cohabiting unions that occurred prior to the date of interview or age 30. The dependent variable is recoded into three categories: zero partners (no cohabitation), one partner (single-instance cohabitation), and two or more partners (serial cohabitation). Over half (53%) of the sample did not cohabit between the ages of 16–30, over one-third (38%) cohabited once, and about 9% cohabited twice or more.

3.2 Relationship Context Indicators

Duration, Marriage Expectations, and Transitions

Three indicators of the relationship context are examined: union duration (in months), marital expectations, and marriage transitions. We show the differences in the percent distributions of these three measures for single-instance vs. serial cohabitors across three birth cohorts. The average sum of months spent cohabiting was calculated by summing the duration of cohabiting unions. All cohabiting relationships that are included in this measure began between the ages of 16–30. Those cohabiting relationships that endured past age 30 were measured up until the date of dissolution. For example, a cohabiting union that began when the respondent was 29 years of age and ended when she was 31 is counted as being two years in length, not one year in length. The percentage of time spent in cohabiting unions was calculated by dividing the sum of months spent cohabiting between the ages of 16–30 by the number of years a woman was unmarried between the ages of 16–30. Women’s marital expectations at the start of their cohabitation were measured by a binary response yes or no question: “At the time you began living together, were you and your partner engaged to be married or have definite plans to get married?” The proportion of women who first married any man, not necessarily their cohabiting partner, before the age of 30 was calculated by using the survey question asking the respondent to report the date she married her first husband. Finally, the average age of first cohabitation was measured using a survey question asking the respondent the month and year they first began cohabiting. Their reported century month of first cohabitation was then converted into an age. A similar strategy was used to ascertain respondents’ age at first marriage.

3.3 Independent Variables

Number of Non-Cohabiting Sex Partners

The number of premarital sex partners women had outside of cohabitation is used as a descriptive measure and is presented in a similar fashion as the other relationship context variables stated above. It is also included as the core independent variable in these analyses. The NSFG asks respondents “Including your (former) husband, how many male sexual partners did you have before you got married (the first time)? Thus, for ever-married women, we were able to create a non-cohabiting sex partner measure by subtracting the number of premarital cohabiting partners from the number of premarital sex partners. Forty-four women in the sample did not report a valid number of sex partners before their first marriage. We replaced these women’s missing data with the weighted average number of premarital sex partners for married women in the sample. A small number of respondents (56) stated that the number of cohabiting partners was greater than their number of sex partners, suggesting that they did not have sex with all of their cohabiting partners. These respondents were recoded as having zero non-cohabiting sex partners. Sensitivity analyses indicate that the results are similar when these respondents are excluded from the analyses.

The number of non-cohabiting sexual partners is more difficult to measure for never-married women because the NSFG does not include a direct question akin to the number of sex partners item prior to first marriage. In light of this limitation, we estimate the number of non-cohabiting sex partners for never-married women prior to age 30. Women who had no sex partners prior to age 30 are coded as zero (n=66). For the remaining 685 never-married women, we generated three estimates, but only include the most conservative estimate in these analyses. We assign to never-married respondents, the weighted average of life time sex partners for never-married, sexually active, 30 year old women. The inclusion of this measure most likely underestimates the number of non-cohabiting sex partners women had prior to age 30. Sensitivity analyses yields similar results at both the zero-order and multivariate level regardless of which of the three estimates is included in the model.

Control Variables

Prior research has examined how sociodemographic characteristics are related to union formation. One of the goals of this paper is to document the change in the occurrence and relationship context of serial cohabitation among emerging adult women, thus the analyses are confined to women’s cohabitation experiences while they were 16–30 years of age. It includes a women’s birth cohort measure: those who were born between 1958–1962, 1963–1967, and 1968–1972. The NSFG does not include dates of educational attainment, so we cannot determine when higher education was pursued. Thus, we only include educational indicators that most likely occurred before cohabitation began and respondent’s education is collapsed into two categories: less than 12 years and 12 years or more of education. Respondent’s mother’s education is measured by four categories: less than 12 years, 12 years, 13 to 15 years, and 16 or more years of education. Twenty-four women reported having no mother figure in their life. These women were recoded into the modal category of 12 years of education. Family structure during childhood is measured as a binary response variable, with respondents falling into one of two categories: grew up in a two-parent household during childhood or respondent grew up in a non two-parent household. Finally, women’s race/ethnicity was recoded into five response categories: white, black, native-born Hispanic, foreign-born Hispanic, and other.

