Decoding reading comprehension from eye movements in narrative and expository texts: A keyword- and whole-text-level...

Area of Science:

• Cognitive Psychology
• Reading Science
• Ophthalmology

Background:

• Eye-movement patterns are recognized predictors of reading comprehension proficiency.
• The influence of text genre and analysis level on this coupling is not well understood.

Purpose of the Study:

• Investigate genre-specific eye-movement patterns in narrative vs. expository reading.
• Evaluate the predictive power of these patterns for comprehension at keyword and whole-text levels.

Main Methods:

• Extracted eye-movement features at whole-text and keyword levels during narrative and expository reading.
• Employed Support Vector Machine Recursive Feature Elimination for predictive modeling and feature selection.

Main Results:

• Distinct eye-movement profiles observed for narrative ('Dual-Stage Balanced Mode') and expository ('Late-Stage Compensatory Mode') reading.
• Keyword-level analysis significantly outperformed whole-text analysis in predicting reading comprehension proficiency.

Conclusions:

• Eye-movement patterns provide a transformative approach to assessing reading comprehension.
• Findings offer heuristics for developing novel reading assessment and instructional tools.