Twin Studies of the Marriage Benefit: Parsing Selection from Causation

PI: Robert Emery
Project Number: 2R01HD056354
Project Dates: 6/1/2011–5/31/2014

Family life is widely viewed as having a profound influence on both children and adults. However, broad changes in family structure have contributed to evolving, and controversial, ideas about what makes a “normal” or “healthy” family. Social scientists contribute both to basic knowledge about family life, and to social programs and political debates, by studying potential causes and consequences of different family experiences. These efforts are exceedingly important, especially given the prospect of altering family life in order to promote individual health and well-being. Our proposed work contributes to this important effort by focusing much of our effort on the question: Does marriage benefit social, emotional, psychological, and economic well-being – and even physical health and longevity?

A large body of research suggests that marriage is correlated with these outcomes, but correlation does not mean causation. Or as two critics of marriage noted in a recent essay, “To say marriage creates wealth is to confuse correlation with causation. If there is more wealth in Manhattan than in Brooklyn, that does not mean that moving to Manhattan will make you wealthier.” We are not only concerned that background and personality affects who gets and stays married, but we also know that genetic factors influence the likelihood that people will experience happy marriage or other key aspects of family life. Even as we raise this concern, however, we offer a method – studying twins – that solves not one but two huge scientific problems. Identical twins share the same genes, and they also share the same upbringing. Thus, when we compare identical twins, we eliminate the problem of nonrandom selection into experience based on background, prior experience, personality, and genes. In fact, the twin method allows us to determine cause and effect better than any available research technique, other than randomly assigning couples to marry, or not, divorce, or not, and so on, which, of course, is logistically and ethically impossible.

We propose to study five unique samples of twins in the U.S., Australia, and Sweden (each sample offers special benefits) to test whether marriage (and single life, cohabitation, divorce, widowhood, and remarriage) really are likely to cause of the outcomes with which they are correlated. We also propose to study how genes and the environment work together to modify the “marriage benefit” for mental and physical health, including studying specific genes recently documented to influence behavior in close relationships. Finally, in order to enrich our highly technical quantitative analyses, we plan to interview identical twins to learn more about what makes them different, particularly how different experiences in marriage, romantic relationships, and relationship dissolution may alter their life course.