Competing theories of multialternative, multiattribute preferential choice.

Area of Science:

• Cognitive Psychology
• Decision Science
• Mathematical Modeling

Background:

• Preferential choice experiments utilize various mechanisms (e.g., lateral inhibition, loss aversion) to explain decision-making dynamics.
• Existing models of context effects are often not rigorously tested quantitatively, leading to a preference for simpler approximations.

Purpose of the Study:

• To quantitatively evaluate the performance of four extant models of context effects using a novel fitting method.
• To investigate the relative importance of different mechanisms within decision-making models by testing all configurations.

Main Methods:

• Applied a recently developed quantitative method to fit four established context effect models to empirical data.
• Utilized data from two experiments featuring consumer goods and perceptual stimuli.
• Systematically tested all mechanism combinations within an overarching model to assess their contribution to performance.

Main Results:

• Lateral inhibition, loss aversion, and pairwise attribute differences were identified as crucial mechanisms enhancing model performance.
• The study found that several mechanisms contribute positively to the predictive power of preferential choice models.
• Model performance varied across different configurations, underscoring the importance of specific mechanisms.

Conclusions:

• Mathematical tractability should not be the primary driver for model development in decision science.
• Model fit, balanced with complexity, is paramount for accurately describing empirical phenomena in preferential choice.
• The findings emphasize the empirical value of specific mechanisms like lateral inhibition and loss aversion in understanding choice behavior.