The Aroma of Canned Beef: Application of Regression Models Relating Sensory and Chemical Data

The purpose of this investigation was to determine general relations between instrumental and sensory aroma data from a reference material consisting of a large set of different types of beef samples analyzed during several years. The relations obtained in this way have been tested on independent “unknown” samples. Different models have been used, basically derived from Stevens' law and formulated in analogy with models used in other psychophysical contexts. From the reference material a great number of highly significant relations‐several with a correlation coefficient greater than 0.90–were obtained for the various odor notes used. Several of these seem to be examples of causative relations. When predicting sensory properties of unknown samples almost all the relations obtained with high correlation coefficient worked very well. These properties could be predicted by the gas chromatographic technique with the same accuracy as when the panel assessed the samples. Therefore, by applying the models in a proper way, the panel service in routine analyses may be supplemented or refined by using a gas chromatographic technique. These methods may, of course, also be used in product and process development work.