iMotions has extensive experience in eye tracking tests and we and our clients have tested thousands of images and on tens of thousands of respondents. Some studies are with even 500-1000 respondents.
When testing on advertisements and packaging, we advice to follow a minimum set of requirements for the physical environment and testing protocol and we have summarized some main points in this blog post, which can hopefully guide you in your eye tracking tests. The recommendations can be used both within Scientific (University) research, market research and other eye tracking relevant areas.
The motivation for these requirements is mainly to keep an “all else equal” balance in the data collection process as well as optimizing the data quality, which is essential when eye tracking is used for business decisions on new package designs etc.
The main assumption taken here is that it is a passive test i.e. there is no need for the operator to interact with the respondent during testing.
Underneath you can find the main recommendations, which include a Gessel chamber (please see the picture in this blog)
No visual or auditory disturbance before, during and after test
– The respondent should not be distracted during any part of the testing process, especially during the calibration and presentation of stimuli.
– Before the test, when the respondent receives instructions, it is important that he/she is not distracted by the environment, so full attention can be paid to the message. Also, write down the instruction, so you are sure that all respondents receive exactly the same message.
– During the test, of course, the respondent must focus on what is on the screen. This requirement is met in practice by having a separate (gessel) room for testing or a physical barrier between the respondent and the operator. If you have a separate room, it does not need to be specifically designed for eye tracking. A normal gessel room setting would be optimal. Here you have a one-way mirror glass and you can instruct the respondent through a microphone and at the same time observe their gestures.
– After the test there may be additional instructions, or a questionnaire, where the respondent should also concentrate. Make sure there are no sources of noise, tvs in the background, ticking clocks on the walls. Even remove posters from the walls.
– As eye tracking equipment is operating in the infrared spectrum, it is important to avoid unintentional sources of infrared light that can introduce noise in the measurements. A typical source of unwanted infrared light is sunlight that comes through the windows of the room. Therefore blind the windows.
Office lights in the ceiling can also create problems, either because they provide an additional unwanted infrared source, or because of the brightness. The best light is ambient light. This can be achieved by placing a couple of lamps behind the respondent. The lamps should be directed toward the wall behind the respondent in order to disperse the light.
– Testing in complete darkness is not recommended as this is uncomfortable for the respondent. Furthermore, in complete darkness the light from the monitor will appear much brighter and therefore less comfortable.
The operator should be separated from the respondent (Gessel Chamber Setup).
– The respondent will always be aware about the presence of other people in the room.
– Unless absolutely necessary, i.e. part of testing protocol, the operator should not be in the room during testing or at least be physically separated. Despite clear and firm instructions, respondents can be tempted to communicate during test and ask questions, if he knows the operator is close-by.
Therefore when the test starts, you as an operator, should sit separated from the respondent.
It is important to tell the respondent how long the test will last, as some respondents quickly get impatient or bored. If they know how long the test will be, their motivation can be maintained.
These are some main recommendations that we have been quite successful with and keeping these settings will make your data even better for benchmarking purposes when you want to compare new studies with prior studies.