Statistical information about reward timing is insufficient for promoting optimal persistence decisions

Area of Science:

• Cognitive Psychology
• Decision Science
• Behavioral Economics

Background:

• Optimal strategies for waiting for delayed rewards vary based on reward timing distributions.
• Heavy-tailed distributions necessitate quitting at a certain point due to high opportunity costs.
• Predictable distributions allow for extended waiting periods.

Purpose of the Study:

• To investigate how individuals learn to optimize waiting times for delayed rewards.
• To determine if understanding the reward timing distribution is sufficient for learning optimal strategies.
• To explore the role of direct experience versus probabilistic reasoning in learning optimal waiting behavior.

Main Methods:

• Participants decided how long to persist for delayed rewards before quitting.
• Information about reward timing distributions was provided through various methods: counterfactual feedback, prior exposure, and explicit description.
• Learning was assessed based on participants' decisions in a feedback-driven context.

Main Results:

• Providing information about reward timing distributions did not eliminate the need for direct, feedback-driven learning.
• Learning optimal strategies for quitting waiting appears to rely on task-specific experience.
• General knowledge of reward timing distributions alone was insufficient for optimal strategy expression.

Conclusions:

• Learning when to quit waiting for delayed rewards is heavily influenced by task-specific experience.
• Probabilistic reasoning based on distribution knowledge alone may not be sufficient for optimizing waiting behavior.
• Future research should focus on the mechanisms of feedback-driven learning in delayed reward contexts.