Skip to content

Pre-marital cohabitation endangers your marriage

weepingwomanBy Charles Foster

Marriage is not well served by its defenders. The loudest and best reported of them are often fundamentalist bigots. It’s a shame, for marriage has a lot going for it.

Even if you think that marriage is an anachronistic/bourgeois/theologically contaminated institution, you’ll probably agree that the breakdown of marriages is best avoided. Of course incurably dysfunctional marriages should be ended, but most people aspire to enduring relationships, and the wrench of marital dislocation is emotionally and financially traumatic. If there are children, marriage breakup is painful for the parents and can be enduringly damaging for the children. There are, in short and quite uncontroversially, some significant harms associated with the breakdown of marriages.

How can marriage breakdown – and hence those harms – be avoided?

Many, many factors of course contribute to marriage failure. But there is a growing body of evidence to suggest that pre-marital cohabitation decreases the quality of marriage and increases the odds of divorce. A meta-analysis (2010)1 came to the following conclusions:

Marital stability: (16 studies): ‘…premarital cohabitors had significantly lower rates of staying married compare to non-cohabitors…[T]he data suggest that premarital cohabitation is negatively associated with subsequent marital stability…’2

Marital quality: (12 studies): ‘…cohabitation before marriage is modestly negatively associated with subsequent marital quality.’3

Spencer James and Brett Beattie summarized the consensus as follows:

‘….results generally [supported] the proposition that, on average, cohabitors tend to report poorer marital quality and experience greater marital instability than those who move directly into marriage.’4

But why? And are these merely correlations, or is pre-marital cohabitation causative of subsequent problems? Several studies have tried to answer these questions.

The obvious question is whether the premaritally cohabiting group is materially different from the non-cohabiting cohort – with those differences (rather than the fact of cohabitation) being responsible for the differences in marital quality and stability.

The detailed answer is complex. But the headline answer, I suggest, is not. James and Beattie examined three factors contributing to marital instability: marital conflict, marital happiness and marital communication.5 So far as marital conflict is concerned, they found ‘no evidence that selection into cohabitation [ie the phenomenon of different cohorts opting or not opting for cohabitation] is associated with marital conflict, net of the experience of cohabitation.’6 Marital happiness was rather different: ‘….the experience of cohabitation is again significantly and negatively related to happiness, with cohabitors reporting somewhat lower happiness than those who married directly. In contrast to conflict, however, selection into cohabitation here exerts a significant negative effect.’7 So far as marital communication is concerned, cohabitation ‘appears to exert a negative effect…’ In relation to two out of the three factors, therefore, it was cohabitation itself, rather than the pre-marital characteristics of the cohabitators, that seemed to be important.

There is, of course, more work to be done in establishing causation. But some of the methodological objections to these conclusions were addressed (to my eyes convincingly) by James and Beattie There was a small but statistically significant negative effect of cohabitation on marital quality.

Many of the studies referred to above had several (expressly acknowledged) methodological difficulties. Perhaps the most fundamental is that they took a snapshot: they looked at the reported marital quality (and so on) just at the moment that it was reported by the study subject, rather than appreciating that views change over time.9

This problem was averted in a recent paper from Spencer L. James, who, using a sequential cohort design, examined data from 35 years of marriage.10 This is a paper about marital quality rather than the probability of divorce, but, unsurprisingly, the link between low quality marriages and divorce is well documented.11 He was interested in the trajectories of marital happiness, communication and conflict. He examined a large number of variables, and so was able to adjust for variables such as the number of children, household income, race, age at marriage, employment, and so on, which in some previous studies had obscured the role of pre-marital cohabitation.

In relation to marital happiness, he identified two groups. The first (66% of his sample) was the ‘High Decline’ group (‘HD’). The subjects in this group started their married lives relatively happily (about half a standard deviation above the overall mean), but this initial happiness declined moderately, in a linear way, as the marriage went on.

The second was the ‘Low Rebound’ (‘LR’) group. They were significantly less happy from the start of the marriage than the members of the HD (about half a standard deviation below them). Their marriages were also much less stable than those in the HD: they had very significant declines in their marital happiness over time – significantly more than the decline in the HD. After about 20 years of marriage their happiness had diminished by more than a standard deviation, although subsequently it recovered modestly (hence the ‘rebound’).

