EDTECH Dissertation Proposal Review

Overview

Once students feel comfortable with their dissertation proposal draft, including formatting of the chapters and a reference section, students upload it to the RapTR tool (Rapid Thesis Review).

RapTR helps to remove the bottleneck that exists for students and dissertation committee members, which can be significant, as students wait for meaningful feedback on their relatively large and important documents.

AI Tools Used

• Multi-agent based AI tools (e.g., Flowise, etc.)

• ChatGPT with multimodal capabilities

Pedagogical Application

Providing a preliminary review based on department standards in terms of format, edits, content, consistency, and APA style standards helps students refine their proposal prior to full committee review, and provides a level of completeness that can assist both students and committee members in ensuring a successful proposal defense.

How It Works

Students submit their dissertation proposal draft and receive feedback based on the rubric provided by the department. The feedback provides a rough “score” to give students an idea on its readiness to share with their doctoral committee. It could be the case that the feedback is less relevant to the student’s particular proposal, depending on the context, however students should be aware of any ideas it provides for possible improvements or revisions.

Future Development

The RapTR tool holds potential beyond its initial application. Different departments across colleges may use it by indicating their own (different) rubrics and expectations for a dissertation proposal. The system currently only uses information based on the expectations of a specific set of requirements, rather than proposal information that exists outside of these guidelines, thus ensuring privacy and confidentiality to the specific student. The tool thereby “resets” after each review instead of using prior reviews as part of its feedback.

Flowise helps link different AI agents for different purposes within its review process depending on the review task (eg. format versus content), and review feedback is then combined into a single assessment.