ANALYSIS OF MULTILINGUAL LABELED SORTING TASKS: APPLICATION TO A CROSS‐CULTURAL STUDY IN WINE INDUSTRY

Verbalization can be used as a complement in a sorting task (also known as categorization). Assessors are required to describe each item or group of items using their own words; thus, a “labeled” categorization is performed. In cross‐cultural studies, several panels may use different languages. We propose to extend hierarchical multiple factor analysis (HMFA) to compare panels from, simultaneously, both categorization and verbalization tasks, even when the latter is performed in several languages. We present this methodology through a study where both a Catalan and French panels performed a labeled categorization on a set of eight Catalan wines. Either Catalan or French is used by the panelists. HMFA allows for comparing the two panels not only globally but also from either the categorization or the verbalization tasks. Although results show a noticeable similarity between both configurations of wines as defined by the panels, textual data highlight interesting differences or viewpoints between Catalan and French. PRACTICAL APPLICATIONS Cross‐cultural studies are increasing in sensory studies, frequently including verbalization that can be performed in different languages. We propose a methodology to tackle the data issued from this kind of studies, including categorical and multilingual textual data. Our application concerns two panels, Catalan and French, who have performed a labeled sorting task, although more panels could be considered. The application shows that textual data nuance the similarities put to the fore between both configurations of wines as defined by the panels. Furthermore, this application uncovers that important concepts in the wine world are not always shared by panels issued from different backgrounds. No translation is required before the treatment.