Towards detection of usability issues by measuring emotions

Abstract

User Experience is one of the most important criteria when designing and testing user interfaces with emotions as its essential element. To assess, how emotions could be used for automatic detection of usability issues, we carried out a user study with a website which included intentionally inserted usability issues. We classified valence of emotions, i.e., negative vs. positive ones based on data from electroencephalography (EEG) and facial expressions recognition. The study results confirmed that usability issues cause negative emotional response of the user and that presence of a negative emotion is a good predictor of a usability issue presence. When detecting negative and positive emotional states from the acquired dataset, we achieved the accuracy of 94% for samples with seconds granularity and 70% for the task granularity.