The fuzzy perspective in statistical analysis is first illustrated with reference to the “Informational Paradigm" allowing us to deal with different types of uncertainties related to the various informational ingredients (data, model, assumptions). The fuzzy empirical data are then introduced, referring to J LR fuzzy variables as observed on I observation units. Each observation is characterized by its center and its left and right spreads (LR1 fuzzy number) or by its left and right “centers" and its left and right spreads (LR2 fuzzy number). Two types of component models for LR1 and LR2 fuzzy data are proposed. The estimation of the parameters of these models is based on a Least Squares approach, exploiting an appropriately introduced distance measure for fuzzy data. A simulation study is carried out in order to assess the efficacy of the suggested models as compared with traditional Principal Component Analysis on the centers and with existing methods for fuzzy and interval valued data. An application to real fuzzy data is finally performed.

This paper deals with optimal partitioning of limited testing time in order to achieve maximum total test score. Nonlinear optimization theory was used to analyze this problem. A general case using a generic item response model is first presented. A special case that applies a response time model proposed by Wang and Hanson (2005) is also presented. Theoretical properties of the optimal solution are derived. Their practical implications to optimal test-taking strategies are also discussed. The theoretical properties are in general agreement with the conventional advice to the examinees on pacing strategy.

A method is presented for generalized canonical correlation analysis of two or more matrices with missing rows. The method is a combination of Carroll’s (1968) method and the missing data approach of the OVERALS technique (Van der Burg, 1988). In a simulation study we assess the performance of the method and compare it to an existing procedure called GENCOM, proposed by Green and Carroll (1988). We find that the proposed method outperforms the GENCOM algorithm both with respect to model fit and recovery of the true structure.

First, I thank Sijtsma, Clark, Kane, and Heiser for taking the time and effort to provide a commentary on my paper, and the Editor for allowing me to respond to them. In general, the commentators agree with the thesis that psychometrics and psychology are, to an extent that must be deemed problematic, disconnected. They further agree with the upshot of this diagnosis: Psychometricians need to work harder to make a difference in psychology, and psychologists need to develop a greater awareness of important psychometric developments. However, the commentators also raise several points of disagreement and criticism.

Taxicab correspondence analysis is based on the taxicab singular value decomposition of a contingency table, and it shares some similar properties with correspondence analysis. It is more robust than the ordinary correspondence analysis, because it gives uniform weights to all the points. The visual map constructed by taxicab correspondence analysis has a larger sweep and clearer perspective than the map obtained by correspondence analysis. Two examples are provided.

Nonparametric item response models have been developed as alternatives to the relatively inflexible parametric item response models. An open question is whether it is possible and practical to administer computerized adaptive testing with nonparametric models. This paper explores the possibility of computerized adaptive testing when using nonparametric item response models. A central issue is that the derivatives of item characteristic Curves may not be estimated well, which eliminates the availability of the standard maximum Fisher information criterion. As alternatives, procedures based on Shannon entropy and Kullback–Leibler information are proposed. For a long test, these procedures, which do not require the derivatives of the item characteristic eurves, become equivalent to the maximum Fisher information criterion. A simulation study is conducted to study the behavior of these two procedures, compared with random item selection. The study shows that the procedures based on Shannon entropy and Kullback–Leibler information perform similarly in terms of root mean square error, and perform much better than random item selection. The study also shows that item exposure rates need to be addressed for these methods to be practical.

Stevens postulated that we can use the responses of a participant in a ratio scaling experiment directly to construct a psychophysical function representing the participant’s sensations. Although Stevens’ methods of constructing measurement scales are widely used in the behavioral sciences, the problem of which scale type is appropriate to describe ratio scaling data is still unresolved. To deal with this problem, we develop a theoretical framework to specify the scale type attained by Stevens’ direct scaling methods. It is shown, under fairly mild background assumptions, that the behavioral axioms presented in this paper are necessary and sufficient for the psychophysical functions to be ordinal-, interval-, log-interval-, or ratio-scales. Furthermore, suggestions on how to test these behavioral axioms are provided.

