PRODUCT TESTING 1: THE DISTINCTION BETWEEN S‐R (STIMULUS‐RESPONSE) AND R‐R (RESPONSE‐RESPONSE) MODELS

There are two approaches to modeling key relations among variables when one tests products. S‐R or stimulus‐response modeling assumes that the researcher controls the antecedent physical variables (such as ingredients or processing), and that these physical variables are the primary cause of product‐to‐product differences. R‐R or response‐response modeling assumes that the researcher can measure co‐varying physical measures of a food, but may or may not have control (or even knowledge) of the antecedent physical variables that generate product differences. S‐R modeling allows for true optimization, in terms of defining the operations needed to maximize an attribute (e.g., acceptance). R‐R modeling allows only a guess as to what particular combination of physical measures would correspond to a maximum level of the attribute. Often S‐R and R‐R modeling and optimization are confused with each other, leading to incorrect conclusions.