Data analyses are presented investigating the relationship of high school grade point average and first year university grade point average. A variety of analyses are presented; the results imply that high school attended is a better predictor of university performance for poorer students than for better students. (Author/JKS)

Limitations of dummy coding and nonsense coding as methods of coding categorical variables for use as predictors in multiple regression analysis are discussed. The combination of these approaches often yields estimates and tests of significance that are not intended by researchers for inclusion in their models. (SLD)

This article reviews some of the properties of criterion-scaled regression analysis, especially as it relates to the analysis of a repeated measures/randomized block design. (Author/LMO)

Among 153 young adult dating dyads, predictions were tested that were derived from a goodness of fit model of the role of temperament in person-social context relations, and from the related filtering/similarity hypothesis. Measured were each dyad member's temperament and expectations for partner's temperament. Results indicated a significant…

A summary and interpretation of the recent literature on the indeterminancy of factor scores is given in simple terms. A good index of factor score determinancy is the squared multiple correlation of the factor with the observed variables. (Author)

A study to demonstrate the use of multiple regression for evaluation of instructional strategies in science. Two groups of students received Audio-Tutorial and conventional instruction. Research design was a treatment by levels with two treatments and three levels of aptitude within each treatment. (PS)

Purpose of this article is to explain the fundamental nature and sources of regression toward the mean. The ultimate goal is that developmental psychologists understand regression effects well enough so that they will not make erroneous interpretations of such effects in their empirical data. (Author)

Authors discovered significant correlations between couse ratings and variables representing commitment and course-end attitudes toward the course; conclusions suggested steps an instructor might take to improve his ratings.'' (Authors)

The authors describe a computer program which deals with sets of variables rather than with one variable at a time. (MM)

Methods for comparing the suitability of alternative models for covariance matrices are examined. A cross-validation procedure is suggested and its properties examined. A series of examples using longitudinal data are examined. (Author/JKS)

Compared six effect size (ES) measures commonly used by gerontological researchers as these measures relate to one another in both the analysis of variance and multiple regression models. Also discusses three other issues involving ES measures: the ES of a contrast; orthogonal and nonorthogonal designs; and partial ESs. (Author/JAC)