When Quality Abounds: The Re-Grounding of Deliberative Theory in Big Data

The grounding of normative theory can be done using tools that might seem to be the antithesis of grounded theory principles, because “big data” is not simply quantitative. On the contrary, it is an abundance of rich, qualitative information, in formats that traditional computing cannot handle, at scales too big for human analysts. This chapter illustrates this with an analysis of two databases, one on the 2014 Scottish independence referendum, the other on a campaign for constitutional recognition of Indigenous peoples in Australia. Using structural topic modeling, our analysis reveals not only the huge range of, and relationships between, topics under discussion but also the crucial importance of what the chapter calls the “meta-conversation,” in which participants set and maintain norms of unmediated online discussion, in context-sensitive ways. This discovery of citizens’ capacity for self-regulation outside moderated deliberative settings should change deliberative scholarship, something that had become too focused on everyday practice.