Experimental Evidence on the Relation Between Network Centrality and Individual Choice

Social interactions shape individual behavior and public policy increasingly uses networks to improve effectiveness. It is therefore important to understand if the theoretical predictions on the relation between networks and individual choice are empirically valid. This paper tests a key result in the theory of games on networks: an individual’s action is proportional to their (Bonacich) centrality. Our experiment shows that individual efforts increase in centrality but at a rate of increase that is lower than the theoretical prediction. These departures from equilibrium are accompanied by significant departures in individual earnings from theoretical predictions. We propose a model of network based imitation decision rule to explain these deviations.