Hedonic responses as predictors of food choice: flexibility and self-prediction.

The association between liking and choices of six cheeses was examined. Particular attention was in hedonic flexibility, i.e. the range of liking ratings, within an individual, leading to a choice. Second, the ability of respondents to predict their future liking and choices was studied. Respondents (62 females, 20-64 years) initially tasted and rated the cheeses blind and then chose three cheeses, 150 g piece of each, for home-use during the following week. They also predicted how their ratings of liking and choices would be next time, after a week. They returned to the laboratory to repeat the tasks twice, with one-week intervals, thus a total of nine choices were made. Expectedly, the association between initial liking and frequencies of choice was strong (aggregate level correlation 0.70). Self-prediction of liking was less successful initially (r=0.56) than in the second session (r=0.68). In the first session, three choices were made within 1.3 points (mean) at the upper end of a 7-point scale, while nine choices were made within 2.7 points of initial ratings. Thus, the choices extended towards the initially less-liked cheeses in the second and third session. Although poor at precise prediction of their choices, the respondents fairly accurately knew which cheeses they would not choose. Hence, an important strategy in food choice may be to screen out unacceptable options and then alternate among the acceptable ones.