In this article, we extend the methodology of the Cut-Score Operating Function that we introduced previously and apply it to a testing scenario with multiple independent components and different testing policies. We derive analytically the overall classification error rate for a test battery under the policy when several retakes are allowed for…

An example demonstrates and explains that summary statistics commonly used to measure test quality can be seriously misleading and that summary statistics for the whole test are not sufficient for judging the quality of the test. (Author/LMO)

This paper is the transcript of a talk given to those who use test information but who have little technical background in test theory. The concepts of modern test theory are compared with traditional test theory, as well as a probable future test theory. The explanations given are couched within an extended metaphor that allows a full description…

The reliability of scores from four forms of the Test of English as a Foreign Language (TOEFL) was estimated using a hybrid item response theory model. It was found that there was very little difference between overall reliability when the testlet items were assumed to be independent and when their dependence was modeled. (Author/SLD)

Techniques derived from item response theory are useful for estimating the reliability of test classification above and below the cutting score. Test developers can construct a test whose information is peaked in the region of the cutting score; users can select a test which provides the most information in this region. The Cut-Score…

It is sometimes sensible to think of the fundamental unit of test construction as being larger than an individual item. This unit, dubbed the testlet, must pass muster in the same way that items do. One criterion of a good item is the absence of differential item functioning (DIF). The item must function in the same way as all important…

Two estimation procedures for the Rasch Model of test analysis are reviewed in detail, particularly with respect to new developments that make the more statistically rigorous conditional maximum likelihood estimation practical for use with longish tests. (Author/JKS)

The critique of the Massachusetts Teacher Tests by W. Haney and others points out some flaws in the tests but ignores the fact that the tests provide some useful information to guide teacher selection decisions. Calls for additional study of these teacher evaluation instruments. (SLD)

Because assessment instruments of the future may well be composed of a combination of types of questions, a way to combine those scores effectively is discussed. Two new graphic tools are presented that show that it may not be practical to equalize the reliability of different components. (SLD)

Two attempts to adjust state mean Scholastic Aptitude Test (SAT) scores for differential participation rates are examined. Both attempts are rejected, and five rules for performing adjustments are outlined to foster follow-up checks on untested assumptions. National Assessment of Educational Progress state data are determined to be more accurate.…

Discusses changes in testing as a result of the availability of extensive inexpensive computing and some recent developments in statistical test theory. Describes the role of the Computerized Adaptive Test (CAT) and modern Item Response Theory (IRT) in ability testing tailored to each student's knowledge and ability. (JAC)