Arguments in big social data analysis : uncovering the hidden rhetoric of sociological data science

Abstract

My research revealed that the values of computer science, behavioral science, and mathematics are implemented unequally, revealing a community preference for building viable analytic technologies rather than producing valid sociological results. Using Kuhn’s idea of the scientific paradigm, the relationship between various values from divergent epistemic communities are explored as they appear in the conference proceedings. The rhetorical concepts of agency, audience, exigency, genre, and invention are explored in relation to how the analysts in the ASONAM community design and implement their research objectives as a communication strategy. While the analysts buy into the conviction that “data speaks for itself,” this dissertation reveals how and when the analyst makes interpretive decisions that constrain and shape the final possibilities for results. The work concludes with recommendations for peer-review questions that recognize the role of interpretation in big social data research and rebalance the importance of fidelity to the original problem being addressed with the need to create analytical technologies that are efficient and productive.

Big social data analysts use data collected from information-communication technologies and process that data for insights into human behavior. In this dissertation, the IEEE/ACM international conference on Advances in Social Networks Analysis and Mining (ASONAM) community was studied using rhetorical analysis, an exploratory survey, and participant observation to understand the interpretive moments embedded in the big social data analyst’s research. To process big social data for insights, the analyst poses a research question and follows a research design to create an argument for the veracity of their results. The research projects as arguments were charted using Toulmin’s argumentation framework, and the arguments were evaluated using the various field-dependent evaluation standards present in the ASONAM community.