Abstract
The elimination-by-aspects (EBA) model assumes that, in a choice situation, alternatives can be clearly dichotomized depending upon whether they include a particular aspect (a desirable feature) or not. In actual life, however, aspects taken account for the choice are often ambiguous; in many instances, alternatives cannot be dichotomized so clearly. In the present study, this ambiguity was represented using membership grade of fuzzy sets theory, and EBA model was generalized to be applicable to fuzzy aspects. This model was termed fuzzy-EBA (FEBA) model, and it was tested by applying it to political choice. For each individual, choice probability was computed by the model, where image ratings of political parties were regarded as grades of membership. The expected number of subjects who select a particular party was estimated from individual choice probabilities, and reasonable agreement with the observed number of subjects was obtained.