3.4 Analytic approach

The analytic method for this current investigation is multinomial logistic regression and it is used to examine the likelihood of women entering zero (no cohabiting unions), one (single-instance cohabitation), or two or more cohabiting relationships (serial cohabitation). This method is appropriate for a categorical dependent variable with more than two response categories (DeMaris 1992). We are most interested in the comparison of serial cohabitors with single-instance cohabitors. Thus, when we examine the odds of serially cohabiting, the reference category is single-instance cohabitation. Our focal independent variable is women’s number of premarital non-cohabiting sex partners. We first estimate zero-order models for this focal independent variable, as well as for each control variable. Next, the premarital sex partners measure as well as all other covariates are included in multivariate multinomial logistic models. The odds ratios presented are exponentiated coefficients. Therefore, an odds ratio of less than one can be interpreted as a negative relationship between the independent variables and the dependent variable and an odds ratio greater than one suggests a positive relationship.

4.1. Descriptive Results

Table 1 shows the cohabitation experiences for all women. On average, women between the ages of 16–30 had .59 cohabiting partners. The average number of cohabiting partners has increased across birth cohorts, from .44 among women who were born between 1958 and 1962 to .73 one decade later (1968 to 1972). In our sample, about half (53%) of women did not cohabit, over one-third (38%) cohabited once, and about 9% cohabited twice or more. We find evidence that serial cohabitation increased within a ten year period. Five percent of the 1958–1962 birth cohort serially cohabited, which increased to 14% among the 1968–1972 birth cohort. Furthermore, among women who have cohabited, the proportion serially cohabiting increased from 14.5% (1958–1962) to about one quarter (25%) one decade later (1968–1972).

Table 1.

Women's Cohabitation Experience during Emerging Adulthood (N= 3,397)

	Birth Cohorts
	Total	1958–1962	1963–1967	1968–1972
All Women Ages 16–30 yrs
Average number of cohabitations while not married	0.59	0.44	0.59	0.73
Serial Cohabitation Measure
Zero	53.13	62.77	52.26	41.59
One	37.63	31.84	39.37	44.52
Two or more	9.24	5.39	8.37	13.89
Ever-Cohabiting Women
Serial Cohabitation Measure
One	80.28	85.53	82.46	74.96
Two or more	19.72	14.47	17.54	25.04

Note: All values are weighted.

Source: 2002 National Survey of Family Growth

Table 2 presents age at first cohabitation and cohabitation duration for serial and single-instance cohabitors. Women who serially cohabit start cohabiting at younger ages (age 20) then women who only cohabit once (age 22). Among recent birth cohorts, the gap in the age at first cohabitation has widened. On average, single-instance cohabitors spend 32 months cohabiting with their only partner, and serial cohabitors spend roughly the same average number of months within their first as in their second cohabiting relationship. As a result, serial cohabitors spend more of their early adulthood within cohabiting relationships than single-instance cohabitors. Serial cohabitors spend almost half of their unmarried young adult years (42%) cohabiting; as opposed to single-instance cohabitors who only spend about 25% of early singlehood within cohabiting unions. There has been little change in the amount of time serial and single-instance cohabitors spend in premarital cohabiting unions across birth cohorts (results not shown).

Table 2.

Women's Cohabitation Duration and age at first cohabitation during Emerging Adulthood (N= 3,397)

	Birth Cohorts
	Total	1958–1962	1963–1967	1968–1972
Age at First Cohabitation
All Single-Instance Cohabitors
Average age at 1st cohabitation	22.25	22.22	21.85	22.22
All Serial Cohabitors
Average age at 1st cohabitation	20.05	20.26	20.33	19.81
Cohabitation Duration
All Women Ages 16–30 yrs
Average duration of all cohabiting relationships while not married	31.26	30.94	33.66	29.44
All Single-Instance Cohabitors
Average sum of months spent in 1st cohabitation	31.68	31.17	34.35	29.57
All Serial Cohabitors
Average sum of months spent in 1st cohabitation	28.24	25.82	29.51	28.42
Average sum of months spent in 2nd cohabitation	29.12	33.14	27.47	28.57

Note: All values are weighted.