A rationale is proposed for approximating the normal distribution with a logistic distribution using a scaling constant based on minimizing the Kullback-Leibler (KL) information, that is, the expected amount of information available in a sample to distinguish between two competing distributions using a likelihood ratio (LR) test, assuming one of them is true. The new constant 1.749, computed assuming the normal distribution is true, yields an approximation that is an improvement in fit of the tails of the distribution as compared to the minimax constant of 1.702, widely used in item response theory (IRT). The minimax constant is by definition marginally better in its overall maximal error. It is argued that the KL constant is more statistically appropriate for use in IRT.

Personality tests often consist of a set of dichotomous or Likert items. These response formats are known to be susceptible to an agreeing-response bias called acquiescence. The common assumption in balanced scales is that the sum of appropriately reversed responses should be reasonably free of acquiescence. However, inter-item correlation (or covariance) matrices can still be affected by the presence of variance due to acquiescence. To analyse these correlation matrices, we propose a method that is based on an unrestricted factor analysis and can be applied to multidimensional scales. This method obtains a factor solution in which acquiescence response variance is isolated in an independent factor. It is therefore possible, without the potentially confounding effect of acquiescence, to: (a) examine the dominant factors related to content latent variables; and (b) estimate participants–factor scores on content latent variables. This method, which is illustrated by two empirical data examples, has proved to be useful for improving the simplicity of the factor structure.

Minimization of the within-cluster sums of squares (WCSS) is one of the most important optimization criteria in cluster analysis. Although cluster analysis modules in commercial software packages typically use heuristic methods for this criterion, optimal approaches can be computationally feasible for problems of modest size. This paper presents a new branch-and-bound algorithm for minimizing WCSS. Algorithmic enhancements include an effective reordering of objects and a repetitive solution approach that precludes the need for splitting the data set, while maintaining strong bounds throughout the solution process. The new algorithm provided optimal solutions for problems with up to 240 objects and eight well-separated clusters. Poorly separated problems with no inherent cluster structure were optimally solved for up to 60 objects and six clusters. The repetitive branch-and-bound algorithm was also successfully applied to three empirical data sets from the classification literature.

Mean comparisons are of great importance in the application of statistics. Procedures for mean comparison with manifest variables have been well studied. However, few rigorous studies have been conducted on mean comparisons with latent variables, although the methodology has been widely used and documented. This paper studies the commonly used statistics in latent variable mean modeling and compares them with parallel manifest variable statistics. Our results indicate that, under certain conditions, the likelihood ratio and Wald statistics used for latent mean comparisons do not always have greater power than the Hotelling T2 statistics used for manifest mean comparisons. The noncentrality parameter corresponding to the T2 statistic can be much greater than those corresponding to the likelihood ratio and Wald statistics, which we find to be different from those provided in the literature. Under a fixed alternative hypothesis, our results also indicate that the likelihood ratio statistic can be stochastically much greater than the corresponding Wald statistic. The robustness property of each statistic is also explored when the model is misspecified or when data are nonnormally distributed. Recommendations and advice are provided for the use of each statistic.

The Candecomp/Parafac (CP) model decomposes a three-way array into a prespecified number R of rank-1 arrays and a residual array, in which the sum of squares of the residual array is minimized. The practical use of CP is sometimes complicated by the occurrence of so-called degenerate solutions, in which some components are highly correlated in all three modes and the elements of these components become arbitrarily large. We consider the real-valued CP model in which p × p × 2 arrays of rank p + 1 or higher are decomposed into p rank-1 arrays and a residual array. It is shown that the CP objective function does not have a minimum in these cases, but an infimum. Moreover, any sequence of CP approximations, of which the objective value approaches the infimum, will become degenerate. This result extends Ten Berge, Kiers, & De Leeuw (1988), who consider a particular 2 × 2 × 2 array of rank 3.

An extension of multiple correspondence analysis is proposed that takes into account cluster-level heterogeneity in respondents’ preferences/choices. The method involves combining multiple correspondence analysis and k-means in a unified framework. The former is used for uncovering a low-dimensional space of multivariate categorical variables while the latter is used for identifying relatively homogeneous clusters of respondents. The proposed method offers an integrated graphical display that provides information on cluster-based structures inherent in multivariate categorical data as well as the interdependencies among the data. An empirical application is presented which demonstrates the usefulness of the proposed method and how it compares to several extant approaches.

