An adaptive test can usually match or exceed the measurement precision of conventional tests several times its length. This increased efficiency is not without costs, however, as the models underlying adaptive testing make strong assumptions about examinees and items. Most troublesome is the assumption that item pools are unidimensional. Truly…

An adaptive testing algorithm is presented based on an alternative modeling framework, and its effectiveness is investigated in a simulation based on real data. The algorithm uses a latent class modeling framework in which assessed latent attributes are assumed to be categorical variables. (SLD)

A method is presented and illustrated for simultaneously generating multiple tests with similar characteristics from the item bank by using binary programing techniques. The parallel tests are created to match an existing seed test item for item and to match user-supplied taxonomic specifications. (SLD)

Mixed integer linear programing models for customizing two-stage tests are presented. Model constraints are imposed with respect to test composition, administration time, inter-item dependencies, and other practical considerations. The models can be modified for use in the construction of multistage tests. (Author/TJH)

A brief overview is provided of the Conger-Lipshitz approach to estimating the reliability of a profile or test battery. A computational example from a recent study shows how canonical reliability can be obtained through existing statistical software. (SLD)

Linear programming models with 0-1 variables are useful for the construction of tests from an item bank. Most solution strategies for these models start with solving the relaxed 0-1 linear programming model, allowing the 0-1 variables to take on values between 0 and 1. Then, a 0-1 solution is found by just rounding, optimal rounding, or a…

Test validity is a concept that has often been ignored in the context of latent trait models and in modern test theory, particularly as it relates to computerized adaptive testing. Some considerations about the validity of a test and of a single item are proposed. This paper focuses on measures that are population-free and that will provide local…

A method is proposed that increases the accuracies of estimation of the operating characteristics of discrete item responses, especially when the true operating characteristic is represented by a steep curve, and also at the lower and upper ends of the ability distribution where the estimation tends to be inaccurate because of the smaller number…

A procedure is introduced for determining the effect of testlet nonequivalence on operating characteristics of a testlet-based computerized mastery test (CMT). The procedure, which involves estimating the CMT decision rule twice with testlet likelihoods treated as equivalent or nonequivalent, is demonstrated with testlet pools from the Architect…

A theoretical framework for mastery testing based on item response theory and Bayesian decision theory is described and illustrated. Implementation depends on the availability of (1) a computerized test delivery system; (2) a pool of pretested items; and (3) a model relating observed test performance to true mastery status. (SLD)

The Rasch Model and various extensions of this model can be formulated as a quasi loglinear model for the incomplete subgroup x score x item response 1 x ... x item response k contingency table. By comparing various loglinear models, specific deviations of the Rasch model can be tested. Parameter estimates can be computed using programs such as…

Derivations of several item selection algorithms for use in fitting test items to target information functions (IFs) are described. These algorithms, which use an average growth approximation of target IFs, were tested by generating six test forms and were found to provide reliable fit. (SLD)

A new test construction model based on the Rasch model is proposed. This model, the cluster-based method, considers groups of interchangeable items rather than individual items and uses integer programing. Results for six test construction problems indicate that the method produces accurate results in small amounts of time. (SLD)

The derivation of several item selection algorithms for use in fitting test items to target information functions is described. These algorithms circumvent iterative solutions by using the criteria of moving averages of the distance to a target information function and simultaneously considering an entire range of ability points used to condition…

A study involving 1,000 simulated examinees compared the partial credit and graded response models in computerized adaptive testing (CAT). The graded response model fit the data well and provided slightly more accurate ability estimates than those of the partial credit model. Benefits of polytomous model-based CATs are discussed. (SLD)