Eric P. S. Baumer. (2017). Toward Human-Centered Algorithm Design. Big Data & Society, 4(2).
As algorithms pervade numerous facets of daily life, they are incorporated into systems for increasingly diverse purposes. These systems’ results are often interpreted differently by the designers who created them than by the lay persons who interact with them. This paper offers a proposal for human-centered algorithm design, which incorporates human and social interpretations into the design process for algorithmically based systems. It articulates three specific strategies for doing so: theoretical, participatory, and speculative. Drawing on the author’s work designing and deploying multiple related systems, the paper provides a detailed example of using a theoretical approach. It also discusses findings pertinent to participatory and speculative design approaches. The paper addresses both strengths and challenges for each strategy in helping to center the process of designing algorithmically based systems around humans.
Eric P. S. Baumer, David Mimno, Shion Guha, Emiy Quan, and Geri Gay. (2017). Comparing Grounded Theory and Topic Modeling: Extreme Divergence or Unlikely Convergence? Journal of the Association for Information Science and Technology (JASIST), 68(6): 1397–1410.
Researchers in information science and related areas have developed various methods for analyzing textual data, such as survey responses. This article describes the application of analysis methods from two distinct fields, one method from interpretive social science and one method from statistical machine learning, to the same survey data. The results show that the two analyses produce some similar and some complementary insights about the phenomenon of interest, in this case, nonuse of social media. We compare both the processes of conducting these analyses and the results they produce to derive insights about each method’s unique advantages and drawbacks, as well as the broader roles that these methods play in the respective fields where they are often used. These insights allow us to make more informed decisions about the tradeoffs in choosing different methods for analyzing textual data. Further- more, this comparison suggests ways that such methods might be combined in novel and compelling ways.
Eric P. S. Baumer and Jed R. Brubaker. (2017). Post-userism. in Proceedings of the ACM Conference on Human Factors in Computing Systems (CHI). (Denver, CO).
Baumer, E.P.S., Xu, X., Chu, C., Guha, S., and Gay, G.K. (2017). When Subjects Interpret the Data: Social Media Non-use as a Case for Adapting the Delphi Method to CSCW. in Proceedings of the ACM Conference on Computer Supported Cooperative Work & Social Computing (CSCW). (Portland, OR).
This paper describes the use of the Delphi method as a means of incorporating study participants into the processes of data analysis and interpretation. As a case study, it focuses on perceptions about use and non-use of the social media site Facebook. The work presented here involves three phases. First, a large survey included both a demographically representative sample and a convenience sample. Second, a smaller follow-up survey presented results from that survey back to survey respondents. Third, a series of qualitative member checking interviews with additional survey respondents served to validate the findings of the follow-up survey. This paper demonstrates the utility of Delphi by highlighting the ways that it enables us to synthesize across these three study phases, advancing understanding of perceptions about social media use and non-use. The paper concludes by discussing the broader applicability of the Delphi method across CSCW research.
Later this week, I’ll be presenting a paper at the Algorithms in Culture conference, organized by an astounding crew of veritable rock stars. I’m eagerly looking forward to it, both in terms of the the numerous fascinating sounding talks, and as a chance to get feedback on some of the questions through which I’m working around how fuse critical studies of algorithmic systems with human-computer interaction design.
I am thrilled to announce that I recently accepted a tenure-track position as Assistant Professor in the Computer Science & Engineering Department at Lehigh University. This position is part of a university-wide initiative called Data X. In my description, Data X acknowledges and incorporates the inherently interdisciplinary nature of data science. Lehigh made several cluster hires this past year as part of Data X, each with one faculty line in CSE and one in another department. For instance, I was part of a cluster hire between CSE and Journalism & Communication. Needless to say, this is an exceptionally perfect niche for me.
I’ll be winding down my position at Cornell and moving to Bethlehem, PA this fall, then starting at Lehigh in January.
Wyche, S.P. and Baumer, E.P.S. (2016). Imagined Facebook: An Exploratory Study of Non-Users’ Perceptions of Social Media in Rural Zambia. New Media & Society.
This article describes an exploratory study of Facebook non-users living in rural Zambia. Drawing on evidence from 37 group interviews with mobile phone owners, we discovered that the majority of our participants were aware of, or ‘imagined’ Facebook, despite never having seen or used the site. Our analysis of how participants perceive Facebook suggests that they are interested in the communication and income-generating possibilities access to the site may provide, but that barriers prevent them from acting on these interests. This study contributes to social media research by making visible the experiences of a population whose non-use of information and communication technologies (ICTs) results from economic, infrastructural, and linguistic sources, as well as from other, hitherto less-explored areas. We discuss the practical significance of these findings, offer future research suggestions, and comment on what our respondents have not yet imagined about Facebook.
Recently, I had two papers accepted for publication an the 2016 ACM Conference on Supporting Group Work (GROUP). One is a full length paper about a long-term study of a “data-as-art” installation at Cornell. The other is a note length paper essentially arguing that machine learning and grounded theory have some deep, compelling resonances. While I played a primarily supporting role on both these, I’m excited about the arguments that each is making (and appreciative of the lead authors for involving me).
Muller, M., Guha, S., Baumer, E.P.S., Mimno, D., and Shami, N.S. (2016). Machine Learning and Grounded Theory Method: Convergence, Divergence, and Combination. in Proceedings of the ACM Conference on Supporting Group Work (GROUP). (Sanibel Island, FL).
Grounded Theory Method (GTM) and Machine Learning (ML) are often considered to be quite different. In this note, we explore unexpected convergences between these methods. We propose new research directions that can further clarify the relationships between these methods, and that can use those relationships to strengthen our ability to describe our phenomena and develop stronger hybrid theories.
Scolere, L., Baumer, E.P.S., Reynolds, L., and Gay, G. (2016). Building Mood, Building Community: Usage Patterns of an Interactive Art Installation. in Proceedings of the ACM Conference on Supporting Group Work (GROUP). (Sanibel Island, FL).
To examine the processes by which appropriation happens around an interactive art installation in an organizational context, this paper presents a qualitative, longitudinal study of an interactive art installation called mood.cloud. While designed to collect and to visually display building occupants’ collective emotion, the installation was not necessarily used or interpreted in this way. Instead, building occupants saw the sensory experience of mood.cloud and the ability to change the display as a way to influence their own feelings, the feelings of others, and the overall workplace ambience. We found that interaction with mood.cloud fostered reflection about the relationship between the individual and the larger collective that the person is a part of. This relationship, between appropriation for individual benefit and appropriation for the benefit of others, afforded participants the opportunity to become more aware of their own contribution as part of a larger community. These findings suggest an opportunity to design systems around the interplay between appropriation for the individual and appropriation for the community.