Data­-Driven Data Research Using Data and Databases: A Practical Critique of Methods and Approaches in “Big Data” Studies


Panelist at Annual Meeting of the International Communication Association (ICA), Seattle, WA

In the past two years, the buzzword "big data" has provoked critiques by a number of social scientists (eg., boyd & Crawford 2011; Bruns & Burgess 2012; Burrell 2012; Baym 2013) on the theories, methodologies, and analysis of large data sources. This panel follows up from last year’s ICA 2013 panel, “Downsizing Data: Analyzing Social Digital Traces,” and draws from the experiential grounded approach of Hargittai’s (2009) “Research Confidential” to bring to light practical critiques of the application of digital data research methods in the study of social media platforms. Namely, the panelists will explore how social scientists can shift away from the rhetoric surrounding “big data” and robustly analyze the use of large-scale, computationally-driven, mixed-methods approaches in digital data research. Again, this panel will not discredit large-scale data approaches; instead, we aim to provide context to researchers who wish to employ them in combination with established methods in the field. The panel brings together five scholars to speak about their successes and failures working on projects that employ large-scale digital data methods and tools, regardless of the size of the data, in addition to their iterative approaches dealing with the practicalities of data collection, sampling, theory, analysis, and especially results. Notably, these projects are not purely quantitative, analytical studies employing large datasets: all participants use largescale data and computational approaches within the context of empirical mixed-methods or even (traditionally) qualitative, interpretive studies. The panelists will also discuss the critical approaches to “big data” that inhabit each project. These projects are all exemplars of an emerging mode of scholarship, and collectively they aim to generate a productive and concrete discussion about methodology and epistemology. After a framed introduction by the moderator, participants will spend 10 minutes each to speak in detail about the methodologies of their projects, after which the latter half of the panel will open to discussion with the audience. This panel also will be paired with a Blue Sky Workshop, provocatively entitled “‘Big Data is Bullshit’: Scoping the Next 5 Years of Digital Data Research.” We aim to use the panel as an expert-driven, experiential methods presentation as well as a launchpad for topics and debates that can be further explored in the workshop session (which will occur at some point following the panel).