Combining automation and expertise: A semi-automated approach to correcting eye-tracking data in reading tasks

Area of Science:

• Cognitive Science
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Eye-tracking during reading is crucial for understanding cognitive processes.
• Data inaccuracies due to fixation drift can invalidate recordings.
• Manual correction is laborious and subjective; automated methods lack precision.

Purpose of the Study:

• Introduce Fix8, an open-source GUI tool for semi-automated eye-tracking data correction in reading tasks.
• Evaluate the efficiency and accuracy of Fix8 compared to manual correction.
• Assess user workload and subjective experience with the Fix8 tool.

Main Methods:

• Developed Fix8, a novel semi-automated algorithm for collaborative eye-tracking data correction.
• Conducted a usability study (N=14) comparing Fix8 to manual correction.
• Measured correction time, accuracy, subjective workload (NASA-TLX), and user preferences.

Main Results:

• Fix8 demonstrated a 44% average time saving compared to manual correction.
• The semi-automated approach maintained accuracy comparable to manual correction.
• Users preferred Fix8, reporting lower workload and higher perceived performance.

Conclusions:

• Fix8 offers an efficient and accurate solution for correcting eye-tracking data in reading studies.
• The collaborative approach enhances usability and reduces researcher burden.
• Fix8 provides additional features for eye-movement analysis and data management.