Presents a generalization and improvement for the definition proposed by E. Dzhafarov (2001) for selectiveness in the dependence of several random variables on several (sets of) external factors. This generalization links the notion of selective influence with that of conditional independence. (SLD)

Uses classical test theory to show that it is the nature of the range restriction, rather than the nature of the available reliability coefficient, that determines the sequence for applying corrections for range restriction and unreliability. Shows how the common rule of thumb for choosing the sequence is tenable only when the correction does not…

This paper studies the general theory of Akaike's Information Criterion (AIC) and provides two analytical extensions. The extensions make AIC asymptotically consistent and penalize overparameterization more stringently to pick only the simplest of the two models. The criteria are applied in two Monte Carlo experiments. (Author/GDC)

Takes several goodness-of-fit statistics as measures of variable selection and develops backward elimination and forward selection procedures in exploratory factor analysis. A newly developed variable selection program, SEFA, can print several fit measures for a current model and models obtained by removing an internal variable or adding an…

A Bayesian approach for simultaneous optimization of test-based decisions is presented using the example of a selection decision for a treatment followed by a mastery decision. A distinction is made between weak and strong rules, and conditions for monotonicity of optimal weak and strong rules are presented. (Author/SLD)

Presents a variable selection heuristic for nonhierarchical (K-means) cluster analysis based on the adjusted Rand index for measuring cluster recovery. Subjected the heuristic to Monte Carlo testing across more than 2,200 datasets. Results indicate that the heuristic is extremely effective at eliminating masking variables. (SLD)

Presented nonparametric tests for testing the validity of polytomous unidimensional ordinal probabilistic polytomous item response theory models along with procedures for testing the comonotonicity of two item sets and for item selection. Describes advantages of the new approach. (SLD)

The Akaike Information Criterion (AIC) was introduced to extend the method of maximum likelihood to the multimodel situation. Use of the AIC in factor analysis is interesting when it is viewed as the choice of a Bayesian model; thus, wider applications of AIC are possible. (Author/GDC)

A linear utility model is introduced for optimal selection where several subpopulations of applicants are to be distinguished. Using this model, procedures are described for obtaining optimal cutting scores in subpopulations in quota-free as well as quota-restricted selection situations. The procedures are demonstrated with empirical data.…

A theoretical framework is developed in which the effects of some common forms of violation of assumptions of linearity of regression and homoscedasticity can be investigated. Simple expressions are derived for the restricted and corrected correlations in terms of the target (unrestricted) correlation in these situations. (SLD)

A procedure for inferring the validity of a selection test as a predictor of some criterion when the available data are limited due to prior selection is described. (Author/JKS)

Multivariate selection can be represented as a linear transformation in a geometric framework. In this note this approach is extended to describe the effects of selection on regression analysis and to adjust for the effects of selection using the inverse of the linear transformation. (Author/BW)

Presents a latent variable representation for multiple-choice items and option characteristic curves, and proposes a criterion for distractors based on distractor selection ratios. Results allow for testing the criterion from observable data without specifying a parametric form for the characteristic curves. (Author/SLD)

Demonstrates mathematically that if every item in an item pool has an equal possibility to be selected from the pool in a fixed-length computerized adaptive test, the number of overlapping items among an alpha randomly sampled examinees follows the hypergeometric distribution family for alpha greater than or equal to 1. (SLD)

Studied whether ordered latent class models can be used as nonparametric item response theory (NIRT) models to scale polytomous models. Simulation findings show that the Bayesian estimation method presented can handle the inequality restrictions on the parameters and the sparseness of the data quite well. (SLD)