Methods for predicting specific college course grades, based on small numbers of observations, were investigated. These methods use collateral information across potentially diverse institutions to obtain refined within-group parameter estimates. One method, referred to as pooled least squares with adjusted intercepts, assumes that slopes and…

Two methods of using collateral information from similar institutions to predict college freshman grade average were investigated. One central prediction model, referred to as pooled least squares with adjusted intercepts, assumes that slopes and residual variances are homogeneous across selected colleges. The second model, referred to as Bayesian…

There are times in survey research when missing values need to be estimated. The robustness of four variations of regression and substitution by mean methods was examined using a 3x3x4 factorial design. The regression variations included in the study were: (1) regression using a single best predictor; (2) two best predictors; (3) all available…

Power and Type I error rates were estimated for contingency tables with small sample sizes for the following four types of tests: (1) Pearson's chi-square; (2) chi-square with Yates's continuity correction; (3) the likelihood ratio test; and (4) Fisher's Exact Test. Various marginal distributions, sample sizes, and effect sizes were examined. (SLD)

Ridge estimates (REs) of population beta weights were compared to ordinary least squares (OLS) estimates through computer simulation to evaluate the use of REs in explanatory research. With fixed predictors, there was some question of the consistency of ridge regression, but with random predictors, REs were superior to OLS. (SLD)

Asserts that, in principle, an analyst should be satisfied with infinite estimates slope in logistic regression because it indicates that a predictor is perfect. Using simple approaches, hypothesis tests may be performed and confidence intervals calculated even when a slope is infinite. Some functions of parameters that are infinite are still…

In research, data sets often occur in which the variance of the distribution of the dependent variable at given levels of the predictors is a function of the values of the predictors. In this situation, the use of weighted least-squares (WLS) or techniques is required. Weights suitable for use in a WLS regression analysis must be estimated. A…

This paper considers the application of Bayesian techniques for simultaneous estimation to the specification of regression weights for selection tests used in various technical training courses in the Marine Corps. Results of a method for m-group regression developed by Molenaar and Lewis (1979) suggest that common weights for training courses…

An alternative is proposed for the Johnson-Neyman procedure (P. O. Johnson and J. Neyman, 1936). Used when heterogeneous regression lines for two groups are analyzed, the Johnson-Neyman procedure is a technique in which the difference between the two linear regression surfaces for the criterion variate (Y) is estimated conditional on a realization…

Proposes a heuristic method for estimating the relative weight of predictor variables in multiple regression that is computationally efficient with any number of predictors and that can be shown to produce results similar to those produced by more complex methods. (SLD)

An evaluation of the rules-of-thumb used to determine the minimum number of subjects required to conduct multiple regression analyses suggests that researchers who use a rule of thumb rather than power analyses trade simplicity of use for accuracy and specificity of response. Insufficient power is likely to result. (SLD)

Discusses determining sample size for estimation of the squared multiple correlation coefficient and presents regression equations that permit determination of the sample size for estimating this parameter for up to 20 predictor variables. (SLD)

The regression coefficients and the associated standard errors in hierarchical regression, when a theoretical basis for the analysis exists, are determined for four regression models. Each reflects different controlling or partialling of the variates. An illustration is presented using data from the Berkeley Growth Study. (SLD)

The predicted scores produced by regression with (1) single best predictor, (2) two best predictors, (3) all predictors with observed values, and (4) all predictors with or without observed values were compared with variable means as estimates of missing values. The study was conducted in a simulation mode on nx8 data matrices using various levels…

This Monte Carlo study explored relationships among standard and unstandardized regression coefficients, structural coefficients, multiple R_ squared, and significance level of predictors for a variety of linear regression scenarios. Ten regression models with three predictors were included, and four conditions were varied that were expected to…