Disentangling the Influence of Knowledge on Processing Strategies in Choice Modelling.

This paper seeks to disentangle the effect of knowledge on processing strategies using data from a discrete choice experiment on cold-water corals in Norway. Cold-water corals are a deep-sea ecosystem for which we have limited scientific knowledge and for which public awareness is low, and consequently is likely to be an unfamiliar good to many members of the public. One simplifying strategy often employed by respondents in a choice experiment is to simply ignore some of the attributes, i.e. attribute non-attendance. After the initial presentation of the good, before answering the choice cards, the respondents were given a quiz over the material covered in the presentation. This provides us with an ex ante measure of their knowledge. We use a combination of discrete and continuous mixture models to disentangle the effects of variations in knowledge about the good. We use a respondent’s quiz score as covariates in the probability function of attending to an attribute. Our results show that knowledge, as measured by the quiz score, has a significant effect on the probability of attending to the attribute for three out of four attributes. This has direct implication for practitioners in that proper information may help avoid the use of simplifying strategies.