Those in the LR had, predictably, a higher chance of divorce than those in the HD.

It is better to be in HD than LR. And women in the LR were more likely to have cohabited premaritally than those in the HD. This was consistent with previous work done by Spencer James12which similarly sought to avoid the ‘snapshot’ objection that had compromised some other studies.

The picture isn’t yet complete. We can quibble about some of the methodology. But it is clear enough that pre-marital cohabitation is a risk factor for marital instability and divorce. The risk is of a grave injury, and seems to be sufficiently large, that, were it a risk of surgery, any competent surgeon would be expected to warn the patient of it.

Even if I’ve overstated the effect of the studies, the precautionary principle should make you cautious of pre-marital cohabitation.


  1. Jose A, O’Leary D and Moyer A (2010) Does pre-martial cohabitation predict subsequent marital stability and marital quality? A meta-analysis. Journal of Marriage and Family 72: 105-116
  2. ibid 109-110
  3. ibid 111-112
  4. James SL and Beattie BA (2012) Reassessing the link between women’s premarital cohabitation and marital quality. Social Forces 91(2): 635-662, 636
  5. ibid
  6. ibid, 652
  7. ibid 652
  8. ibid 654-656
  9. Smock P.J. (2000) Cohabitation in the United States: An appraisal of research themes, findings and implications. Annual Review of Sociology 26: 1-20
  10. James SL (2015) Variation in trajectories of women’s marital quality. Social Science Research 49: 16-30
  11. Huston TL, Caughlin JP, Houts RM, Smith SE, George LJ (2001) The connubial crucible: newlywed years as predictors of marital delight, distress and divorce. J. Pers Soc Psychol 80: 237-252
  12. Spencer James (2014) Longitudinal patterns of women’s marital quality: the case of divorce, cohabitation and race-ethnicity: Marriage and Family Review, 50(8):738-763




Share on

13 Comment on this post

  1. I like the idea of data-driven human psychology, and work like this might well prove to be very valuable as part of a larger, empirical project on human happiness. But at the moment the methodology is all a bit ad hoc. I think you perhaps need to be a bit stricter with yourself on what you can allow as a valid result.

    So in this post you attempt to collate empirical evidence on marital happiness and divorce. But there are gaps in the data, and these are glossed over with qualitative assumptions: “you’ll probably agree that the breakdown of marriages is best avoided…There are…some significant harms associated with the breakdown of marriages.”; “the link between low quality marriages and divorce is well documented”.

    If we’ve learned one thing from the field of medical statistics, it’s that that is not OK. Causal links are not necessarily transitive; looking at secondary endpoints does not always get you the right conclusion.

    Given the state of the data that we have, these fudges are inevitable, and it doesn’t do any harm to advance the empirical evidence bit by bit. But I wouldn’t go making any life decisions based on it, no.

  2. Marriage is a purely religious institution. It is cross-cultural, and cannot be linked to one single religion, but where it exists it always has it roots in some religion. I don’t know any example of a society where people get married, without a pre-existing religious tradition of marriage. When societies secularise, that results in tensions, because the institution of marriage may shift from its religious roots – the Catholic Church opposes divorce, most Christians and Muslims oppose gay marriage, and so on.

    The obvious solution to this problem is that the state should cease to recognise marriage entirely, and leave it to individual values and beliefs. Catholics can have divorce-free marriage, Christian churches can have marriage between a man and a woman, a gay man can marry a gay man, brother can marry sister, humans can marry their dog if they like, and so on. It would all have no legal consequences, and no official status.

    Now Charles Foster might not agree with that policy, but he can not deny that it is a possibility. Nothing prevents the state from ceasing to recognise marriage entirely. And that has consequences for his argument. If marriage in the legal sense only exists because the state has chosen to legally recognise and institute it, then the distinction between post-marital and pre-marital cohabitation is artificial. If the state and the law no longer recognise marriage, then there is no pre-marital sex either, no ‘pre-marital’ anything.

    In fact, if the countries where the research was done retrospectively de-recognised and de-instituted marriage, then all the statistics quoted by Charles Foster become invalid. As evidence, they simply evaporate. Their governments could do the same to the data by retrospectively recognising cohabitation as marriage. If so then, hey presto, all the pre-marital cohabitation becomes post-marital, and all the conclusions of the studies evaporate.

