Tim Oren points to an HP paper on information flow in social groups to predict my impending doom in the Mayfield-Shirky Cage Match.  The paper studies email use by 30 clients inside the organization and 10 outside.  However, the study combines both datasets to simulate a Power-law for the purpose of the study, information flow.  The paper identifies that it does not capture the outbound email of the external dataset, skewing the distribution.  I believe this was intentional to simulate the scale-free network characteristics that exist in large email networks, to allow detailed analysis down to individudal messages. 

I'll continue to assert that weak tie networks are scale-free while strong tie networks, segmented by the capacity constraints of people (12, 150) have a more even distribution of connections. 

What the study does show is that social networks are not epidemic in distributing information.  There is a low probability that a given message will be widely distributed even in a simulated scale-free network.

When people are nodes in a network, they are selective about what information they pass on.  In contrast to embedded information, such as advertisements in Hotmail messages, every forwarding decision exists in social context constrained by the fitness of the message, norms and relational reciprocity. 

Within formal groupings of the organization, information flow is dense and ties are strong, and the study implies great difficulty spreading information at epidemic porportions through weak ties.  The stronger the tie the greater the information flow.