Yalun Huang: Comparing Just-About-Right Scaling and Satisfaction-Based DFA Methods for Sensory Attribute Diagnostics in Coffee

However, standard JAR penalty analysis exhibited diagnostic ambiguity with notable reversed penalties (e.g., "too low" acidity or bitterness associated with higher liking than JAR), suggesting JAR deviations can reflect consumer preference directions rather than true product defects. Furthermore, agglomerative hierarchical clustering (k = 2) revealed that JAR primarily split consumers by general scale-use level (higher- vs. middle-scale users with flat sample differentiation), whereas satisfaction-based DFA effectively separated consumers into distinct, actionable taste-preference profiles (balance-oriented vs. dark-roast-oriented).

These findings demonstrate that satisfaction-based DFA provides clearer attribute-level diagnostic profiles (d’_A) and richer consumer segmentation without relying on the assumption that deviations from "just about right" inherently represent liking penalties. </p>