    Research which compares two groups, distinguished only by arbitrary and shifting criteria, does not seem to have much value. Data on ‘marriages’ from countries which introduced gay marriage during the data collection period is also inherently suspect, because the definition changed and the marriages are non-comparable.

  3. Hello Charles,
    Thaks for your interesting post, which merits a far longer comment.
    But here’s a brief one : Spencer James writes the following in his discussion (Jan 2014) –
    “The results stressed the importance of focusing on variation in trajectories of marital quality. The results clearly illustrate that levels of and changes in marital quality are not uniform across the married population. Notably, I found no evidence of differences in trajectories of marital quality between women who married directly and those who cohabited.”
    Why do you draw the opposite conclusion ?

  4. Phil H: Thank you. You write: these are glossed over with qualitative assumptions: “you’ll probably agree that the breakdown of marriages is best avoided…There are…some significant harms associated with the breakdown of marriages.”; “the link between low quality marriages and divorce is well documented”. As to the first quote: is there anyone, anywhere, who disagrees? I thought I was being embarrassingly trite. As to the second quote: I’ve given the reference.

  5. Paul Treanor: Thank you. You say: ‘Now Charles Foster might not agree with that policy [of the state not recognising marriage], but he can not deny that it is a possibility.’ I think it is a splendid idea.
    But you then go on to say: ‘In fact, if the countries where the research was done retrospectively de-recognised and de-instituted marriage, then all the statistics quoted by Charles Foster become invalid. As evidence, they simply evaporate.’ I don’t see how that follows at all. Whether or not a state labels a particular relationship ‘marriage’ won’t have any effect on the stability of the relationship, or the reasons for its stability/instability, or the consequences of its dissolution.

    1. You are perfectly right. A relationship is not the same as a marriage. So stop labeling some relationships ‘pre-marital’, it does not make sense.

    1. Sorry, I didn’t spend the $35.95 to read the 2015 version ! Hope the new evidence is worth the price ……
      (Do I recall a post on the reproducibility problem ?)

    1. To be a little more serious : two points, Charles .

      1. Shouldn’t we compare not only the trajectory of married couples who have cohabited with married couples who have not, but also married couples who have cohabited with couples who continue to cohabit without marrying ?

      2. If it is indeed true that cohabiting before marriage is associated with higher breakdown than marriage without cohabitation, we also need to ask why a couple having cohabitated should decide to marry. It is presumably because they have a reason : ie, that something will change.
      This something may be practical (eg, for tax reasons, to facilitate getting a mortgage, or because their rich moralising uncle is about to write his death-bed will), in which case the redefined relationship becomes merely a consequence of these external factors.
      Or it may be more “internal”, emotional or symbolic ( eg, “we now want to commit ourselves more deeply”), in which case the decision creates new expectations which can lead to later disillusion. So it’s not that surprising that the relationships don’t all last.
      However, they may well last longer than by cohabitation alone (my point 1).

  6. Anthony: many thanks. Both of your queries are indeed important, but I don’t see that any conceivable answer to either or both of them would disturb the very limited point that I made in the post.
    As to (1): as a general rule, cohabitors report poorer relationships than married partners: Brown SL and Booth A (1996) Cohabitation versus marriage: A comparison of relationship quality, Journal of Marriage and the Family 58, 668-678; Nock SL (1995) A comparison of marriages and cohabiting relationships, Journal of Family Issues 16; 53-76; Brown S: Relationship Quality Dynamics of Cohabiting Unions (2003) Journal of Family Issues 24(5): 583-601.
    As to (2): don’t know. I suspect that there are two main types of reason: (a) The ‘Four Weddings and a Funeral’ reason: you live with someone for a while, run out of things to say, and get married because, for a while, it gives you something to talk about, or because it fosters the comfortable delusion that you’re going somewhere: humans don’t like stasis. And (b) Children. So there are reasons that predispose to instability, and reasons that predispose to stability. But the reasons don’t matter for the purposes of my argument because, in the cohorts examined in the cited studies, and presuming that the subject selection was methodologically sound, all reasons were no doubt represented. The end result was what it was, having been generated by – and thus taking into account – all those reasons.

  7. Could the fact that religious couples are less likely to live together before marriage and less likely to divorce be skewing the results?

    I’d like to see the same study results but only including secular couples.


Comments are closed.