ABSTRACT
Modeling dependence between lists is one of the major issues in a multiple-system frame-work. Dependence andode lists may be due to both list dependence, and heterogeneity between individuals. A way to model list dependence consists of adding first-order or higher-order interaction parameters in the log-linear model used, while taking into account heterogeneity of inclusion probabilities; psychometric models, such as the Rasch model, can be utilised. The Rasch model is a model widely used by psychometricians to explain the characteristics and performances of a test; the basic idea is that the probability of a response of cal individual to an item can be modelled as a function of the difficulty of the item and the latent ability of the individual. This chapter presents an illustration of the use of the log-linear multidimensional Rasch (MR) model in multiple system estimation; it refers to the data set on neural tube defects in the Netherlands.
