The utility of the three-form design coupled with maximum likelihood methods for estimation of missing values was evaluated. Simulation studies demonstrate that maximum likelihood estimation and multiple imputation methods produce the most efficient and least biased estimates of variances and covariances for normally distributed and slightly…

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)

Assessed five procedures for estimating a common risk difference in a set of independent 2 x 2 tables through Monte Carlo simulation in terms of bias, efficiency, confidence level adjustment, and statistical power. The maximum likelihood estimator showed best performance, followed closely by the Cochran (W. Cochran, 1954) and Mantel-Haenszel (N.…

Provides asymptotic standard errors of the estimates of equated scores from several types of item response theory (IRT) true score equatings. Equating designs considered cover those with internal or external common items and separate or simultaneous estimation. Uses marginal maximum likelihood estimation for the estimation of item parameters. (SLD)

When the item response theory (IRT) model uses the marginal maximum likelihood estimation, person parameters are usually treated as random parameters following a certain distribution as a prior distribution to estimate the structural parameters in the model. For example, both PARSCALE (Muraki & Bock, 1999) and BILOG 3 (Mislevy & Bock,…

The likelihood ratio test statistic G[squared](dif) is widely used for comparing the fit of nested models in categorical data analysis. In large samples, this statistic is distributed as a chi-square with degrees of freedom equal to the difference in degrees of freedom between the tested models, but only if the least restrictive model is correctly…

When the underlying responses are on an ordinal scale, gamma is one of the most frequently used indices to measure the strength of association between two ordered variables. However, except for a brief mention on the use of the traditional interval estimator based on Wald's statistic, discussion of interval estimation of the gamma is limited.…

Using a set of scenarios derived from actual evaluation studies, this simulation study examines the reported influence of evaluation information on decision makers' potential actions. Each scenario described a context where one of three types of evaluation information (large-scale study data, case study data, or anecdotal accounts) is presented…

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…

When latent class parameters are estimated, maximum likelihood and Pearson chi-square statistics can be derived for assessing the fit of the model to the data. This study used simulated data to compare these two statistics, and is based on mixtures of latent binomial distributions, using data generated from five dichotomous manifest variables.…

This study investigated the efficiency of item selection in a computerized adaptive test (CAT), where efficiency was defined in terms of the accumulated test information at an examinee's true ability level. A simulation methodology compared the efficiency of 2 item selection procedures with 5 ability estimation procedures for CATs of 5, 10, 15,…

The analytically derived expected asymptotic standard errors (SEs) of maximum likelihood (ML) item estimates can be predicted by a mathematical function without examinees' responses to test items. The empirically determined SEs of marginal maximum likelihood estimation/Bayesian item estimates can be obtained when the same set of items is…

Type I error rates for the likelihood ratio test for detecting differential item functioning (DIF) were investigated using Monte Carlo simulations. Type I error rates for the two-parameter model were within theoretically expected values at each alpha level, but those for the three-parameter model were not. (SLD)

The robustness of the likelihood ratio difference statistic to the violation of a regularity condition when used to assess differences in fit provided by pairs of latent class models was investigated. Recommendations are made regarding the use of the statistic under violation of the regularity condition. (SLD)

An extension to the partial credit model, the linear partial credit model, is considered under the assumption of a certain linear decomposition of the item x category parameters into basic parameters. A conditional maximum likelihood algorithm for estimating basic parameters is presented and illustrated with simulation and an empirical study. (SLD)