4.6 Homophily

Relative and Expected Rates EI Homophily Index Blau’s Heterogeneity Index

Gower and Legendre (1985) \[ \begin{equation} \sqrt(\frac{a}{(a+b)}-\frac{c}{(c+d)})(\frac{a}{(a+c)}-\frac{b}{(b+c)}) \end{equation} \]

Famous Network Studies: Kandel 1978

Kandel conducted one of the first longitudinal studies of the role of homophily in driving friendship choices. At the beginning of a school year, Kandel asked high schoolers to provide a self assessment of their personal misbehavior at school, use of marijuana, anticipated level of educational completion, and political beliefs and also to nominate their best friend. Kandel then asked the students to complete the same questions at the end of the year.

Since nearly everyone in the high schools did this, Kandel was able to compare the similarity of the nominated friendship pairs along the four dimensions.

She found that friendships where use of marijuana, misbehavior, and educational intentions were similar were likely to persist into the second time period, while friendships where such values were far apart were likely to dissolve. Furthermore, the new best friendships were closer on these assessed values than the original friendship. Thus, low homophily predicted friendship dissolution and higher homophily predicted new friendship formation. Political leanings were not as useful in estimating tie formation or dissolution.

Kandel, Denise B. 1978. “Homophily, Selection, and Socialization in Adolescent Friendships.” \(American\) \(Journal\) \(of\) \(Sociology\) 82 (20): 427-436.