A use of the information function in tailored testing

Several important and useful implications inlatent trait theory, with direct implications for individualizedadaptive or tailored testing, are pointedout. A way of using the information function intailored testing in connection with the standard errorof estimation of the ability level using maximumlikelihood estimation is suggested. It is emphasizedthat the standard error of estimationshould be considered as the major index of dependability,as opposed to the reliability of a test. Theconcept of weak parallel forms is expanded to testingprocedures in which different sets of items arepresented to different examinees. Examples aregiven.Researchers have tended to use latent trait theoryrather than classical test theory in research on individualizedadaptive or tailored testing. This isquite natural, since latent trait theory has definitemerits over classical test theory in many crucial matters.Because of the lack of opportunities to reallylearn the theory, however, these researchers tend tooverlook some important implications in latent traittheory. As a result, its full use has not yetmaterialized. Not only are information functionsseldom used to maximum advantage, but also thosewho have tried to use latent trait theory still usesome popular concepts in classical test theory, suchas reliability. expanded to testingprocedures in which different sets of items arepresented to different examinees. Examples aregiven. In this paper, the author points out some importantimplications in latent trait theory which arenot fully understood and appreciated among researchers,and gives some practical suggestions forits use.