ABSTRACT

This chapter deals with response models for dichotomous items that specify item difficulty parameters as a linear function of a set of parameters for the features of the item. Since Fischer's (1973, 1983) seminal work on the linear logistic test model (LLTM), these explanatory item response models (Volume One, Chapter 33) have always been present at the psychometric forefront. Several models have been developed that extend the basic LLTM. Besides, an empirical line of research has been established dealing with the effects of item features on item difficulty. For instance, both developments have led to research on automated item generation (Volume One, Chapter 26), in which the item features found to influence item difficulty are integrated in a set of item-generating rules.