2026 ASEE Annual Conference & Exposition

Teaching contextual reasoning as a key skill in data science education

Presented at Liberal Education/Engineering & Society Division (LEES) Technical Session 5

Teaching contextual reasoning as a key skill in data science education

Contextualization is a theme of recurring significance in engineering education. It is highly relevant to the emerging interdisciplinary field of data science/data analytics, which brings together computing, statistics, data systems engineering, and real-world applications. Data analytics and algorithmic decision-making systems have major impacts in our world today akin to other engineering practices. The models, algorithms, and other products created by data science practitioners derive from data and assumptions that inherently originate in particular contexts. Yet such tools abstract from those circumstances to create generalizable mechanisms and analyses that are meant to be cut loose from particularity and travel across contexts. In fact, this kind of abstraction is arguably intrinsic to technical training and specifically to computational and inferential thinking, which are defined as core skills in the data science education literature. Data science is thus a good setting in which to examine the kinds of work that “context” can do in technical education, including integrating key strategies into pedagogy, examining impact in students’ learning, building scaffolded curriculum to achieve key goals toward robust and responsible professional practice, and fleshing out a substantial place for a form of ethics that itself goes beyond abstraction.

“Contextual reasoning,” as used in this paper, is a cognitive skill that moves counter to abstracting from particular circumstances. In teaching data science, contextual reasoning centers skills of systematically analyzing where data comes from, how it is made, and what assumptions it may harbor; how the intermediate and end products of technical practice (including data and algorithms) respond to or carry traces of the human contexts from which they are drawn; and how the actual impact of data-driven solutions in the real world depends on how they end up embedded in concrete settings and contexts. As implemented in a data science curriculum that from the start has included “human contexts and ethics,” contextual reasoning has proven to be essential in advanced data science courses that consider real-world problems. Developing contextual reasoning is an ongoing project within this curriculum. This paper provides a mid-project examination of how contextual reasoning serves technical students and how it can be taught as a core skill across an undergraduate curriculum in ways that are responsive to growing discussions about contextualization and sociotechnical integration as broader strategies in technical education.

Authors

Cathryn Carson University of California, Berkeley [biography]

Cathryn Carson is a historian of contemporary science and technology and scholar of STS (Science, Technology, and Society) at UC Berkeley. She has been active in interdisciplinary collaborations in undergraduate and graduate education, including nuclear engineering and data science, since the 2000s. She currently co-leads the Human Technology Futures group, a proposed department in the College of Computing, Data Science, and Society (CDSS), and chairs the Department of History. Trained as a historian of 20th-century physics, she co-chaired Berkeley's Data Science Education Rapid Action Team, which framed the broad-based, integrative vision its undergraduate data science curriculum; and chaired the Faculty Advisory Board that developed the blueprint for Berkeley’s new College of CDSS. She is part of the teaching team for Data C104, Human Contexts and Ethics of Data, which is the largest humanities or social science class at UC Berkeley, at a scale of about 700 students each semester. Her current research is on disciplines and interdisciplinarity in contemporary universities and on the intellectual history of the philosophy of technology.
Lisa Yan http://orcid.org/0009-0007-2310-3060 University of California, Berkeley [biography]

Lisa Yan is an Assistant Teaching Professor of Electrical Engineering and Computer Sciences at the University of California, Berkeley. She is a computer science and data science educator and researcher with expertise in data analysis of student learning in large courses, computing ethics education, and teaching assistant education. She received her Bachelor of Science degree in Electrical Engineering and Computer Sciences from UC Berkeley and her Master of Science and PhD degrees in Electrical Engineering from Stanford University.
Ari Edmundson University of California, Berkeley [biography]

Ari Edmundson is a lecturer in the Department of History at the University of California, Berkeley. He teaches and develops Human Contexts and Ethics curriculum for data science majors and minors.
Joshua Grossman University of California, Berkeley [biography]

Josh Grossman is assistant teaching professor of statistics and data science at UC Berkeley. He has conducted research on discrimination law, college admissions, pretrial detention, and policing. His current research focus is AI-assisted grading in data science courses. He is also interested in how computerized testing centers can be used to provide in-person, low-stakes, and high-frequency quizzing at-scale in large courses. Prior to his academic career, Josh worked as a product manager at IXL Learning, a K–12 education technology company. Josh completed his Ph.D. in Computational Social Science at Stanford University’s Department of Management Science & Engineering (MS&E) and his bachelor’s degree at Harvard University.

Download paper (443 KB)

Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.

» Download paper

« View session