Sometimes it seems as if researchers are just drawing the wrong lines:

The new research, part of a marriage survey of 22,000 men and women, suggests times have changed from the days when living together signaled poor chances for a successful marriage later. …

The study found those who were engaged and living together before the wedding were about as likely to have marriages that lasted 15 years as couples who hadn’t lived together.

But what about the couples who were living together but weren’t engaged? The new study found marriage was less likely to survive to the 10- and 15-year mark among couples who weren’t engaged when they lived together — findings similar to earlier research.

It sounds like the news is really that couples are behaving as if marriage starts at proposal. Of course, one would need the additional data showing how many engagements actually lead to marriage, but based on this finding, alone, it appears that moving in together makes less of a difference to the longevity of a marriage if the couple takes the step after their engagement.

Looking at the actual report, it’s interesting to note the large differences in the results pending on whether the survey respondent was a man or a woman:

The red lines indicate survey respondents who were women, the blue is for men.  Solid lines indicate marriages that involved no cohabitation before the wedding day; the dashed lines indicate couples who were engaged before they moved in together; and the dotted lines are couples that moved in together before they were engaged or married.  Just from the data, one might be tempted to conclude that men have longer-lasting marriages, but obviously, every line on the graph corresponds to one man and one woman.  Gender should therefore make no difference.

We therefore must look to the methodology and other variables.

The first notable factor is that all interviews were conducted by females.  However, there is no significant difference percentages of each cohabitation group between the male and female respondents, which is where that bias would be expected to manifest.

A more likely factor is that potential participants received information prior to the interview, and there was a three-percentage-point difference between the number of female respondents (78%) and the number of male respondents (75%) who went forward.  It’s plausible that men who were divorced were less likely to want to talk about it.  Note that “no cohabit” lines are almost precisely parallel at 2% difference, and the “cohabit” lines, while rougher, aren’t much different.

The intriguing lines are the “engaged cohabit” ones.  Men who said that they lived together with their fiancés  tracked pretty closely with men who did not, and both were well above men who cohabited prior to engagement.  Engaged cohabiting women, on the other hand, wound up, after 20 years, not far from those who were not engaged when they moved in with their future husbands.

A likely candidate for explanation can be found in another variable:  The survey’s subjects are all describing their first marriages, but 13.7% of female respondents had married men who had, themselves, been married before.  For the male respondents, the percentage is 11.9%.  Perhaps engagement is less of a factor when one of the spouses has been married before, although it appears to make a difference early on.