Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Theory of Adaptive Scaffolding for LLM-Based Pedagogical Agents

Clayton Cohn, Surya Rayala, Namrata Srivastava, Joyce Horn Fonteles, Shruti Jain, Xinying Luo, Divya Mereddy, Naveeduddin Mohammed, Gautam Biswas

TL;DR

This paper addresses the lack of theoretical grounding in LLM-based pedagogical agents by proposing a framework that fuses Evidence-Centered Design (ECD) with Social Cognitive Theory (SCT) to enable adaptive scaffolding guided by Zone of Proximal Development (ZPD). The Inquizzitor system implements this framework with an assessment module (ECD-driven) and an adaptive decision module (ZPD+SCT), evaluated in a real middle-school Earth Science setting with 104 students over three weeks. The evaluation demonstrates strong scoring accuracy and faithfulness to theoretical constructs, along with positive student perceptions, though goal-setting support and some off-task behaviors require further attention. The work highlights the potential of theory-driven LLM integration for adaptive, principled instruction and outlines concrete directions for future research, including knowledge-graph–based ZPD tracking and ZPD-aware RLHF.

Abstract

Large language models (LLMs) present new opportunities for creating pedagogical agents that engage in meaningful dialogue to support student learning. However, the current use of LLM systems like ChatGPT in classrooms often lacks the solid theoretical foundation found in earlier intelligent tutoring systems. To bridge this gap, we propose a framework that combines Evidence-Centered Design with Social Cognitive Theory for adaptive scaffolding in LLM-based agents focused on STEM+C learning. We illustrate this framework with Inquizzitor, an LLM-based formative assessment agent that integrates human-AI hybrid intelligence and provides feedback grounded in cognitive science principles. Our findings show that Inquizzitor delivers high-quality assessment and interaction aligned with core learning theories, offering teachers effective guidance that students value. This research underscores the potential for theory-driven LLM integration in education, highlighting the ability of these systems to provide adaptive and principled instruction.

A Theory of Adaptive Scaffolding for LLM-Based Pedagogical Agents

TL;DR

This paper addresses the lack of theoretical grounding in LLM-based pedagogical agents by proposing a framework that fuses Evidence-Centered Design (ECD) with Social Cognitive Theory (SCT) to enable adaptive scaffolding guided by Zone of Proximal Development (ZPD). The Inquizzitor system implements this framework with an assessment module (ECD-driven) and an adaptive decision module (ZPD+SCT), evaluated in a real middle-school Earth Science setting with 104 students over three weeks. The evaluation demonstrates strong scoring accuracy and faithfulness to theoretical constructs, along with positive student perceptions, though goal-setting support and some off-task behaviors require further attention. The work highlights the potential of theory-driven LLM integration for adaptive, principled instruction and outlines concrete directions for future research, including knowledge-graph–based ZPD tracking and ZPD-aware RLHF.

Abstract

Large language models (LLMs) present new opportunities for creating pedagogical agents that engage in meaningful dialogue to support student learning. However, the current use of LLM systems like ChatGPT in classrooms often lacks the solid theoretical foundation found in earlier intelligent tutoring systems. To bridge this gap, we propose a framework that combines Evidence-Centered Design with Social Cognitive Theory for adaptive scaffolding in LLM-based agents focused on STEM+C learning. We illustrate this framework with Inquizzitor, an LLM-based formative assessment agent that integrates human-AI hybrid intelligence and provides feedback grounded in cognitive science principles. Our findings show that Inquizzitor delivers high-quality assessment and interaction aligned with core learning theories, offering teachers effective guidance that students value. This research underscores the potential for theory-driven LLM integration in education, highlighting the ability of these systems to provide adaptive and principled instruction.

Paper Structure

This paper contains 12 sections, 3 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Theoretical Framework
  3. Study Design
  4. Methodology
  5. Assessment Module
  6. Adaptive Decision Module
  7. Evaluation
  8. Scoring Accuracy (RQ1)
  9. Faithfulness (RQ2)
  10. Insights from Students' Perception Survey
  11. Related Work
  12. Discussion and Conclusions

Figures (3)

  • Figure 1: Framework for LLM agent adaptive scaffolding.
  • Figure 2: Inquizzitor's key components. The blueassessment module applies ECD to generate mastery evidence from formative assessments; the greenadaptive decision module uses this evidence to scaffold student feedback.
  • Figure 3: Sequences of Inquizzitor utterance turns (x-axis) for three case study students during FA3, annotated by theoretical and teacher constructs (y-axis).