This report summarizes the research findings of a four year contract investigating the applicability of item response theory and tailored testing to criterion-referenced measurement. Six major areas were studied on the project: (1) techniques for forming unidimensional item sets; (2) techniques for calibrating items; (3) item parameter linking…

This report describes the stratified adaptive (stradaptive) test as a strategy for tailoring an ability test to individual differences in testee ability; administration of the test is controlled by a time-shared computer system. The rationale of this method is described as it derives from Binet's strategy of ability test administration and…

The utility of item response theory (IRT) models in computerized adaptive tests is considered. Measurement questions that have been answered using IRT, and those that might be overlooked because of IRT, are reviewed. Areas in which fuller use of IRT could improve adaptive testing practices are identified. (SLD)

A study was conducted to evaluate the effects of choice of item response theory (IRT) model, parameter calibration group, starting ability estimate, and stopping criterion on the conversion of an 80-item vocabulary test to computer adaptive format. Three parameter calibration groups were tested: (1) a group of 1,000 high school seniors, (2) a…

The four purposes of this study were: (1) To compare two versions of a tailored testing model similar to one suggested by Kalisch (1974); (2) To identify levels of the variables within the two versions, which produce an efficient tailored testing procedures; (3) To compare, within each version, the results obtained when employing relatively small…

Two tailored testing models, specifying procedures by which the correctness of examinees' responses to a fixed number of test items are predicted by presenting as few items as possible to the examinee, were compared for their efficiency. The models differ in that one requires reconsideration of each prediction whenever additional information is…

Items with the highest discrimination parameter values in a logistic item response theory (IRT) model do not necessarily give maximum information. This paper shows which discrimination parameter values (as a function of the guessing parameter and the distance between person ability and item difficulty) give maximum information for the…

The quality of fit between the data and the measurement model is fundamental to any discussion of results. Fit has been the subject of inquiry since as early as the 1920s. Most early explorations concentrated on assessing global fit or subset fits on fixed length, traditional paper and pencil tests given as a single unit. The detection of aberrant…

Expert systems can be used to aid decisionmaking. A computerized adaptive test is one kind of expert system, although not commonly recognized as such. A new approach, termed EXSPRT, was devised that combines expert systems reasoning and sequential probability ratio test stopping rules. Two versions of EXSPRT were developed, one with random…

The use of computers to build diagnostic inferences is presented in two contexts: (1) closed world, exemplified by the space shuttle launch monitoring system; and (2) open world, represented by computerized diagnostic testing of reading comprehension. The analysis shows that the closed world provides a substantially cleaner environment within…

An application of the two-paramenter logistic (Rasch) model to tailored testing is presented. The model is discussed along with the maximum likelihood estimation of the ability parameters given the response pattern and easiness parameter estimates for the items. The technique has been programmed for use with an interactive computer terminal. Use…

Test-item writing efforts typically results in item pools with an undesirable correlational structure between the content attributes of the items and their statistical information. If such pools are used in computerized adaptive testing (CAT), the algorithm may be forced to select items with less than optimal information, that violate the content…

When an examination consists, in whole or in part, of constructed response items, it is a common practice to allow the examinee to choose among a variety of questions. This procedure is usually adopted so that the limited number of items that can be completed in the allotted time does not unfairly affect the examinee. This results in the de facto…

A brief overview of item response theory is provided, and a 186-item bibliography of books and articles on the subject dating from 1953 to June 1989 is presented. The overview includes a definition of the theory, a discussion of its development and application, and comparisons with classical test theory. All publications in the bibliography were…

A tailored testing model employing the beta distribution, whose mean equals the difficulty of an item and whose variance is approximately equal to the sampling variance of the item difficulty, and employing conditional item difficulties, is proposed. The model provides a procedure by which a minimum number of items of a test, consisting of a set…