From Weak Ties to Strong Ties: Community vs. Social Networks 3

Good food for thought Michael,

I'd add a fourth point, namely that strong ties and weak ties are fluid notions based on time and context. You yourself use terms such as "at some point in my life", "frequent engagement", "time spent together". It's like the person in your College group of friends that never got married and the rest of the group that did; at one time that person will ahve more difficulties to relate to the life of the others and what was once a strong tie can peter out. And there are hundreds of thousands of examples like that

In terms of online communities, it may be interesting to look at how ties evolve over time as interests evolve or simply the lifestage changes: a tie is not set in stone (but I guess you will cover this in stage 3...)., so it may be just as interest to research how strong ties devolve into weak ones - and maybe do something about it...  :)

Cheers,

Mark

Hello Mark and Mike

Thank you for the comments.

It is defnitely true that communities are dynamic entities, so are social networks. Ties form, grow in community, and then weaken over time when members leave the communities. It is the natural course of things. That is why there need to be 3 stages of relationships as mentioned in my earlier post. So weak ties or strong ties are definitely based on time and context.

I hope I didn't give the impression that they are static. Thank you for pointing this out. Good observation.  :)

However, nearly all the social network analysis (SNA) out there do not take into account of time. In fact, SNA inherently eliminated the temporal dimension (essentially freezing time) for simplicity. But there is another field called Dynamic Network Analysis (DNA) that takes into account of the temporal evolution of the network. DNA essentially combines SNA with time series analysis from statistics, and it is much more powerful than SNA.

You guys are way ahead of the crowd. Next time, I will discuss precisely what happen when you leave a community due to change interest, moving on to different stages of life, job changes etc. That should finalize this introductory mini-series that explore the anthropological roles of communities vs social networks.

Evolution of tie strength is a bit tricky, but we can certainly do it, since we compute the tie strength between every interacting pair of users in our community. Will try to put some analysis together and see if the results are interesting enough to merit a blog post. :)

Great suggestion, Mark. I see what you mean, and would also be interested to hear Michael's thoughts on that. Some obvious ones that jump out are proximity and frequency of interaction. If someone leaves the community (online or off), or frequents it (or contacts their friends in other ways) less often, ties tend to weaken.

Hello Hans,

Thanks for the affirmation of the value of this little research project of mine.

Yes, the Internet + web 2.0 technologies definitely make the creation of weak ties much easier (either via community or social network). There are actually some social psychology theory that can explain why that is the case. This particular aspect has to do with a sense of safe distance. This is a topic that I promised that I will cover after I finish exploring the social anthropological roles of community and social network in human history.  See previous post when I talked about The Desire to Connect is a Basic Requirement to move through the 3 stages of relationship lifecycle.

I also like to take the opportunity here to address some of your question in email, so others can pitch in if they like.

Having been through so many discipline of science and research. I've come to the realization that reality is usually never as simple as any model that people can build or any theory that people can come up with. People can always build  models and hypothesize theories, but none will ever be complete. Good models/theories that tend to stay are those that can explain a lot with very few simle underlying principles. And great models and theories are those that have predictive power.

The key is to understand when will a theory/model work, and under what conditions will the theory/model break down. For example, Newton's law of motion works great on earth. But when you are traveling near light speed or near very massive objects that are millions times larger than the our sun, Newtonian Mechanics fails! And you need Einstein's Theory of Relativity to make sense of things. In the ultra-tiny regime, Newtonian Mechanics also fail, and Shroedinger, Dirac and other bright modern physicist develop Quantum Mechanics to deal with small particles like atoms, electrons, and other subatomic particles. There are so many more case that I can think of.

But to your point, social media is very new and many people are still exploring and testing their own hypothesis. We are all still learning. Lots of new theory will be proposed and refuted in this pre-paradigm era before the good/great one emerge and stay. So whether it is social objects, social network analysis (SNA), or dynamics network analysis (DNA), they all have their validity. None of them explains everything, but that doesn't mean the theory is bad. Some expains more, but that doesn't meant it's better either (if the theory is too complex).

Newtonian Mechanics, although incomplete, still governs all the civil/mechanical engineering that builds the building we live and work in, and the cars we drive every day. Relativity, although more complete is too complex that most people only use it when they have to. You get the idea, is the Occam's Razor, the simplest theory that explain the most.

With respect to DNA, not many people use it out there, because it is a more complex theory than SNA and SNA is already quite daunting for some people already. Since most data are stored by date, many people just apply SNA for a series of dates (e.g. daily, weekly or monthly). This method seem to fits better with the data architecture within most organization. I know LinkedIn uses SNA. But I am not sure if FB use SNA, although I'm quite certain that they are not using DNA. However, DNA is more powerful when you are looking at many years of network data all at once, rather than week by week or month after month. Researchers and scientists tend to do that lot more than enterprises, which focuses more on the daily, weekly, monthly operational metrics.

Alright, this reply is getting a bit long... But thank you for dropping by and commenting.

Hello Carolyn,

Thank you for commenting.

I am not 100% sure if I get your point. But my understanding of the situation you described can be sufficiently analyzed simply by social network analysis (SNA). Using SNA you will be identified as a bridge (or connector, boundary spanner, hidden influencers, etc) between the multiple communities. That is, if we compute the betweenness centrality using SNA, you will have extraordinarily high value for this SNA metric. I've written about an example of such analysis in an earlier post "Social Network Insights from Unconventional Graph Metrics."