The assumptions of the model for factor analysis do not exclude a class of indeterminate covariances between factors and error variables (Grayson, 2003). The construction of all factors of the model for factor analysis is generalized to incorporate indeterminate factor-error covariances. A necessary and sufficient condition is given for indeterminate factor-error covariances to be arbitrarily small, for mean square convergence of the regression predictor of factor scores, and for the existence of a unique determinate factor and error variable. The determinate factor and error variable are uncorrelated and satisfy the defining assumptions of factor analysis. Several examples are given to illustrate the results.

We describe a principled way of imposing a metric representing dissimilarities on any discrete set of stimuli (symbols, handwritings, consumer products, X-ray films, etc.), given the probabilities with which they are discriminated from each other by a perceiving system, such as an organism, person, group of experts, neuronal structure, technical device, or even an abstract computational algorithm. In this procedure one does not have to assume that discrimination probabilities are monotonically related to distances, or that the distances belong to a predefined class of metrics, such as Minkowski. Discrimination probabilities do not have to be symmetric, the probability of discriminating an object from itself need not be a constant, and discrimination probabilities are allowed to be 0’s and 1’s. The only requirement that has to be satisfied is Regular Minimality, a principle we consider the defining property of discrimination: for ordered stimulus pairs (a, b), b is least frequently discriminated from a if and only if a is least frequently discriminated from b. Regular Minimality generalizes one of the weak consequences of the assumption that discrimination probabilities are monotonically related to distances: the probability of discriminating a from a should be less than that of discriminating a from any other object. This special form of Regular Minimality also underlies such traditional analyses of discrimination probabilities as Multidimensional Scaling and Cluster Analysis.

The present paper shows that the usual factor analytic structured data dispersion matrix Λ Ψ Λ’ + Δ can readily arise from a set of scores y = Λ η + ε, where the “common” (η) and “unique” (ε) factors have nonzero covariance: Γ = Cov(ε,η) ≠ 0. Implications of this finding are discussed for the indeterminacy of factor scores, and for the issue of invariance of factor analytic covariance models. The size of the problem is explored with numerical examples.

Component loss functions (CLFs) similar to those used in orthogonal rotation are introduced to define criteria for oblique rotation in factor analysis. It is shown how the shape of the CLF affects the performance of the criterion it defines. For example, it is shown that monotone concave CLFs give criteria that are minimized by loadings with perfect simple structure when such loadings exist. Moreover, if the CLFs are strictly concave, minimizing must produce perfect simple structure whenever it exists. Examples show that methods defined by concave CLFs perform well much more generally. While it appears important to use a concave CLF, the specific CLF used is less important. For example, the very simple linear CLF gives a rotation method that can easily outperform the most popular oblique rotation methods promax and quartimin and is competitive with the more complex simplimax and geomin methods.

When a simple random sample of size n is employed to establish a classification rule for prediction of a polytomous variable by an independent variable, the best achievable rate of misclassification is higher than the corresponding best achievable rate if the conditional probability distribution is known for the predicted variable given the independent variable. In typical cases, this increased misclassification rate due to sampling is remarkably small relative to other increases in expected measures of prediction accuracy due to samplings that are typically encountered in statistical analysis.

This issue is particularly striking if a polytomous variable predicts a polytomous variable, for the excess misclassification rate due to estimation approaches 0 at an exponential rate as n increases. Even with a continuous real predictor and with simple nonparametric methods, it is typically not difficult to achieve an excess misclassification rate on the order of n−1. Although reduced excess error is normally desirable, it may reasonably be argued that, in the case of classification, the reduction in bias is related to a more fundamental lack of sensitivity of misclassification error to the quality of the prediction. This lack of sensitivity is not an issue if criteria based on probability prediction such as logarithmic penalty or least squares are employed, but the latter measures typically involve more substantial issues of bias. With polytomous predictors, excess expected errors due to sampling are typically of order n−1. For a continuous real predictor, the increase in expected error is typically of order n−2/3.