Revue Algérienne des Sciences du Langage
Volume 11, Numéro 1, Pages 328-341
2026-06-10

Examining The Impact Of Structured Prompt Frameworks On The Pedagogical Quality And Cognitive Level Of Ai-generated Research Methodology Lesson Plans

Authors : Boukhelouf Athina .

Abstract

This study uses a 2x2 factorial design to examine the effects of AI model (Chat GPT vs Copilot), prompt framework (CRISPE vs TAG) and their interaction on the pedagogical quality and cognitive demand of lesson plans. Twenty research methodology lesson plans were generated under the four conditions and evaluated using an analytic evaluation rubric. The latter was designed and validated to assess five major dimensions, namely learning objectives and pedagogical alignment, lesson structure and clarity, cognitive demand and higher order thinking skills, practical usability, learner engagement and motivation. The two- way analysis of covariance (ANOVA) indicates that both AI model and prompt framework significantly influence the pedagogical quality of research methodology lesson plans. Lesson plans generated using CRISPE achieved substantially higher pedagogical quality scores than those generated by TAG, across both AI models. Particularly, lesson plans produced by Chat GPT under the CRISPE framework proved well-structured and pedagogically aligned. On the other hand, TAG lesson plans revealed multiple weaknesses relevant to alignment, clarity and learner engagement. Across all conditions, lesson plans failed to account for higher order thinking skills consistently. Overall, this apparent gap implies that AI lesson plans cannot serve as pedagogical alternatives to human lesson plans. Indeed, the human pedagogical expertise is crucial to ensure that AI lesson plans are fully comported with pedagogy, cognitive rigor and education standards.

Keywords

prompt framework ; pedagogical quality ; cognitive demand ; higher order thinking skills ; Chat GPT ; Copilot