From computer-assisted to data-driven: Journalism and Big Data

Abstract

Despite claims of continuity, contemporary data journalism is quite different from the earlier tradition of computer-assisted reporting. Although it echoes earlier claims about being scientific and democratic, these qualities are understood as resulting from better data access rather than as being something achieved by the journalist. In the context of Big Data in particular, human subjectivity tends to be downgraded in importance, even understood as getting in the way if it means hubristically theorising about causation rather than working with correlation and allowing the data to speak. Increasing ‘datafication’ is not what is driving changes in the profession, however. Rather, the impact of Big Data tends to be understood in ways that are consonant with pre-existing expectations, which are shaped by the broader contemporary post-humanist political context. The same is true in academic analysis, where actor–network theory seems to be emerging as the dominant paradigm for understanding data journalism, but in largely uncritical ways.

Hybrids, materiality, and black boxes: Concepts of actor‐network theory in data journalism research

“…They advise against treating ANT “as a convenient argument to highlight the importance of technology” (Primo & Zago, , p. 44) due to its heavy implications for research agendas. As they consider technological artefacts “full‐blown social actors, with transforming roles” (p. 39), Hammond (, p. 418) endorses the idea “to reassess our view of the human subject as the central, active agent.” Many of these theorisations of data journalism through the lens of ANT work towards raising awareness of nonhuman entities by putting forward sociotechnical frameworks in a reaction to epistemological challenges posed by data journalism.…”

“…Third, Lewis and Westlund (, p. 26) conceptualise audiences as passive recipients, active participants as well as commodities in order to address “normative, commercial, and cultural functions alike.” There is still a lack of empirical findings drawn from audience research other than theorisations of the audience, for instance by Hammond (, p. 411), who describes the audience as active participants of data‐driven investigations under the notion of democratising effects of open data and data journalism that allows readers to explore and access data themselves. Appelgren (), in contrast, considers visualisations as delegating very little actual control to their recipients, only offering “the illusion of interactivity” (p. 322), thus depriving the audience of its agency.…”

“…Such paternalistic tendencies may be beneficial to convey complex matters. Still, Appelgren () argues, stripping the audience of their active role (as proposed by Hammond, ) and by not making clear production processes that led to paternalistic designs would take away “a level of transparency that [the audience] need to form their own judgements about the information presented” (Appelgren, , pp. 322 and 323).…”

Quantitatively Oriented Forms of Journalism and Their Epistemology

“…In his 2015 article, he also recommends the Actor Network Theory (ANT), therefore analyzing how data are embodied in social and material objects such as databases, documents, or surveys. The ANT is consequently often used to analyze innovation in the field (Parasie and Dagiral ; Parasie ; Weiss and Domingo ), and it can be seen as a dominant paradigm for understanding data journalism, though it has not rarely been used in uncritical ways (Hammond ).…”

“…Computational journalism generally adopts a data‐driven hypothesis. The knowledge that it produces relies on correlation and is based on the completeness of the data that can reveal unexpected stories (Hammond , 4). Following Parasie and Dagiral's (2013) acute interpretation, in computational journalism, the programmers have political values about open source and open government, which assume democratic values cutting across immense quantities of data.…”