Multiattribute Dyadic Choice: Models and Tests

Abstract

Three models for resolving a cooperative, dyadic choice problem are formulated and solved. Solutions are presented in n -dimensional attribute space (versus the utility plane) to focus on cognitive conflict between decision makers rather than on goal conflict. Forecasts from the models are compared in a choice experiment. All three models outperform chance in predicting a dyad's first choice, but the nonzero reference point model best predicts second through n th places—a desirable feature for product positioning applications. Recommendations for incorporating the model into choice simulators and expert negotiating systems are provided.