Assessment in Team Problem-Solving Exercises in Computing Education

2026-07-21 · Source: Artificial Intelligence · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Expert, quick

Summary

This paper presents methods for assessing student teams in tabletop exercises (TTXs) within computing education, addressing the challenge of delayed or incomplete feedback in complex, open-ended scenarios. Researchers compared two post-TTX assessment approaches—clustering and large language models (LLMs)—using an original dataset from 81 participants across two countries. The methods were evaluated against instructor scores based on standardized rubrics. Clustering effectively grouped teams with similar task approaches, facilitating faster, targeted feedback with low computational demands, proving both valid and reliable. For LLM-based assessment of team communication, GPT-4o frequently diverged from instructor scores, whereas GPT-5.2 demonstrated considerably lower error. These assessment methods have been integrated into INJECT, an open-source TTX learning platform, and all datasets, software, and a full TTX scenario are publicly shared to encourage community adoption.

Key takeaway

For educators designing or implementing team problem-solving exercises, this research suggests integrating computational assessment methods to improve feedback efficiency. You should consider clustering algorithms to quickly group teams with similar approaches, enabling targeted and timely feedback. Additionally, if automating communication assessment, prioritize advanced LLMs like GPT-5.2, as their accuracy significantly surpasses earlier models. This approach can scale assessment efforts and enhance learning outcomes in your computing education programs.

Key insights

Computational methods like clustering and advanced LLMs can enhance and accelerate team assessment in educational tabletop exercises.

Principles

• Clustering identifies similar team approaches for targeted feedback.
• LLM performance in rubric-based assessment varies significantly by model generation.
• Open-source integration and public data sharing foster method adoption.

Method

Compare clustering and LLMs for post-TTX team assessment against instructor scores and standardized rubrics, then integrate validated methods into an open-source platform.

In practice

• Apply clustering to group student teams for faster, tailored feedback.
• Utilize advanced LLMs for automated, rubric-aligned communication assessment.
• Explore the INJECT platform for open-source TTX assessment tools.

Topics

• Team Assessment
• Tabletop Exercises
• Large Language Models
• Clustering Algorithms
• Computing Education
• INJECT Platform

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• Towards Just-in-Time Adaptive Feedback: Enhancing Student Learning via Knowledge-Grounded LLM — Knowledge-grounded LLMs can provide effective Just-in-Time adaptive feedback, significantly improving student learning outcomes.
• An Integrated System for Real-Time Student Assessment and Career Guidance Using Neural Networks in Computing Disciplines — An integrated AI system can accurately guide students toward suitable careers by aligning abilities and interests.
• The Effects of Structured LLM-Generated Feedback on Programming Assignment Performance — LLM-generated feedback, especially less guided forms, significantly reduces programming assignment solution time for students.
• CRAFT: Clustering Rubrics to Diagnose Weak LLM Capabilities and Generate Targeted Fine-Tuning Data — CRAFT diagnoses LLM weaknesses at the rubric criterion level to generate targeted fine-tuning data, improving model performance.
• GTBench: A Curriculum-Grounded Benchmark for Evaluating LLMs as Mathematical Research Assistants in Graph Theory — LLMs struggle with mathematical reasoning, particularly graduate-level proofs, with GPT-5 significantly outperforming other models.
• Automatically Inferring Teachers' Geometric Content Knowledge: A Skills Based Approach — Integrating explicit skills information significantly improves LLM-based classification of teachers' Van Hiele geometric reasoning levels.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence.