ABSTRACT
In this chapter, I will apply a macro approach to model the patterns of homophily and heterophily using the position-generated network data in the U.S. and China. I will follow the methods used in Marsden (1988) to capture (1) the strength of tendencies toward homophily or inbreeding for the formation of in- groupties and (2) the extent of heteroph ily or social distanced iferentiation for the formation of out-group ties. There is voluminous evidence in the literature that the principle of homophily exists in social networks (Fischer 1982a; Homans 1950; Laumann 1966, 1973; Laumann and Pappi 1976; Lazarsfeld and Merton 1954; Lin 1982; Marsden 1981, 1988, 1990; McPherson et al. 2001). Over a century ago, the phenomenon of homophily was observed by Tarde: “Social relations, I repeat, are much closer between individuals who resemble each other in occupation and education” (1903). In recent years, Lazarsfeld and Merton (1954, p. 303) defined the concept of homophily as “a tendency for friendships to form between those who are alike in some designated respect” and heterophily as “a tendency for friendships to form between those who differ in some designated respect.” They further distinguished two types of homophily: status homophily, which includes the major socio-demographic dimensions that stratify society—ascribed characteristics such as race/ethnicity, sex or age, and acquired characteristics such as religion, education, occupation or behavior patterns, and value homophily, which includes the wide variety of internal states presumed to shape our orientation toward future behavior. The degree of homophily varies over different characteristics. For instance, a network can be homophilous over some characteristics but not over others. Moreover, the extent of homophily varies for each respondent, which may be one of the sources that differentiations in status attainment come out.
