Using think aloud to boost eye tracking usability research results
There is no question that eye tracking can provide valuable data on visual behavior, including visual interest and attention. On their own, however, the only solid conclusion these data points provide is where a person looked. Any attempt to interpret further meaning from these results using solely the coordinates and durations of the fixations is nothing more than speculation. That doesn’t mean there is no value in using eye tracking for usability or behavioral research; in fact, it has proven quite effective. The key to abstracting value from eye tracking research is to combine the quantitative results with qualitative information. The Think Aloud protocol has shown to be useful for providing researchers with the verbal feedback that makes sense out of the scattered points on a gaze plot.
Think aloud methods are commonly used in usability research to add a qualitative element to the eye tracking results. Think aloud gives subjects the opportunity to explain their behaviors, thought processes, and opinions either while they are using the eye tracker or after the results have been recorded. There are some criticisms about this method, such as the fact that many cognitive processes happen unconsciously, but the general consensus is that the insight it provides is more helpful than none at all.
Concurrent think aloud (CTA) is when the subject verbally describes their actions and thoughts as they do the proscribed tasks. CTA has some limitations that may alter the eye tracking data. Verbalizing tasks may cause the subject to do them slower, since verbal processes are much slower than cognitive processes. In addition, some subjects may forget to speak about their thoughts if they get distracted by a difficult interaction or may make something look more difficult than it is because they are concentrating on trying to speak and navigate at the same time.
The most effective method of think aloud is retroactive think aloud (RTA). This method has subjects describe their experience after the tasks have been completed. This can be done in several ways, either an un-cued verbal recount of their actions and thoughts or a video or image-cued narration. Critics of this method say that it relies on long-term memory which is often unreliable, and it also leaves open the possibility of omitted or fabricated information. Research conducted by Tobii concluded that RTA outcomes using a video playback with a gaze plot overlay resulted in the most detailed comments and more frequent mention of usability problems than the un-cued method.
No matter which think aloud method is used, it is apparent that having qualitative input to back up the quantitative data is valuable. The absolute accuracy is up for debate; however, if the goal of your usability study is to identify problems and collect feedback from users, then eye tracking combined with think aloud is a good way to go.