A study was conducted to investigate whether augmenting the calibration of items using computerized adaptive test (CAT) data matrices produced estimates that were unbiased and improved the stability of existing item parameter estimates. Item parameter estimates from four pools of items constructed for operational use were used in the study to…

This paper suggests several item selection criteria for adaptive testing that are all based on the use of the true posterior. Some of the ability estimators produced by these criteria are discussed and empirically criticized. (SLD)

Used simulations of computerized adaptive tests to evaluate results yielded by four commonly used ability estimation methods: maximum likelihood estimation (MLE) and three Bayesian approaches. Results show clear distinctions between MLE and Bayesian methods. (SLD)

Expected a posteriori has a number of advantages over maximum likelihood estimation or maximum a posteriori (MAP) estimation methods. These include ability estimates (thetas) for all response patterns, less regression towards the mean than MAP ability estimates, and a lower average squared error. R. D. Bock and R. J. Mislevy (1982) state that the…

In a variable-length adaptive test with a stopping rule that relied on the asymptotic standard error of measurement of the examinee's estimated true score, M. S. Stocking (1987) discovered that it was sufficient to know the examinee's true score and the number of items administered to predict with some accuracy whether an examinee's true score was…

Derives three reliability estimates for the Bayes modal estimate (BME) and the maximum-likelihood estimate (MLE) of theta in computerized adaptive tests (CATs). Computes the three reliability estimates and the true reliabilities of both BME and MLE for seven simulated CATs. Results show the true reliabilities for BME and MLE to be nearly identical…

Extended the use of a beta prior in trait estimation to the maximum expected a posteriori (MAP) method of Bayesian estimation. This new method, essentially unbiased MAP, was compared with MAP, essentially unbiased expected a posteriori, weighted likelihood, and maximum-likelihood estimation methods. The new method significantly reduced bias in…

Proposes a procedure for empirical initialization of the trait (theta) estimator in adaptive testing that is based on the statistical relation between theta and background variables known prior to test administration. Illustrates the procedure for an adaptive version of a test from the Dutch General Aptitude Battery. (SLD)

Some alternative criteria for item selection in adaptive testing are proposed that take into account uncertainty in the ability estimates. A simulation study shows that the likelihood weighted information criterion is a good alternative to the maximum information criterion. Another good alternative uses a Bayesian expected a posteriori estimator.…

Two simple empirical approximate Bayes estimators are introduced for estimating domain scores under binomial and hypergeometric distributions respectively. Criteria are established regarding use of these functions over maximum likelihood estimation counterparts. (SLD)

This paper describes a study, now in progress, of new methods for representing the sampling variability of Mantel-Haenszel differential item functioning (DIF) results, based on the system for categorizing the severity of DIF that is now in place at the Educational Testing Service. The methods, which involve a Bayesian elaboration of procedures…

R. J. Owen (1975) proposed an approximate empirical Bayes procedure for item selection in adaptive testing. The procedure replaces the true posterior by a normal approximation with closed-form expressions for its first two moments. This approximation was necessary to minimize the computational complexity involved in a fully Bayesian approach, but…

The effect of dimensionality on an adaptive test's ability estimation was examined. Two-dimensional data sets, which differed from one another in the interdimensional ability association, the correlation among the difficulty parameters, and whether the item discriminations were or were not confounded with item difficulty, were generated for 1,600…

F. M. Lord's flexilevel, computerized adaptive testing (CAT) procedure was compared to an item-response theory-based CAT procedure that uses Bayesian ability estimation with various standard errors of estimates used for terminating the test. Ability estimates of flexilevel CATs were as accurate as were those of Bayesian CATs. (TJH)

In this study some alternative item selection criteria for adaptive testing are proposed. These criteria take into account the uncertainty of the ability estimates. A general weighted information criterion is suggested of which the usual maximum information criterion and the suggested alternative criteria are special cases. A simulation study was…