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¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · q2
Currency · newest study 2005 · not yet checked against its sources · not reviewed by a person · overdue for a recheck since 2008: tool and model accuracy changes quickly

Subclaims¶

q2 i? For implicitSE assessment, time alone gave 67.9% sensitivity versus 62.3% for gaze shifts alone, while gaze shifts gave higher specificity (67.8% vs 64.8%); the combined predictor outperformed either on its own (73.0% combined). → Christina Merten and Cristina Conati 2005 q2 i? The same complementary pattern held for exploration assessment, with the combined predictor yielding the highest accuracy for each measure (75.1% combined). → Christina Merten and Cristina Conati 2005 (2)

Evidence¶

Christina Merten and Cristina Conati 2005¶

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

q2 · i? · design · r2

Model 3 log files from the new study were rerun twice, once withholding eye-tracking data and once withholding time data (Table 7). The authors note these findings match the original user study and are consistent with time overestimating self-explanation.

"information on time alone generates higher sensitivity than information on gaze shifts alone, while the latter generates higher specificity"

Christina Merten and Cristina Conati 2005 (2)¶

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

q2 · i? · design · r2

Similar evidence-withholding analysis for exploration nodes (Table 8): time-only combined accuracy 72.1%, gaze-only 71.6%, both 75.1%. Each single predictor "seems to succeed where the other fails", which the authors link to the combined predictor's high accuracy.

"It should also be noted that each si ngle predictor seems to succeed where the other fails so this comp limentary behavior likely contributes to the high accuracy of the combined predictor."

Discussion¶

Related Claims¶

• Adding eye-tracking evidence to ACE's student model significantly improves assessment of implicit self-explanation over a time-only model
• More accurate assessment of self-explanation significantly improves the model's assessment of student learning through exploration
• Leave-one-out cross-validation over 36 students confirms significant accuracy and stability gains from eye-tracking in both self-explanation and exploration assessment — related
• 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 — related