Title: Bayesian analysis of cognitive diagnostic models for continuous response data
Authors: Eduardo Schneider Bueno de Oliveira - Federal University of Sao Carlos (Brazil)
Xiaojing Wang - University of Connecticut (United States) [presenting]
Jorge Luis Bazan - University of Sao Carlos (Brazil)
Abstract: The use of nondichotomous response models for assessment is increasing each time. Different scoring methods may be used to evaluate aspects of interest in many research fields, with the continuous responses being one of these. In the cognitive diagnosis models literature, much effort has been done to develop dichotomous and polytomous models in the past but, recently, the possibility of using continuous responses has been brought to discussion. However, there is no Bayesian approach for this class of models considering continuous responses. A first Bayesian framework for the continuous DINA is proposed. The good performance for parameter recovery is shown through a simulation study and also an application with continuous responses for risk perception is presented.