An experiment was performed to investigate the effect of various nitrogen sulfer combinations on the quality of poinsettias. After various physiological measurements were taken, commercial growers, retailers, and consumers were asked to evaluate the salability of the plants. In order to avoid evaluator fatigue, only a limited number of plants could be evaluated. This presented both experimental design and data analysis problems. In view of these constraints, and in order to obtain meaningful results, an unreplicated 7 x 8 factorial design was used. Data were analyzed using the method of half-normal plots in conjunction with a modification of the analysis of variance procedure. Rationale and alternative designs will be presented, as well as the step-by-step procedure for using this method as contrasted with the standard ANOVA technique.