Research Dataset: Evaluating Instructional Reliability of Large Language Models in Curriculum-Constrained Learning Objective Generation: Toward Quality Education (SDG 4)

This dataset contains the evaluation results of 160 learning objectives generated by Large Language Models (LLMs) under curriculum-constrained instructional prompts. The dataset was developed as part of a research study investigating the instructional reliability and quality of AI-generated learning objectives. The evaluation focuses on the alignment of AI-generated learning objectives with established instructional design principles, including Bloom’s Taxonomy and the ABCD (Audience, Behavior, Condition, Degree) framework. The dataset contains the analyzed outputs and evaluation results used to compare the instructional reliability of the generated learning objectives. The dataset is intended to support research on Generative AI in education, particularly the use of Large Language Models for instructional design and learning objective generation. It may also be used for further analysis, replication, and comparative studies concerning prompt design, curriculum constraints, and AI-generated educational content.