The shape of the response surface associated with a discriminant analysis provides insight into the value of the derived optimal discriminant variates. A procedure for the determination of "indifference regions," presented in this article, allows the assessment of the degree of flatness of the response surface for any analysis.…

Five external criteria were used to evaluate the extent of recovery of the true structure in a hierarchical clustering solution. The results of the study indicated that the Hubert and Arabie adjusted Rank index was best suited to the task of comparison across hierarchy levels. (Author/LMO)

The use of generalized discriminant analysis as a descriptive technique which can be employed outside of the traditional analysis of variance studies is discussed. Examples based on real data are provided. (Author/JKS)

The logic and procedures underlying a disturbance term regression test of logical consistency for structural models are reviewed for recursive and nonrecursive designs. It is shown that in a simple three-variable, complete mediational case the test procedure is mathematically equivalent to a part correlation. (Author/LMO)

In order to make parallel analysis more accessible to researchers employing principal component techniques, regression equations are presented for the logarithms of the latent roots of random data correlation matrices with unities on the diagonal. (Author/LMO)

Twenty-three different methods of cluster analysis were compared in a four-stage sequential validation design. Results demonstrated that the solution given by Ward's method of cluster analysis was particularly powerful in comparison to solutions yielded by other techniques. (Author/JKS)

Two applications of Kristof's theorem on traces of matrix products are presented in order to highlight their utility for psychometric theory and studies. (Author/JKS)

Principal component analysis, image component analysis, and maximum likelihood factor analysis were compared to assess the effects of variable sampling. Results with respect to degree of saturation and average number of variables per factor were clear and dramatic. Differential effects on boundary cases and nonconvergence problems were also found.…

In multigroup covariance structure analysis with structured means, the traditional latent selection model is formulated as a special case of phenotypic selection. Illustrations with real and simulated data demonstrate how one can test specific hypotheses concerning selection on latent variables. (SLD)

Three methods of transforming unordered categorical response variables are described: (1) analysis using dummy variables; (2) eigenanalysis of frequency patterns scaled relative to within-groups variance; (3) categorical variables analyzed separately with scale values generated so that the grouping variable and the categorical variable are…

Decomposable models, which allow for the interdependence of structure among observable variables, are described. Each model is fully characterized by a set of conditional interdependence restrictions and can be visualized with an undirected as well as a special type of directed graph. (Author/JKS)

An approach to determining the number of factors to be retained from a factor analysis using split-half factor comparabilities is presented. The use of this approach in determining proper factor rotation is also discussed. (JKS)

The misleading character of the correlation coefficient was investigated in two studies of intuitive statistical behavior: subjective estimation of partial correlation and subjective estimation of the minimum possible correlation between two variables given their equal correlation with a third. (Author/JKS)

The present study compares the performance of phi coefficients and tetrachorics along two dimensions of factor recovery in binary data. These dimensions are (1) accuracy of nontrivial factor identifications; and (2) factor structure recovery given a priori knowledge of the correct number of factors to rotate. (Author/LMO)

Three estimates of the probabilities of correct classification in predictive discriminant analysis were computed using mathematical formulas, resubstitution, and external analyses: (1) optimal hit rate; (2) actual hit rate; and (3) expected actual hit rate. Methods were compared using Monte Carlo sampling from two data sets. (Author/GDC)