Michael Macy: "Complex Contagion and the Weakness of Long Ties"

Teleconferenced lecture to be held on Monday June 12, 2006, 18:15.
Room D 16.2, Main Building, ETH Zurich. Event broadcasted from , in Cambridge, Massachusetts.

The strength of weak ties is that they tend to be long – they connect socially distant locations. Recent research on “small worlds” shows that remarkably few long ties are needed to give large and highly clustered populations the “degrees of separation” of a random network, in which information can rapidly diffuse. We test whether this effect of long ties generalizes from simple to complex contagions – those in which the credibility of information or the willingness to adopt an innovation requires independent confirmation from multiple sources. Using Watts and Strogatz’s original small world model, we demonstrate that long ties may not only fail to speed up complex contagions, they can even preclude diffusion entirely. Results suggest that the spread of collective actions, social movements, and risky innovations benefit not from ties that are long but from bridges that are wide enough to transmit strong social reinforcement. Balance theory shows how wide bridges might also form in evolving networks, but this turns out to have surprisingly little effect on the propagation of complex contagions. We find that hybrid contagions, which have high thresholds for some nodes and low-thresholds for others, can propagate on perturbed networks if the latter are sufficiently numerous. However, for purely complex contagions, propagation depends decisively on wide bridges, a characteristic feature of spatial networks. This may account in part for the widely observed tendency for social movements to diffuse spatially.

Michael Macy is Professor and Chair of Sociology at Cornell. In a series of studies funded by the National Science Foundation, his research team used computational models and laboratory experiments with human subjects to explore how threshold effects in network interactions might generate familiar but enigmatic social patterns, such as the emergence and collapse of fads, the spread of self-destructive behaviors, and the polarization of opinion. Macy pioneered the use of agent based models in sociology to explore the effects of heterogeneity, bounded rationality, and network structure on the dynamics and stability of social systems (http://hsd.soc.cornell.edu/Macy.htm). He now heads a team of social, information, and computer scientists who are building tools that will make the Internet Archive (http://www.archive.org/) accessible for research on social and information networks (http://www.news.cornell.edu/stories/Sept05/NSFcybertools.dea.html). He also leads a Cornell initiative to promote cross-disciplinary collaborative research and teaching on social and information networks.