COMPUTERIZED RULES OF THUMB FOR FACTORIAL ANALYSIS OF VARIANCE TABLES

IN factorial analyses of variance, researchers are faced with the problem of finding the correct formulae for the entries in the analysis of variance (ANOVA) tables. These computational for-mulae differ for various factorial ANOVA designs depending on the characteristics of the variables in the design. An article by Millman and Glass (1967) entitled &dquo;Rules of Thumb for Writing the ANOVA Table&dquo; provided a much-needed generalized collection of rules of thumb for finding the appropriate formulae. These rules allow for crossing or nesting of random, finite, or fixed factors. The authors stated their hope that mastery of such rules would cause researchers to consider these more complex designs as useful and even convenient. The Millman and Glass procedure is accurate, but cumber-some. Researchers must study the article and follow the rules very cautiously to ensure the correctness of the formulae. In order to eliminate tedious work and the danger of making errors, this computer program was written to generate the ANOVA table, complete with all necessary formulae. Input This program is appropriate for factorial ANOVA models which have two, three, or four factors, some of which may be nested. To run this program, one only needs from four to six input data cards, which describe the factors in the design, including the replications factor, and which signal the end of the input set