Fall clusters of coding-question students trend toward higher exam scores than report-writer clusters¶

Claim · All claims
Evidence · 1 study · associational r2 · q2
Currency · no dated study recorded · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? In fall-semester K-means clusters, coding-question groups trended toward higher average exam scores (89±7.3) than report-writers (83.8±11.7), though cluster membership added no statistically significant explanatory power. → cluster-exam

Evidence¶

cluster-exam¶

Hunter McNichols, Fareya Ikram, and Andrew Lan. 2026. The StudyChat Dataset: Analyzing Student Dialogues With ChatGPT in an Artificial Intelligence Course. In LAK26: 16th International Learning Analytics and Knowledge Conference (LAK 2026). https://doi.org/10.1145/3785022.3785029

q2 · i? · associational · r2

K-means clustering (k=4, chosen by elbow method and silhouette analysis) of normalized dialogue act feature vectors, with fall cluster outcomes compared. All clusters had comparable average assignment scores; the authors suggest targeted coding questions beat extensive writing assistance, "which may circumvent the learning objectives of the assignments."

"coding-question groups trend towards a higher average score on the exams89± 7.3while report-writers tend to have a marginally lower average score83.8± 11.7"

Discussion¶

Related Claims¶

• PTA-derived traits significantly improve explained exam-score variance over a prior-outcome baseline in one semester of StudyChat but not the other — related
• No significant differences in AI-free assessment performance across GenAI usage clusters in an OOP course — related
• No group effects on learning and performance were found: a non-significant intra-group correlation — related
• High-frequency LLM users show higher minimum course outcomes and reduced score variability, though interaction level is not a dominant outcome predictor — related
• The six clusters differ substantially in academic performance: Final Grade ANOVA η² = 0.770, with C0 exceeding C1 by 40.17 exam-score points — related
• In the first deployment, exam scores on the induction question were visibly lower than in previous semesters taught without the tool — related