This study examines whether premarital cohabitation increases one’s likelihood of future marital dissolution. We use data from the 2015–2017 National Survey of Family Growth, which surveyed 5554 women aged 15–49. Premarital cohabitation predicts a substantially higher rate of marital dissolution, with a 31.4% divorce rate among those who had cohabited before marriage compared to a 25.9% rate among those who had not. However, some of this relationship is driven by the fact that cohabiting couples were also less religious and less likely to come from an intact family, factors which independently predict divorce. Our regression analysis controlling for these and other differences shows that cohabitation leads to a smaller, but still statistically and practically significant, 4.6 percentage point increase in the rate of marital dissolution.
In today’s society, it is becoming the norm to cohabit prior to marriage, or even to cohabit with a significant other whom you have no serious intention of marrying. Our analysis of the National Survey of Family Growth (NSFG) shows that among married Americans aged 15–49, 60% had cohabited before marriage. Many conservatives are against this phenomenon, as it goes against traditional societal norms and certain religious beliefs. In addition to traditional moral and religious reasoning, some empirical research finds negative effects of premarital cohabitation including a higher likelihood of divorce (Rosenfeld and Roesler (2018)). However, in recent years divorce rates have fallen even as premarital cohabitation rates have increased. Thus, our analysis attempts to account for other factors that may affect divorce rates, using regression analysis in order to determine the true relationship between premarital cohabitation and marital dissolution.
We analyze data from the most recent National Survey of Family Growth. While several articles have previously considered the relationship between premarital cohabitation and marital dissolution, we are the first to do so using the 2015–2017 NSFG survey or using any data more recent than 2015. While some previous work argued that the effect of cohabitation on divorce was diminishing towards zero over time, our work with the latest data shows that it remains substantial. We also find that divorce is more likely for women who are poor, from a non-intact family, or non-white, while it is less likely for women who are religious, married young, or have children (whether they were born before or during the marriage).
Jose et al. (2010) provide a full meta-analysis of prior work on whether cohabitation predicts martial stability. Reviewing 16 papers that had been published on the subject prior to 2010, they find that cohabitation had a significant negative association with marital stability, except when the cohabitation was with the eventual marriage partner.
Studies cited in the meta-analysis identified additional variables that sometimes predict marital stability. For instance, Teachman (2003) estimates the association between premarital relations and subsequent marital dissolution using data from the 1995 National Survey of Family Growth. He finds that premarital sex or premarital cohabitation with a woman's husband is not associated with an increased risk of marital dissolution; however, women who have more than one premarital relationship including premarital sex and/or cohabitation have an increased risk of marital dissolution. Lillard et al. (1995) use data from the National Longitudinal Study of the High School Class of 1972 to show that even those cohabiting with their eventual spouse are more likely to divorce, but that this effect is not causal. Instead, some of the same variables that predict someone is more likely to cohabit (multiple previous marriages) also predict that they are more likely to divorce, and some of the variables that predict they are less likely to cohabit (currently pregnant, intact family, rural) predict that they are less likely to divorce.
While Jose et al. (2010) argue that the negative effect of cohabitation on marital stability has been steady over time, Reinhold (2010) argues that it has changed. Using pooled data from the 1988, 1995, and 2002 National Survey of Family Growth (NSFG), he finds that the positive relationship between premarital cohabitation and marital instability has weakened in more recent birth and marriage cohorts, to the point that cohabitation may even be a stabilizing factor for recent marriages. Lamidi et al. (2019) show that cohabitation itself has changed over time (from the 1988 to the 2011–2013 NSFG), with cohabitations lasting longer and being less likely to convert to marriages.
More recent studies identify additional variables that affect cohabitation and marital instability. Manning and Cohen (2012) use data from the 2006–2008 National Survey of Family Growth. They find a positive association between premarital cohabitation and marital instability for women who were not engaged with their spouse at the outset of cohabitation. However, women who premaritally cohabited with their spouse and were not engaged shared similar odds of marital instability as women who never cohabited. Rosenfeld and Roesler (2018) use data from the National Surveys of Family Growth from six different surveys between 1970 and 2015. They show that women in their first year of marriage who cohabited before marriage have a lower marriage dissolution rate than those who did not cohabit before marriage. However, after five years of marriage, those who cohabit before marriage have a greater likelihood of marriage dissolution than those who did not cohabit. Thus, cohabitation before marriage has short run benefits but long run negative consequences. These negative effects were especially large when cohabitation was with partners other than the future spouse.
Our study differs from all of these studies in various ways, but the key difference is that we analyze the most recent data, the 2015–2017 National Survey of Family Growth.
