The effect of opponent type on human performance in a three-alternative choice task

Area of Science:

• Behavioral Psychology
• Cognitive Science
• Reinforcement Learning

Background:

• The generalized matching law describes choice behavior in response to reinforcement schedules.
• Human decision-making can be influenced by perceived social contexts.

Purpose of the Study:

• To investigate how perceived opponent type (computer vs. human) affects sensitivity to reinforcement schedules in a three-alternative choice task.
• To test the applicability of the generalized matching law in scenarios with varying perceived opponent identities.

Main Methods:

• Adults played computerized
• Paper Scissors Rock
• against a programmed opponent.
• Participants were randomly assigned to believe they played against a computer or another human.
• Reinforcement schedules were manipulated across trials, and sensitivity was analyzed using the generalized matching law.

Main Results:

• Participants who believed they were playing against a computer showed significantly higher sensitivity to the opponent's response distribution.
• Sensitivity to reinforcement schedules was lower when participants believed they were playing against another human.

Conclusions:

• Perceived opponent type is a crucial factor influencing the efficacy of reinforcement contingencies on human behavior.
• Findings suggest caution when generalizing results from human-computer interactions to human-human interactions.
• The generalized matching law adequately describes human three-alternative choice behavior, but its parameters are modulated by external factors like perceived opponent type.