At the beginning of the semester, we discussed the concept of homophily – its pervasiveness in social networks and its role in triadic closure (the idea that if two people have a mutual friend they are more likely to become friends themselves at some point). Simply by examining our close friendships we can verify that homophily often holds up in real life. That is, our friends tend to be similar to us in terms of basic characteristics such as ethnicity, interests and life experiences. This occurs both because we are attracted to people with whom we share things in common and because, over time, we become increasingly more similar to our friends. Thus, our immediate social networks tend to be composed of people who are like us. What might be the implications of homophily for certain social networks?

A 2004 study by Levendosky, Bogat, Theran, Trotter, von Eye and Davidson (see website below) used social networks data to study pregnant female victims of domestic violence. A major aspect of their investigation involved examining the relationship between the homophily of the women’s social networks and the quality of the social support they received. In the study, a woman with a high homophily index possessed a social network that was heavily composed of friends who had also experienced domestic violence. The results suggested that women with higher homophily indexes received less emotional support. Perhaps, the authors suggest, homophilous friends have a diminished capacity to provide psychological support due to their own history of abuse. The study also found that a greater presence of non-homophilous supporters (friends who had not experienced domestic violence) within one’s social network was associated with better mental health.

It seems like abused women would be more likely to be socially isolated than women who are not abused. Here, social isolation is defined as having few supporters within one’s social network. We could think of social isolation as being equivalent to having few strong ties to others or possibly few connections to others at all. In fact, the researchers in this study found that battered women did not experience a higher level of social isolation compared with non-battered women; there was no significant difference between the two groups.

Maybe when someone is in a situation as complex and psychologically stressful as domestic violence, it’s not about how many people you know, but rather, the types of people you know and the quality of support that they can offer. It may be extremely important for victims of domestic violence to form “edges” to people in different social circles (or components), who may be able to expose the victim to non-violent social norms and more sturdy sources of psychological resilience. As the researchers propose, these non-homophilous connections could be facilitated by agencies, which could help battered women become involved in various social groups outside their immediate social network. These agencies would effectively be creating new bridges and local bridges. As we learned in class, homophily is a powerful force in social networks; but perhaps it can and should be countered by social intervention…especially when the consequences are negative.

The Social Networks of Women Experiencing Domestic Violence (Levendosky et al 2004) was published in the American Journal of Community Psychology. The abstract and full article (if you have access) can be found at http://www.springerlink.com/content/w87627452795731g/