Computer adaptive testing for small scale programs and instructional systems
Lawrence M. Rudner, Graduate Management Admission Council
Based in item response theory, much of the research and development for computer adaptive testing has been has been restricted to programs with large item banks and large numbers of examinees. This paper will present approaches to CAT that lift these requirements. These measurement decision theory approaches are simple to implement and are well suited for small certification programs and as embedded tools in intelligent instructional systems. Publically available software for simulating data, calibrating questions, and scoring examinees will also be presented.