Repeated measures designs often involve dichotomization of a continuous variable in order to be amenable to the analysis of variance nature of such designs. An alternative to that approach wherein the independent variable is kept continuous is presented. (JKS)

The construction and interpretation of confidence intervals for the prediction of new cases in multiple regression analysis is explained. An example is provided. (JKS)

The use of multiple regression analysis was compared to the use of discriminant function analysis in the prediction of college faculty rank. The multiple regression technique was shown to be generally superior in this instance. (JKS)

Problems associated with the use of gain scores, analysis of covariance, multicollinearity, part and partial correlation, and the lack of rectilinearity in regression are discussed. Particular attention is paid to the misuse of statistical techniques. (JKS)

A Monte Carlo simulation was employed to determine the accuracy with which the shrinkage in R squared can be estimated by five different shrinkage formulas. The study dealt with the use of shrinkage formulas for various sample sizes, different R squared values, and different degrees of multicollinearity. (Author/JKS)

A series of multiple regression analyses was used to investigate a salary equity policy in a statewide system of institutions of higher education. Faculty rank, number of publications, and teaching effectiveness were among the variables examined. (JKS)

Historical regression uses the assumption that without intervention, growth at post-test will proceed at the same rate as at pre-test. Several methods of historical regression are compared with an illustrative example. (JKS)

With even the simplest bivariate regression, least-squares solutions are inappropriate unless one assumes a priori that reciprocal effects are absent, or at least implausible. While this discussion is limited to bivariate regression, the issues apply equally to multivariate regression, including stepwise regression. (Author/CTM)

The direct and indirect effects of neighborhood characteristics on school-level reading achievement were investigated via path analysis in this study. Race, family structure, income, and neighborhood density were among the variables investigated. (Author/JKS)

The purpose of this study was to use multivariate techniques in a federally-regulated validation study, and compare the results obtained from zero-order correlations and multiple correlations with the results obtained using factor scores and canonical correlation. The subjects consisted of 51 applicants for the position of patrolman. (Author/CTM)

Of 50 students taking a graduate course in multiple regression analysis at a particular university, they tended to use multiple regression in their dissertations. (Author/JKS)

A Type VI error results from inconsistency between the researchers' question of interest and the statistical procedures employed to analyze the data. An example of a research problem is analyzed to show the increase in statistical power resulting from improved research design, using multiple regression instead of analysis of variance. (CTM)

A model for research on teacher effects in which the variance in student outcome post-test performance is attributed to pre-test perfomance; to separate construct domains of student, instructional, and teacher variables; and to interactions between variables in these three construct domains is presented and tested. (Author/JKS)

The inclusion of unmeasured variables in path analyses in educational research is discussed. The statistical basis for inclusion is presented, along with several examples. (JKS)

The relative magnitudes of R-squared values computed through multiple regression models using grade equivalent scores, raw scores, standard scores, and percentiles as both predictor and criterion variables are compared. Grade equivalents and standard scores produced the highest R-squared values. (Author/JKS)