Treatment of "don't know" responses in the consumers' perceptions about sustainability in the agri-food sector

In several studies the analysis of consumers' perceptions has become of paramount strategic value in order to understand consumption behaviour. The aim of this paper is to determine the effectiveness of two alternative strategies for analyzing the "don't know" responses in questionnaires with items having ordered response categories. Both strategies rest on a statistical model, denoted as cub (Combination of Uniform and shifted Binomial random variables), designed to measure the agreement of consumers with the latent trait measured by the questionnaire together with the level of uncertainty in choosing the ordinal response category. In the first strategy, the fraction of don't know responses is considered as a measure of the item difficulty and used, together with some personal characteristics, as covariate in the cub model. The second strategy is a recent extension of the standard cub, consisting in adjusting the uncertainty measure to take account of the fraction of don't know responses, considered as a source of further uncertainty. We applied the two proposed strategies on real data from a survey about the consumers' perceptions on the importance of sustainability issues in the agri-food sector.