|
Abstract:
|
Computerized adaptive testing (CAT) has become a popular mode of educational assessment. The most important component in CAT is the item selection procedure for selecting items during the course of the test. According to Lord (1970), an examinee is measured most effectively when test items are neither too difficult nor too easy. Several sequential item selection methods have been proposed, including a version modified from the Robbins-Monro process, Lord's maximum Fisher information method, and Owen's Bayesian method. Recently, Chang and Ying proposed the a-stratified method with an objective to limit the exposure on any given item by using that item at the most advantageous point in testing. One of the advantages to use the a-stratified method is that it attempts to equalize the item exposure rates for all the items in the pool. Various statistical issues will be addressed in this paper
|
