Answers are grounded in this paper and cite their sources.
Eye tracking is a key method in educational sciences, for example, in studying teachers' professional vision. However, currently used infrared (IR) eye trackers confine research to small laboratory samples. Webcam-based eye tracking could enable large online studies, but its data quality for educational contexts such as classroom video has to be empirically determined. Seventy-nine participants viewed five classroom video vignettes twice, with a webcam-based system (Labvanced) and an IR system (Tobii Pro Spectrum). Webcam recordings took place in the laboratory (n = 45) or at home (n = 34). Webcam precision was markedly worse than IR precision (laboratory medians within fixations: 0.37° vs. 0.12°; dz = 2.97). With identical processing, fixation counts, fixation durations, and saccade amplitudes were not equivalent across systems. A more tolerant webcam threshold equalized fixation counts but not durations, and a noise-robust algorithm reversed the duration difference, so these measures reflect noise rather than viewing behavior. After offset correction, dwell-time distributions on coarse grids agreed across systems (r = .73 for 4 × 4 cells). It rested on regions all viewers attended to; individual deviations from this pattern were reliable within each system but did not correspond across systems. At home, angular precision did not decrease, but the signal was noisier relative to the stimulus, and usability ratings were lower and more variable. Webcam-based eye tracking can thus describe how attention is distributed across dynamic videos at a coarse scale, but neither individual differences in dwell time nor fixation-based measures are currently interchangeable with IR data.
No takes yet. Share an insight, caveat, or question.
Kosel et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: