STATISTICAL PLANNING AND DESIGNING OF EXPERIMENTS TO DETECT DIFFERENCES IN SENSORY EVALUATION OF BEEF LOIN STEAKS

Abstract

ABSTRACTThe main objective of a sensory experiment is to determine the effect of certain treatments on the response of a group of people. The primary obstacle in accomplishing this objective is that responses of sensory groups are extremely variable making the detection of true differences in response due to treatments difficult. It was the objective of this study to perform sensory experiments so that true differences in the main effects of interest could be detected. Consequently, achieving this objective necessitated the completion of the following three facets of an exemplary sensory study. Initially, the experimenter should carefully plan the experiment so as to control variation by rigorously training the panelists and controlling the objective and subjective error components of experimental error. Secondly, the experimenter should choose a restricted experimental design which stratifies the experimental units into homogeneous subgroups and accounts for all known sources of variation. The known but uncontrollable sources of variation should be nullified through randomization. A sufficient number of replications should be run to attain the desired level of precision. The principles necessary for the efficient planning and designing of experiments are presented. Lastly, the experimenter must collect accurate data and employ theoretically sound procedures to analyze that data. A sensory experiment was conducted invoking the above guidelines which made possible the detecting of true differences in steak palatability due to cooking endpoint temperature and marbling level.