Can a Flow-Network Approach Shed Light on Children's Problem Solving?

Abstract

A complete theory of human behavior must capture both the apparent randomness of behavior (its strong context dependence) as well as its stability (a surprising resistance to changes in behavior despite salient changes in the context). Recent efforts in cognitive science have made important discoveries toward such a theory, emphasizing the idea that skilled behavior seeks to balance overregular tendencies with tendencies that are overrandom. The hallmark of these efforts is the idea of self-organized criticality, the state of a system poised toward maximally adaptive behavior, characterized by pink-noise pattern of variability. In this article, we expand these efforts, looking for a new measure to capture the balance between order and disorder, one that can be applied to small data sets of categorical performance. The proposed measure borrows ideas from information theory, previously applied to the stability of energy flow in an ecosystem. Using published data on a problem-solving task with preschoolers, we describe ways in which this measure could be applied. Results are promising, opening the possibility for studying the trade-off between randomness and stability in children's reasoning.