While eye tracking is gaining strong momentum in various research fields all over the world, its increasing popularity has also made poor eye tracking practices more common place. An article on ITWeb discusses the bad habits that are beginning to slip into eye tracking studies. Researchers who do not completely understand eye tracking are opting for these simpler, less expensive methods without realizing the unreliability of their results. Aga Bojko, the associate director at User Centric, pointed out three major mistakes researchers make.
Virtual Eye Trackers
Some companies offer computer-simulated eye tracking studies and claim that the results are 75%-90% accurate with user-based studies. Researchers can upload images and a computer generates a heat map prediction of where people would look first at the image. Unfortunately, no computer can ever truly emulate human reaction. In a study Bojko performed comparing these computer-generated results with real-world results, the heat maps generated were very different.
People often assume that eye tracking studies do not require a large number of participants. While it may be possible to get decent results with fewer people, there is no set maximum participant number to get accurate results. Like any other scientific test, these numbers rely not on the equipment doing the measuring, but on the variables and desired outcomes of the specific test. Unfortunately, many researchers opt for the smaller participant count due to budget constraints.
Webcam Eye Tracking
A new big thing in eye tracking is webcam eye tracking studies. This is a cost effective alternative to legitimate lab testing. It does not require expensive equipment or special lighting. Unfortunately, this also means webcam eye tracking studies are extremely inaccurate. Without being able to control the environment, a researcher cannot be sure the tracker is accurately calibrated. Additionally, there is no way to monitor a webcam participant to be sure they are following directions correctly.
Eye tracking is an extraordinarily valuable research tool when used correctly. When used incorrectly, though, researchers will yield inaccurate or misleading data. In the end, researching the correct equipment and experiment specifications prior to performing a study will help you perform a sounder, more stable test.