VIEW ARTICLE    DOI: 10.1094/ASBCJ-52-0155

Statistical Analysis with Weighting Factors: Hop Aroma in Beer. Xiaogen Yang and Max Deinzer, Department of Agricultural Chemistry, Oregon State University, Corvallis, OR, and Cindy Lederer and Mina McDaniel, Department of Food Science and Technology, Oregon State University, Corvallis, OR. J. Am. Soc. Brew. Chem. 52:0155, 1994.

Because the olfactory sensitivity and consistency of sensory subjects are different for a given aroma note, their sensory ratings need to be weighted, based on the subjects' performance, to increase the panel's ability to discriminate among treatments. The F ratio and the corresponding confidence coefficient S for the significance test of treatment differences from a subject's ratings are measures of the effectiveness of the subject's sensory response. The correlation coefficient r reflects agreement of individual subject's ratings with the panel's average. Three statistical measures, F, S, and (S X r), were examined for the prospective application of weighting factors for analysis of sensory data. The application of weighting sensory data was studied with the example of beer aroma evaluation. Weights were given to the sensory scores of individual subjects based on their perception and performance. Data weighting enhanced the discrimination of differences among products and reduced the effect of less reliable sensory scores from insensitive subjects, as shown by the results of analysis of variance. The sensory ratings from a descriptive sensory panel for pilot beers brewed with hop fractions were weighted before cluster and factor analysis. As a result, replicates were clustered more closely together, and the treatments were separated further apart. This made it possible to differentiate among four different beer samples. Floral, grassy, spicy, and citrus were the main aroma characteristics of beer brewed with Hallertauer Hallertauer hop pellets. Beer made with hop oil only had a strong soapy-minty aroma. Beer made with hop pellets was described as having floral, grassy, spicy, and citrus aromas. The results demonstrate that weighting sensory data is an effective method for accentuating differences among treatments.

Keywords: ANOVA, Cluster analysis, Hop aroma, Hop oils, PCA, Sensory analysis, Statistical analysis