In matrix comparison, the performance of four vector matching indices (the coefficient of congruence, the Pearson product moment correlation, the "s"-statistic, and kappa) was evaluated. Advantages and disadvantages of each index are discussed, and the performance of each was assessed within the framework of principal components…

Several methodological issues in the multidimensional scaling of coarse dissimilarities were studied, examining whether it was better to scale dissimilarity data directly or to scale a new matrix derived from the original by row comparisons. Findings support an alternative row-comparison measure based on the Jacard coefficient. (SLD)

A theorem of Spjotvoll (1972) was used to determine and apply a confidence interval for the difference of two multiple correlations based on observations from a single sample. Spjotvoll's method of comparing regression functions is also applicable to a comparison of dependent multiple correlations, an unsolved problem posed by Hotelling in 1940.…

Matrix differencing of data vectors is introduced as a method for computing test variance and is compared to traditional analysis of variance. Applications for computer assisted instruction, provided by supplemental computer software, are also described. (Author/GDC)

Rowwise matrix correlation, based on the weighted sum of correlations between all pairs of corresponding rows of two proximity matrices, is discussed. Rowwise and columnwise indices are particularly suited for evaluating different types of conjectures of a similar pattern of entries across the two matrices. (SLD)

H. F. Kaiser, S. Hunka, and J. Bianchini have presented a method (1971) to compare two matrices of factor loadings based on the same variables, but different groups of individuals. The optimal rotation involved is examined from a mathematical point of view, and the method is shown to be invalid. (SLD)

A method is proposed for comparison of distinct partitions of the same set of n objects through a simple cross-product index defined between corresponding entries from two proximity matrices providing particular a priori codings of the within-class and between-class relationships for the partitions. (SLD)

Various analytic choices in principal components and common factor analysis are discussed. Differences and similarities among these extraction methods are explained, and aids in interpreting the origin of detected effects are explored. Specifically, the nature and use of structure and pattern coefficients are examined. Communalities and methods…

A relatively simple technique for assessing the convergence of sets of variables across method domains is presented. The technique, two-step principal components analysis, empirically orthogonalizes each method domain into sets of components, and then analyzes convergence among components across domains. (Author)

Monte Carlo procedures were used to investigate the properties of a nonmetric multidimensional scaling algorithm when used to scale an incomplete matrix of dissimilarities. Recommendations for users wishing to scale incomplete matrices are made. (Author/RC)

Standard errors of pooled mean estimate in multiple matrix sampling were compared for two procedures. The data were from tests involving items with and without replacement. The two procedures involve the formulations of Madow and Lord, and Novick; the former permits sampling of item, with or without replacement, whereas the latter is to be used…

Results of empirical comparisons of an inferential model of order analysis with factor analytic models were reported for two sets of data. On the prestructured data set both order and factor analytic models returned its dimensions of length, width and height, but on the random data set the factor analytic models indicated the presence of…

A rapid method for hierarchically clustering the n objects of a matrix which portrays the interrelation of every object to every other object, where n equals any number up to 1,000 and even larger, is developed and discussed. Results compare favorably with those from other excellent methods. (Author/BJG)

Both the Rule Space Method (RSM) and the Neural Network Model (NNM) are techniques of statistical pattern recognition and classification approaches developed for applications from different fields. RSM was developed in the domain of educational statistics. It started from the use of an incidence matrix Q that characterizes the underlying cognitive…