Treatment Comparisons

The purpose of this paper is to provide concrete illustrations of the efficacy of the multiple regression approach to the analysis of experimental results. A bridge between theoretical and specific applications is provided by parallel multiple regression and analysis of variance solutions for two typical educational designs. Detailed illustrations are given in the techniques of writing linear constraints and determining degrees of freedom. The major advantages of the multiple regression approach are its adaptability to unusual designs and the facilitation of meaningful interpretation afforded by the provision of regression weights in addition to the usual Fratios.