In Search of the Preference Reversal Zone.

Area of Science:

• Behavioral economics
• Decision-making science
• Psychology

Background:

• Preference reversals, where choices between smaller-sooner (SS) and larger-later (LL) rewards change as they approach, are common.
• Hyperbolic delay-discounting models are often cited to explain these reversals due to their simplicity.
• Previous research has not reliably predicted the timing of preference reversals, potentially due to overlooking the impact of reward magnitude.

Purpose of the Study:

• To test the predictive accuracy of hyperbolic and exponential discounting models for preference reversals.
• To investigate the influence of reinforcer magnitude on discounting rates and reversal prediction.
• To compare the goodness-of-fit for hyperbolic versus exponential models with empirical data.

Main Methods:

• A pre-registered study assessed individual discount rates for two different reinforcer magnitudes.
• Participants made choices between SS and LL rewards at varying temporal delays.
• Individualized predictions of preference reversal timing were generated for each participant and model.

Main Results:

• Little evidence supported the ability of either hyperbolic or exponential models to predict the temporal location of preference reversals.
• Hyperbolic models showed slightly better predictive performance than exponential models.
• Exponential models provided a superior fit to the observed choice data compared to hyperbolic models, contrary to some prior studies.

Conclusions:

• Current hyperbolic and exponential discounting models have limited success in predicting the precise timing of preference reversals.
• The influence of reinforcer magnitude requires careful consideration when modeling delay discounting.
• Further refinement of discounting models is needed to accurately capture human choice behavior, with exponential models showing promise in fitting observed data.