Leveraging ChatGPT for Scoring Students' Subjective Tests

Abstract

Assessing students' understanding of circle-related material through subjective tests is effective, though grading these tests can be challenging and often requires technological support. ChatGPT has shown promise in providing reliable and objective evaluations. Many teachers in Indonesia, however, continue to face difficulties integrating technology into assessments, highlighting the need for research on leveraging ChatGPT for scoring subjective tests. This study explores the potential of ChatGPT in mathematics education, specifically for evaluating students’ answers on subjective tests related to circles. Using a qualitative case study approach, the research investigates how ChatGPT analyzes student responses. The findings reveal that ChatGPT effectively processes the answers of 15 vocational school students, delivering results that are logical and well-structured. ChatGPT aids teachers by offering a comprehensive evaluation focused on accuracy, clarity, completeness, and technical terminology. This approach supports effective feedback and objective assessment, making ChatGPT a valuable tool for understanding and addressing students' mathematical comprehension issues.