|
Activity Number:
|
203
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 10, 2004 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics and Marketing
|
| Abstract - #300341 |
|
Title:
|
Representing Brand and Category Personality with Circumplex Models
|
|
Author(s):
|
Peter Lenk*+ and Rajeev Batra and Michel Wedel
|
|
Companies:
|
University of Michigan and University of Michigan Business School and University of Michigan Business School
|
|
Address:
|
701 Tappan, Ann Arbor, MI, 48109-1217,
|
|
Keywords:
|
Bayesian Inference ; latent variables models ; marketing applications
|
|
Abstract:
|
Consumers often anthropomorphize brands by associating them with personality characteristics. This provides brands with durable and distinctive traits that differentiate brands within a category. The personality of a brand and its category are often confounded, caused by the transfer of perceptions across these two domains. This paper develops a hierarchical Bayes model for brand and category personality that separates the two. The model is based on a circumplex correlation structure for personality items. The circumplex model consists of two bipolar factors where the loadings are restricted to the unit circle. Personality items are ordered according to their angles. The model is applied to a survey of three categories (cars, jeans, and magazines) with 10 brands in each category. The empirical study detects substantial transference of category personality to the brands within the category and finds a strong relation between the subjects' feeling towards a brand and its perceived personality.
|
