Regularized Mislevy-Wu Model for Handling Nonignorable Missing Item Responses

Missing item responses are frequently found in educational large-scale assessment studies. In this article, the Mislevy-Wu response model is applied for handling nonignorable missing responses. This allows that missingness of an depends on itself and a further latent variable. However, with low to moderate amounts responses, parameters mechanism difficult estimate. Hence, regularized estimation using fused ridge penalty stabilize estimation. The function separately defined multiple-choice constructed items because previous research indicated mechanisms strongly differed two types. simulation study, it turned out improves stability parameter method also illustrated international data from progress reading literacy study (PIRLS) 2011 data.