A Learning-Based Model for Imputing Missing Levels in Partial Conjoint Profiles

Abstract

Respondents in a conjoint experiment sometimes are presented with successive partial product profiles. First, the authors model how respondents infer missing levels of product attributes in a partial conjoint profile by developing a learning-based imputation model that nests several extant models. The advantage of this approach over previous research is that it infers missing levels of an attribute not only from prior levels of the same attribute but also from prior levels of other attributes, especially ones that match the attribute levels of the current product profile. Second, the authors provide an empirical demonstration of their approach and test whether learning in conjoint studies occurs; to what extent; and in what manner it affects responses, partworths, and the relative importance of attributes. They show that the relative importance of attribute partworths can shift when subjects evaluate partial profiles, which suggests that consumers may construct rather than retrieve partworths and are sensitive to the order in which the profiles are presented. Finally, the results show that consumers' imputation processes can be influenced by manipulating their prior information about a product category. This research is of both theoretical and practical importance. Theoretically, this research sheds light on how customers integrate different sources of information in evaluating products with incomplete attribute information; practically, it highlights the potential pitfalls of imputing missing attribute levels using simple rules and develops a better behavioral model for describing and predicting customers' ratings for partial conjoint profiles.