ALTERNATIVE REPRESENTATIONS OF DUMMY VARIABLES AND THEIR INTERPRETATIONS

Abstract

The objective of the paper was to bring together some of the preexisting theoretical treatments ~f dummy variables in regression analysis and to present them in such a way as they could be more effectively used and interpreted. This Will enable researchers to select the particular representation that best suits their hypotheses. Four alternatives were exhibited for both a one-way analysis of variance and a one-way analysis of variance with a covariate. The predicting equations were presented in the text of the paper along with numerical examples and their interpretations. The derivations of the predicting equations and variable definitions were presented in the Appendix.