Reflections on social network analysis as a research method (SOCRMx: assessed post)

There are a number of methods I could use to gain a deeper understanding of my core interests – social inequality and higher education, as separate areas, or to study their relationship with each other. But, as I intend to narrow my field of inquiry by investigating the potential role of technology as a mediator, certain research methods start to become particularly interesting, especially social network analysis and learning analytics. So, for my third research method analysis, I have selected Social Network Analysis.

Although social mapping activities have a much longer history, digital technologies now enable us to study all sorts of relationships between individuals and groups, across global networks, which could bring any number of benefits and risks depending on the intentions of those who use this information. In his TEDx talk ‘Ending up on the wrong side of the tracks‘, social network analyst, Valdis Krebs illustrates how this might affect certain communities and individuals (Krebs, 2012). The use of learning analytics to support students and prevent disengagement is already quite widely adopted, but attentions are starting to turn to the admission process where the potential to ‘predict whether students will succeed and graduate’ (Felton, 2015) has not gone unconsidered. I came across quite a few other potential areas where this type of technology could become a force for good or bad when it comes to tackling educational inequality, with some good examples found on the Herchinger Report website.

With this rapidly-evolving educational concern in mind, I have looked at a range of sources to find out what SNA approaches might be particularly useful to my endeavours:

What three (good) research questions could be answered using this approach?

As SNA could have so many potential applications, it was an initial challenge to determine what would constitute a ‘good’ research question. I ended up taking quite an interesting online dérive between journals, wikis, marketing and social media analytics pages, and government and university websites.

I found Grunspan, Wiggins, and Goodreau’s (2014) ‘Primer for Social Network Analysis in Education Research‘ particularly helpful in responding to this and the other questions in this assignment. Their section on ‘Ties as Predictors of Performance’ immediately caught my attention as relevant to some of my aforementioned areas of interest:

‘Understanding study group formation and evolution is both interesting and important, but we are not limited to questions focused on network formation. As educators, we are inherently interested in what drives student learning and the kinds of environments that maximize the process. We can start addressing this broad question by integrating student performance data with network data.’

Personally, I would be keen to exchange their ‘student performance’ with ‘student experience’ in the way that Lee and Bonk (2016) studied relationships in their journal article: Social network analysis of peer relationships and online interactions in a blended class using blogs (see final section). In their investigation, the “environment” would be reflective blogs, and the “experience” would be a sense of community and ‘perceived closeness’. It is also quite relevant to my professional practice, as a learning technologist in HE, as it ties into our Domain of One’s Own project, and the NSS theme of ‘learning community’.

So, after some consideration, my three “good” research question examples are:

  1. Using a position generator approach, what can we learn from Facebook networks about the social, cultural and economic capital of university applicants?
  2. Using the same technique within Facebook networks, how does the social, cultural and economic capital of university students change between application and graduation?

In the above examples, I liked the idea that one question could lead to another, thus breaking down the process to focus on specific areas. The third choice looks at a different topic area:

  1. Is there any correlation between students’ public political affiliation and engagement on social media networks and academic performance?

With any of the example questions I’ve written for this course, I can see ways to expand or narrow their scope, depending on the purpose.

What assumptions about the nature of knowledge (epistemology) seem to be associated with this approach?

I think that SNA seems most at home within a constructivist view of society, particularly when social media is involved. Social media profiles and networks tell us a lot about how people see themselves in relation to each other and the world around them. Social media provides individuals with so many tools and languages to curate external representations of themselves and their world, which one might argue could only ever really be subjective. By analysing this information, a researcher is ascribing value and meaning to social phenomena, power relationships, and the constructed meanings of individuals and groups.

Of course, SNA data or research could be compared to quantitative research, using a positivist paradigm, in an attempt to demonstrate the existence of an objective reality beyond the façade of social media.

What kinds of ethical issues arise?

Although it would theoretically be possible to gain consent of some target groups and individuals to analyse their network behaviours, and unless the research involved a closed group e.g. employee intranet, the global and dynamic nature of social networks would make it very difficult to gain informed consent from everyone. Given the public nature of a lot of social media content, it may not be necessary to gain any consent from individuals who have arguably agreed for their public content to be used. That does not mean that SNA is free from ethical scrutiny, however. For example, despite minimum age requirements of some sites, there may be children participating in social networks, and it is important to note that vulnerable adults may also be present.

The findings of any research is usually designed to have an impact on the community, or issues affecting the community, so this should also be borne in mind. Even if the research sets out to show something that could benefit the community, the results might not support the theory and could unearth unintended findings that could have a negative impact. Using my third example question (above) to illustrate, the aim might be to show that students who are interested and engaged in politics perform better academically. But what if the research found that this only applied to one type political affiliation. What might educational institutions do with this information?

What would “validity” imply in a project that used this approach?

Validity can have a number of meanings, but it tends to evaluate whether the findings effectively answer the question. In terms of how the findings of SNA research might be applied, one thing to consider is how the data was sourced. If this was primarily from social media networks, then its interpretation may be limited by its exclusion of individuals who do not have the ability, or prefer not to engage in social media. This could mean that findings reveal more about people who use social media generally than the actual community being studied.