Data and data analysis
Our data are taken from the 2015–2017 National Survey of Family Growth. The sample includes 5554 women ages 15 to 49. This is cross-sectional data consisting of individuals from across the USA. Our main dependent variable is whether or not the first marriage ended in divorce or separation. This is a dummy variable coded equal to zero if the respondent is still in his or her first marriage and one if the respondent’s first marriage has ended in divorce, annulment, or separation. Our key independent variable is coded equal to zero if the respondent did not participate in premarital cohabitation prior to their first marriage and equal to one if the respondent did participate in premarital cohabitation prior to their first marriage. Our controls will include ethnicity (0 if white and 1 if non-white); education (years of schooling, bottom-coded at 9 and top-coded at 19); a binary indicator for intact status of childhood family; premarital fertility (measured using an indicator of whether the respondent had a child before his or her first marriage); marital fertility (whether the respondent had a child in the first marriage); whether the respondent’s income is at the level of poverty at the time of the survey; age at first marriage; and religious affiliation (coded zero for no religion, one for any religion). Table 1 below gives the summary statistics for the dependent variable, independent variable, and controls for the 2582 women in the sample who are or have been married. The overall rate of marital dissolution is 29.2%, and 60.2% cohabited before marriage.
Table 2 shows how the sample means differ across ever-divorced and still-married women. Divorced women are 6.4 percentage points (pp) more likely to have cohabited before marriage, 9.2 pp more likely to be non-white, 17.1 pp more likely to come from a non-intact family, 11.8 pp more likely to have had a child before marriage, 11.9 pp less likely to have had a child while married, 5 pp less likely to be religious, had 1 year less education, and were 1.9 years younger at first marriage. All differences are statistically significant at the 5% level.
Table 3 shows how the sample means differ across women who did and did not cohabit before marriage. Women who cohabited before marriage are 5.5 pp more likely to divorce, 17 pp more likely to come from a non-intact family, 23.5 pp more likely to have had a child before marriage, 23.5 pp less likely to have had a child while married, 13.6 pp less likely to be religious, had 0.27 years less education, and were 2.41 years older at first marriage. All differences mentioned are statistically significant at the 5% level.Footnote 1
Our model is an Ordinary Least Squares regression with the form:
Our dependent variable is whether or not the first marriage ended in divorce or separation. Our key independent variable is whether or not the respondent participated in premarital cohabitation in their first marriage. Our controls will include ethnicity, education (number of years of schooling), intact status of childhood family (representing whether or not the respondent has two biological or adoptive parents from birth), premarital fertility (measured using an indicator of whether the respondent had a child before his or her first marriage), marital fertility (whether the respondent had a child in the first marriage), whether the respondent’s income is at the poverty level at the time of the survey, age at first marriage, and religious affiliation which represents whether the respondent identifies as Catholic, Protestant, other religion, or no religion.
The controls account for factors other than premarital cohabitation that may a couple’s likelihood in later marital dissolution. The control for religious affiliation is particularly important because many religions are against divorce and premarital cohabitation; an association between divorce and cohabitation may be due to religion reducing both, rather than cohabitation increasing divorce. One area in which the model may fall short is the fact that we are using only one survey, so there could be some political, cultural, or social circumstances in 2015–2017 that may affect these variables temporarily. In addition, our poverty income control variable estimates poverty based on the average American family (a three-person family) between the years 2015–2017, so this may not be accurate if the respondent has an exceptionally large (or small) family.
Estimation and hypothesis testing
Table 4 shows the regression results. After controlling for other factors, we find that premarital cohabitation predicts a statistically significant 4.6 pp increase in the likelihood of marital dissolution (divorce or separation). Other statistically significant findings are that non-white women are 9.9 pp more likely to divorce, those with a non-intact childhood family are 9.7 pp more likely to divorce, those in poverty are 19.2 pp more likely to divorce, religious women are 8.8 pp less likely to divorce, women who married later are less likely to divorce (1.5 pp less likely per year of age at first marriage). Those with children are substantially less likely to divorce, including if the children were born before marriage.Footnote 2 While more education is associated with a lower divorce rate, education does not have a statistically significant effect on divorce after controlling for other factors.
Table 5 shows the results of a logistic regression for which controls predict premarital cohabitation. Our statistically significant findings are that non-whites were 9.6 pp less likely to cohabit, more educated women were slightly (0.9 pp per year of education) less likely to cohabit, women from a non-intact childhood family were 13.9 pp more likely to cohabit, women with children (born either before or within marriage) were substantially less likely to cohabit, those who married later were more likely to cohabit (2.1 pp per year of age), and religious women were 20.2 pp less likely to cohabit before marriage.
