The free-linking task: A graph-inspired method for generating non-disjoint similarity data with food products

“Free sorting”, in which subjects are asked to sort a set of items into groups “most similar” items, is increasingly popular as technique for profiling sets foods. However, free sorting implies an unrealistic model sample similarity: that similarity purely binary (is/is not similar) and fully transitive (similarities {A, B} {B, C} imply C}). This paper proposes new method rapid testing—the “free-linking” task—that solves both problems: linking, draw graph they connect pairs samples with line if similar, according the subject’s individual criteria. simple task provides more realistic allows degrees through distance metric does require similarity. In two pilot studies spice blends (10 samples, 58 subjects) chocolate bars 63 subjects), linking evaluated compared using DISTATIS, RVb, parameters degree, transitivity, connectivity; also indicated their preferences ease-of-use tasks. studies, first dimensions DISTATIS consensus were highly comparable across tasks; however, provided discrimination three four. RVb stability was equivalent methods. Graph statistics had greater power: on average made groupings lower higher connectivity studies. did overall find easier liked it more, indicating cognitive difficulty linking. The free-linking task, therefore, robust, maps at cost panelist effort, should prove valuable alternative sensory assessment product sets.