Judgment Based Marketing Decision Models: An Experimental Investigation of the Decision Calculus Approach

Abstract

Managerial judgment is frequently required to estimate many of the parameters of decision calculus models and the quality of these judgmental inputs may substantially affect model-based decisions. This paper suggests that if model builders are to rely upon managerial judgments in building models, research should be directed at understanding when managerial judgments will be valid and the types of biases that might be expected. A quasi-experimental design is used to explore managers' abilities to estimate the parameters of a decision-calculus model (ADBUDG) and to examine the value of this model in decision-making. The results are consistent with previous evidence of the existence of biases in human judgment. More specifically, they indicate that a manager's experience in a limited region of a nonlinear response function does not enable him to accurately predict decision outcomes or parameters in the unfamiliar regions and that model usage may, in certain situations, actually lead to poorer decisions.