WIP: A Pedagogical Prompt Engineering Framework for LLM-Based Feedback in Higher Education (PPE-LLM)

This Work-in-Progress (WIP) paper introduces PPE-LLM, a structured framework for prompt engineering that enables LLM -generated feedback to be pedagogically meaningful and tailored to novice programming students in introductory pro-gramming education. Although large language models (LLMs) show promise for feedback generation, existing research lacks a structured methodology that integrates pedagogical principles, student needs, best-practice prompting, and technical consid-erations. PPE- LLM addresses this gap by providing a theory-driven framework that supports the systematic design of prompts to generate formative feedback, enhancing learning tailored to students' learning levels and needs. This paper presents the key components of the framework, along with a practical example prompt. It also analyses the alignment of recent CER studies with the PPE- LLM components. The results indicate that studies with limited alignment often produced inconsistent or lower-quality feedback, whereas more substantial alignment was found to be pedagogically effective in AI-generated feedback. Future work includes expert validation and student evaluation in real-world settings.