Tversky's product similarity model has some interesting implications. First, similarity judgments between two product categories may be asymmetrical if one is considered a prototype and the other a variant. Second, a brand association may reverse the asymmetry in similarity judgments. This may happen if a particular brand has a prototype product that is not considered prototype outside the context of the brand. Third, brands can, by themselves, increase the perceived similarity between two products. This may occur either by increasing the salience of the common feature, or by adding a common feature to both products (most obviously, the brand itself can be perceived as a common feature). Contrast models have been revealed for consumer products (Johnson 1986), but the asymmetric effects of brand on product similarity are unfamiliar to the academic literature.

1.2. Structure

Barsalou (1983; 1985) emphasized the role of motivation and context in judgments of conceptual similarity. Sets of similar items, labeled "special" or "intentional" categories, with some properties found in more general categories, such as prototype examples. However, the most typical parts of these classifications (such as "Foods not to be eaten on a diet") are determined not by specific characteristics, but by extreme values ??on a single scale (such as calories). Barsalou notes that intentional categories often seem to violate the structural interrelationships of the environment, often including parts from several common taxonomic categories (e.g., ice cream and French fries may both be typical "weight loss period items"). Foods that cannot be eaten"). His findings underscore the flexibility of the classification process and the ease with which specific classifications can be generated for specific purposes. Medin, Goldstone, and Gentner (1993) used a construct view to demonstrate the same kind of asymmetric similarity judgments reported by Tversky. They suggested that factors within the contrasting process itself were responsible for the asymmetry. In the consumer behavior literature, both product typicality judgments and reference product use have been found to be context- and goal-dependent (Huffman, et al. 1990; Ward and Loken 1988).

Humanly translated.