2026 ASEE Annual Conference & Exposition

A Comparative Analysis of Machine Learning Algorithms for Assessing Engineering Students’ Problem-Solving

Assessing undergraduate and graduate students’ problem-solving is essential for engineering education and requires considerable human effort. Existing Systems focus on the correctness of final answers rather than reasoning or processes; there remains a critical gap in artificial intelligence (AI) tools capable of analyzing the problem-solving process itself, a reflection of cognitive sophistication utilized during solution generation. Conventional automated assessment systems integrated into Learning Management Platforms (e.g., Canvas, Moodle, Blackboard Learn) predominantly emphasize the correctness of final answers, often overlooking the underlying reasoning and solution processes. This oversight can result in inaccurate evaluations, where correct solutions may arise from flawed logic, or well-structured approaches may yield minor computational errors.
To overcome this problem, a hybrid human-machine dual assessment solution for LMS platforms is proposed in this study and a comparative analysis of supervised machine learning algorithms (Logistic Regression, Naïve Bayes, K-Nearest Neighbors, Decision Tree, Random Forest, etc.) are performed to predict student performance based on their problem-solving processes in the context of electrical circuit.
A dataset comprising 363 solution events was collected from the course Fundamental Electronics for Engineers at a mid-sized land-grant university in the Western United States. Since the dataset showed some imbalance across problem types and performance categories, an over-sampling method (SMOTEN) was used to balance the data. Each solution was qualitatively evaluated by two independent researchers using five process-oriented constructs adapted from Docktor et al.'s (2016) rubric: Useful Description, Engineering Approach, Specific Application of Engineering Principles, Mathematical Procedures, and Logical Progression.
Solutions were classified into three performance categories—Correct, Partially Correct, and Incorrect—based on expert comparisons. The machine learning models were trained and evaluated using key performance metrics: accuracy, precision, recall, F1-score, specificity, AUROC, mean cross validation scores, etc. The findings provide evidence of AI utility in assessing cognitive processes associated with engineering problem-solving. This study also has several implications across STEM disciplines, workforce management and development.

Authors

Mr. Rifatul Islam Himel http://orcid.org/0000-0002-0415-4774 Utah State University [biography]

Rifatul Islam Himel is a doctoral candidate in Engineering Education at Utah State University (USU), Logan, Utah, US. His doctoral research focuses on understanding human cognition in engineering problem-solving. He holds a Bachelor's degree in Electrical and Electronics Engineering and a Master's degree in Electronics and Telecommunication Engineering. With over six years of professional experience in the Information Technology industry and a strong background in teaching and research, Rifatul brings a unique interdisciplinary perspective to his work. His research interests include Human and Machine Cognition, Applied Artificial Intelligence in Engineering Education, Information Processing, and the study of Consciousness.
Dr. Oenardi Lawanto Utah State University [biography]

Dr. Oenardi Lawanto is a professor in the Department of Engineering Education at Utah State University, USA. He received his B.S.E.E. from Iowa State University, his M.S.E.E. from the University of Dayton, and his Ph.D. from the University of Illinois at Urbana-Champaign. Dr. Lawanto has a combination of expertise in engineering and education and has more than 30 and 14 years of experience teaching engineering and cognitive-related topics courses for his doctoral students, respectively. He also has extensive experience in working collaboratively with several universities in Asia, the World Bank Institute, and USAID to design and conduct workshops promoting active-learning and life-long learning that is sustainable and scalable. Dr. Lawanto’s research interests include cognition, learning, and instruction, and online learning.
Dr. Yashin Brijmohan http://orcid.org/0000-0003-2485-9808 Utah State University [biography]

Yashin Brijmohan is a registered professional engineer who is currently appointed as Chairman of Engineering Education Standing Technical Committee of the Federation of African Engineering Organizations, Executive committee member of the Commonwealth Engineers Council, Board Member of the UNESCO International Centre for Engineering Education, and Co-Chair of the Africa Asia Pacific Engineering Council.

He was the founding Executve Dean of Business, Engineering and Technology at Monash South Africa, former Vice President of the World Federation of Engineering Organizations, and led several committees in the engineering profession.

Yashin has both leadership and specialist experience within the engineering power industry and education sectors and is known for his thought leadership in capacity building and engineering education.
Dr. Wade H Goodridge http://orcid.org/0000-0002-5811-7629 Utah State University [biography]

Wade Goodridge is a tenured Associate Professor in the Department of Engineering Education at Utah State University. His research often has a focus on what drives us to learn, how we learn and how well we do it, and what can improve that learning. This work is often directed with a spatial thinking emphasis. Dr. Goodridge has investigated different learning modalities (fully online, hybrid, blended courses), cognitive and metacognitive learning processes (self-regulation, engineering problems solving, neural efficiency in engineering problem solving) learning motivators (engineering self-efficacy), and effective and personalized engineering education interventions (tactile engineering curriculum, physical teaching models, XR environments and AI informed systems of delivering and monitoring engineering learning). Dr. Goodridge has created a valid and reliable instrument to assess spatial ability in blind and low vision (BLV) learners thus opening new areas of research centered on tactile spatial ability. He has created tactile and visual learning tools and has offered professional development experiences on how to improve the education of tomorrows engineers.
Dr. Zain ul Abideen Utah State University [biography]

Zain ul Abideen holds a Ph.D. in Engineering Education from Utah State University, where he also served as a Graduate Researcher. He is a computer engineer and earned his bachelor’s degree in computer engineering and master’s degree in engineering, and brings over a decade of experience in teaching and research with undergraduate engineering students. He actively contributes to several technical committees and serves as a reviewer for leading journals and international conferences. His doctoral research focused on cognitive engagement, AI in education and engineering design, self-regulated learning, and mobile learning. His work seeks to deepen our understanding of how these elements influence learning and success in engineering education. With a strong commitment to educational advancement and a robust academic background, Zain is recognized as a dedicated and emerging scholar in the field of engineering education.
Dr. Talha Naqash Dickinson State University [biography]

Dr. Talha Naqash completed his doctoral studies in Engineering Education at Utah State University. With a diverse academic foundation, he also holds a Master’s degree in Telecommunication and Networking. He is currently serving as a faculty member at Dickinson State University in the School of Applied Sciences. Dr. Naqash’s research focuses on cognitive engagement analysis, assessment methods in engineering education, and the role of facial expressions and emotions in the learning process. His scholarly contributions include numerous publications and presentations at prestigious venues such as IEEE, ASEE conferences, and leading Wiley and Springer journals. He is an active member of several technical committees and serves as a reviewer for reputable journals and international conferences, including the ASEE and IEEE Transactions on Education. Dr. Naqash’s dedication to advancing educational research, combined with his broad academic and professional experience, underscores his commitment to fostering innovation and excellence in engineering education.
Dr. Assad Iqbal http://orcid.org/https://0000-0002-8060-7384 Ohio State University, Columbus Ohio [biography]

Assad Iqbal (Sid) is a Senior Lecturer at the Department of Engineering Education in the College of Engineering, the Ohio State University. Assad has been working with the National Science Foundation (NSF) funded national collaboration research project, "Engineering For Us All (e4usa) since October 2022. The core mission of this project is to increase student and teacher access to engineering with a focus on reaching traditionally underrepresented populations in this field.

With a BS in Computer Information Systems Engineering, a Master of Science in Engineering Management, and a Ph.D. in Engineering Education, Assad's long-term research goal is to converge his varied research experiences and endeavors to explore the possibilities of an inclusive instructional design to promote self-directed, self-regulated, life-long learning among pre-college and undergraduate engineering students.

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

For those interested in:

computer science
engineering
engineering technology
information technology