Academy of Mathematics and Systems Science, CAS
Colloquia & Seminars

Speaker:	Associate Professor Xiaojing Wang, University of Connecticut
Inviter:
Title:	New Insights on Bayesian Model Assessment to Quantify the Necessity of Jointly Modeling Multi-Resources Data
Language:	Chinese
Time & Venue:	2023.07.12 09:00-10:00 思源楼802 腾讯会议：962-4818-2651
Abstract:	Computerized assessment provides rich multi-resources data including trial-by-trial accuracy, response time (RT) measures and paper-and-pencil scores. Then, a natural question is how to quantify the necessity to incorporate additional source of data besides the response accuracy (the key focus) in the modeling. To address this, we first propose a joint model for consisting of a two-parameter IRT model for the dichotomous response accuracy data, a log-normal model for the continuous RT data, and a normal model for corresponding paper-and-pencil scores. Then, we propose several new model assessment criteria based on the decomposition of deviance information criterion (DIC) and the logarithm of the pseudo-marginal likelihood (LPML). These proposed criteria can quantify the improvement in the fit of one piece of the multi-resources data given the others. In order to examine the empirical performance of the proposed model assessment criteria, we have conducted several simulation studies and finally, we have illustrated the application of these criteria using a real dataset from AppRISE computerized educational assessment program.