Model Selection with Incomplete Data Using Adjusted Variance
*Ashok Chaurasia, University of Connecticut
Keywords: Incomplete Data, Multiple Imputation (MI), Model Selection, AIC, BIC, Multiple Regression
Model selection is a task that comes up often in applied research. Various model selection criteria have been proposed over the years. Two of the most commonly referenced model selection procedures are the Aikaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC). The advantages and disadvantages of model selection procedures for complete data are well documented in literature. However, when dealing with incomplete data the task of model selection still remains elusive. In this paper we propose and investigate the performance a model selection procedure based on adjusted variance estimates when dealing with multiply imputed data in the multiple regression setup.
Important Dates & Deadlines
- October 9 - 11, 2013