This presentation is based on a study aimed at characterising the discourse of ‘ordinary citizens’ who become influential in Twitter. The study is motivated by two gaps in knowledge. Firstly, social network science research has shown that open web systems develop in ways whereby small groups of users (10% - 20%) attract inordinate levels of attention (Cha et al 2010). The discourse features of these influential users, however, remain largely unknown. Secondly, understanding social media influence is important in the context of controversial social policies, such as those regarding the adoption or otherwise of living (rather than legally stipulated) minimum wages, on which our work centres (Lawton & Pennycook 2013). Yet, to our knowledge, there is no discourse-based research on this issue.
Methodologically, the study draws upon corpus linguistics, discourse analysis and social media analytics techniques in order to collect, collate and interrogate an analytic corpus of 32,000 tweets and a reference corpus of 1.6 million tweets, both collected over a ten-day period (26th May – 4th June, 2014). Social media analytics software / methods (Twittonomy and Klout) were used to derive a set of thematically relevant hashtags (n= 12; e.g. #poverty, #lowwage, #minimumwage and #austerity) and the most influential ordinary user within each hashtag, filtering out celebrities’ and organisations’ Twitter profiles. The reference corpus was collected via the Twitter API facility (through the use of Twitter4j). Software-enabled [Wmatrix (Rayson 2009) / Antconc (Anthony 2014)] analyses of semantic domains, cluster, collocation and Key-Word-In-Context (KWIC) concordances were conducted for each of the twelve influential user sub-corpora. These analyses were complemented with close reading of full tweets in order to better understand emerging patterns around three discourse analytic categories: frames (Goffman, 1974), stance (Du Bois 2007) and impoliteness (Culpeper, 2011).
By ‘funnelling’ down (Marchi 2010) the corpus through these CADS techniques, four main features of the discourse of influential citizens in Twitter were revealed: (i) limited originality coupled with high participation rate; (ii) varying degrees of thematic engagement, (iii) emotionality, particularly expressed through impoliteness; and (iv) stance-taking acts that conveyed full confidence in one’s views. Additionally, our study resulted in a fuller appreciation of the challenges and opportunities inherent to interdisciplinary research in CADS, such as being able to identify ‘suitable’ databases (of influential users) by going back and forth between corpus and social media analytic techniques and to transform data (Twitter data, with its distinctive affordances) into analysable corpora.
Bibliography: Anthony, Laurence (2014). AntConc (Version 3.4.3) [Computer Software]. Tokyo, Japan: Waseda University. Available from http://www.laurenceanthony.net/ Cha, Meeyong, et al “Measuring user influence in Twitter: The million software fallacy.” ICWSM no. 30 (2010): 10-17. Culpeper, Jonathan. Impoliteness. Cambridge: CUP, 2011.
Du Bois, John W. “The stance triangle.” Stancetaking in discourse: Subjectivity, evaluation, interaction (2007): 139-182.
Goffman, Erving. “Frames analysis. An Essay in the Organization of Experience.” Northeastern University Press. (1974).
Lawton, Kayte, and Matthew Pennycook. “Beyond the Bottom Line.” (2013) http://www.ippr.org/files/images/,edia/files/publication/2013/01/beyond-bottom-line_livong-wage_Jan2013_10162.pdf?nodirect=1