A difficult result to interpret in Computerized Adaptive Tests (CATs) occurs when an ability estimate initially drops and then ascends continuously until the test ends, suggesting that the true ability may be higher than implied by the final estimate. This study explains why this asymmetry occurs and shows that early mistakes by high-ability…

A general structural equation model for measuring residualized true change and studying patterns of true growth or decline is described. This approach allows consistent and efficient estimation of the degree of interrelationship between residualized change in a repeatedly assessed psychological construct and other variables. (SLD)

Model checking in item response theory (IRT) is an underdeveloped area. There is no universally accepted tool for checking IRT models. The posterior predictive model-checking method is a popular Bayesian model-checking tool because it has intuitive appeal, is simple to apply, has a strong theoretical basis, and can provide graphical or numerical…

Path Analysis of Covariance Matrix (PACM) is described as a way to separately estimate measurement and structural models using standard least squares procedures. PACM was empirically compared to simultaneous maximum likelihood estimation and use of the LISREL computer program, and its advantages are identified. (SLD)

The form of item log-likelihood surface was investigated under two-parameter and three-parameter logistic models. Results confirm that the LOGIST program procedures used to locate the maximum of the likelihood functions are consistent with the form of the item log-likelihood surface. (SLD)

Analysis of hypothetical data with dependent responses demonstrates how information loss caused by the overt response scale has an unknown influence on effect sizes in moderated regression analysis. The number of scale steps measuring the dependent variable results in a form of systematic error that alters interaction effect sizes. (SLD)

Inferential conditions in the statistical detection of measurement bias are discussed in the contexts of differential item functioning and predictive bias in educational and employment settings. It is concluded that bias measures that rely strictly on observed measures are not generally diagnostic of measurement bias or lack of bias. (SLD)

The reliability of derived measures from 4 cognitive paradigms was studied using 19 Navy enlisted men (aged between 18 and 24 years). The paradigms were: graphemic and phonemic analysis; semantic memory retrieval; lexical decision making; and letter classification. Results indicate that derived scores may have low reliability. (SLD)