IRT ESTIMATION OF DOMAIN SCORES

In classical test theory, a test is regarded as a sample of items from a domain defined by generating rules or by content, process, and form at specifications. If the items are a random sample of the domain, the n the percent-correct score on the test estimates the domain score, th at is, the expected percent correct for all items in the domain. When the domain is represented by a large set of calibrated items, as in it em banking applications, item response theory (IRT) provides an altern ative estimator of the domain score by transformation of the IRT scale score on the rest. This estimator has the advantage of not requiring the rest items to be a random sample of the domain, and of having a si mple standard error. We present here resampling results in real data d emonstrating for uni-and multidimensional models that the IRT estimato r is also a more accurate predictor of the domain score than is the cl assical percent-correct score. These results have implications for rep orting outcomes of educational qualification testing and assessment.