This work presents the full research design of a pre-registered quasi-experimental pilot study on the responsible integration of large language models (LLMs) into Legal English instruction.
Unlike most studies in this rapidly evolving field, our design was publicly registered prior to data collection (AsPredicted №278663), ensuring methodological transparency and eliminating post-hoc hypothesis revision.
The study compares LLM-assisted and traditional instructional approaches, measuring:
– Legal English competencies (writing, argumentation, communication)
– academic integrity attitudes
– pedagogical effectiveness of structured AI integration
We see this preprint as part of a broader effort to build rigorous, transparent, and reproducible research on AI in education.
We welcome feedback, discussion, and potential collaboration.