Testing Expectancy Models: A Conjoint Measurement Approach

Abstract

The term expectancy theories refers to a category of social cognitive theories that use the same underlying Subjective Expected Utility (SEU) model for describing goal directed behavior. However, the mathematical SEU formulation has in many expectancy theories a dubious validity. Expectancy theorists lack an adequate measurement model to test their formal theory in a valid way. The conjoint measurement approach is proposed as a possible solution for this problem. In two experiments we tested an expectancy model respectively by means of the UNICON scaling algorithm for conjoint measurement and by means of an axiomatic approach. Evidence was found for the functional form of the SEU model, which implies that the subjective probability (expectancy) and the utility of an outcome of an action combine multiplicatively while the resultant products combine additively. In the discussion, it is argued that only the descriptive validity of a model has been proven. More research is needed to identify the cognitive processes that can be described by addition and multiplication.