Nathan TeBlunthuis

Postdoctoral Research Fellow

Information School University of Michigan

I am a Postdoctoral Research Fellow in the Information School at the University of Michigan. I was previously at the Department of Communication Studies at Northwestern University and completed my PhD in Communication at the University of Washington. I am a computational social scientist who studies online collective action in projects like Wikipedia, online communities like Reddit, and social movements. Here is a link to my CV.

An important and serious methodological problem affects a huge range of research in computational social science: When you use a machine learning algorithm to automatically classify a variable, and then you use that variable as an input to a statistical procedure such as a regression, the errors and biases of the algorithm contaminate the downstream analysis. Predictiveness metrics such as precision and recall do not guarantee validity. My article forthcoming in Communication Methods and Measures with Dr. Valarie Hase and Dr. Chun-Hong Chan shows that statistical methods can use validation data to correct the downstream analysis. Try out the R package.

Many people invoke “ecosystem” as metaphor to emphasize complexity and interdependence in communication systems like the Internet. However, there is also a huge natural science called “ecology” which successfully learns about biological ecosystems. Organizational sociologists and communication scientists have already appropriated theories, models and methods from ecology to understand interdependence between human organizations like firms and social movements. I draw both from these social science literatures and from bio-ecology to understand how environmental contexts and interdependence between online communities shapes their growth, survival and organizing processes.

In my ICWSM 2022 paper with Benjamin Mako Hill, I use time series analysis to infer when overlapping online communities hosted on Reddit are competitors or mutualists and find that mutualism is much more common than competition. To help explain why overlapping communities are often mutualists, my collaborators and I interviewed members of these overlapping online communities. As described in our CSCW 2022 paper, people seek different types of benefits from online communities such as a like-minded community, specific information and an audience for their content. Tensions between the different benefits mean that "no community can do everything," but one can obtain a greater range of benefits from a portfolio of overlapping communities.

My research also investigates quality control and machine learning systems in peer production. My paper published in CHI 2017 with Aaron Shaw and Benjamin Mako Hill shows that quality control systems often become increasingly hostile to newcomers and difficult to change over time. These mechanisms, which limit Wikipedia's growth and diversity, are common in a range of digitally mediated forms of organizing. Online organizations increasingly use data science methods as a part of their quality control systems. In my CSCW 2021 paper with Benjamin Mako Hill and Aaron Halfaker, I used a regression discontinuity analysis to evaluate how algorithms for flagging encyclopedia edits shape the fairness of Wikipedia moderation. I also introduce an improved machine learning method for measuring Wikipedia article quality in my OpenSym 2021 paper .

My academic work has received generous support from the National Science Foundation through the Graduate Research Fellowship Program (GRFP) and through the cyber-human systems program.

In the past, I worked at Microsoft where I developed search suggestions for Bing multimedia and as a technician at the Pacific Northwest National Laboratory (PNNL) where I measured foaming molten glass as part of a project to store nuclear waste. I went to college at Whitworth University where I received a double-B.S. in Mathematics and Computer Science in 2012.

Publications

Peer-reviewed publications

  • Misclassification in Automated Content Analysis Causes Bias in Regression. Can We Fix It? Yes We Can!

    Nathan TeBlunthuis, Valerie Hase, Chung-Hong Chan

    Communication Methods and Measures [DOI] [Arxiv Repository]

  • Identifying Competition and Mutualism Between Online Groups

    Nathan TeBlunthuis, Benjamin Mako Hill

    International AAAI Conference on Web and Social Media — ICWSM 2022 [Proc. ICWSM] [Arxiv Repository]

  • No Community Can Do Everything: Why People Participate in Overlapping Online Communities

    Nathan TeBlunthuis, Charles Kiene, Isabella Brown, Laura (Alia) Levi, Nicole McGinnis, Benjamin Mako Hill

    Proceedings of the ACM on Human-Computer Interaction — CSCW 2022 [ACM DL] [Arxiv Repository]

  • Measuring Wikipedia Article Quality in One Dimension

    Nathan TeBlunthuis

    Proc. of the 17th International Symposium on Open Collaboration — OpenSym 2021 [ACM DL] [Arxiv Repository]

  • Effects of Algorithmic Flagging on Fairness: Quasi-experimental Evidence from Wikipedia

    Nathan TeBlunthuis, Benjamin Mako Hill, Aaron Halfaker

    Proceedings of the ACM on Human-Computer Interaction — CSCW 2021 [ACM DL] [Arxiv Repository]

  • Dwelling on Wikipedia: Investigating Time Spent by Global encyclopedia Readers.

