Relation of Preference Panel and Trained Panel Scores on Dry Whole Milk

Abstract

Summary Dry whole milk samples manufactured by the EURDD Dairy Products Laboratory, Washington, D. C., were scored by an Oregon State University trained panel, EURDD Dairy Products Laboratory, Washington, D. C., trained panel, EURDD Engineering and Development Laboratory, Wyndmoor, Pennsylvania, trained panel, and by an OSU preference panel of from 100 to 160 student judges. After preliminary testing of 63 samples by the DPL and OSU panels, two series of 50 samples each for a total of 100 samples were tested by all four panels, except the Wyndmoor panel, which tested only 96 samples. Two of the three trained panels used modifications of the American Dairy Science Association score-card for dry whole milk. The Oregon trained panel used a simple six-factor ballot. The student preference panel scored the samples on the nine-point hedonic scale. All data were statistically analyzed for correlation and regression coefficients. Results indicated that total scores from the trained panels could be used as a basis for a scoring system which would predict consumer preference scores. The trained panels’ scores for individual scoring factors, i.e., cooked, oxidized, stale, astringent flavors, did not satisfactorily predict preference scores, but could be used as an indication of possible flavor defects in dry whole milk samples.