A modified and improved inductive inferential approach to evaluate item discriminations in a conditional maximum likelihood and Rasch modeling framework is suggested. The new approach involves the derivation of four hypothesis tests. It implies a linear restriction of the assumed set of probability distributions in the classical approach that…

In standard item response theory (IRT) applications, the latent variable is typically assumed to be normally distributed. If the normality assumption is violated, the item parameter estimates can become biased. Summed score likelihood based statistics may be useful for testing latent variable distribution fit. We develop Satorra-Bentler type…

Standard approaches for estimating item response theory (IRT) model parameters generally work under the assumption that the latent trait being measured by a set of items follows the normal distribution. Estimation of IRT parameters in the presence of nonnormal latent traits has been shown to generate biased person and item parameter estimates. A…

A simulation study was conducted to explore the performance of the linear logistic test model (LLTM) when the relationships between items and cognitive components were misspecified. Factors manipulated included percent of misspecification (0%, 1%, 5%, 10%, and 15%), form of misspecification (under-specification, balanced misspecification, and…

The usefulness of the l[subscript z] person-fit index was investigated with achievement test data from 20 exams given to more than 3,200 college students. Results for three methods of estimating ? showed that the distributions of l[subscript z] were not consistent with its theoretical distribution, resulting in general overfit to the item response…

We propose a new limited-information goodness of fit test statistic C[subscript 2] for ordinal IRT models. The construction of the new statistic lies formally between the M[subscript 2] statistic of Maydeu-Olivares and Joe (2006), which utilizes first and second order marginal probabilities, and the M*[subscript 2] statistic of Cai and Hansen…

Compared approaches to determining the precision of gain scores: (1) the asymptotic normal distribution of the maximum likelihood estimator of the person parameter; and (2) the exact conditional distribution of the gain score. Use of three data sets illustrates that these methods yield more relevant and more detailed information than traditional…

The extent to which properties of the ability scale and the form of the latent trait distribution influence the estimated item parameters of item response theory (IRT) was investigated using real and simulated data. Simulated data included 5,000 ability values randomly drawn from the standard normal distribution. Real data included the results for…

Studied the extent to which three appropriateness indexes, Z(sub 3), ECIZ4, and W, are well standardized in a Monte Carlo study. The ECIZ4 most closely approximated a normal distribution, and its skewness and kurtosis were more stable and less affected by test length and item response theory model than the others. (SLD)

Previous work on the mathematical and implementation details of the marginalized maximum likelihood estimation procedure is extended to encompass the marginalized Bayesian procedure for estimating item parameters of R. J. Mislevy (1986) and to communicate this procedure to users of the BILOG computer program. (SLD)

The sensitivity of marginal maximum likelihood estimation of item and ability (theta) parameters was examined when prior ability distributions were not matched to underlying ability distributions. Thirty sets of 45-item test data were generated. Conditions affecting the accuracy of estimation are discussed. (SLD)

A model developed by G. Rasch that assumes scores on some attainment tests can be realizations of a Poisson process is explained and expanded by assuming a prior distribution, with fixed but unknown parameters, for the subject parameters. How additional between-subject and within-subject factors can be incorporated is discussed. (SLD)

Detecting differential item functioning (DIF) on test items constructed to favor 1 group over another was investigated on parameter estimates from 2 item response theory-based computer programs--BILOG and LOGIST--using data for 1,000 White and 1,000 Black college students. Use of prior distributions and marginal-maximum a posteriori estimation is…