LLM-Generated Basketball Lesson Plans Evaluated by Experts

28 quality criteria were used to evaluate basketball layup lesson plans generated by three LLMs (GPT-4o, Claude Sonnet, and Google Gemini). Teaching experts rated the plans on 5-point Likert scales, with the most frequent median rating being 'acceptable'. Benedikt Meixner published this research on figshare in March 2026.

Use Cases

• Benchmark LLM performance for lesson planning based on expert ratings
• Analyze the quality of AI-generated lesson plans based on 28 predefined criteria
• Compare the outputs of GPT-4o, Claude Sonnet, and Google Gemini for a specific teaching task

Strengths

• Evaluation based on 28 predefined quality criteria
• Expert ratings from teaching professionals on a 5-point Likert scale
• Direct comparison of three major LLMs (GPT-4o, Claude Sonnet, Google Gemini)

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download
• Scale is limited to a single physical education topic (basketball layup)

Provenance

Source

figshare

Collection Method

Generated by submitting a prompt to three LLMs and surveying teaching experts.

Freshness

Last updated 2026-03-19 06:48:13