Group decision-making in animals

Summary of: Group decision-making in animals

Author(s) / Editor(s)

Analytical results from modeling the fitness consequences of two decision-making mechanisms, despotism and democracy, shows that generally despotic models leads to higher costs than democratic models because despotism produces more extreme decisions than democracy.

Disciplines

Publication Reference

Published in/by
Nature, Vol 421, 1/9/2003, pp. 155-158
Date
2003

Findings

  • Although animal behavior in groups is usually presumed to follow despotic models, there is mounting empirical evidence of democratic behavior demonstrated through body postures, ritualized movements, and vocalizations, as well as vote counting in the form of summing up votes, integration of votes up to an intensity threshold, and averaging of votes.

Analytical results from modeling the fitness consequences of two decision-making mechanisms, despotism and democracy, shows that generally despotic models leads to higher costs than democratic models because despotism produces more extreme decisions than democracy. "Even when the despot is the most experienced group member, it only pays other members to accept its decision when group size is small and the despot's average error is lower than the average median error of all other group members."

Research has largely assumed despotism because social structures among animals are commonly hierarchical and the ability to vote and to count votes is not obvious. However, there is mounting empirical evidence of voting in the animal world by body postures, ritualized movements, and vocalizations, as well as vote counting in the form of summing up votes, integration of votes up to an intensity threshold, and averaging of votes.

An important context in which social animals have to make group decisions is activity synchronization, e.g. red deer herds have to decide when to end rumination and move on. The model assumes that (1) synchronization costs increase linearly with the difference between when an individual would have preferred to stop and when the group actually stops, and (2) costs of stopping to early or too late are symmetrical. Even when relaxing assumption (1) costs are still higher for despotic than for demographic groups in most cases. When relaxing assumption (2) a democratic majority rule different from simple majority that reflects the asymmetry between "too early" and "too late" costs is least costly.

These results are fairly robust with respect to group heterogeneity, energy needed for enforcement, and individuals having incomplete information about their own optimal activity duration. The model predicts that democracy gives groups a competitive advantage and due to natural selection should be quite common in social groups of animals.