2nd Module SNA Proposal- Noemi

(I will be taking the second part of the module.)

I would possibly like to use SNA for my thesis work, which is, broadly, on mobilization through social media. My case study is Tunisia, and my research question (still under way) is the following: How can new media help sustain the democratic momentum in the Tunisian post-revolutionary context? I will be focusing on human rights advocacy through new/social media (thus dwelling on the latter’s capacity to mobilize people, and more particularly the youth). I will probably be conducting field work in April, possibly evaluating the work of a NGO/ project using “online advocacy”.  However, I still have to collect the data, and thus exploring this topic in the next few weeks reveals difficult. However, I hope to exploit the skills I’ll learn during this term and in this class to conduct a SNA, and will be working on that project in parallel to the module.

On a completely different topic, my SNA proposition for this class:

“Birds of a feather flock together”: testing homophily in SNA

General topic, research questions

At the beginning of the term, all Fletcher students were submitted to a personality test (Myers-Briggs) and were issued a "personality type".

My research questions are the following: Do the friendships people develop have anything to do with their personality types? i.e. Do people tend to develop friendship with people that are similar to them?  If yes, is there any consistency with similar interests (i.e. class)? If not, are there relevant patterns?

Hypotheses

1) Rather than similar, complementary: I hypothesize that it is not very likely we will be observing “personality clusters”. This does not mean patterns are random,  or that “opposites attract”. Rather, complementarily is the hypothesis I make: such a concept exists in the model developed by Myers-Briggs called “mirror personalities”, i.e. certain personalities that tend to “match better” but are quite different from one another. 

2) Friendships: personalities over interests: It can be argued that students in schools mainly socialize through specific interests, which are translated in the courses they take. At Fletcher as well, we are encouraged to share ideas, join interests groups, and discuss topics in class. However, I hypothesize that friendships are not necessarily formed over common interests, and I’d be interested to see what role “personality type” may play in that matter. In sum, I hypothesize that friendships bridge the interests-clusters, and that personality compatibility plays a greater role in developing friendships than common interests.

Methodology

The data I will use will be possibly based on the results of 2^nd year students. They indeed may be more accurate in the sense real friendships have had the time to form over the year, thus giving more depth to the analysis. However, this is an assumption that is worth examining.

I would start by conducting a survey (by mail), asking different questions:

1/ Personality type (that I will use as an attribute)

2/ Identify 3 to 5 persons you spent most time with (i.e. interact on a daily basis)

3/ Specializations at Fletcher (i.e. areas of interest, also used as an attribute)

Addressing hypotheses

1.  I hypothesized that “personality groups” will be formed according to “mirror-personalities”. I will first identify patterns in the network (organizing it by personality types) by measuring density and identifying possible clusters.  Other measures will include eigenvector degree of personality types to identify further patterns (are certain personalities more “popular”? outliers? ... )

2.  I hypothesized personalities are more important than interests in the development of friendships. I could start by measuring the eigenvector degree of interests (as nodes) to see if this data is relevant with the patterns identified in step 1, i.e. if people that were identified as friends are also interested in the same area of studies.

Interpretation and application

As Professor Tunnard suggested, I might need a sociologist or psychologist to help me draw interesting conclusions from the data analyzed, and see if they are consistent with theory.  I am sure the network will provide us with interesting, if not answers,  at least questions for further research: is there a “personality type” fit for academia?  Do personality types at Fletcher reflect the general distribution of these types among tested populations? Is there any subject/interest that draws certain personalities?  How can these data be used by, for instance, career services to better advise the students on their career paths?

Limitations

The personality type test cannot be expected to offer a comprehensive account of one’s actions and behaviors. The analysis thus can only be made through general patterns, and personality types cannot be the sole explanation for the friendships developed over the year. Further, a complete survey (which would need way more time to analyze than what is imparted to us in this module) would need to address the following, subjective, questions: how do you qualify your relation to the persons you identified (friends “like peas in a pod”, good friends,  person I enjoy, acquaintance) ; how did you meet these persons (in class, at work, at informal events)…

I however believe the general patterns that will be discovered through this SNA will offer interesting, interdisciplinary, avenues of reflection and I’m looking forward to testing my hypotheses.

See you in class on Friday!

Noemi