Evaluate data quality for scientific reports

[Peter]

I am using the Tobii 4C eye tracker to move parts of study participants’ digital representations in Unity. In scientific reports using eye tracking, I’m used to reports of data quality reports, including values such as calibration accuracy and calibration precision.
As an experimentator I appreciate the calibration visualizations in both Tobii Eye Tracking Core Software and Tobii Pro Eye Tracker Manager. They are useful to quickly assess similiarity of actual and tracked fixation point.
What I was wondering about is what data I can base my data exclusion decisions on: How precise and accurate are the samples based on the calibration, how much data was lost during recording etc. How do I go about exporting such data quality measures with Tobii’s Unity (Pro) SDK?

Best regards
Peter

[Grant [Tobii]]

Hi @treep, for the Tobii EyeX and C4 range of eye trackers which are targeted for interaction and gaming purposes, we do not currently have a guide for an accuracy and precision report for the Tobii C4 Tracker.

The current industry standard for measuring performance of eye trackers comes from research environments where eye trackers have historically been most common. Metrics are typically made in controlled enviornments where the light conditions do not change, measurements are made very soon after calibration, and the user typically does not move very much. Presenting such metrics for a dev kit intended for day to day consumer usage, gaming etc can easily be misleading, so we have historically chosen not to measure our eye trackers intended for consumer usage in this way.

That being said, we are currently working on providing a detailed report for our Pro SDK users that should satisfy your need for quantifiable accuracy. We expect this to be available relatively soon (2/3 Weeks) so please check back with us at that time.. if it ready before then, I will inform you.

With respect to data lost, you can use the Gaze Data Validity Codes as an easy way to calculate the Percent of Time that Gaze Data is captured during recording but of course one should bear in mind that Gaze Data may not be captured under such circumstances and blinking, looking down at the keyboard or indeed looking away from the screen.

I hope this answers your query and please let us know if there is anything else we can help you with!

[Peter]

This is great to hear! Thanks for your helpful response, Grant.
Are you allowed to disclose the current status of said accuracy quantification integration to the Pro SDK?

[Grant [Tobii]]

Hi @treep, I have some information ready for you and have sent a PM with download details.

[Author]

Posts