Source: 2002 National Survey of Family Growth

Table 3 presents the marital expectations, percent married, and age at first marriage for serial and single-instance cohabitors. While over half (53%) of single-instance cohabitors expected to marry that partner, a much lower proportion of serial cohabitors expected to marry their first (20%) and second (32%) cohabiting partner. In addition, serial cohabitors have experienced a decline in the percent who expected to marry their second partner. Furthermore, on average serial cohabitors marry at older ages (27 vs. 24 years old) than single-instance cohabitors. Even though the vast majority of single-instance cohabitors and serial cohabitors marry, a lower percentage of serial than single-instance cohabitors were married by age 30 (86% vs. 91.5%). The gap in age at marriage and proportion marriage has declined, with recent birth cohorts of serial and single-instance cohabitors experiencing similar patterns of marriage. Taken together, serial cohabitors start forming unions at earlier ages, but each serial cohabiting union is on average just as long as single-instance cohabiting unions. However, serial cohabiting relationships start with a much different purpose and are less likely to become marriages.

Table 3.

Women's Marital Expectations, Percent Married by Age 30, and Age at First Marriage Throughout Emerging Adulthood (N= 3,397)

	Birth Cohorts
	Total	1958–1962	1963–1967	1968–1972
Marital Expectations
All Single-Instance Cohabitors
Percent expect to marry 1st cohabiting partner	52.75	50.64	53.83	53.32
All Serial Cohabitors
Percent expect to marry 1st cohabiting partner	20.84	18.73	22.65	20.57
Percent expect to marry 2nd cohabiting partner	32.20	38.72	35.12	27.96
Proportion Married by Age 30
All Single-Instance Cohabitors
Proportion that married by age 30	91.49	88.09	92.14	93.45
All Serial Cohabitors
Proportion that married by age 30	85.79	80.13	79.36	91.79
Age at First Marriage
All Single-Instance Cohabitors
Average age at 1st marriage	23.94	24.21	23.6	24.06
All Serial Cohabitors
Average age at 1st marriage	27.2	28.09	27.91	25.94

Note: All values are weighted.

Source: 2002 National Survey of Family Growth

Table 4 incorporates the sexual experiences of women. On average, women between the ages of 16–30 had about 4.1 sex partners and 3.5 non-cohabiting sex partners. The average number of sex partners has increased over time. Women lived with only 20% of their sexual partners (results not shown). The proportion of sexual partners that women lived with has risen from 16% among women who were born between 1958 to 1962 to 23% among the 1968–1972 birth cohort (results not shown). Single-instance cohabitors have a lower average number of sex partners and non-cohabiting sex partners than serial cohabitors. Single-instance cohabitors had on average 4 non-cohabiting sex partners compared to 7 partners among serial cohabitors. While the average number of sex partners has increased among single-instance cohabitors, the average number of sex partners and non-cohabiting sex partners has decreased for serial cohabitors. Still, among all birth cohorts, single-instance cohabitors have a lower average number of sex partners and non-cohabiting sex partners than serial cohabitors.

Table 4.

Women's Sexual Experiences during Emerging Adulthood (N= 3,397)

	Birth Cohorts
	Total	1958–1962	1963–1967	1968–1972
All Women Ages 16–30 yrs
Average number of sex partners	4.10	3.54	4.10	4.65
Average number of non-cohabiting sex partners	3.53	3.12	3.53	3.93
Never Cohabitating Women
Average number of sex partners	2.58	2.26	2.64	2.93
Single-Instance Cohabitors
Average number of sex partners	4.99	4.56	4.85	5.43
Average number of non-cohabiting sex partners	4.02	3.64	3.88	4.45
Serial Cohabitors
Average number of sex partners	9.26	12.30	9.71	7.84
Average number of non-cohabiting sex partners	7.01	10.12	7.37	5.62

Note: All values are weighted.

