Abstract:

Discrimination methodologies are used to investigate similarities and differences among variants of manufactured products. Applications typically involve cost savings, storage and product improvements. A researcher often requires more information than a simple test of significance which can be readily changed by simply altering the sample size used. In many cases, estimation of the degree of difference between pairs of products is more important. Two indices obtained from discrimination methods have been largely used in industry: d’ and the R-Index. R-Indices are usually obtained from the Same-Different and A/Not A (yes/No task) with sureness judgments or using a ranking procedure. d’ values can be estimated from the same procedures as well as from a variety of discrimination and scaling methods (triangle, duo-trio, 2-alternative forced choice or 2-AFC, category scaling, etc.) Both indices measure degrees of difference for which a greater value indicates a larger sensory difference. R-Indices take values between 50% (no difference) and 100% (maximum differences), while d’ values vary between 0 and infinity. Usually d’ values between 0 and 2 are encountered when using discrimination and scaling methodologies. Both measures have the same goal. However, are they interchangeable?