LLM-Based Educational Simulation: Evaluating Temporal Student Persona Stability Across ADHD Profiles

TL;DR

This study evaluates the stability of student personas in LLM-based educational simulations across extended interactions, highlighting the importance of structured prompts for maintaining behavioral coherence in ADHD profiles.

Contribution

It introduces a dual-assessment framework to measure persona stability in LLM simulations and demonstrates that scripted interactions enhance behavioral consistency.

Findings

Self-reported characteristics are stable at high intensities.

Unscripted dialogues show behavioral drift in ADHD personas.

Structured prompts eliminate behavioral drift.

Abstract

Student simulation with Large language models (LLMs) offers a scalable alternative for educational research and teacher training. Yet, its validity depends on whether models maintain stable personas across extended interactions. We test this prerequisite using a dual-assessment framework measuring self-reported characteristics and observer-rated behavioral expressions. Across two experiments testing four clinically-grounded ADHD persona conditions, five LLMs, and three prompt designs, we quantify between-conversation stability (N=4,968) and within-conversation stability (N=3,952 across 9 turns). Self-reported characteristics remain stable for high intensities, constituting a necessary prerequisite for valid behavioral simulation. Observer-rated behavioral expression reveals selective instability: within-conversation drift occurs in unscripted dialog for high and moderate ADHD personas. Scripted interactions with explicit task prompts eliminate this drift entirely. Stable, persona-aligned simulated learners benefit from a structured interaction design to maintain behavioral coherence, which holds significant implications for teacher training, adaptive tutoring, and any application requiring sustained, path-dependent learner interactions.