When groups should not imitate their most successful members

Abstract

The imitation of successful peers is often heralded as an intelligent shortcut to reduce individual learning costs. Using computer simulations, we demonstrate that this advice can be ill-founded and harmful in a cognitive inference task involving continuous learning. In particular, success-based imitators perform worse than both learners who integrate the learning experience of all group members and isolated learners. We report on sensitivity analyses for this phenomenon and offer explanatory mechanisms.


Back to Table of Contents