How to detect indeterminacy in preference data analysis

The paper is focussed on the indecision component in self-reported preferences concerning happiness, well-being, trust, etc. The proposal takes into account the difficulties that some respondents may experience in sharply declaring their status with respect to this kind of items. More specifically, we implement CUB models to assess both uncertainty and indecision of rating responses and check the usefulness of this methodological approach on some empirical data. We exploit observational data arising from a research on well-being, leisure and interpersonal relationships of people living in the metropolitan area of Naples, Italy. Results confirm the hypothesis that indeterminacy in responses to such surveys is a prominent issue to be investigated and the proposed modelling approach is able to measure the contribution of this latent variable.