A connectionist model of brand-quality associations

Abstract

Consumers use brand names and product features to predict the performance of products. Various learning models offer hypotheses about the source of these predictive associations. Spreading-activation models hypothesize that cues acquire predictive value as a consequence of being present during the acquisition of product performance information. Least mean squares connectionist models hypothesize that any one cue acquires predictive value only to the extent that it can predict differences in performance that are not already predicted by other available cues. Five studies in the context of portfolio-branding strategies provide evidence supporting a least mean squares connectionist model. As predicted by this model, results show that subbranding and ingredient-branding strategies can protect brands from dilution in some situations but can promote dilution in other situations.