Steven Johnson
Assistant Professor,
Fox School of Business,
Temple University
February 25, 2011
Speakman Hall 200, 1000am – 1130am
Seminar Title : How do power law distributions arise in online communities?
Abstract
Power law rank/frequency distributions appear ubiquitous in online communities but the mechanisms of their formation are not well understood. This study models online communities and multiple network formation mechanisms that can lead to the emergence of power distributions. First, we establish the presence of power law distributions in twenty-eight online communities. Next, we develop a simulation model of the formation of thread-based asynchronous online communities and provide results based on over 4,500 runs of the model simulating a total of over 3,200,000 messages generated by over 340,000 participants. Finally, we evaluate if these network formation models generate simulated networks with power law distributions. To validate that these models are consistent with the observed networks we use multiple measures of network structure: the power law distribution degree, network density, mutuality index and clustering coefficient. This study contributes to our understanding of online communities and other social communication networks by illuminating the relationships between specific behavioral tendencies of participants and emergent structural network characteristics.
We find no evidence that preferential attachment explains the presence of power laws in online communities but instead that a generalized social exchange mechanism is the participant behavior most consistent with observed power laws.
Please email me for a copy of the full paper (swattal@temple.edu).