ABSTRACT
In this chapter, we discuss recent research and possible future directions for difficulty modeling and automatic item generation (AIG) of quantitative items. For a summary of recent research on item generation in reasoning, verbal comprehension, and mathematics, refer to Bejar (2010). First, we define some background terminology used to discuss item modeling and AIG. Next, we consider the relationship between cognitive modeling and difficulty modeling, and review research that has identified features of quantitative items that have an impact on difficulty. We argue that an iterative approach to item design, evaluation, and revision can guide the transition from weak to strong cognitive theory and that, as a technology, AIG can facilitate an iterative approach. We explore difficulty modeling of quantitative items from the perspective of transfer of learning, and review several psychometric methods that support difficulty modeling. Finally, we discuss advances in AIG that have the potential to assist difficulty modeling, instructional diagnosis, and automatic scoring.
