作者: Jean-Laurent Rosenthal , D. Roderick Kiewiet , Michael Alvarez , Inés Levin
DOI:
关键词:
摘要: Existing survey-based studies of political participation can be grouped into those that model in each activity independently (with most them focusing on voter turnout), and overall based aggregate indicators civic engagement. In this paper I develop apply a new approach for the study engagement simultaneous modeling multiple activities. use finite mixture to allow parameters vary across latent citizen types classify survey respondents apathetic activist classes citizens. Specifically, Bayesian methods estimate models where skewness linked function is allowed classes, multilevel coefficients Using data from 1990 American Citizen Participation Study, find even after controlling comprehensive set individual attributes, large heterogeneities still remain. Also, impact attributes such as education varies considerably depending characteristics activity.