How to get past “I have more MySpace friends than you”

Online social networks are a big interest of mine (I haven’t stopped writing about the topic in months). I have often bemoaned the lack of understanding about how we operate in these social networks. Without this understanding online social networks (or Social Network Services – SNS) are little more than shopping malls, coffee houses or beauty/barbershops. A place for some good conversation, but highly susceptible to being overly influenced by crowd mentality – who is popular now, how many friends does s/he have – and centered around gossip. But social networks have the potential for so much more. They offer us the possibility of understanding group behavior (or Social Network Analysis – SNA) in ways that were impossible before. Why is this important? Applying SNA to SNS can give us ways to interact in a rich context with people all over the world. Orgnet.com has a great list of applications of SNA in the real world. Here are just a few:

• Uncloak criminal network amongst slumlords
• Reveal how hospital-acquired infection[HAI] spread amongst patients and medical staff
• Map and weave nationwide volunteer network
• Improve the innovation of a group of scientists and researchers in a worldwide organization
• Unmask the spread of HIV in a prison system

There has been a lot of rigorous work done to understand offline social networks using highly academic methodologies. But, I have yet to see this methodology fully applied to online social network services. Let me define my terms here so that we don’t get confused. Social Network Services (SNS) are the online tools, Facebook, Orkut, MySpace, etc. Social Network Analysis (SNA) pertains to research social scientists are doing to understand behavior within networks (mostly offline to date, but increasingly this research is including online networks).

My brother, who is getting his PhD in Government, is writing about disenchantment with Mexico’s emerging democracy. As such he has had a lot of opportunity to view the research that has gone into understanding how people behave in social groups as that behavior relates to voting patterns. He pointed me to a post by David Lazer on Complexity and Social Networks Blog that clearly articulates the findings of a recent study detailing contagion of voters whose household members were targeted in a Get Out The Vote (GOTV) campaign. Even if the household members were not the recipient of the GOTV message the contagion effect made them more likely to vote.

ABC News has a Facebook group that includes a poll of who people are supporting in the Presidential race. Obama is at 63% compared to Clinton’s 19%. Ron Paul is at 36%. Obviously these do not represent the populace in general, so one would assume that perhaps a certain age or sex demographic might come into play. However, looking at the demographics we find that the breakdown is not skewed enough to make that large of a difference. So what does this mean? To be honest I have no idea. But, I believe that SNA could help us better understand these trends and without it, these numbers are little more than an understanding of how if only Facebook voters could vote what the election would turn out to be. In short – the data is meaningless. Now we could get off here on a tangent of how polls only show the horse race and have the potential to negatively affect voter turnout, but we won’t do that here. Imagine, however, that we used Facebook information for good rather than evil to analyze unlikely voters, identify likely voters and use the likely voters to influence the unlikely voters. Yes, I realize the privacy implications here. But that question aside, such analysis would bring about a pretty large social benefit through a better understanding of Facebook behavior.

At WWW2007 a paper was presented by a number of professors at the Korea Advanced Institute of Science and Technology. The paper was one of the most advanced analyses of SNS based on SNA methodology I have seen. But what hampered their ability to continue the research beyond simply understanding limited examples of friend relationships into centrality, flow and other SNA metrics (see below) was a lack of available data. I certainly hope that this research is going on in the depths of the online services, but I doubt it could be carried out to nearly the degree that academics could accomplish. Let me add that there are companies that are doing this research within the organization or corporation. But this research is often highly silo’d and does not pertain to SNS as much as software and content behind company firewalls. Online services should make their data available open to researchers. Obviously the researchers can be required to sign strict non-disclosure agreements. But in order to develop social networks beyond the high-school mentality of “how many friends do you have” we need to build more structure in how we understand online behavior.

The Wikipedia article on Social Network Analysis identifies metrics that I believe are essential to building a better understanding of behavior and therefore more useful SNS.

Betweenness

Degree an individual lies between other individuals in the network; the extent to which a node is directly connected only to those other nodes that are not directly connected to each other; an intermediary; liaisons; bridges. Therefore, it’s the number of people who a person is connected to indirectly through their direct links.

Closeness

The degree an individual is near all other individuals in a network (directly or indirectly). It reflects the ability to access information through the “grapevine” of network members. Thus, closeness is the inverse of the sum of the shortest distances between each individual and every other person in the network.

(Degree) centrality

The count of the number of ties to other actors in the network. See also degree (graph theory).

Flow betweenness centrality

The degree that a node contributes to sum of maximum flow between all pairs of nodes (not that node).

Eigenvector centrality

a measure of the importance of a node in a network. It assigns relative scores to all nodes in the network based on the principle that connections to nodes having a high score contribute more to the score of the node in question.

Centralization

The difference between the n of links for each node divided by maximum possible sum of differences. A centralized network will have many of its links dispersed around one or a few nodes, while a decentralized network is one in which there is little variation between the n of links each node possesses

Clustering coefficient

A measure of the likelihood that two associates of a node are associates themselves. A higher clustering coefficient indicates a greater ‘cliquishness’.

Cohesion

The degree to which actors are connected directly to each other by cohesive bonds. Groups are identified as ‘cliques‘ if every actor is directly tied to every other actor, ‘social circles‘ if there is less stringency of direct contact, which is imprecise, or as structurally cohesive blocks if precision is wanted.

(Individual-level) density

the degree a respondent’s ties know one another/ proportion of ties among an individual’s nominees. Network or global-level density is the proportion of ties in a network relative to the total number possible (sparse versus dense networks).

Path Length

The distances between pairs of nodes in the network. Average path-length is the average of these distances between all pairs of nodes.

Radiality

Degree an individual’s network reaches out into the network and provides novel information and influence

Reach

The degree any member of a network can reach other members of the network.

Structural cohesion

The minimum number of members who, if removed from a group, would disconnect the group.^[10]

Structural equivalence

Refers to the extent to which actors have a common set of linkages to other actors in the system. The actors don’t need to have any ties to each other to be structurally equivalent.

Structural hole

Static holes that can be strategically filled by connecting one or more links to link together other points. Linked to ideas of social capital: if you link to two people who are not linked you can control their communication.

Some SNS have adopted a few of these. LinkedIn, for example, shows you Betweenness and Closeness. Other metrics such as Radiality are hard to program. Obviously some research is being done into how to easily graph influence or reach (rating engines, friend maps). But these are still very rudimentary, and we need to build tools to encompass the whole kit and caboodle.