Source: 2002 National Survey of Family Growth

4.2 Cohabitation during Emerging Adulthood

Table 5 presents the multivariate results predicting serial cohabitation. The zero-order and multivariate findings mirror one another so we focus here on the multivariate results. The contrast group for the first set of findings is women who never cohabited during emerging adulthood, thus the first column shows the odds of single-instance cohabitation versus never cohabiting. The contrast group for the next set of findings is single-instance cohabitation; thus, these results present the odds of serial cohabitation versus single-instance cohabitation. The number of non-cohabiting sex partners is associated with greater odds of cohabitation in both the zero-order and full models. In the full model, for every additional non-cohabiting sex partner, the odds of a woman single-instance cohabiting versus never cohabiting increased by 7%. Each additional non-cohabiting sex partner increased the odds of serial cohabitation by 6%. The number of sex partners is associated with greater odds of serial cohabitation than single-instance cohabitation in both the zero-order and full model.

Table 5.

Zero Order and Multivariate Logistic Regression of Women's Number of Cohabiting Relationships during Emerging Adulthood (N=3,397)

	Zero Order				Full Model
Variable	Compared to Never Cohab Single-Instance		Compared to Single-Instance Serial		Compared to Never Cohab Single-Instance		Compared to Single-Instance Serial
	Odds Ratio	SE	Odds Ratio	SE	Odds Ratio	SE	Odds Ratio	SE
Non-Cohabiting Sex Partners	1.08 ***	0.01	1.06 ***	0.01	1.07 ***	0.01	1.06 ***	0.01
Birth Cohort (reference = 1958–1962)
1963–1967	1.49 **	0.18	1.26	0.24	1.42 **	0.08	1.21	0.23
1968–1972	1.84 ***	0.22	1.97 ***	0.32	1.76 ***	0.19	1.80 **	0.30
Education (reference = 12 years or more)
<12 years	0.96	0.12	0.98	0.23	0.84	0.12	1.04	0.28
Mother's Education (reference = 12 years)
<12 years	1.09	0.15	0.71	0.15	1.27	0.19	0.70	0.16
13 to 15 years	1.39 **	0.19	0.69 *	0.14	1.31 *	0.18	0.63	0.14
16 or more years	1.67 ***	0.27	0.78	0.17	1.60 **	0.26	0.72	0.16
Family Type During Childhood (reference = two-parent household)
Non two-parent houseold	1.69 ***	0.18	1.96 ***	0.34	1.55 ***	0.17	1.82 **	0.33
Race/ethnicity (reference = white)
Black	1.16	0.15	1.32	0.26	1.04	0.15	1.22	0.28
Native-Born Hispanic	0.95	0.13	1.29	0.29	0.89	0.14	1.34	0.33
Foreign-Born Hispanic	0.57 ***	0.08	0.75	0.19	0.57 **	0.10	1.03	0.29
Other	0.64	0.16	0.83	0.29	0.61	0.15	0.94	0.34

^*p<.05;

^**p<.01;

^***p<.001

Note: Binary logistic regression analysis is weighted.

Source: 2002 National Survey of Family Growth

Several of the control variables are related to serial cohabitation. Birth cohort is related to greater odds of cohabiting. Women in the more recent birth cohort (1968–1972) had 80% greater odds of serial cohabitation than women of comparable ages ten years earlier (1958–1992). Graduation from high school is not associated with single-instance cohabitation or serial cohabitation. Mother’s education is significantly related to the odds of cohabitation. Women whose mothers have some college experience or a college degree have higher odds of single-instance and serial cohabiting than women who have a mother with a high school degree. Family structure during childhood is also significantly associated with single-instance and serial cohabitation. Women who did not grow up in a two parent household have 82% higher odds of serially than single-instance cohabiting than their counterparts raised in intact households. Foreign-born Hispanic women have 43% lower odds than whites of single-instance cohabiting. Whites, blacks and native-born Hispanics share similar odds of single-instance and serial cohabitation.

5. Discussion

The proportion of emerging adult women who serial cohabit nearly tripled from 5% among the 1958–1962 birth cohort to 14% among the 1968–1972 birth cohort. The average number of cohabiting partners has increased and nearly one-quarter (23%) of women who cohabited during emerging adulthood cohabited with more than one partner (serially cohabited). Our results support Lichter et al.’s (2010) findings of an increase in serial cohabitation. Given the rising popularity of premarital cohabitation, this means that increasing numbers of women are experiencing serial cohabitation in emerging adulthood and we expect these trends to continue.

