Webcams can be used to study eye movements during reading

Eye movements have revealed a wealth of information about how readers process written language. Modern eye trackers can record eye movements in the lab with high accuracy. However, such equipment is expensive, and participant testing is labour-intensive. Kaduk et al. (Behav Res Methods 56(5):5002-5022, 2023) recently reported that webcam eye tracking may be a viable alternative, showing that webcams achieved correlations with laboratory eye trackers of ~0.83 during a free-viewing task. We conducted three experiments to test whether webcams can be used to study eye movements during reading. Participants read short sentences with a target word frequency manipulation while their eye movements were recorded with both an EyeLink system and a webcam in the laboratory (co-registration, Experiments 1 and 2) or online with their own webcam (online study, Experiment 3). In the lab, the webcam-based system recorded raw gaze positions that were highly correlated with the EyeLink (median r = 0.805), replicating Kaduk et al.'s results. We also found evidence for word frequency effects in the webcam data, but there was more noise than in the EyeLink data. In the online study, data quality was much poorer, but we still found reliable effects that were consistent with those observed in the laboratory. In summary, webcam-based eye tracking may be viable for studying reading, especially under laboratory conditions. However, obtaining useful data may require collecting many more observations when using webcams than with high-accuracy eye trackers.