Much has been written about the possibilities of computer adaptive testing (CAT), but little practical success has been reported. Most proposals are so complicated and fraught with expected problems that one gets the impression the idea is hopeless. In fact, CAT enabled by the application of probabilistic conjoint measurement works well and the resulting data has shown that most of the expected problems do not necessarily appear. This chapter explains how probabilistic conjoint measurement enables CAT, reviews the results of some applications to medical certification problems in the U.S., and outlines how CAT leads to practical and useful computer assisted instruction.
ASJC Scopus subject areas