Modeling Framing Effects: Comparing an Appraisal-Based Model with Existing Models

One significant challenge in creating accurate models of human decision behavior is accounting for the effects of context. Research shows that seemingly minor changes in the presentation of a decision can lead to shifts in behavior, phenomena collectively referred to as framing effects. This work presents a computational modeling analysis comparing the effectiveness of Context Dependent Utility, an appraisal-based approach to modeling the multi-dimensional effects of context on decision behavior, against Cumulative Prospect Theory, Security-Potential/Aspiration Theory, the Transfer of Attention Exchange model, and a power-based utility function. To contrast model performance, a non-linear least-squares analysis and subsequent calculation of Akaike Information Criterion scores, which take into account goodness of fit while penalizing for model complexity, are employed. Results suggest that multi-dimensional models of context and framing, such as Context Dependent Utility, can be much more accurate in modeling decisions which similarly involve multi-dimensional considerations of context. Furthermore, this work demonstrates the effectiveness of employing affective constructs, such as appraisal, for the encoding and evaluation of context within decision-theoretic frameworks to better model and predict human decision behavior.