Computerized Adaptive Testing

To use computerized adaptive testing, we must first calibrate a bank of test items using an item response theory (IRT) model that relates properties of the test items (e.g. their difficulty and discrimination) to the ability (or other trait) of the examinee. Item selection rules derived from IRT and adaptive testing can explicitly use concepts of item information. IRT procedures for estimating an individual’s trait level are applicable to the adaptive testing process. In addition, maximum likelihood and Bayesian estimation procedures also provide individualized standard errors of measurement, for each trait level. Finally, adaptive testing procedures developed in accordance with IRT can take advantage of a number of different procedures for terminating an adaptive test. Traditional mental health measurement has been based on classical test theory, in which a patient’s impairment level is estimated by a total score, which requires that the same items be administered to all respondents.