Despite numerous efforts in developing glare indices through human assessment studies, predicting visual comfort in indoor environments still poses important challenges in design. A major limitation in discomfort glare indices is that they all ignore its dependencies on view direction. In this study we adopted eye-tracking methods in a series of human assessment experiments in order to record actual visual response when experiencing discomfort glare. We set up an experiment where the view directions distributions were monitored as the participants were working in a side-lit office with three different task-supports - monitor, paper and phone - on a standardized office task sequence. The participants were allocated randomly to two groups where they were exposed to two different views from the window. The results show that the “view outside the window” is the main determinant of view direction bias whenever the participant is not focused on any cognitive or visual office task procedure.