Children's school milk consumption: How realistic are lexicographic preferences?

Although dairy products are part of a balanced diet for children, their consumption declines quite often and will become insufficient over the course of years. Focusing on school milk in principle, the paper analyses differences in preferences and willingness to pay for varying school milk products on the part of youths and parents. Special attention is paid to the recent phenomenon, discussed with increasing frequency, of lexicographic preferences. If these preferences are present, respondents choice behaviour follows very strict rules. Data was obtained by online-surveys filled in by parents of school children and youths in Germany. In the survey, choice experiments are conducted to analyze willingness to pay and preferences of conventional and novel school milk products as well as novel yoghurt products as different alternatives. The varying attributes are price, fat content, or sweetening agent. A Nested Logit Model is used to estimate willingness to pay for these different attributes. Within this model, those respondents who always choose the conventional school milk are separated from the others who at least once buy a novel product. Differences between these groups are explained by socio-demographic and psychometric variables. The analysis shows that parents’ and youth’s product utility declines with increasing fat content, the use of an artificial sweetener, and increasing price. Novel products raise overall product utility. Parents, who tend to choose the novel products, more often judge supplementary calcium and vitamins in dairy products as useful, they think that fat-reduced milk products are healthier, and more often estimate their children to be overweight. Youths preferring novel products more often choose products with lower fat or sugar content, they more often show a positive attitude to milk products, and more often estimate themselves as overweight. Interpreting the estimates we are also able to detect lexicographic preferences. Our reasonable doubts about this finding are explained in detail.