Comparison of best-worst and hedonic scaling for the measurement of consumer wine preferences

Abstract

Background and Aims: Best–worst scaling (BWS) is compared to standard hedonic scaling for measuring consumer wine preferences. BWS is a relatively new method for producing ratio-level scales and has gained recent attention for application in sensory research, but has not been applied to wine. Methods and Results: Regular wine consumers (112) evaluated eight designed wines with both scaling methods in an intra-subject design over two test periods. The methods did not result in comparable product liking results. The eight wines could almost be differentiated on an aggregated level with hedonic ratings (P = 0.076); there was no significant difference with BWS. Latent class analysis was used to identify two clusters, which differed on the preferences for the designed sensory components. The BWS design had to be split into several blocks, so no complete individual measures were available, which prevented analysing heterogeneity for this method. Conclusions: BWS needs more wines to be assessed per person in order to discriminate between red wines and to allow modelling of consumer preference heterogeneity. Respondents would have to accomplish complete individual BWS designs, which requires repeated exposure to the same set of wines over several tasting sessions. Significance of the Study: This study demonstrates that BWS is not as suitable for sensory consumer preference measurement of red wine as hedonic rating. While BWS has shown a higher discriminative ability for different products and in non-sensory research, the factors of alcohol, tannin and memory fatigue make it less practical for red wine sensory measurement compared to hedonic rating.