Visual fixations and level of attentional processing

Abstract

The analysis of visual fixations can provide rich information about the user's attention. We will show that it is possible to distinguish between preattentive scanning and cognitive elaboration using parameters of fixation duration. It is, however, not without difficldty to interpret fixation durations solely on the basis of averages. First of all, fixation durations show a log-normal rather than a normal distribution, with a more or less strong leftward skew. Van Zandt & Ratcliff [13 ] stipulated that frequency distributions contain information that is not always represented by the usual first order statistics. This is palticularly the case when the distribution is multimodal, but also when experimental factors provoke changes that do not affect the whole distribution, which is usually the case with fixation durations. Moreover, as will be shown below, the same mean fixation duration may result from quite different distributions. Furthermore, the mode is very stable, and can only be altered under few conditions. Finally, the distribution of fixation durations is not unimodal. In most of our experiments, a distinct bimodality can be observed with the second mode lying around 60 ms. These fixations are often ignored, possibly because of the low temporal resolution of the most recording devices. We found these express-fixations to be selectively sensitive to task conditions.