Feedback beyond accuracy: Using eye-tracking to detect comprehensibility and interest during reading

Area of Science:

• Information Science and Technology
• Human-Computer Interaction
• Cognitive Science

Background:

• Understanding user information needs is crucial for effective information retrieval systems.
• Implicit feedback, such as user interactions, is vital for systems to learn user preferences.
• Existing implicit feedback methods are often limited and difficult to interpret.

Purpose of the Study:

• To investigate the utility of eye-tracking as a method for gathering implicit feedback.
• To explore whether eye behavior can elucidate the complexities of user relevance judgments.
• To determine if eye-tracking data can predict user comprehension and interest in content.

Main Methods:

• A user study involving 30 participants reading 18 news articles.
• Collected eye-tracking data (e.g., gaze patterns, fixation durations).
• Compared eye behavior metrics with subjective ratings of content comprehensibility and interest.

Main Results:

• Eye-tracking signals significantly explained variance in user-appraised comprehensibility (49.93%) and interest (30.41%).
• Statistical analysis confirmed the significance of these findings (p < .001).
• Eye behavior provides a rich source of implicit feedback beyond simple accuracy metrics.

Conclusions:

• Eye behavior serves as a powerful, implicit signal for understanding user information needs and engagement.
• Eye-tracking data can enhance personalized information systems by enabling more nuanced adaptation and interaction support.
• This research opens new avenues for leveraging physiological responses in human-computer interaction.