A recent paper in PNAS discusses work by examining how human social networks can reach consensus. I discuss some of it at A Man With A PhD.
The key is that they have an experimental system to examine different social network architectures and how they are able to arrive a decisions. It should be a useful tool.