作者: Jing Gao
DOI:
关键词: Estimator 、 Inference 、 A priori and a posteriori 、 Bayesian inference 、 Best fitting 、 Operations research 、 Computer science 、 Model selection 、 Contingency table 、 Econometrics 、 Log-linear model
摘要: Title of Dissertation: THE IMPACT OF MODEL SELECTION ON LOGLINEAR ANALYSIS CONTINGENCY TABLES Jing Gao, Doctor Philosophy, 2009 Dissertation directed by: Professor C. Mitchell Dayton Department Measurement, Statistics and Evaluation It is common practice for researchers in the social sciences education to use model selection techniques search best fitting models carry out inference as if these were given a priori. This study examined effect on framework loglinear modeling. The purposes (i) examine consequences when behavior ignored; (ii) investigate performance estimator provided by Bayesian averaging method evaluate usefulness multi-model opposed single