Why some search tasks take longer than others: Using eye movements to redefine reaction times

Instead of simply describing search in terms of the time needed to detect a target, this behavior might also be understood as an interaction between two underlying processes: the number of search movements preceding a judgment (the variable number model) and the time taken to initiate these movements (the variable duration model). We introduce these models to address a specific question: why do some search tasks take longer than others?Search times may take longer in a difficult task due to either an increased number of movements or because of longer durations between search movements. Since a reaction time measure lacks sufficient resolution to answer this question, the oculomotor variables of saccade number and fixation duration were used to estimate how these processes contribute to search behavior. Subject's eye movements were recorded as they searched for a target among either 4 or 16 distracters in an easy “parallel” task or a harder “serial” task. The results showed more eye movements and longer initial saccade latencies in the serial task relative to the parallel. Display size differences were also observed for both oculomotor variables in the parallel task, but only for initial latency in the serial condition. These findings provide partial support for both variable number and duration interpretations of search. They also suggest that eye movements may be a useful means in which to study the spatio-temporal dynamics of search behavior.