Collaborative Research: Software Engineering Tutor - A Cohesive, Integrated Feedback Engine to Support Inquiry-Based Learning of Software Engineering Concepts

This project will serve the national interest by helping more undergraduate students become skilled and careful software developers. Learning to write software is hard, and students rely heavily on automated tools that check their work, such as programs that grade assignments, scan code for mistakes, or look for security weaknesses. These tools are good at pointing out that something is wrong, but they rarely explain why it matters or how to think about fixing it, so students often guess and patch their code without really understanding the ideas behind the problem. This project will create a freely available teaching tool, the Software Engineering Tutor, that takes the dense technical messages produced by checking tools and turn them into clear, guiding questions that help students discover the underlying concepts for themselves. The significance of this work is its potential to change the feedback students receive in computing courses from simple error-flagging into genuine learning, across topics such as writing high-quality code, testing, and security. Its importance also lies in the national need for a larger computing workforce: by sharing the tool, its training materials, and faculty workshops openly, the project will bring strong, modern teaching methods to many colleges and universities, including those with fewer resources. This Engaged Student Learning Level 3 project will help students move from simply correcting mistakes toward understanding the concepts beneath them. The project’s goals are to advance understanding of how automated, theory-grounded feedback affects conceptual learning in computing and to investigate the conditions under which inquiry-based feedback improves student outcomes. Its scope spans the undergraduate computer science curriculum, from introductory programming through data structures, software testing, and software security, across three partnering institutions with distinct educational contexts. The project plans to operationalize established learning theories (e.g., cognitive load theory, the zone of proximal development, constructivist and inquiry-based learning, and metacognition) as concrete design principles within an artificial-intelligence feedback engine and a plug-in architecture that integrates external tools such as static analyzers, coverage tools, and security scanners. The project plans to study the effectiveness of this approach through a phased, mixed-methods research design that includes think-aloud formative studies, randomized and within-subjects pilot studies with pre- and post-tests, learning analytics, and longitudinal, cross-institutional analyses; assessment and evaluation are planned with an external evaluator, and results, datasets, instruments, and open-source software are planned for dissemination through workshops, peer-reviewed publications, and public repositories. By linking tool feedback, student responses, and learning outcomes at scale, the project plans to add new knowledge about how students develop programming expertise, which scaffolding progressions are most effective, and how artificial-intelligence-generated feedback compares with traditional feedback. The NSF IUSE: EDU Program supports research and development projects to improve the effectiveness of STEM education for all students. Through the Engaged Student Learning track, the program supports the creation, exploration, and implementation of promising practices and tools. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.