Investigates how students interpret and analyze multivariate data organized in tables and the nature of external visual displays that they tend to create and use for this purpose. Reports that: (1) students based their conclusions on only part of the data; (2) students did not use either efficient or sufficient visual representations; and (3)…

Provides a nontechnical explanation of the family of multivariate statistical techniques known as multidimensional scaling (MDS). Explains the purpose of MDS, and the advantages of using these techniques in the study of musical perception. Discusses two approaches to the interpretation of MDS solutions. Lists examples of musical studies using MDS.…

A response to comments on internal correlation for statistical analysis, as proposed by the present authors (1989), is provided. Focus is on issues raised by W. W. Rozeboom (1989). Comments by J. H. Schuenemeyer (1989) and R. Bargmann (1989) are briefly considered. (TJH)

Both doubly multivariate and multivariate mixed models of analyzing repeated measures on multivariate responses are reviewed. Given multivariate normality, a condition called multivariate sphericity of the covariance matrix is both necessary and sufficient for the validity of the multivariate mixed model analysis. (SLD)

Large sample multivariate methods for estimating and comparing effect sizes across independent samples within a single study are presented. Procedures for pooling treatment effects are provided to allow determination of overall study effect sizes for treatments with similar effects prior to implementation of meta-analyses. (TJH)

Data from four studies were used to analyze the hypothesis that performance evaluation variables can be constructed with a rater's facet and a trait's facet. The regularity of the pattern of intercorrelations across different rater groups was remarkable. The use of facet analysis before confirmatory analysis is discussed. (SLD)

Recognizes some of the limitations of the field of behavioral genetics, but argues that the methods employed in multivariate behavior genetics and developmental behavior genetics have become the dominant paradigms in the field. (MDM)

Gender differences in cognitive abilities were studied in 1,224 13-year olds using a multivariate latent variables approach. No gender differences were found in the structure of cognitive abilities, but mean differences in favor of females were found for general intelligence, with higher means for males on most spatial dimensions. (SLD)

Conceptual rationales from five research contexts are presented, which all posit unmeasured variables that interact with observed independent variables to produce a complete model of the dependent variable. Strategies for overcoming related difficulties are outlined, including increased longitudinal assessment, oversampling of levels of…

The analysis of contingency tables via set correlation allows the assessment of subhypotheses involving contrast functions of the categories of the nominal scales. The robustness of such methods with regard to Type I error and statistical power was studied via a Monte Carlo experiment. (TJH)

Purposes of multivariate analyses are discussed, focusing on the primary purposes of prediction and structure identification and the secondary purpose of response variable ordering. The sound initial choice of response variables and the advisability of simpler analyses when feasible are discussed. (SLD)

Reviews the use and interpretation of canonical correlation analysis in 1978-89 "Journal of Educational Research" articles. Wide variation in the use of terminology was found across 10 articles reporting 11 studies. Because supplementary analyses are now available, researchers are encouraged to use the model more fully when interpreting…

This article presents a didactic discussion on the role of asymptotically independent test statistics and separable hypotheses as they pertain to issues of specification error, power, and model misspecification in the covariance structure modeling framework. A small population study supports the major findings. (SLD)

Proposes a procedure for generating multivariate nonnormal distributions. The procedure, an extension of the Fleishman power method (A. Fleishman, 1978), generates the average value of intercorrelations much closer to population parameters than competing procedures for skewed and heavy tailed distributions and small sample sizes. Reports Monte…

Examines the relation of gender role conflict to certain aspects of gay men's psychological experience. Gay men with less gender role conflict had a more positive view of seeking psychological help and reported fewer symptoms of anger, anxiety, and depression. (Author/JDM)