Sentiment … ality?

During my pilot studies, a couple of findings suggested areas for further exploration I’d not previously considered. One of these was the degree to which people talking or writing about Twitter seemed to be ‘affected.’ Although it was not a topic I had gone looking for, nor had asked questions about, and although people rarely mentioned it explicitly, the language and terms they used implied some element of emotional response. Before I could take this much further, I needed to return to the literature and see how people have discussed and/or researched the affective side of teacher learning.

Although there is some literature on emotions within teaching and learning in general, there is less which is specifically under the professional learning umbrella. That is both good and bad for me of course; it gives me little upon which to build, though of course potentially offers a gap here for me to exploit? I have however, been able to turn to the small body of literature which has begun to explore the emotional underpinnings of the teaching process and what it is to be a teacher.

Before proceeding further, I ought to acknowledge first that I’ve used two terms rather loosely so far: affect and emotion. There are a wide range of definitions in the literature, suggesting concepts which do not yield easily to simple definitions; indeed authors often settle instead for descriptions (Fried et al). The OECD (2010:94) tells us that

Emotions refer to a wide range of affective processes, including feelings, moods, affects and well-being.

But perhaps we need to break that down further. Rosenburg (1998) distinguishes between traits and states. Affective traits are more stable, enduring aspects of our personality and predispose us to particular ways of responding emotionally. Affective states can be subdivided into moods and emotions. Emotions are ‘acute, intense, and typically brief responses’ to circumstances we encounter, whereas moods occupy a halfway-house; longer lasting, more diffuse and have a background influence on our responses, but one which we are  also often aware of on a conscious level. The significance of the emotions for the individuals concerned is taken up by Van Veen and Sleegers (2006) who focus on the relations between those individuals and the environments they experience. This might be aligned with the relational ontology I’d adopting, so might offer me a way forward.

It is perhaps self-evident that a teacher’s trait or state will be linked with their well-being and consequently their sense of effectiveness (Zembylas and Schutz, 2009). Here then may be a useful point of entry for me. Perhaps Twitter provides positive emotional stimuli which improve teachers’ sense of well-being and thereby has an impact on their effectiveness? It’s no more than an idea at the moment, so I need some way of exploring that more robustly as my research moves forward. But how?

Arising from a review of the literature, Fried et al (2015) developed a conceptual model which brings together both the functions of emotions and the influences on them. The five functions of emotions are:

  • information provision – about oneself and others, within a broader social, cultural and political context.
  • giving quality to experience – although principally the ‘experience’ teachers offer to their students, I’d extrapolate this to their own learning experiences.
  • cognition – both negative and positive emotions have effects which influence our cognitive processes.
  • regulation – emotions help to regulate our internal systems and also have an external effect on the emotions of others
  • motivation – emotions can also have intra- and interpersonal effects our our motivation and that of others.

And the three influences are:

  • Personal characteristics including identity, beliefs, values and personality traits
  • Appraisal – our emotional responses follow an internal assessment of the emotional meaning of information that we receive from interacting with the environment.
  • Social, cultural and political factors.

How I might operationalise this model is another matter entirely, although perhaps at this stage, I’m delving too deeply? To return to my opening remarks, it was simply a sense that engaging in Twitter brought forth certain emotional responses. My next steps perhaps ought to be looking for harder evidence that that actually is the case. Being aware that there might be an emotional component means I can actively look for evidence as I gather my data, but alternatively, I can also go out looking for it. Just recently, doubtless prompted by a tweet, the principle of sentiment analysis bubbled back to the top of my consciousness, which sparked a connection not previously apparent. Only marginally aware of the underpinning principles, I’m not sure how helpful it might be, but am keen to explore the potential. Having cast around for tools that might help, it soon became clear that you’re not short of options, although since they’re often pitched at business and enterprise, many come with a hefty price tag. There are free options though and the simplest and quickest specifically aimed at Twitter seemed to be sentiment viz. The intention is to produce ‘a visualization that presents basic emotional properties embodied in the text, together with a measure of the confidence in our estimates.’ There are more details here, but that basic idea of being able to take multiple tweets, generate a sense of their emotional content and then display it in multiple ways was very appealing. It does the heavy lifting of initial analysis which I’d usually have to do by manual coding, and allows you to jump to the more complex analysis and interpretation – what is this data telling me. I particularly like being able to get the bird’s eye views, but also being able to quickly drill down to the detail and explore those boundary data for example.

A couple of hashtag chats were taking place just at the right moment, so I popped one of the tags into the search box to see the outcome.

[click to visit the full-sized photo]
As the support resources explain “Each circle’s colour, brightness, size, and transparency visualize different details about the sentiment of its tweet,” which are ascribed using Russell’s (1980) model of emotional affect. We’re immediately able to see the generally positive nature of the 354 tweets posted and where they cluster, but importantly, by clicking on any circle, we can interrogate those data, see the tweets which generated them and the calculated values which positioned them. We’re therefore able to quickly see why tweets are where they are and if they’ve been misplaced. (These kinds of tools rarely cope well with sarcasm for example) There’s also a range of other tabs which, at a click, display the data differently and allow you to access the data in different ways. I think you can take your interpretations quite some way using this tool, but unfortunately, you can’t download the data, nor can you export the visualisations.

There are other tools which, like sentiment viz, are also produced as part of research initiatives. SentiStrength, produced as part of the CyberEmotions project, allows simple online searches, but can be downloaded to undertake much more detailed analyses. Chorus is a package which enables data to be visualised in multiple ways and facilitates both quantitative and qualitative interpretations.

I’m clearly not short of tools, although I need to think through how coherent these approaches are with my chosen methodology and what, if anything, they will add to my study. Given the limited time I can spare to a full and detailed plan, might I still be able to undertake a robust and defensible mini-study which informs my wider research? Is it a flight of fancy, or a legitimate way of shedding some light on the emotional aspects of teacher learning?

Fried, L., Mansfield, C., & Dobozy, E. (2015). Teacher emotion research: Introducing a conceptual model to guide future research. Issues in Educational Research, 25(4), 415-441.
Rosenberg, E. L. (1998). Levels of analysis and the organization of affect (Vol. 2, No. 3, p. 247). Educational Publishing Foundation.
Russell, J., & Hogan, Robert. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39(6), 1161-1178.
Van Veen, K., & Sleegers, P. (2006). How does it feel? Teachers’ emotions in a context of change. Journal of Curriculum studies, 38(1), 85-111.
Zembylas, M., & Schutz, P. A. (2009). Research on teachers’ emotions in education: Findings, practical implications and future agenda. In Advances in teacher emotion research (pp. 367-377). Springer US.

One thought on “Sentiment … ality?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s