The relationship context for serial cohabitors differs somewhat from single-instance cohabitation. Like their single-instance counterparts, many serial cohabitors are also on the road to marriage, albeit theirs is a somewhat long and winding one. The analyses of women’s experiences throughout emerging adulthood indicate that serial cohabitors start cohabiting at younger ages, but spend about as much time in each of their cohabiting unions as single-instance cohabitors. Serial cohabitors are much less likely to expect to marry their cohabiting partners at the outset of cohabitation. Half of single-instance cohabitors expect to marry their partner in contrast to one-fifth or one-third of serial cohabitors’ first or second cohabiting union. Furthermore, serial cohabitors marry at older ages and are less likely to marry before the age of 30. Taken together, these findings imply that serial cohabiting unions are not short-term or fleeting relationships, but often do not start out with marriage as the intention. Serial cohabitors’ lower marital expectations and comparably lower marriage transitions (vs. single-instance cohabitors) suggest that they may be aware of the structural barriers that exist to marriage or they may be cohabiting for reasons other than eventual matrimony. Further analysis of the marital transitions of serial cohabitations will reveal additional details about the odds of marrying a particular cohabiting partner and attention should be paid to cohabitation histories in studies of union formation.

Differences according to cohabitation experience in the age at marriage and eventual marriage are diminishing across birth cohorts, while the gap in the average age at first cohabitation is widening somewhat. Recent cohorts of serial cohabitors are marrying at earlier ages compared to older cohorts, but they begin their cohabitation careers earlier over time. These trends suggest that for recent cohorts of serial cohabitors, the period of emerging adulthood may be lengthening, not because of a delay in marriage, but because of the earlier onset of cohabitation.

Unlike most prior work on young adults, this study incorporates both sexual relationships that are coresidential (cohabitation) and those that are not (non-cohabiting sex partners). We find that the majority (80%) of sexual relationships that women have with men are not taking place within the context of co-residential unions. Indeed, a growing proportion of women are experiencing a higher number of premarital relationships. Cherlin (2009) coins the phrase the “marriage-go-round” to illustrate the plethora of unions and union types an adult may enter in their lifetime: dating, premarital cohabitation, divorce, postmarital cohabitation, and remarriage. The term “relationship go-round” may more accurately describe emerging adults’ contemporary courtship process, which involves entering and exiting multiple sexual and cohabiting relationships prior to first marriage.

Women who have cohabitation experience prior to marriage, on average, have a greater number of non-cohabiting sex partners than those who did not cohabit before marriage. Women who serial cohabit are more likely to have a greater number of non–cohabiting sex partners than single-instance cohabitors. Emerging adult women with more non-cohabiting sex partners prior to first marriage are more likely to experience serial cohabitation. One source of the greater number of serial cohabitors’ sexual partners may be their earlier age at first sex (Lichter et al. 2010). Serial cohabitors’ higher number of sexual and cohabiting partners suggests that they have a longer history of dissolved relationships -- i.e., sexual, (most likely dating) and cohabiting relationships – that they bring to their cohabiting and later marital relationships. This relationship experience may affect the quality and stability of their cohabiting relationship and the odds of marrying their cohabiting partners. Consistent with Teachman (2003), who found that both sexual and cohabiting partnerships significantly predicted the odds of marital dissolution, our findings suggest that studies of union formation and stability should consider the full range of sexual experiences in early adulthood.

As serial cohabitation becomes a more common occurrence among emerging adults, it is likely that serial cohabitors may become a less select group. The results suggest that the difference between serial and single-instance cohabitors regarding their number of non-cohabiting sex partners and the percent who married before age 30 narrows as serial cohabitation becomes more popular. While serial cohabitors, on average, have sex with almost 7 more partners outside of cohabitation than single-instance cohabitors born between 1958 and 1962, this substantial difference reduces to about 1 sex partner among the 1968–1972 birth cohort. Similarly, the difference in the proportion of women who marry by age 30 decreases from 8% (1958–1962) to less than 2% among women born ten years later.

