Skip to main content



Bow Tying the Social Network

While we have talked about relational graphs and power of social network, we might not have observed their connections to the network’s structure. According to the article “How Close Are You Really?” from the MIT Technology Review, social scientists have been trying to measure the strength of ties based on the network structure, combining major concepts in class. The author explores the work of Heather Mattie at Harvard University and her colleagues and reports that the team found a specific pattern: the bow-tie structure. In their research, they studied the strength of links in two completely different settings. For the first, they survey 70,000 people in 75 rural villages in India, asking “which friends and relatives visit the respondent’s house, which people the respondent would borrow from or receive medical advice from or go to the temple with, and so on” (“How Close Are You Really”). After gathering data, they connected 37,000 between 17,000 people who responded and used the number of social links from the survey to measure the social ties between people. In addition, Mattie’s team also constructed a social network between mobile-phone users in a European country. Choosing a random sample size of 500,000, they constructed a social network based on the number and length of calls between them and assumed that the link between two people was stronger if they called each other and the total amount of time spent talking was long. From the graph visualization, they were able to discover that the number of common friends between two individuals were strongly correlated with the strength of the tie between them. These findings supported two prominent hypotheses, one from Elizabeth Bott and one from Mark Granovetter. Bott hypothesized that the closer you are with a group of people, the less strong links you have outside of the group, while Granovetter claimed that the stronger the ties between two individuals, the higher the fraction of friends in common. The bow tie framework captured these ideas, suggesting that the structure of their social network can help measure the strength of links between people.

Although we have talked about bow tie structures in the Web, we have not seen an example where we could apply the structure to a social network. In Chapter 13, we covered that there is a central “core” containing prominent pages within the Web, along with other nodes that lie upstream, downstream, or to the side of the core. The picture can give us a global view of the Web, yet they provide limited information about the patterns and connections within the different parts of the structure. However, in the article, we have cases where we could analyze the strength of ties between individuals just by the structure, something we cannot do with the Web. It is possible that because there is a direct interaction within the social network, the strengths and patterns are more apparent. If we see the Web in terms of human interaction, we could conclude that each page is a form of communication between people. However, since the page is not a direct way to communicate with others, it becomes more difficult to analyze the relationships between pages. Web communities might be considered directly communicating with others, but the users are usually anonymous, have little information about others, , etc. therefore, these limitations create this difficulty to observe the details of links in the Web, while Mattie can use the structure of the bow tie to reveal information about how strongly connected you are with friends. Mattie’s research makes a valid connection between the structure of a social network and the strength of ties between the network, but these settings are still minimal – she must continue to observe similar patterns with other situations and environments. And looking at different social connections within the Web and social media could contribute to the analysis of the different patterns of parts within the Web.

Source:

 

Comments

Leave a Reply

Blogging Calendar

November 2018
M T W T F S S
 1234
567891011
12131415161718
19202122232425
2627282930  

Archives