FAQ: What is the difference between Drift Check and Drift Correction?

[Sam]

The Drift Correction and Drift Check procedure are very similar; the participant fixates a single target, then either the participant or more typically the experimenter presses a button or key and the system calculates the deviation of the recorded gaze position of the participant and the drift correction target. The difference between the two operations is what the system does with that information subsequently.

• For a Drift Check, the system simply checks if the participant's gaze position is < 2° of visual angle from the presented target. If so, the participant may proceed with the experiment, if not the participant must try again.
  
• For a Drift Correction, the system attempts to apply a linear correction to the calibration model to compensate for this deviation.
  

For the EyeLink I and II systems, an actual Drift Correction is applied by default - the fixation error is used to shift the calibration map. This linear adjustment often improves the overall accuracy for upcoming recordings because the drift is usually caused by slippage of headband.

For the EyeLink 1000, EyeLink 1000 Plus and EyeLink Portable Duo systems, performing an actual Drift Correction and shifting the calibration map generally does not improve gaze accuracy (in fact, it can actually reduce the accuracy of your tracking if the correction was applied when the participant was actually not properly fixating). Therefore, the default behavior in the Pupil-CR mode in these systems is to simply report the calculated fixation error without altering the calibration map in any way (i.e. a Drift Check).

If the participant repeatedly fails the Drift Check, it's best to press the "ESC" key to enter Camera Setup and recalibrate the participant. Once you have recalibrated and validated, pressing "Output / Record" on the Host PC will take you back to the same Drift Check where you left off without disrupting the course of your experiment.