Collective Problem Solving on Small World Networks

(with Daniel Diermeier)

Project Description

How does the social structure of a group, organization, or society impact its collective performance? Recent work in network science puts forth the idea that a certain class of networks, called “small world” networks, may provide unique performance advantages. Small world networks are characterized by densely connected subgroups that are loosely connected to each other through a small number of nodes who serve as bridges. The theoretical appeal of small worlds comes from the following finding: if one starts with a world where people are only connected to their immediate neighbors, exchanging local connections for bridges to more distant parts of the network decreases the average number of steps it takes for people to reach each other much more quickly than it deteriorates the local structure which gives it its “community” nature. Consequently, small world networks are thought to provide the advantages of increased access to resources and information from different groups, without sacrificing the benefits of local structure, or “community.”

A limited amount of empirical work has attempted investigated the correlation between small world network properties and the performance of teams, organizations, or industries, resulting in mixed conclusions.  But even if such work were more conclusive, testing for the theorized benefits of small worlds, and perhaps ultimately applying such findings to improve real social systems, requires going much further.  This is because it is the mechanisms implied by the network characteristics of small worlds that impact performance, and not the network itself. 

In this project, we investigate the underlying mechanisms thought to give rise to performance advantage in small worlds through the use of a laboratory-controlled experiment. More specifically, we will investigate how quickly and accurately large groups solve an “information aggregation” problem – one that requires coordination to discover a piece of information that they all share – under different network topologies for communication (i.e., varying who can communication with whom).