Individual Submission Summary
Share...

Direct link:

Scaling Ethnography: Studying Cases with Machine Learning and Interviews on reddit.com

Sat, June 11, 11:00 to 12:15, Fukuoka Hilton, Navis C

Abstract

This paper draws from my dissertation using mixed computational and ethnographic methods to investigate and analyze news production on reddit.com. Using a corpus of all of the over 2 billion messages posted to reddit.com, I am resituating traditional ethnographic observation and interviews (e.g., Burawoy 1998) within a computational analytical framework, moving from macro-scale trends – elucidated via statistical and social network analysis, machine learning, and natural language processing – to micro-scale, ethnographic interpretation. This paper therefore builds off of the integration of “big data” tools in social science research and examines their benefits and drawbacks at their intersection with traditional qualitative empirical studies, In particular, the use of these tools for interpretive ethnographic research. Some scholars have discussed “trace” or “scalable” ethnography (Geiger & Ribes 2011; Rotman et al. 2012), though few have discussed how advanced computational techniques can benefit from ethnographic insight, and vice versa.

The presentation will focus on the “back-and-forth” between the qualitative, interpretive and quantitative, empirical analyses of how people report on breaking news within reddit.com’s sociotechnical platform and community. I draw from experience doing similar work to study and macro- and micro-scales the 2012 Presidential debates on Twitter (e.g., Driscoll, Guth, Leavitt, & Bar 2013; Leavitt, Guth, Driscoll, & Bar 2015). I will explain how using topic modeling, network analysis, and database querying leads to identifying participants for interviews, and how new computational questions such as machine learning classification of particular news behaviors were explored after discovering particular news practices from interviews. These insights point back to issues of how social scientists should conceive of frameworks for implementing big data tools and techniques into their research, particularly regarding the various skill sets required for doing this kind of work, especially within “holistic” data contexts.

References

Burawoy, M. 1998. “The Extended Case Method.” Sociological Theory, 16 (1).
Driscoll, K., K. Guth, A. Leavitt, & F. Bar. 2013. “Big Bird, Binders, & Bayonets: Humor & Live-Tweeting During the 2012 Presidential Debates.” Paper presented at the 2013 Association of Internet Researchers Conference: Denver, CO.
Geiger, R.S., & D. Ribes. (2011). Trace Ethnography: Following coordination through documentary practices. In Proceedings of the 44th Annual Hawaii International Conference on Systems Sciences, HICSS ’11.
Leavitt, A., K. Driscoll, K. Guth, & F. Bar. 2015. “Beyond Big Bird: The Role of Humor in the Aggregate Interpretation of Live-Tweeted Events.” Paper presented at the 2015 Association of Internet Researchers Conference: Phoenix, AZ.
Rotman, D., Preece, J., Yurong, H., and Druin, A. (2012). Extreme Ethnography: Challenges for Research in Large Scale Online Environments. In Proceedings of 2012 iConference.

Author