Comparing rank-dependent, subjective weight, and configural weight utility models: Transitivity, monotonicity, coalescing, stochastic dominance, and event-splitting independence

This study compared three models that attempt to predict decision making behavior. One hundred subjects were instructed to choose between gambles that were defined in terms of the probabilities to win different amounts of cash. Three models, Configural Weight Theory (CWT), Cumulative Prospect Theory (CPT), and Subjective Weighted Utility Theory (SWU), were tested for their ability to describe choices between gambles. Three properties that distinguish these models were tested: Coalescing, Stochastic Dominance, and Event Splitting Independence (ESI). Coalescing implies that if two or more outcomes have the same value within a gamble, they can be combined by adding their probabilities, and the combined gamble will be perceived as equivalent to its split form. CPT assumes coalescing whereas SWU and CWT predict violations. The data showed violations of coalescing when the high outcome was split, contrary to CPT. Stochastic dominance implies that coalescing, monotonicity, and transitivity have been satisfied. CPT implies stochastic dominance, whereas SWU and CWT predict violations. The data showed violations of stochastic dominance under conditions predicted by CWT, contrary to CPT. ESI assumes that the effect of splitting any event is independent of the value and relative position of the outcome associated with it. Since CPT does not predict violations of coalescing, the theory does not make a prediction for ESI. Although SWU predicts violations of coalescing, SWU implies ESI. CWT predicts violations. The data, however, were inconclusive for ESI. Further research is called for in order to test ESI.