The time dimension is a characteristics distinction between Dynamic network analysis (DNA) and SNA, because all DNA applications deal with changes of the network over time. That is why it is called Dynamic. However, not all problems required the temporal dimension to reach a good solution. So DNA is often not used because most of the time SNA is sufficient. Just as Einstein's Relativity is rarely used for problems we encounter on earth, because Newtonian Mechanics are plenty sufficient for that, even though Relativity is a much more complete theory and much more powerful.

DNA as opposed to SNA will show its superior strength when dealing with complex network systems where the entities (the nodes and vertices in a graph) and relationships (edges in a graph) can both change over time. An example of change of relationship is when two person were friends today, but become enemies tomorrow. An example of change of entities is where a person change from being healthy to being infected. Another example would be if a caterpillar metamorephosed into a butterfly. As you can see in my examples, DNA is used frequently in conflict and resolution studies and their simulations, as well as disease propagation models in epidemiology, or animal population dynamics in ecology.

So having multiple communities interacting with each other or with actors mediating the interactions between them would probably not be considered as a kind of DNA. Nevertheless, you have posed a very interesting SNA problem that I've recently analyzed. If this subject interests you, I highly recommend that you take a look at "Social Network Insights from Unconventional Graph Metrics." The discussion following that post may also be interesting to you, if my memory still serves me well.  :)

Glad to have been able to add my 2c here :)

I had never heard of Dynamic Network Analysis before, it seems like a very interesting field of research, thanks for drawing my attention to it!

We have been talking about finding the influencers, but if their influence is waning over time, or on the contrary, we miss that their influence is increasing, we miss an opportunity to provide 'added support' to them.

Keep up the good work!

Cheers,

Mark

@MarkTamis

Thanks Michael for sharing a highly valuable blog post with the community.

I would like to add that it is much easier and less time consuming to create a large number of weak ties online. The treshold (social barrier) to initiate online interaction with strangers is much lower compared to the offline world. Most of us never talk to strangers in the subway simply because we don't know if we share the same interest. The online community help human beeings to find likeminded people. The total time you invest in a relationship (social capital) will always be taken into account. That´s why it is so easy to reconnect with people you havn't seen for ages.

Hans Leijström

Community Evangelist

@iceman65 / LinkedIn

Hey Mark,

Glad I can bring up an interesting subject to you. It happens all the time for me due to my interdisciplinary background: Math, Physics, Neuroscience, Statistics/Machine Learning and now Social Analytics. I often find solutions to a difficult problem in one domain from a totally different domain of knowledge.

With regards to influencers, time (temporal relevance) is definitely very important. Not sure if you read my social media influence model. One important factor in this model that many people overlooked is timing. This post actually give an example using SNA at different temporal slices of the data. This is another way (besides DNA) to incorporate temporal dimensions into the static nature of SNA. Essentially we do many static SNA over time, so you get the temporal evolution of the SNA metrics. This allow you to monitor people's influence over time and see when they are rising and when they are falling. This way you can pick the most relevant influencer for reaching your target.

Thanks again for the comment.

Hello Mike, something I am interested in, is  DNA when time and events within a group of communities perpetuate the making of strong ties because of the event and the timing of it.  Is not my social network enhanced when two or more communities come together around this event and it's timing? For instance, if I am interested in a particular LEED standard test analysis that a local green organization, Build it Green has to offer, I might join their community and listen in. Then I might join a similar dialogue going on within the United States Green Building Council, a national community, on the same topic, all because of a seminar on the subject coming to my area. When one program lends it's expertise to another, say in a local testing service for the national standards, then the possibility of my linking to both communities and the interconnectivity between the two communities, allows me greater chances for my own social networking and those people coming together for that event, at that point in time Does that make sense?

I am a novice at this kind of thinking, being that I am a designer by trade, yet the prospects of using the 'community to community' linking fascinates me. Isn't it sort of the way you describe the familiarity accessed by one individual because of their community affiliations or interests?   But instead, one community is bolstered because of events and their timing, by those following such events and seeking information from a multitude of communities. Would that not also be a form of Dynamics Network Analysis?

Carolyn

Hello again Mike. Thank you for explaining to me the difference between this SNA example and how DNA is different. I enjoyed reading your unconventional graph metrics although at times it was a bit over my head. I do understand how one connector can make all the difference and not be a pro and that is very interesting. What I didn't ask correctly, in my last note is this: because of certain information around a time frame, wouldn't communities shift, and dynamics shift and because of this better exposure is possible? Remember, I am not a social scientist, so I'm out of my comfort zone here. I appreciate your patience with me for this reason.

The reason I am interested in these various dynamics within a given time period,  is this: it is my belief, that if I were to track green design events and connect to people before, during and after these events, I have a better chance of making contacts and creating bridges, giving me a better network with new communities, because of the frequency most connections are made around these times. Is there any data that you know of pertaining to this idea?

Connecting to people around design events might mean 1. people gain more information, assimilate information faster and communicate more  2. the communities they connect to, shift, and morph, based on these interactions. 3. the communities change in context to the events sometimes just as a butterfly comes out of a cocoon.  

This is more accurate I'm sure for certain professions of course. The design communities often alter considerably after a seasonal show of products and styles. This in turn offers new dialogue and in turn new possibilities. Sometimes even new communities develop because of a particular show. I know this is true for a green festival where more LED lighting was shown than ever before. Now new communities develop around LED lightiing products. Is there more possibility for connection and forming strong ties because of this product and in turn because of this event, is my question.

I enjoy reading your articles Mike. You make your point so well and with good diagrams and theories. Thank you.

Carolyn