Not All Marriages Are Equal at Protecting Longevity

December 07, 2016

On November 30, 2016, a post by Robert H. Schmerling MD appeared on the Harvard Health Blog, detailing health benefits of marriage and why tying the knot has been associated with advantages that include longevity. Sociologist Linda Waite and journalist Maggie Gallagher had made similar arguments in The Case for Marriage, originally published in 2000.

Ten years later, Bella dePaulo offered Psychology Today a careful analysis of the studies upon which that treatise was based, and concluded that several analytic flaws had led Waite and Gallagher to overestimate the benefits of marriage, underestimate its hazards – including its ending in divorce or widowhood – and ignore the advantages of single living.

In 2002, my colleagues Stanislav V. Kasl and Amy Darefsky and I argues that all marriages were not equal.  We looked at the effects on mortality of marriage, and of variations in the quality of marriages, from yet another angle.  We were lucky to have access to six years of longitudinal data from both spouses in 305 older married couples.  We also had data that might otherwise explain or contribute to mortality— age, education, financial strain, race, chronic health problems, disability, cognitive impairment, depressive symptoms, underweight or extreme overweight, drinking alcohol, a history of smoking (or current smoking), self-perception of how healthy they were, and many more dimensions that might be operating more subtly.

What we were interested in was the extent to which qualities of closeness in the marriage, defined as the extent to which each spouse relied on the other as a primary confidant or major source of emotional support, identified a style of relationship that influenced survival over time.

Bingo! The approach yielded dramatic results.  Four clearly different types of marriages had four clearly different survival rates for both husbands and wives.  Even better, the four types of marriages—defined through our dry but impressive epidemiological data—mapped beautifully onto the four styles of stable marriages that Judith Wallerstein and Sandra Blakeslee had found in their wonderful qualitative study of couples who had been married for at least ten years, The Good Marriage. 

We found that, for our heterosexual couples in which both spouses had been born no later than 1917, a marital partnership in which the woman relied on the man as her go-to confidant or emotional rock, husbands lived longer than in any other marital style and wives tended to live longer than those in mutually dependent couples.  If the wife had been  a mother, the survival advantage of this style was enormous.

We had a lot of additional (Established Populations for the Epidemiologic Study of the Elderly) data available through which we could try to better understand these findings.  No matter how we tried to otherwise explain our results – health conditions, whether it was the first marriage or a later one, length of marriage, whether there was a possibility one spouse might have overheard the other spouse’s interview, the number of other close friends and additional social support. We could examine who did what when, how religiouseach spouse was, and so on. The only additional influence that mattered in predicting mortality was whether a person had children, and even that was irrelevant for husbands.

For wives, however, having children changed everything.  When we separated couples who had had a child from those who had not, wives who had never been a mother  were massively protected by being able to rely on their husbands compared to wives in any of the other three marital styles.  Even though our numbers were getting smaller as we sliced and diced the data—only 46 wives had never had children and only a few of those 46 had died in the six years of the study—for childless women, the husband relying on them was protective.

Fast forward 14 years.  Last summer, Jennifer Roebuck Bulanda, J. Scott Brown and Takashi Yamashita put more of the pieces together.  They included married people who later divorced or were widowed in their mortality analyses.  Even better, the people in their study were younger when it began (over 50 instead of over 65), were born later (up until 1942 instead of 1917), and were followed for a longer period (16 years instead of 6).  They also had a much larger number of participants, although they did not have data from both spouses as we had been fortunate enough to have, and so they could only look at individual reports and not marital styles.  These careful researchers underscored a main point in Bella dePaulo’s earlier argument:  Consequences of losing a marriage through death or divorce are negative indeed.  By shining a light on ways in which mortality is influenced by becoming coupled and uncoupled along with the more subtle issue of the quality of the relationship, they underscored that thinking of marriage categorically does a disservice to people’s individuality.

It’s complicated.  All marriages are not equal and they do not equally protect or threaten. Arguments can be made for how and when and why we choose to commit our destinies to another person.  For a romantic like me, there is little choice:  the richness of life filled with interpersonal meaning trumps pain that may follow a relationship’s loss.  But that is, after all, a very individual choice.

Copyright 2016 Roni Beth Tower


Tower, R. B., Kasl, S. V., & Darefsky, A. (2002). Types of Marital Closeness and Mortality Risk in Older Couples.  Psychosomatic Medicine; 64:644-659.

Bulanda, J. R., Brown, J. S., & Yamashita, T. (2016). Marital quality, marital dissolution, and mortality risk during the later life course.  Social Science and Medicine;  165, 119-127.

Leave a reply
How will you know when you met your soulmate?

Leave Your Reply

Your email address will not be published. Required fields are marked *