Eye tracking can be split into two distinct categories; screen-based eye tracking (also called stationary or desktop eye tracking) and mobile eye tracking.
In a stationary setup, the eye tracker is positioned near the object to be tracked, usually, a screen (a monitor or projector), and the respondents will be placed in a stationary position in front of the screen-based eye tracking system.
For observations of two-dimensional structures, this setup is particularly practical and precise. In a mobile setup where respondents have to move around to capture things from different angles, however, eye movements are tracked with eye-tracking glasses that are fitted near the eyes of the respondents.
This setup is also known as a head-mounted eye tracking and is especially helpful when the object to be observed has a three-dimensional structure, or respondents need to be able to move freely as in the cases for product testing or shopping behavior studies.
Precision and accuracy differences
Precision and accuracy are critical in research using eye movements. The outcome highly depends on the tracking accuracy of the device, so using a low-quality system may be detrimental when extracting high precision data.
A common misconception is that all screen-based eye trackers have a higher level of precision and accuracy than eye tracking glasses.
The precision depends a lot on the type of research the tracker is used for. Eye tracking glasses are fixed to the eye, where no head movement can interfere with data capture which is great for research where the respondents have to move around (e.g. in-store studies).
A screen-based eye tracker has only a certain window in that the head can move without losing connection and with its accuracy and precision.
Eye tracking glasses, on the other hand, compensate for movements if calibrated properly and therefore can be very precise.
However, researchers must take into account the intrusive nature of such hardware when designing an experiment. Therefore, a screen-based tracker might be better for stationary setups on a screen (e.g. website testing).
Two-dimensional vs three-dimensional stimuli
Most screen-based eye trackers work on the basic assumption that the object is two-dimensional. Screen-based eye tracking is also possible for three-dimensional objects, however, only to a certain degree. It is most useful when the respondents are exposed to stimuli on a screen as well as to some physical objects that are two-dimensional such as smartphones and tablets.
Using a screen-based eye tracking system for three-dimensions is a more advanced and difficult setup. Furthermore, experiment respondents are required to sit still while being measured.
The need for respondents to remain motionless can be eliminated by eye tracking glasses. Since the tracker is fixated close to the eye, respondents can move their head around freely, move to the side or move their entire body around.
This setup also enables the researcher to capture the entire scene that a respondent is looking at. Eye tracking glasses are therefore ideal for eye tracking in a physical environment such as a store or interaction with physical objects such as a package.
Situations suitable for the different setups
Generally, screen-based eye tracking is mostly recommended for screen-based stimuli, meaning anything that can be shown on a screen or computer monitor. This includes stimuli like still images, videos, computer games, websites and much more.
However, it also works well for books, magazines, small shelves, etc. Eye tracking glasses work best for three-dimensional environments (e.g. in a shop) or when the respondent needs to move (movability element). In limited cases, screen-based eye tracking could also be used with an environment like a shelf, but it has the limitation that the respondents need to stand still and within a specified proximity so that their eyes can be captured correctly.
In limited cases, screen-based eye tracking could also be used with an environment like a shelf, but it has the limitation that the respondents need to stand still and within a specified proximity so that their eyes can be captured correctly.