A quantile regression analysis of acquaintance volume

Abstract

Acquaintance volume, total size of personal/egocentric network including relatively sheer, inactive relationships, is a key concept in some important theory of social science such as subcultural theory of urbanism, emancipation theory of trust, and social capital. However factors correlating with acquaintance volume have not been identified well. In the present study, we measured acquaintance volume with a telephone book method and explored effects of some predictors on median and wide range of quantiles of acquaintance volume by quantile regression. Results showed positive effects of age, education, household income, use of cellular phones, and general trust on median of acquaintance volume. From the comparison of coefficients of regression models predicting q90, q50, and q10, we identified four predictors of network hubs: age, education, city size, and interpersonal anxiety. While age and education increased q90 more than q50 and q10, city size and interpersonal anxiety decreased q90 more than q50 and q10.