Within the framework of item response theory (IRT), there are two recent lines of work on the estimation of classification accuracy (CA) rate. One approach estimates CA when decisions are made based on total sum scores, the other based on latent trait estimates. The former is referred to as the Lee approach, and the latter, the Rudner approach,…

The purpose of this simulation study is to investigate the effects of anchor subtest composition on the accuracy of item response theory (IRT) likelihood ratio (LR) differential item functioning (DIF) detection (Thissen, Steinberg, & Wainer, 1988). Here, the IRT LR test was implemented with a free baseline approach wherein a baseline model was…

Douglas, Roussos, and Stout introduced the concept of differential bundle functioning (DBF) for identifying the underlying causes of differential item functioning (DIF). In this study, reference group was simulated to have higher mean ability than the focal group on a nuisance dimension, resulting in DIF for each of the multidimensional items…

This article describes SIMREL, a software program designed for the simulation of alpha coefficients and the estimation of its confidence intervals. SIMREL runs on two alternatives. In the first one, if SIMREL is run for a single data file, it performs descriptive statistics, principal components analysis, and variance analysis of the item scores…

Rudner (2001, 2005) described an expected classification accuracy index for determining the asymptotic expectation of accuracy of classifications of examinees into score categories. This article expands on that exposition by evaluating the index as it is likely to be used in practice (as a point estimate of classification accuracy), provides a…

Mokken scale analysis can be used for scaling under nonparametric item response theory models. The results may, however, not reflect the underlying dimensionality of data. Various features of Mokken scale analysis--the H coefficient, Mokken scale conditions, and algorithms--may explain this result. In this article, three new H-based objective…

In Ramsay-curve item response theory (RC-IRT), the latent variable distribution is estimated simultaneously with the item parameters of a unidimensional item response model using marginal maximum likelihood estimation. This study evaluates RC-IRT for the three-parameter logistic (3PL) model with comparisons to the normal model and to the empirical…

This study compares the performance of three nonparametric item characteristic curve (ICC) estimation procedures: isotonic regression, smoothed isotonic regression, and kernel smoothing. Smoothed isotonic regression, employed along with an appropriate kernel function, provides better estimates and also satisfies the assumption of strict…

Conducted a simulation study to determine how well two models for local item dependency, the constant order interaction and the dimension-dependent interaction models, could be distinguished. Results suggest that when there is good reason to suspect a dimension-dependent interaction, it is possible to detect it if the sample size is at least 500.…

Varied the number of items and categories per item to explore the effects on estimation of item parameters in the nominal response model. Simulation results show that increasing the number of items had little effect on item parameter recovery, but increasing the number of categories increased the error variance of the parameter estimates. (SLD)

Studied results from five linking methods under the graded-response model using simulated data. Results show that differences in the linking coefficients are small. The five methods yielded similar results for longer common-item links with large sample sizes and when the distribution of item-location parameters matched the underlying trait…

Investigated the relationship between sample size, latent trait distribution, and item parameter estimation with the nominal response model through simulation. Results suggest guidelines for reasonable item parameter estimation. (SLD)

General guidelines are developed to assist practitioners in devising operational computerized adaptive testing systems based on the graded response model. The effects of the following major variables were examined: item pool size; stepsize used along the trait continuum until maximum likelihood estimation could be calculated; and stopping rule…

Investigated Type I error rates of the likelihood-ratio test for the detection of differential item functioning (DIF) using Monte Carlo simulations under the graded-response model. Type I error rates were within theoretically expected values for all six combinations of sample sizes and ability-matching conditions at each of the nominal alpha…