2025 ASEE Annual Conference & Exposition

Writing proficiency is an essential skill for engineers, particularly during capstone design projects, which serve as a culmination of their academic journey. These projects require students to communicate complex technical ideas effectively. However, the conventional approach to evaluating writing in capstone projects is predominantly summative, offering feedback only after the submission is complete. In the past, writing tools such as Grammarly, ProWritingAid, and MY Access! have been developed to provide automatic corrective feedback, primarily addressing lower-level writing issues like grammar and spelling. While these tools have been effective in tackling surface-level language errors, they fall short in supporting students with more advanced writing skills. Specifically, they lack the capacity to offer in-depth feedback on critical aspects such as coherence, argumentation, and organizational structure. Furthermore, in academic and technical writing contexts, these tools are often unable to provide the task-specific guidance that students need to meet assignment requirements. This reactive method limits the potential for students to improve their writing during the project itself.

To address this gap, we introduce "FeedCap", a tool designed to deliver real-time, rubric-specific feedback, helping students refine their writing as they work. By providing timely, formative feedback, FeedCap supports the development of students' writing abilities, ensuring their work aligns with the assignment's specific criteria.

The development process for FeedCap followed a multi-phase approach. We first engaged with faculty to identify challenges in current capstone project assessments and key areas for improvement. From there, we created a digital framework that integrates seamlessly with existing educational platforms. FeedCap uses advanced natural language processing, leveraging large language models and document retrieval techniques to assess student submissions against rubric criteria. FeedCap focuses on areas such as clarity, conciseness, correctness, and compelling communication (the four Cs).

At the current state, the tool has been designed to address feedback in writing skills for students enrolled in a Capstone Design Project in Construction Management in a Southeastern institution in the United States. However, the tool is continuously evolving towards a more generalizable approach in engineering design. Ongoing work aims to refine the feedback algorithms, improve responses to more complex student inputs, and deliver human-like evaluations. Long-term studies will assess the tool’s user engagement and impact on student writing skills across multiple academic terms, with the goal of fostering a deeper integration of continuous, constructive feedback into the capstone experience.