July 5, 2015

Publication

Methodology Brief: Social Network Analysis

1) Basic information:

Method: Social Network Analysis

Author: Jordan Soukias Tchilingirian, PhD candidate in Sociology, University of Cambridge and Research Manager Public World

Contactjst35@cam.ac.uk

Twitter@JordanTchil

2) Description of method and how to use it:

Objectives 

Research about think-tanks can be divided into four broad categories: historical studies of a polity’s think-tank tradition, studies which attempt to classify these diverse organisations, studies which seek to understand their role and impact on party politics or specific policies, and reflections on the actual ideas created by think-tanks and what the rise of these hybrid intellectuals might mean for expertise in the public sphere. My research is closer the final strand as it asks “how do think-tanks make knowledge?”

To answer this question I developed a relational theoretical and methodological approach. This means the unit of analysis was the networks which emerge from the process of making knowledge. Rather than developing a comparative case study, my approach placed greater attention on the wider context in which think-tanks are situated. Thus my study was a case study of a social space. This is a space between professions and is populated by a plurality of actors and interested groups from across the knowledge-policy nexus. Therefore I sought to understand how think-tanks enrol the worlds of media, politics, business and academia in the creation of their intellectual interventions. To ground this study I focused on British think-tanks.

Think-tanks produce a number of publicly focussed interventions (e.g. blogs, tweets, Oped pieces and so on) all of which represent important sites a one could study. I chose to focus on policy reports as they do not just contain an argument and findings. They can also be viewed as an artefact of the wider social world which it is embedded within. Scan the first few pages of a think-tank document and you are likely to find names of authors, funders and those people whose help was acknowledged as important in bringing the document to publication. You will also notice these people range from colleages, members of other think-tanks, academics, civil servants, people from NGOs, practitioners and businesses. Therefore within a typical policy report we are able to locate, not just the think-tank’s (or individual’s) ideas, but the people and organisations the researchers has interacted with during the production process.

Description of the method

Social Network Analysis (SNA) is the systematic study of relationships between any set of actors.

  • Quantitative social network analysis uses the mathematical methods to analyse and visualise the structure of relationships and position of actors (or nodes) within these relationships.
  • Qualitative social network analysis focuses on similar issues but seeks to understand the meaning, processes, histories, and social life which constitute a relationship.

In this presentation I will only be focusing on the former.

The quantitative element was based upon data gathered from 27 purposefully sampled British think-tanks from across the ideological/political and organisational spectrum. From these think-tanks I was able to construct data based on the names of the funders, authors and advisors from 2,875 individual policy reports. This process produced a list of 501 funders, 2,842 authors and 4,219 advisors. I then cleaned the data and constructed basic profiles (education, career, gender) of each individual using search engines, professional social media sites such as LinkedIn, official directories (e.g. Whose Who?) and organisational websites. Using Excell I cleaned and prepared the data for descriptive and statistical analysis in the specialist social network analysis software UCInet.

Quantitative analysis enabled me to study the space of British think-tanks from two perspectives.

  1. The first examined the relational structure of the 27 British think-tanks in order to explore the patterns of collaboration and competition between these organisations. I was able to test whether the networks I had gathered corresponded with patterns of ideological polarisation or whether think-tanks cluster into specific groups based on certain organisational attributes.
  2. The second approach focused on the wider social space think-tanks are embedded within and focused on the actual people who are active in this world. Following the work of Kathryn Oliver (2012), who used social network analysis to discover powerful and influential actors in public health policy, my analysis followed two routes.

The first was to establish whether there is core of actors within the network. The existence of a core would suggest there are a number of individuals who the majority of think-tanks select or consult during the production of their policy reports. This would suggest that the knowledge these actors’ poses is of some value to think-tanks. I found this was the case.

The second sought to uncover identity of these ‘core individuals’ by using measures of network centrality and brokerage. These measures consider which individuals hold special positions within a network: central actors are those who receive the most nominations by the largest number of think-tanks and brokers are actors who are nominated by think-tanks from across the ideological and organisational divide. Once these individuals were identified I could focus on their career and personal biographies and consider the professions they come from and resources they hold. This approach offered a helpful proxy by which to comprehend the types of knowledge and skills that are perceived as valuable within the space.

Practicalities

Although these approaches offered insights more traditional methods could not provide this method raises several key issues. The first is time. Following the exact approach I used is both time consuming and tedious. It required me cutting and pasting the names of every author, funder and acknowledged individual from nearly 3,000 individual documents. Web-scraping could be an option, however as I was using PDFs it was harder to consistently use this method to harvest information. If you can stomach this (or can find a quicker method!) the results are worthwhile.

A second problem is more methodological. Using quantitative methods makes it impossible to understand meaning and dynamics of these relationships. Do central actors or brokers somehow coordinate the think-tanks they are nominated by, or does the think-tank have a level of agency in the relationship? How does the information flow into the think-tank? How are relationships established, and what exactly is the role of the think-tank researcher in all of this? Answering these questions requires qualitative methods to explore the social life which constitute these networks. This is exactly the approach I took and this will be explored in in my second presentation

3) Further resources:

Introduction to social network analysis:

Borgatti, S. P., Everett, M., & Johnson, J. C. (2013). Analyzing social networks. London: SAGE,.

Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods. Riverside, CA: University of California, Riverside. A free online e-book which can be found at http://faculty.ucr.edu/~hanneman/

Kadushin, C. (2012). Understanding Social Networks: Theories, Concepts, and Findings. New York: OUP USA.

Knoke, D., & Yang, S. (2008). Social Network Analysis (2nd ed). Los Angeles ; London: Sage.

Prell, C. (2012). Social Network Analysis: History, Theory and Methodology. London: Sage.

Scott, J. (2012). Social Network Analysis (Third Edition edition). Los Angeles: Sage Publications Ltd.