Six Degrees: The Science of a Connected Age

Stephanie Viggiano
3 April 2009

Response to Six Degrees:  The Science of a Connected Age by Duncan J. Watts

I chose this book because I was interested in how people are connected in today’s global village, how fascinating it is that we can be perhaps six degrees from the President of the United States.  I felt that the book wasn’t as “easy” to read as Blink or The Tipping Point.  Watts being an expert in network theory contained more figures and scientific data in his book which was a harder to read than other books which seemed to be more psychological.  In fact, a book review online suggests that you have a good math background as Watts speaks in graphs and graph theory very often.  This essay gives a little taste of what I found in Watts’ book.

To begin, the Six Degrees of separation experiment was the Small World Experiment (1967) by Stanley Milgram, who was a psychologist at Harvard University.   His Small World experiment was supposed to see how people were connected to each other.  The goal for his test subjects was to send an envelope to a stockbroker in Massachusetts only by forwarding it to people they knew that would have more of a chance of possibly knowing the stockbroker.  Consequently, in a few articles that I have read, Milgram did not follow up on all of the envelopes, which then brought the controversy:  is Six Degrees really legitimate?

The author of Six Degrees: The Science of a Connected Age is Duncan Watts, is a sociologist at Columbia University who also works at Yahoo! Research in the Human Social Dynamics group.  Watts focused on covering different types of networks and how they interact, as well as different types of situations, some explained below.  To make a comment about Evan’s book, The Tipping Point—I’ve learned that Watts opposed Gladwell’s idea that a few people essentially manage social connectivity.

Watts talked about a number of different models.  I will be discussing here what models I understood in context.  He discussed isolated networks (much like a biological example of how a new species would form and separate) and interconnected networks (how a species would evolve from a few different types of species to another hybrid kind) – which “locally-occurring” events affect a whole entire network.  Another kind of network that he discussed was randomness (and random graphs) in social networking/the small world.  Watts and his friend Steve Strogatz made the Alpha model to show how mutual friends are connected and the likelihood that they would be connected in graph form.  He described the model that they created as being very close to the mathematician Anatol Rapoport’s random-biased networks.  Both models basically showed that through one “single tunable parameter” as Watts put it, the likelihood of one person meeting another could increase or decrease based on that.  Hand in hand with the Alpha model, a page later he goes into explaining about Path length as a function of alpha.  At the top of the graph (the critical alpha value), clusters are formed in small areas that in turn make the network shrink.

There is also another model that Watts and Steve Strogatz created called the Beta Model.  The Beta model shows more effectively than the Alpha model how “close” the nodes are to each other.  Watts also spoke about the theory of decision making and how this deals with networking.  As mentioned in many other presentations that we’ve had during the past few weeks, the Kevin Bacon Number experiment is a good example of what degrees people are associated with Kevin Bacon.  Another fascinating model is the Affiliation model which basically is the graph of how people associate with different social groups/ethnicities.  The more people that share and overlap groups, the more likely that they will be friends.  Just discussed in my social psychology class, people are more likely to be friends with people that are more like themselves (the distance between the nodes/people in the model are contingent upon the associations that are given from each person).

Watts also talked about searching networks—communicating between nodes.  The broadcast type of search sends out signals to all nodes, to find the quickest path.  This has been proven not as efficient as a directed kind of search.  Milgram used a direct searching network frame.  He told his subjects to send their envelopes, with certain criteria as to how to send them.  Unlike the broadcast type of search, it may not get there quickly, but it still functions.

Focusing on recent events, Microsoft has now been a participant of the Six Degrees experiment, as well as European mobile service provider O2.  An article from TechDirt (http://www.techdirt.com/blog.php?tag=stanley+milgram), a popular technology blog, says that O2, the UK mobile service provider, is postulating that we have moved to three degrees of separation because of how powerful our technology of communication is;  O2’s results are not to be compared with Milgram’s experiment though.  O2 based their Three Degrees of Separation on a certain pool of “shared” user experiences.   This article and another article (given above in the hyperlink) further mentions Microsoft’s involvement in the Six Degrees experiment.  Microsoft claims that Six Degrees may not be entirely wrong.  The article states:  “The researchers looked at data on how people use Microsoft's MSN Instant Messaging software, and discovered that the average chain length to connect any two users on the software was 6.6, and that 78% of all random pairs could be connected in fewer than 7 hops. Of course, what isn't accounted for is whether or not this has changed in the 40 years since Milgram's experiment, during which technology may have made connectivity much easier.”  The article makes a good point; we become “friends” with people that are acquaintances and vice versa.  It is even common to add someone that you may have met one time somewhere.  Does that really mean that we are that “close”?

Through reading this book, I have also gained interest in emergence in today’s technological community.  Like we have discussed in class, the internet has no central growth and therefore has a different kind of network structure.  It comes out of many places (servers, etc.) that connect all around the world.  Reading this book has given me more of an insight as of how networks can be formed (what different kinds of networks there are), and what kinds of criteria the nodes in the network determine a connection.


Sources:
Six Degrees: The Science of A Connected Age, Duncan J. Watts, Norton Publishing:  2003.
http://www.techdirt.com/articles/20080819/0257472025.shtml
http://www.techdirt.com/articles/20080804/0052121879.shtml
Wikipedia, Wikipedia.org - Stanley Milgram, 2009, <http://en.wikipedia.org/wiki/Stanley_Milgram>. Wikipedia, Wikipedia.org - Stanley Milgram, 2009, <http://en.wikipedia.org/wiki/Stanley_Milgram>.