Latent Dirichlet Analysis of Categorical Survey Responses
提出贝叶斯分层潜在类别模型，分析分类调查响应中的信念共动与异质性，对应信息获取的经济结构模型，并用密歇根调查和教育回报率两个例子展示其应用。
Beliefs are important determinants of an individual's choices and economic\noutcomes, so understanding how they comove and differ across individuals is of\nconsiderable interest. Researchers often rely on surveys that report individual\nbeliefs as qualitative data. We propose using a Bayesian hierarchical latent\nclass model to analyze the comovements and observed heterogeneity in\ncategorical survey responses. We show that the statistical model corresponds to\nan economic structural model of information acquisition, which guides\ninterpretation and estimation of the model parameters. An algorithm based on\nstochastic optimization is proposed to estimate a model for repeated surveys\nwhen responses follow a dynamic structure and conjugate priors are not\nappropriate. Guidance on selecting the number of belief types is also provided.\nTwo examples are considered. The first shows that there is information in the\nMichigan survey responses beyond the consumer sentiment index that is\nofficially published. The second shows that belief types constructed from\nsurvey responses can be used in a subsequent analysis to estimate heterogeneous\nreturns to education.\n