SNA, NS, CSS: what are we doing?

01Jul12

The rise of social media has brought a new life to the academic field of social network analysis (commonly referred to as SNA). Traditionally grounded in sociology with applications to neighbouring fields such as management and education, it has now expanded to a variety of other disciplines including economics, geography, psychology, science studies and even history. In parallel the mathematical basis of SNA, graph theory, is thriving with new developments that today’s computational power, and the increasing availability of “Big Data” make possible. Applications are spanning disciplines far beyond the social sciences, including physics and biology. Analogies across these various fields suggest grouping their shared formal underpinnings under a single label such as “Network Science” (which I’ll refer to as “NS” for brevity, though this acronym is admittedly uncommon). The massive use of computing power, both to explore large datasets and to build models (sometimes borrowed from the hard sciences) has led some social scientists to think of today’s SNA as an incarnation of Computational Social Science (CSS, again for brevity), a category that is sometimes taken to include other specialisms such as agent-based simulation.

I was thinking a lot about these matters at the latest conference of the UK Social Networks Association in Bristol. Is SNA changing its nature today -losing its own tradition and specificity as a result of its very success? Is it being subsumed into a more multidisciplinary, but less sociologically informed, NS? Is it discharging its “old”, non-computationally intensive past -all those studies done with tiny-sized, offline networks such as Padgett’s Florentine families of the Renaissance (16 actors) or Sampson’s Monastery (18 actors)? And if yes, what of this tradition can still be preserved, and what can insights can it bring that can nourish today’s reflections on millions of Facebook users or Tesco customers?

For one thing, my feeling is that SNA scholars have been slow in taking notice of these changes. The latest UKSNA conference featured three sessions and one round table on social media and online communications -but only two years ago, I counted only two papers! Not to mention earlier conferences.

On the other hand, there are (too) many papers on social media around, that do not integrate any SNA element at all…

The key point is that it is still unclear how SNA may contribute and respond to today’s challenges. This became very clear at a round table on SNA and social media at the conference, chaired by Bernie Hogan with a group of panelists including Dimitris Christopoulos, Martin Everett, Daniele Quercia and myself. SNA has great potential, but we still need to understand the extent to which its earlier results and findings are applicable to today’s data and questions. It is not just a matter of devising metrics, measures and software that work with large datasets; my point is that there is an unmet and unrecognised need for theories. We don’t know how social influence works in online networks. How social media help to form social capital, and how comparable this is to the social capital we knew offline. Whether local or global network effects are prevalent. How online/offline multiplexity works. And so on.

I do not mind if a more all-encompassing NS or CSS ultimately emerges as a new, more comprehensive paradigm beyond old SNA. But before this happens, we still need to draw the lessons of SNA and see how we can build upon them, or we will just reinvent the wheel.

—

My little contribution to the development and spread of the SNA tradition in an “Introduction to SNA” workshop I ran on 27th June. The slides are here. Together with theoretical elements, I used Gephi to familiarize participants with network data tools and visualisations.

I also presented a paper on Culture and Privacy in Social Media, co-authored with A.A. Casilli and Y. Sarabi and part of my THEOP project.