Abstract
The present article discusses the interpretability of beta weights in terms of their definition, technical aspects and the research philosophy guiding the use of multiple regression. The major conclusion is that variable importance and variable ordering can not be ascertained by examining beta weights. Additionally, it is recommended that discussion of variables as a group without identification of singular variable importance would more appropriately match the multivariate purpose of multiple regression.

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Copyright (c) 1992 Ric Brown (Author)
Downloads
Download data is not yet available.