Call me if you can : an experimental investigation of information sharing in knowledge networks

In the public promotion of R&D cluster and network formation, the following situation typically arises: An initial network structure has developed over a long time span and policy measures affect the structure of links between the actors. This new network structure influences the effectiveness of the information flow in a way that is not clear from the beginning. As analyzing the effects of a change in the network structure is difficult in the field, this paper uses a laboratory experiment to analyze how information is distributed in four different network structures. Networks are modeled as five-actor groups. Every individual represents a node and possesses some private information. The experimental results suggest that the different network structures do indeed influence the way information is exchanged. Both too many possible links (causing a coordination problem) and too few possible links (introducing bottlenecks) are harmful. The participants in all network structures learn over time and achieve a faster exchange of information in the later rounds. These results suggest that when influencing communication structures, one has to be careful to balance the positive and negative effects of adding more communication possibilities. -- Network ; communication ; laboratory experiment ; information flow