In conclusion, we find that premarital cohabitation is associated with a statistically significant 4.6 pp (15.8%) higher chance of marital dissolution for the first marriages in our sample. This relationship held through the inclusion of many control variables, though the possibility of remaining omitted variable bias means this may not be a causal effect. Some additional controls that would benefit future research are the couple’s intention before entering cohabitation (whether they had plans for marriage or engagement at the outset of cohabitation), measures of premarital relationship quality (such as conflict, emotional security, and the ability to cope with difficulties), work hours, more precise measures of religion, and the age difference of the couple (past studies have shown that a wider age gap may increase the likelihood of marital dissolution). Also, while we use the most recent available data from the USA, the effect of cohabitation on divorce may differ across time and place; for instance, Hiekel and Fulda (2019) show how this effect in West Germany worked differently from the USA.
Overall, since this paper determines that premarital cohabitation is positively correlated with marital dissolution, couples should consider carefully whether the benefits of cohabitation exceed the costs for them, recognizing that one of the costs is a moderately higher probability of a future divorce. Our results contrast with some previous research, such as Reinhold (2010), which argued that the association of cohabitation with marital dissolution was shrinking over time to the point that cohabitation might become a positive for marital stability. Instead, the relationship seems to be fairly stable over time, as our finding that cohabitation is associated with a 15.8% higher likelihood of divorce is almost identical to the finding in the meta-analysis by Jose et al. (2010) that the first marriages of those who had cohabited were 16% less stable than those who had not.
Availability of data and material
All data generated or analyzed during this study are included in this article’s supplementary information files.
Stata code for this study is included in this article’s supplementary information files.
There were no statistically significant differences by race and poverty, which are included in the table but not mentioned above. The lack of a significant association with poverty is surprising, given the economic benefits of combining households (Jamison 2018).
The regression results differ most sharply from the simple cross-tabulation of Table 2 for premarital births; while women who had a child before marriage are substantially more likely to divorce, this appears to be entirely driven by confounding factors, so that the actual causal effect of the premarital birth is to reduce divorce.
Hiekel N, Fulda BE (2019) Love. Break up. Repeat: The prevalence and stability of serial cohabitation among West German women and men born in the early 1970s. Demogr Res 39(30):855–870. https://doi.org/10.4054/DemRes.2018.39.30
Jamison TB (2018) Cohabitation transitions among low-income parents: a qualitative investigation of economic and relational motivations. J Fam Econ Issues 39:73–87. https://doi.org/10.1007/s10834-017-9546-3
Jose A, O’Leary KD, Moyer A (2010) Does premarital cohabitation predict subsequent marital stability and marital quality? A meta-analysis. J Marriage Fam 72:105–116. https://doi.org/10.1111/j.1741-3737.2009.00686.x
Lamidi EO, Manning WD, Brown SL (2019) Change in the stability of first premarital cohabitation among women in the United States, 1983–2013. Demography 56:427–450. https://doi.org/10.1007/s13524-019-00765-7
Lillard LA, Brien MJ, Waite LJ (1995) Premarital cohabitation and subsequent marital dissolution: a matter of self-selection? Demography 32(3):437. https://doi.org/10.2307/2061690
Manning WD, Cohen JA (2012) Premarital cohabitation and marital dissolution: an examination of recent marriages. J Marriage Fam 74(2):377–387. https://doi.org/10.1111/j.1741-3737.2012.00960.x
Reinhold S (2010) Reassessing the link between premarital cohabitation and marital instability. Demography 47(3):719–733. https://doi.org/10.1353/dem.0.0122
Rosenfeld MJ, Roesler K (2018) Cohabitation experience and cohabitations association with marital dissolution. J Marriage Fam 81(1):42–58. https://doi.org/10.1111/jomf.12530
Teachman J (2003) Premarital sex, premarital cohabitation, and the risk of subsequent marital dissolution among women. J Marriage Fam 65(2):444–455. https://doi.org/10.1111/j.1741-3737.2003.00444.x
The authors received no funding for this research.
Conflict of interest
The authors have no conflicts of interest to report.
About this article
Cite this article
Kerrigan, S., Bailey, J. Does premarital cohabitation increase the likelihood of future marital dissolution?. SN Soc Sci 1, 123 (2021). https://doi.org/10.1007/s43545-021-00146-1
- Premarital cohabitation
- Marital dissolution
- National Survey of Family Growth