    Nathan TeBlunthuis, Tilman Bayer, Olga Vasileva

    Proc. of the 15th International Symposium on Open Collaboration — OpenSym 2019 [ACM DL]

  • All Talk: How Increasing Interpersonal Communication on Wikis May Not Enhance Productivity.

    Sneha Narayan, Nathan TeBlunthuis, Wm Salt Hale, Benjamin Mako Hill, and Aaron Shaw

    Proceedings of the ACM: Human-Computer Interaction — CSCW 2019 [ACM DL]

  • Revisiting "The Rise and Decline" in a Population of Peer Production Projects

    Nathan TeBlunthuis, Aaron Shaw, Benjamin Mako Hill

    Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems — CHI 2018 [ACM DL]

  • WINACS: construction and analysis of web-based computer science information networks

    Tim Weninger, Marina Danilevsky, Fabio Fumarola, Joshua Hailpern, Jiawei Han, Thomas J Johntson, Surya Kallumadi, Hyungsul Kim, Zhijin Li, David McCloskey, Yizhou Sun, Nathan E TeGrotenhuis, Chi Wang, Xiao Yu

    ACM SIGMOD International Conference on Management of data — SIGMOD 2011 [ACM DL]

  • Conversion of batch to molten glass, II: Dissolution of quartz particles

    Pavel Hrma, José Marcial, Kevin J Swearingen, Samuel H Henager, Michael J Schweiger, Nathan E TeGrotenhuis

    Journal of Non-Crystalline Solids — 2010 [ScienceDirect]

  • Cluster Formation of Silica Particles in Glass Batches During Melting

    Michael J Schweiger, Pavel Hrma, Carissa J Humrickhouse, José Marcial, Brian J Riley, Nathan E TeGrotenhuis.

    Journal of Non-Crystalline Solids — 2010 [ScienceDirect]

  • Effect of Glass-Batch Makeup on the Melting Process

    Pavel Hrma, Michael J Schweiger, Carissa J Humrickhouse, J Adam Moody, Rachel M Tate, Timothy T Rainsdon, Nathan E Tegrotenhuis, Benjamin M Arrigoni, Jose Marcial, Carmen P Rodriguez, Benjamin H Tincher.

    Ceramics-Silikaty — 2010 [Ceramics Silikaty]

Published Extended Abstracts

  • Density Dependence Without Resource Partitioning: Population Ecology on Change.org.

    Nathan TeBlunthuis, Aaron Shaw, and Benjamin Mako Hill

    Companion of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing — CSCW ’17 Companion [ACM DL ]

Book Chapters

  • Big Data

    Nathan TeBlunthuis, Andrea Ceron (Ed.)

    Encyclopedia of Technology & Politics — Edward Elgar Publishing — 2022 [preprint pdf ]

Theses

  • Ecology of Online Communities

    PhD Thesis, University of Washington — 2021 [UW Digital Libraries ]

  • Density Dependence without Resource Partitioning on an Online Petitioning Platform

    Master's Thesis, University of Washington — 2017 [UW Digital Libraries ]

Teaching

Interpersonal Media: Online Communities (COM 482)

Online communities are central parts of each of our daily lives and have an important impact on our cultural, social, and economic experience of the world and each other. This course combines an in-depth look into several decades of research into online communities and computer-mediated communication with exercises that aim to give students experience applying this research to the evaluation of, and hands-on participation in, online communities.

University of Washington. Winter 2021, Spring 2019 [Sample syllabus]

Contact

The best way to contact me is over email. I sometimes check my DMs on Twitter and Mastodon. Facebook messages, LinkedIn messages, etc. are most likely to go unanswered.