For computerized test systems to be operational, the use of item response theory is a prerequisite. As opposed to classical test theory, in item response models the abilities of the examinees and the properties of the items are parameterized separately. Hence, when measuring the abilities of examinees, the model implicitly corrects for the item properties, and measurement on an item-independent scale is possible. In addition, item response theory offers the use of test and item information as local reliability indices defined on the ability scale. In this chapter, it is shown how the main features of item response theory have given rise to the development of promising procedures for computerized testing. Among the topics discussed are procedures for item bank calibration, automated test construction, adaptive test administration, generating norm distributions, and diagnosing test scores.