Social Network Dysfunction - Part I

You've probably heard that "a chain is only as strong as its weakest link". While that's true if I'm trying to pull my Bobcat 763 out of mud, it’s not true of social networks. Although social networks are comprised of both nodes (which are hubs when highly connected) and links, the fabric of the network itself is somewhat self-healing. This is because traffic going from point A to point E in the network has numerous optional paths available. Sometimes the network itself can be programmed to take various paths between points for efficiency, reliability or speed.

Consider the electric power grid where on-demand electricity is generated and delivered within seconds across a dynamic path across many hubs and links determined by the operational policy of clusters of independent system operators.
	U.S. Electric Power Grid map courtesy of the U.S. Department of Energy

Another example is the U.S passenger airline system. Passengers can choose various routes and levels of service in air transportation, based on their budget, schedule or airline preference.

But networks do have dysfunctional components, and although the impact of dysfunction is negligible to member nodes, the dysfunction has an impact in the aggregate quality and overall effectiveness of the network.

The concept of node fitness was introduced by Barabasi and is discussed in the book, Linked. Understanding a node's fitness, leads us to our first dysfunction, a dysfunction that I call node dystrophy or node dysfunction, which in technical terms would be equivalent to a node having no fitness, as opposed to a node having great fitness. In an online social network like LinkedIn, this would be like entering "Bob B" for a name and nothing else in his profile. If "Bob B" was connected to anyone, his dystrophy would have a minimal impact on the network as a whole, but maintaining a network link costs the network something. If there were 100,000 atrophied nodes in the network, the network performance and effectiveness would surely be noticed. It would be similar to having thousands of potholes on an interstate highway, avoidable, but quite a nuisance and it would affect one's speed and comfort of travel. Interestingly, if "Bob B" wasn't connected to anyone, there wouldn’t be ANY impact on network traffic or performance at all, as no path would ever be chosen that would pass through "Bob B".

Another dysfunction is what I call link dystrophy or link dysfunction, which occurs because two nodes are not connected well enough, or don't know each other well enough to add value to the transmission through the link. In network science lingo, this would be the condition where coupling strength is low or near 0. This is the potential problem with "link spam" where people just try to get lots of links, not considering the quality of each link connection.

We’ll go into some other symptoms and causes of social network dysfunction in Part 2.