Home » Eye Tracking » Framework for Eye Tracking Patterns and Usability Problems: Pt 4

Framework for Eye Tracking Patterns and Usability Problems: Pt 4

Posted on Jul 19, 2010 | 0 comments

Framework for Eye Tracking Patterns and Usability Problems: Pt 4 This week at Eye Tracking Update we’ve been discussing the lack of a concrete, rigorous framework by which eye tracking usability metrics can be interpreted. Typically, interpretations are done by usability experts or in some of the more D.I.Y studies, whomever is doing the testing. We’ve centered much of our discussion around an academic paper published by two researchers, Ehmke and Wilson, from City University, London. In it, they suggest that there needs to be a more standardized model for interpreting eye tracking metrics in usability studies. Today we focus on their method and process.

The researchers conducted an empirical study to collect data from which to analyze correlations between eye tracking patterns and usability problems. It’s more of an exploratory study, and so they write that it was important to identify a diverse set of usability issues rather than collecting a bunch of data for a few.

Participants completed tasks in either one or two websites, and Ehmke and Wilson collected verbal descriptions (Think Aloud processes) and observational data identifying the usability problems the users had. Eye tracking data, of course, was also monitored and the correlations between the two were then analyzed. Nineteen internet-savvy participants were recruited for the job. Each session started with the participant navigating their way through the BBC websites and another called thetrainline.com while using a Tobii eye tracker. Each participant spent around 20 minutes accomplishing a list of tasks on either screen, and they were recorded both visually and with a microphone. Software was also used to record the screen, mouse and keyboard input, and the eye tracker was set to record every 20ms.

The researchers goal was to extract a list of frequent usability problems using the raw data collected such as the visual observations, concurrent protocols, and retrospective protocols. The usability problems were then matched across participants to determine where more than one participant had experienced the problem, and then again matched across protocols to determine is they were unique to each task or site.

The usability problems? They were described in terms of the number of the specific participant, a unique number for each problem, a description of the usability problem (such as “Participant mentions that the holiday page does not contain any useful data for her task “finding information about surfing in South Wales”). Location or URL of the site was also important as problems naturally varied across several pages. Then the problems were divided into two categories and labeled either a “page problem” or a “site problem.”

Overall, they successfully identified 48 unique problems on the BBC site and 44 on thetrainline.com

Identifying Web Usability Problems from Eye-Tracking Data

Framework for Eye Tracking Patterns and Usability Problems: Pt 4

Categories