ACE's Dynamic Bayesian Network student model with gaze-shift and time nodes as independent predictors of implicit self-explanation

Design · All designs
Evidence · 5 claims (5 for) · 1 study (1 design), q2 · 0 of 1 report an effect size · 5 claims rest on one study

Description

ACE's student model is a Dynamic Bayesian Network that assesses exploratory behavior at multiple levels of granularity (exploration cases, exercises, units, and categories) together with knowledge nodes. The revised model separates the causes of implicit self-explanation from its effects, encoding gaze shifts and time on action "as independent predictors as in a naïve Bayesian classifier". This structure is highly modular, allowing the gaze shift and time nodes to be used or ignored as needed, and facilitates learning the conditional probability tables from data.

Design Implications

Context

Requirements

• Conditional probability tables for the time and gaze shift nodes derived from user-study data frequencies
• A threshold over implicitSE node probabilities, set via ROC analysis over training data

Constraints

• The structure assumes independence between time and the presence of gaze shifts, which the data suggests is not strictly true (a small positive correlation)
• The model does not currently include the coach's hint-to-self-explain nodes because no hint data was available from the user study

Target Learners

• University students learning mathematical functions through exploration

Learning Goals

• Effective exploratory learning of mathematical functions
• Self-explanation of instructional material

Claims

• Eye Tracking Improves Implicit Se Assessment [+M]
• Se Assessment Improves Exploration Assessment [+M]
• Time and gaze-shift evidence are complementary predictors of self-explanation: time alone yields higher sensitivity, gaze shifts alone higher specificity, and the combination the best accuracy [+W]
• In an initial eye-tracking study, time on action predicted self-explanation (16-second threshold) and gaze shifts were most reliable at detecting its absence [+W]
• Leave-one-out cross-validation over 36 students confirms significant accuracy and stability gains from eye-tracking in both self-explanation and exploration assessment [+W]

Related Designs

Examples

-

Key Sources

• Christina Merten and Cristina Conati. (2005). Eye-Tracking to Model and Adapt to User Meta-cognition in Intelligent Learning Environments. https://www.cs.ubc.ca

Page metadata

Field	Value
type	design
id	ace-dbn-student-model-gaze-time-predictors
title	ACE's Dynamic Bayesian Network student model with gaze-shift and time nodes as independent predictors of implicit self-explanation
description	ACE's student model is a Dynamic Bayesian Network that assesses exploratory behavior at multiple levels of granularity (exploration cases, exercises, units, and categories) together with knowledge nodes.
status	review
generated	by process:settle-candidates on 2026-10-16
trust	unverified — no one has confirmed this page's content
sources	1 — christina-merten-and-cristina-conati-2005

Frontmatter is the page's machine-readable record — see the Schema & Guide.