Identifying temporal drivers of liking and satiation based on temporal sensory descriptions and consumer ratings

•A panel described temporal perception and consumers rated liking/expected satiety.•Two clusters of consumers were retained according to their expected satiety.•Penalty-lift analysis applied to sequential time points to find temporal drivers.•Textures differently impact satiety expectations in two groups of consumers.•Particle size attributes (Gritty vs Sandy) were found to be important classifiers. Capturing temporal sensory changes has been the focus in recent research to better understand how consumers perceive food products. This information can be linked to consumer expectations (e.g., liking, satiety) to study the sensory drivers throughout the eating experience, namely temporal drivers. This study explores the use of penalty-lift analyses for each time point in the temporal sensory description to identify the temporal drivers of liking/satiety for different groups of consumers with different patterns in their expectations of satiety. Eight yoghurt samples formulated based on an experimental design, with identical composition, varying in textural properties, were used in the study. Temporal Check-All-That-Apply (TCATA) was used to describe dynamic sensory profiles. Consumers (n = 101) tasted each yoghurt and rated their liking and expected satiety. Cluster analysis of variables around latent variables (CLV) method was applied to cluster consumers based on their expectations of satiety, detecting two relevant clusters. Penalty-lift analysis was used for each time point. Also, the false discovery rate (FDR) was applied to correct p-values for multiple tests responding to sequential time points. Differences were found related to how particle size attributes and flavour intensities drove liking for each cluster at different time points. For cluster 1, while Gritty was positive driver from the middle to the end, Sandy was negative driver in the middle; and Vanilla was positive driver of liking throughout the mastication. For cluster 2, only Sweet was pointed as positive driver at the beginning, and Dry as negative driver in some time points at the middle of the mastication. With regards to expected satiety, main difference was that Gritty (or Sandy) was considered as positive (or negative) driver for cluster 1, but not for cluster 2; significant over the entire time period. These findings demonstrate that the temporal driver approach appears as a suitable method to unveil the drivers of liking/satiety during the eating process in groups of consumers with