Examining learning coherence in group decision-making: triads vs. tetrads.

Area of Science:

• Cognitive Psychology
• Computational Social Science
• Behavioral Economics

Background:

• Group decision-making effectiveness is influenced by group size.
• Previous research has not fully explored the impact of group size on learning dynamics in small groups.
• Understanding learning properties in groups is crucial for optimizing collective performance.

Purpose of the Study:

• To investigate whether group size (three vs. four members) affects performance and learning in group decision-making.
• To test the hypothesis that larger groups (tetrads) exhibit learning incoherence compared to smaller groups (triads).
• To apply a computational reinforcement learning framework to analyze group learning behavior.

Main Methods:

• Utilized a reinforcement learning framework with Q-learning algorithms.
• Estimated learning parameters to quantify group learning dynamics.
• Compared performance and learning properties between groups of three (triads) and four (tetrads).

Main Results:

• Tetrads demonstrated underperformance compared to triads, supporting the hypothesis.
• Evidence of learning incoherence was observed in tetrads, potentially due to the inability to apply majority rule with subgroup emergence.
• The computational approach successfully identified underlying learning parameters in group decision-making.

Conclusions:

• Group size significantly impacts learning coherence and decision-making performance.
• Triads are more effective than tetrads in group decision-making tasks requiring coherent learning.
• Computational modeling offers a valuable tool for dissecting learning mechanisms in small group dynamics.