This post originally appeared on the HASTAC website at: https://www.hastac.org/blogs/kjordan/2014/06/01/exploring-network-structure-academic-social-networking-sites
When I started my PhD, I was interested in studying academic social networking sites (such as Academia.edu, for example), but needed to find a focus for my research. A literature review identified three themes: development of academic identity; communication; and collaboration. Each would have provided ample opportunities for further research, but I needed to refine it further.
However, I was troubled by the fact that none of the existing studies had looked at network structure in academic social networking sites, when this has been shown to be critical in other contexts in terms of determining the types of functions that the network can perform. In order to address this and help determine the direction of my main project, I carried out a pilot study to explore network structure. I sampled academics from my home institution (The Open University UK, OU) across three platforms – Academia.edu, Mendeley, and Zotero – and visualised the network of academics and their connections. I chose to sample academics from my home institution so I could follow up on the reasons behind trends in network structure.
Emerging characteristics of the network structure
The networks of OU-affiliated academics on each platform were entered into Gephi, in order to visualise the networks and carry out basic social network analysis metrics. It wasn’t possible to use the Zotero data for network analysis; although there was a sizeable sample of OU academics using the site, there were very few connections between them (this suggests that OU academics at least aren’t using Zotero as a social networking site). From the network analysis, two trends in network structure emerged, which were consistent across both the Mendeley and Academia.edu datasets. Subject areas appeared to play an important role in defining the structure of communities within the network, while academic seniority affects the position of individuals (with more senior academics having more connections, and occupying a more central position, than more junior colleagues). Network graphs colour-coded according to discipline and further information about how degree and centrality varied according to seniority can be found in this presentation.
Delving into network formation
In order to explore the network structure trends and makes links to the themes from my literature review, I followed up on the network analysis by circulating an online survey to the OU academics in my sample. The survey asked participants about their level of use of academic social networking sites, and contained Likert scale items in relation to the themes of identity, communication and collaboration, and also network forming behaviour. The network formation items showed differences according to academic seniority, with more junior academics being more open to making connections (see slide 12 for the full list of items). Differences also emerged according to how active academics reported themselves to be, with more active academics finding sites more useful. There is also a link between academic seniority and level of activity, with the most senior academics being the least active. This is an interesting contrast with the network analysis; while the most senior academics seem to occupy the most privileged positions in the network, they are the least engaged.
It would be interesting to repeat the study using samples from a range of different institutions, to see whether different types of institutions foster different network structures. Future work could expand the network to include links between institutions, for example by sampling and mapping all of the UK Higher Education academics.
In terms of finding the direction for the main part of my PhD study, the differences according to discipline and academic seniority suggest a role in academic identity development, so I plan to look at the personal networks (ego networks) of a sample of academics across a range of sites that they use.
If you would be interested in participating in my study (which will involve completing an online survey in the first instance, in the Autumn), please register your interest using this online form, and I will invite you to take part when the survey is live.