Answers are grounded in this paper and cite their sources.
The problems of accurately gauging the public's perception of a new product or service and their preferences for alternative designs of the new product are problems constantly being faced by corporate and government organizations. Often because of the costs involved or the possibility of alerting competition to the details of a new idea, it is not always practical actually to build prototypes and let sampled individuals and test market communities respond to the actual goods or service. The organization with such a product, be it a magnetic stove or a new public transportation system, has a critical need for information on how the eventual customer perceives alternative formulations of the product or service and the customer's current preference or ranking of the alternatives. To obtain generalizable information, an organization usually generates what it hopes is an all encompassing set of attributes that describe the alternatives. In practice the number of attributes employed in marketing concept testing is usually nontrivial (n > 10) [1]. Since one doesn't really believe a person thinks in terms of 10 or more independent dimensions in evaluating a product or service concept, an approach, consistent with some psychological theories, is to reduce the attribute space into a more manageable number of variables. Two general sets of psychometric techniques for accomplishing this purpose have been proposed in the marketing literature. First, there are the multidimensional scaling techniques (MDS). The second method is a covariance analysis approach commonly called factor analysis. Utilizing one of these two general approaches the researcher attempts to identify underlying or latent factors. These underlying factors often serve as an aid to the researcher in reducing the original extensive list of attributes to a subset of relatively independent attributes which become inputs to a choice model. Alternative concepts can then be compared in terms of a choice model. For example, there is the standard linear multi-attribute attitude model which has
No takes yet. Share an insight, caveat, or question.
Gensch et al. (1975) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: