Which statistical hypothesis are tested by SAS in the presence of missing plots?

Abstract

The interpretation of the tested hypothesis through variance analysis of balanced agricultural data, can be made, in general, without great difficulties, specially hi the case of well designed and well conducted experiments. If, however, data is not balanced and missing plots are present, the interpretation of the tested hypothesis through the sums of squares given by the available statistical packages, may be extremely difficult for statisticians, and practically impossible for profissionals of applied sciences, which use these packages. In this study, the interpretation of the most common tested hypothesis is discussed through the General Linear Models (GLM) procedure of the Statistical Analysis System (SAS), with the objective of alerting users about the problems related to the choice of the hypothesis that best reflects the objectives of his research.