PROBLEMAS Y MÉTODOS COMUNES DEL ANÁLISIS DE EXPERIMENTOS FACTORIALES

Abstract

In the agronomic research, factorial experiments are frequently carried out. These are tools that make an optimus use of resources; produce more precise estimates of contrasts of treatment means, and make it possible the study of the interactions among factors. The full release of these advantages, however, is not always obtained due to one or several of the following common problems: Poor interpretation of the interaction concept; abuse of multiple comparison procedures; subutilization and even omission of regression techniques, orthogonal contrasts and polynomials, etc. To contribute to solve these problems and thereby to improve the quality of the scientific articles to be published, in this paper several meanings of the interaction concept, including since its etymological aspects until its meaning as a source of variation in the analysis of variance, were analyzed. In addition, cases related with the pertinence, use and advantages of regression, contrasts, and orthogonal contrast and polynomial techniques were analyzed in the context of factorial experiments. To make a more objective presentation, hypothetical cases of factorial experiments were considered. It is hoped that the reader’s conceptual perception and capacity to successfully associate cases of factorial experiments and statistical methods to analyze them properly will be improved and thereby will enable them to increase the quality of the interpretation of the results and their scientific publications.