IDeaS 2019 Conference: Perspectives (II)

By Bandita Deka Kalita

Day 1

Kyle Murray and Dev Jennings flagged off Day 1 of the IDeaS conference by talking about how big data can transform organizations and how it can be interpreted. On the topic of interpreting big data, Dev emphasized a theme that resonated throughout the two days of the conference - that rendering corpora is inherently interpretive in nature. He then invited Laura Nelson, and Tim Hannigan, to share insights from their respective projects.

Laura Nelson presented her ongoing project on rendering social movement strategy. She explained that in empirical studies of social movement, count data of events are generally used. However, she reminded the audience that care must be taken when interpreting various forms of a certain action, because they may not mean the same thing. In fact, in emphasizing that meanings lie in discourses, Laura laid the groundwork for the core tenet undergirding much discussion in the forthcoming sessions of the IDeaS conference. The CHC’ (Computation-Human interpretation-Computation) technique she discussed during her talk also prompted discussion around the implications of the sequence of computation and human interaction (“CHC’ or HCH’?”) for particular research contexts. Laura, and several other researchers in the room, stressed the importance of context in meaning making.

Tim Hannigan’s work on the rendering of the kernel of the British MP scandal provided a nice segway into further deep discussions about meanings and the tools available to social scientists to uncover meanings from textual data. Tim provided a candid description of the various challenges and opportunities that presented themselves in the course of the project, and how eventually the team uncovered clusters of meaning from the visualization of the “kernel” of the scandal: the first seven days of textual data from newspapers. Overall, the work of Tim and colleagues highlighted the utility of topic modeling as an emerging approach enabling qualitative and quantitative researchers with powerful technology for exploring data and theorizing in management.

Mark Kennedy’s presentation was themed around hopes for a growing community between scholars of qualitative and quantitative research methodologies respectively. In emphasizing the “back and forth between counts and stories”, Mark reminded the audience of the fact that it is the stories we tell with numbers that really matter. His spirited presentation was aptly titled “We could be Friends”.

The participants enjoyed meeting and discussing projects with colleagues from various places over the coffee and networking sessions throughout the two days of the conference. There were six simultaneously occurring round table discussions on the late morning of the first day with these diverse scholars working on their respective projects using similar methods. These discussions were respectively themed around qualitative studies on deviance & crowds, qualitative studies on open strategy, blockchain and distributed trust, topic modeling and field emergence, applied topic modeling, and quantitative studies on entrepreneurship and technology battles.

After lunch, Dev Jennings moderated a session on sharing the solutions that emerged from the above described round table discussions. Discussion was generated in the room around the question of qualitative data archives and the sharing of these data. Other critical issues that were raised were around big data copyright matters, practices around citation of codes, and the utility of topic modeling and semantic analysis in coding qualitative data. Marc-David Seidel moderated the next session on publishing papers with computational methods. He triggered interesting discussions around topics such as human limitations around interpreting data, and how computational methods can bridge those limitations.

Joe Porac’s presentation on meaning and interpretation was themed around the difference between big data and small data. His fascinating talk on the embeddedness of meaning in the social world took the audiences to several places – from Peter Winch’s theses of rules, practicality, and participation, and Jacques Derrida’s “aporias” of undecidability, a point where a text has been deconstructed to such an extent that the meaning is undecipherable.

Day 2

The second day of the conference started with an address by Trish Reay, who made a poignant reference to the different history of big data usage by the First Nations of Canada, as she continued with thought strands about marginalized contexts from the first day. She introduced keynote speaker Wendy Espeland who delivered a fascinating talk on “Governance by Numbers”. She spoke about the increasing pressure to code meanings and value in terms of numbers, in the context of university ranking systems. The association of desired traits with measurement and quantification has implications for meaning, and hence for governance systems, and Wendy’s work investigates into matters of rationality and inclusion arising from these implications. She described some of her work with governance issues in the Yavapai community, and associated matters of power, and accountability.

Chris Steele moderated a panel on “the politics of data”, the panel consisting of Wendy Espeland, David Kirsch, Dev Jennings, and Joel Gehman. As a precursor to the panel discussion, Chris also presented some of his work around the “ecology of facticity”: how is it that certain things come to be accepted as fact? The panel discussion that ensued picked up on some of the questions inspired by Chris’s and Wendy’s presentations. Some of the many interesting points of conversation were around matters pertaining to the “what” of the data to be interpreted - how institutionally or bureaucratically freighted they may be, and what implications for stories of power and marginalization could these allude to?

Marc-David Seidel presented a session on HIBAR research, and described the suitability of HIBAR research towards making a difference by solving socio-technical problems. He urged the participants to engage in answering the following question: if you had the access to all of Google Scholar’s data, what data would you like to see included in evaluation of “metrics” for individual academics?

The two optional sessions in the afternoon of the second day registered an impressive turnout by participants. These sessions were respectively on “Deep Learning”, led by Muhammed Abdul-Mageed, and a “practicum on creating corpora, topics, and artifacts”, led by Tim Hannigan and Rodrigo Valadao.