Applying Large Language Models and Chain-of-Thought for Automatic Scoring
This addresses the challenge of making AI-based automatic scoring tools more accessible and explainable for researchers and educators in educational assessment, though it is incremental as it builds on existing LLM and CoT methods.
This study tackled the problem of automatic scoring of student-written science assessments by applying large language models (LLMs) like GPT-3.5 and GPT-4 with Chain-of-Thought (CoT) prompting, finding that few-shot learning (accuracy = .67) outperformed zero-shot learning (accuracy = .60) by 12.6%, and CoT with contextual item stems and rubrics increased accuracy by up to 13.44%.
This study investigates the application of large language models (LLMs), specifically GPT-3.5 and GPT-4, with Chain-of-Though (CoT) in the automatic scoring of student-written responses to science assessments. We focused on overcoming the challenges of accessibility, technical complexity, and lack of explainability that have previously limited the use of artificial intelligence-based automatic scoring tools among researchers and educators. With a testing dataset comprising six assessment tasks (three binomial and three trinomial) with 1,650 student responses, we employed six prompt engineering strategies to automatically score student responses. The six strategies combined zero-shot or few-shot learning with CoT, either alone or alongside item stem and scoring rubrics. Results indicated that few-shot (acc = .67) outperformed zero-shot learning (acc = .60), with 12.6% increase. CoT, when used without item stem and scoring rubrics, did not significantly affect scoring accuracy (acc = .60). However, CoT prompting paired with contextual item stems and rubrics proved to be a significant contributor to scoring accuracy (13.44% increase for zero-shot; 3.7% increase for few-shot). We found a more balanced accuracy across different proficiency categories when CoT was used with a scoring rubric, highlighting the importance of domain-specific reasoning in enhancing the effectiveness of LLMs in scoring tasks. We also found that GPT-4 demonstrated superior performance over GPT -3.5 in various scoring tasks when combined with the single-call greedy sampling or ensemble voting nucleus sampling strategy, showing 8.64% difference. Particularly, the single-call greedy sampling strategy with GPT-4 outperformed other approaches.