Discrimination of Leaders Members and Non-Members in Neighborhood Block Organizations

Abstract

Three sets of variables, including a set of traditional demographic variables (Set A), a set of cognitive social learning variables (Set B), and a combined set of demographic and cognitive social learning variables (Set AB) were used to discriminate among leaders, members, and non-members in neighborhood block organizations. Each set was analyzed using both the stepwise and direct methods of multiple discriminant function analysis (Wilk's-Bryan Method). These analyses were first performed on an initial sample (n=216), then applied to a classification analysis (Geisser Method) for both the initial sample and a cross-validation sample (n=205). Subjects were 421 adult residents, 18 years or older, on 17 blocks with active block organizations, living in a transitional neighborhood in Nashville, Tennessee in 1978. Subjects were randomly assigned to either the initial or cross-validation samples. Data were collected in 45 minute interviews with each respondent. Set A analysis revealed that age, education, and homeownership were the most significant discriminators of leaders, members, and non-members. In the Set B analysis, perception of personal skills, block importance, and block satisfaction were the most significant. The combined AB Set showed age, education, block importance, block satisfaction, and personal skills, respectively, to be the most efficient discriminating variables. There was a high degree of consistency in percent correct classification between samples and across methods