On Selecting the Best Set of Regression Predictors

Publisher Website
Google Scholar
Abstract
The three traditional methods of backward, forward, and stepwise selection of variables to be included in a “best” regression equation were compared to a method designed to maximize weight validity. With student achievement as the criterion, and aptitudinal variables manifesting considerable multicolinearity as predictors, the subset of variables selected by the traditional methods performed poorer than the one selected by the weight validity maximization method. Implications for constructing regression equations for prediction are discussed, with consideration of the weight validity maximization method recommended in crucial situations.

This publication has 6 references indexed in Scilit: