EMA: Automated eye-movement-driven approach for identification of usability issues

Abstract

The work described in this paper presents an automated, eye movement-driven approach (EMA) that allows for the identification of time intervals in which a user is experiencing difficulties in locating interface components required for completion of a task. Due to the substantial amount of visual search exhibited during these time intervals, this type of the user behavior is referred to as excessive visual search (ES). In this work we propose and evaluate several ES detection algorithms as part of the EMA. Empirical results indicate that it is possible to identify ES with a certain degree of accuracy (51-61% on average), warranting future research that would allow for increased accuracy in ES identification and reduction of misclassification errors. Practical application of EMA should allow the reduction of the amount of time required for manual detection of usability problems present in graphical user interfaces. © 2011 Springer-Verlag.