Theories drawn from more traditional social network research may not transfer well to the language, behaviour and social norms of social media and virtual spaces. Howison, J. Wiggins, A. and Crowston, K. (2011) raise concerns about ‘the manner in which SNA concepts are translated to research using digital trace data’, arguing that ‘digital trace data are of a different nature than those used in earlier studies using SNA’. Any interpretations of the findings need to be judged against how the community itself might understand and apply them. Here’s one example of how this kind of ‘credibility’ has been defined, to illustrate this point:

‘the match between an evaluator’s representation and the ‘constructed realities’ of respondents or stakeholders’ (Guba and Lincoln, cited in Coe 2017, p46)

Guba and Lincoln also listed five types of authenticity as criteria for representing and honoring ‘the beliefs, values and understandings of all participants appropriately’: ‘fairness’, ‘ontological authenticity’, ‘educative authenticity’, ‘catalytic authenticity’, and ‘tactical authenticity’ (ibid, p47)

What are some of the practical or ethical issues that would need to be considered?

Depending on the sources of data, it is likely that SNA is one of the most practical research methods. If social media and online environments are the sources, then research can probably take place anywhere with a stable internet connection and access to relevant analysis tools. If the study was going to take place within a private online space, then access to data may need to be provided by the organisation or site moderator, and with consent of the community. There may also be some context-specific practical challenges, e.g. countries where internet activities are controlled/monitored by government may require a lot more negotiation and preparation.

Find and reference at least two published articles that have used this approach. Make some notes about how the approach is described and used in each paper, linking to your reflections above.

1) Curtis and Bonk (2016) Social network analysis of peer relationships and online interactions in a blended class using blogs (also discussed briefly above). The research followed the online interactions of a class of students to see whether/how blogging affects their network and relationships, in particular ‘perceived closeness’ and whether this had any effect on their offline relationships. The concept of ‘perceived closeness’ aligns the study with a constructivist, qualitative paradigm. They identified a sense of ‘ownership’ as a means of reducing blogging anxiety (ethics). Data was gathered and analysed in different ways, which, in addition to network diagrams, included a paper survey to measure things like ‘perceived closeness’ (mixed method).

2) Toubøl , J. and Larsen, A.G. (2017) Mapping the Social Class Structure: From Occupational Mobility to Social Class Categories Using Network Analysis. Toubøl and Larsen developed new clustering algorithm, the Mobility Network Clustering Algorithm (MONECA). They used this to analyse Danish intra-generational mobility, and they feel this differs from most studies, which tend to look at class formation rather than mobility. They have mapped occupational shifts occurring between 2001 and 2007, and later reflect on other social factors like gender, income and education. They identify two ‘interrelated limitations’ in their study:

  • ‘the occupational classification may weaken the validity of the results if the categories do not fit the actual job-positions in the economy’ and;
  • ‘the number of observations in the data imposes limits to how disaggregated a level we can start from. A too sparsely populated mobility table lowers the validity of the subsequent analysis’.

They also note the limitation of any universal class scheme or theory given  ‘significant variation between different societies’. They gathered their data from official records, so ethical implications mainly lie within the impact and application of their findings. Although this was not a study of online data, I feel something similar could be achieved through publicly-accessible social media data.

I came across the latter of my two selected papers while revisiting a BBC project that was published back in 2013 and fascinated me ever since – The Great British Class Calculator: What class are you? While reading about SNA for this course, and with my field of interest in mind, I thought about the criteria the BBC project had used to determine someone’s class and it occurred to me that many of their indicators of social, cultural and economic capital could probably be identified within social networking profiles, through “likes”, group affiliations, hashtags, and relationships with others. This informed the first two question examples I gave earlier on in this post, including reference to Nan Lin’s ‘position generator technique’ (Lin 2001, cited on BBC website 2013).

I had some technical issues getting some of the Facebook network mapping apps to work, but had better luck with Twitter tolls like SocioViz. I’m quite excited about exploring SNA practice further, as the course progresses.

screenshot of Facebook analysis


BBC website (2013, April 3rd) How do you identify new types of class? Available from

Coe, R.J. (2017) Inference and interpretation in research. In Arthur, J. Waring, M. Coe, R.J. and L. Hedges (eds) Research Methods and Methodologies in Education. London: Sage. pp.44-56

Felton, E. (2015, August 21st) ‘The new tool colleges are using in admissions decisions: big data’, Retrieved from

Grunspan, D. Z., Wiggins, B. L., & Goodreau, S. M. (2014). Understanding Classrooms through Social Network Analysis: A Primer for Social Network Analysis in Education Research. CBE Life Sciences Education13(2), 167–178.

Howison, J. Wiggins, A. and Crowston, K. (2011) Validity Issues in the Use of Social Network Analysis with Digital Trace Data, in Journal of the Association for Information Systems (JAIS.) Retrieved from

Krebs, V. (2012, June 14th). Ending up on the wrong side of the tracks [Video file]. Retrieved from

Lee, J. and Bonk, C.J. (2016) Social network analysis of peer relationships and online interactions in a blended class using blogs, in The Internet and Higher Education, Volume 28, pp 35-44, ISSN 1096-7516,

Toubøl , J. and Larsen, A.G. (2017) Mapping the Social Class Structure: From Occupational Mobility to Social Class Categories Using Network Analysis. Sage Journals: Sociology.  . DOI: 10.1177/0038038517704819

1 Trackback / Pingback

  1. #SOCRMx Week 4: ‘half-way’ tasks and reflections – Jenny Connected

Leave a Reply

Your email address will not be published.