2026 ASEE Annual Conference & Exposition

Enhancing Conceptual Understanding in Data Science through Peer Instruction: An Empirical Study Using Poll Everywhere

Presented at Computers in Education (CoED): Learning, Engagement & Inclusion (3 of 9) -- M508A

In the dynamic landscape of STEM education, active learning strategies have become central to promoting student engagement, critical thinking, and long-term retention. In this study we investigate the application of Peer Instruction (PI), a structured peer-to-peer discussion method, in an undergraduate data science classroom, specifically within the module "Problems Faced When Handling Large Data & General Techniques for Handling Large Volume of Data". The goal was to assess PI's impact on performance across different cognitive levels and gauge student perception of difficulty for complex data science topics.

Sixteen third-year undergraduate students participated in the structured activity. They were divided into four balanced groups. The intervention utilized the digital response system Poll Everywhere to administer ten multiple-choice questions (MCQs): five targeting recall (factual knowledge) and five targeting conceptual understanding (abstract models and trade-offs). Each question followed a three-step PI cycle: an initial individual vote (Pre-discussion Vote), a 5–7 minute peer discussion in groups of four, and a final individual vote (Post-discussion Vote). Learning gain was calculated by comparing pre- and post-discussion accuracy. A dedicated perception poll was also conducted to determine which question type students perceived as more difficult.

Our results demonstrate that PI significantly enhances student performance, particularly on cognitively demanding material.
• Learning Gain (RQ1): While both question types improved, conceptual questions showed a greater average learning gain of +21.8%. The highest single improvement was a +50% gain for the abstract CAP theorem question, followed by substantial gains in questions related to MapReduce, caching, and batch processing. Recall-based questions saw an average gain of +17.6%. The question related to the Parquet file format, an unfamiliar term for many, saw a notable +56% gain, highlighting the effectiveness of PI even for factual recall under ambiguity.
• Perception (RQ2): The perception poll revealed that 75% of students found conceptual questions to be more difficult, confirming their higher cognitive load and the necessity for collaborative reasoning.
• Student Feedback (RQ3): Student feedback indicated increased conceptual clarity, boosted confidence, and a clear preference for the collaborative approach. Students particularly valued peer explanations and reported that discussion made abstract content easier to grasp, expressing strong interest in adopting this model for future sessions.

This study confirms that Peer Instruction, when integrated with real-time polling via Poll Everywhere, is a highly effective pedagogical strategy for enhancing deep learning in data science education. The consistent performance improvement and positive student feedback reinforce that PI successfully bridges the gap between surface-level recall and the essential conceptual understanding required for technical, abstract subjects like distributed computing. These findings strongly support embedding PI into the data science curriculum to support metacognitive engagement and conceptual clarity.

Authors

Dr. Divya Mohan http://orcid.org/0000-0002-5988-3851 Department of Computing Technologies, School of Computing, SRM Institute of Science and Technology, Kattankulathur, Chengalpattu, India-603203 [biography]

Divya Mohan is an Assistant Professor in the Department of Computing Technologies at SRM Institute of Science and Technology, India. She completed her Ph.D. in Computer Science and Engineering from Vellore Institute of Technology. Her research interests span wireless network scheduling, soft computing, machine learning, and active learning strategies in computing education. She has authored multiple peer-reviewed journal and conference publications and actively integrates research-driven and student-centered pedagogies into her teaching practice. Dr. Mohan is particularly interested in applying artificial intelligence and data-driven approaches to enhance engineering education outcomes and learner engagement. She has participated in several faculty development programs and workshops focused on innovative teaching, curriculum design, and experiential learning. Her work aims to bridge the gap between emerging technologies and effective engineering education practices
Dr. Bhuvana Gopal http://orcid.org/0000-0002-1796-2789 University of Nebraska - Lincoln [biography]

Dr. Bhuvaneswari (Bhuvana) Gopal is an Assistant Professor of Practice in the School of Computing at the University of Nebraska-Lincoln where she teaches Software Engineering, Software Security in Practice, the industry internship course, and leads the Learning Assistant Program that serves several computing courses at the School of Computing. Dr. Gopal has extensive experience in the software industry, where she spent 14 years in various roles, including Software Architect and Lead software engineer positions before switching to full time academia. She is also a Microsoft certified professional, with an MCPD certification. Her research work is focused on evidence-based, active learning pedagogies to improve software engineering education. She has published papers specifically on Peer Instruction and Process Oriented Guided Inquiry-based Learning-like pedagogical approaches in the undergraduate software engineering classroom, both in person and virtual. She holds a Bachelors degree in Physics from the University of Madras, Chennai, India. She holds two Masters degrees - one in Computer Science from the University of Nebraska-Lincoln, and one in Medical Physics from Anna University, Chennai, India. She holds a Ph.D. in Computer Science from the University of Nebraska-Lincoln.
Byrav Ramamurthy http://orcid.org/0000-0002-4104-7513 University of Nebraska - Lincoln [biography]

Byrav Ramamurthy is currently a Professor in the Department of UNL School of Computing at the University of Nebraska-Lincoln (UNL). He is the author of the book "Design of Optical WDM Networks - LAN, MAN and WAN Architectures" and a co-author of the book "Secure Group Communications over Data Networks" published by Kluwer Academic Publishers/Springer in 2000 and 2004 respectively. He served as the Chair of the IEEE Communication Society's Optical Networking Technical Committee (ONTC) during 2009-2011. He served as the IEEE INFOCOM 2011 TPC Co-Chair. He is currently the Editor-in-Chief for the Springer Photonic Network Communications (PNET) journal.

His research areas include optical and wireless networks, peer-to-peer networks for multimedia streaming, network security and telecommunications. His research work is supported by the U.S. National Science Foundation, U.S. Department of Energy, U.S. Department of Agriculture, NASA, AT&T Corporation, Agilent Tech., Ciena, HP and OPNET Inc.

Note

The full paper will be available to logged in and registered conference attendees once the conference starts on June 21, 2026, and to all visitors after the conference ends on June 24, 2026

« View session

For those interested in:

computer science
undergraduate