Fig. 2.

Characteristics of eye-tracking data quality. A Horizontal gaze position (in Fick, 1854, coordinates, see Haslwanter (1995)) of the right eye as a function of time. The gaze position was recorded from an adult participant with an EyeLink 1000 by Hooge et al., (2015). Call-outs indicate the relatively precise gaze-position signal (compared with panel B). B Horizontal gaze position in Fick coordinates of the right eye as a function of time. The gaze position was recorded from an infant participant with the Tobii TX300 by Hessels et al., (2016). Call-outs indicate the relatively imprecise gaze-position signal (compared with panel A), short gaps in the gaze-position signal (data loss), and an extreme gaze position reported by the eye tracker. The extreme gaze position is interesting because it can be considered an aspect of eye-tracking data quality not captured in the measures accuracy, precision, or data loss. C, D Gaze position signals (black dots) in a 2D representation, i.e. as if on a screen. Gaze position signals were recorded from adult participants by Hooge et al., (2019). Gaze position samples with high velocity were removed such that saccades are not visible. Orange markers represent validation targets. They are positioned to illustrate good/poor accuracy and do not correspond to the location of the actual validation targets in the experiment by Hooge et al., (2019). Call-outs indicate validation targets with corresponding precise and accurate, precise and inaccurate, imprecise and accurate, and imprecise and inaccurate gaze position signals, respectively. Note that the qualifications ‘precise’, ‘imprecise’, ‘accurate’, and ‘inaccurate’ are relative here and are often quantified