The limitations to this study include the measure of sexual histories among never-married women as well as the recall of cohabitation experience. We rely on estimates of the number of sexual experiences among never-married women because we could not discern the number that had occurred by a specific age. We obtain similar findings with the lifetime estimate and an estimate that assigned the mean number of partners. Further inquiry into the timing of sexual partnerships during early adulthood would contribute to analysis of the sexual lives of emerging adults. Some research documents bias in the recall of cohabitation (Hayford and Morgan 2008) and may be an issue in this paper which relies on retrospective reports of cohabitation. This would suggest that we have under-estimated serial cohabitation and the levels may in fact be greater than reported. However, longitudinal data may be the best way to capture the experiences of young adults today. Because the NSFG did not collect detailed sexual and cohabitation histories for male respondents, we could not present a parallel analysis for men. Future research should focus on male serial cohabitors’ sexual histories and relationship context to ascertain whether they differ from their female counterparts.

Nock’s (1995) article lays the groundwork for research on cohabitation with the 1987–1988 wave the National Survey of Families and Households (NSFH). His work is pivotal in demonstrating the substantive differences between cohabitation and marriage, especially in regards to relationship quality and commitment. During the era of the NSFH (late 1980s), less than two-fifths of women had ever cohabited, and only about 7% had serially cohabited (Lichter et al. 2010). We expect that Nock would take great interest in the growing serial cohabitation phenomenon and reflect on how serial cohabitors’ complex relationship history may have implications for their subsequent odds of marriage and divorce and marital quality. Future studies should include a measure of serial cohabitation to help researchers understand the relationship between cohabitation and a variety of predictors, including marriage transitions, child well-being, and adult mental and physical health. In particular, distinguishing between serial and single-instance cohabitors in contemporary research exploring the quality and stability of cohabiting and marital relationships will aid scholars’ understanding of the meaning and context of serial cohabitation as a step on the road to marriage.

Nock (2009) argues that one of the key distinctions between cohabitation and marriage is that only marriage is a contract governed by default exit rules (i.e. divorce). There is no legal equivalent to divorce when a cohabiting union dissolves. The results of this study show that a small, yet select, proportion of women exit at least one cohabiting union prior to their first marriage and subsequently enter another (i.e. serial cohabitors). Future research should consider the implications of cohabitation dissolution or “premarital divorce” (McManus and McManus 2008). Women who have experienced a premarital divorce may share similar losses with adults who have experienced a marital dissolution: loss of partner, residence, income-pooling, and emotional support. Indeed, Avellar and Smock (2005) found that premarital divorce reduced women’s economic standing, leaving a larger proportion of women in poverty. Premarital divorcés may have a more difficult time coping than marital divorcés because cohabitation does not offer the same legal protections and social support. Alternatively, premarital divorce may be a simpler process precisely because there is no stress from such legal entanglements or societal pressure. Further inquiry into serial cohabitation may benefit greatly from examining the dissolutions of such unions, identifying the circumstances in which these unions dissolve and how serial cohabitors move on economically and emotionally.

Appendix

Table 6.

Distribution of Covariates for Women (N=3,397)

	All Women	Never Cohabiting	Single-Instance Cohabitors	Serial Cohabitors
	(N=3,397)	(N=1,751)	(N=1,291)	(N=355)
Variable
Birth Cohort	32.99	38.98	27.91	19.23
1958–1962	33.41	32.86	34.95	30.26
1963–1967	33.61	28.16	37.14	50.51
1968–1972
Education
< 12 years	13.79	13.51	14.06	14.31
12 years or more	86.21	86.49	85.94	85.69
Mother's Education
< 12 years	26.84	28.11	26.08	22.58
12 years	39.4	41.63	35.35	43.14
13 to 13 years	19.25	18.05	21.29	17.87
16 or more years	14.51	12.21	17.28	16.42
Family Type During Childhood
Two-parent household	70.6	77.02	66.51	50.36
Non two-parent houseold	29.4	22.98	33.49	49.64
Race/Ethnicity
White	69.07	67.84	71.01	68.29
Black	12.91	11.41	13.87	17.64
Native-Born Hispanic	7.08	5.51	5.50	6.82
Foreign-Born Hispanic	5.63	8.88	5.33	3.86
Other	5.31	6.37	4.29	3.40

Note: All values are weighted.

Source: 2002 National Survey of Family Growth