Concept Testing New Products: Errors vs. Segmenting Variables

Abstract

Utilizing data from an actual concepts testing experiment, this paper presents a statistical analysis of the impact of problem- and benefit-oriented concepts statements as well as several “variables-in-audience” such as: Product Involvement, Use of the Product Category, Level of Compliance, and Level of Education. These predictors are evaluated in terms of their impact on the criterion variables, operationalized as components of an emerging attitude, and are discussed as either the elements of the segmentation scheme, or as errors biasing the representativeness of the sample.