The study of social networks is a powerful tool for understanding how behaviors spread across groups. Researchers are using a large networked dataset originally designed to study heart disease to ask a different question: Is divorce contagious? Rose McDermott, James Fowler, and Nicholas Christakis used the Framingham Heart Study to look at how breakups flow across social ties. The study of nearly 5,000 residents in a single town began in 1948, spans several generations, and includes information on how individual residents know one another (e.g., “spouse,” “sibling,” “coworker”) and how closely they are connected.

Within the network, divorce appeared in clusters and spread across friends, extending to two degrees of separation. In other words, a person’s tendency to divorce depends not just on his friends’ divorce, but also his friends’ friends’ divorce. The full network shows that participants are 75% more likely to be divorced if a person they are directly connected to is divorced.

Divorce patterns affect other qualities of the network, too. People who go through divorce experience a small drop in the number of people who name them as friends in the study, and divorcees are more likely to immerse themselves in denser social groups with fewer ties outside their groups. If they remarry, those who are divorced are very likely to partner with another divorcee.

The beauty of network analysis is that it can reveal how our social ties might shape our individual decisions. This study shows that the most “popular” people in the network (those most oft named as a friend by other study participants) were the least likely to divorce. The researchers think having this strong social support system points to other qualities that might help a marriage: good social skills, friends to rely on, and people other than one’s spouse to vent to. Though divorce is an intensely personal process, it seems the people around us shape our ability to make a marriage work and our embrace of alternatives.