Parallel attentive visual search: Testing of serial processing assumption during visual search and implications to visual search models

Kyongje Sung, Purdue University


The assumption of serial attentive processing during visual search (Treisman & Sato,1990; Wolfe,1994) was tested in terms of interaction contrast of cumulative distribution functions of the response times, which allows us to test the serial and parallel processing systems more directly. In the current paper, a similar logic of response time analysis in Schweickert (1978) was adapted to investigate the effects of two experimental factors, which are intended to increase the number of possible candidates for the target and therefore increase the overall response times. In experiments, subjects were asked to find a known target among various distracters. In some trials, two existing distracters were independently replaced by another two new distracters which were more similar to the target than the replaced ones. In experiment I and II (fixed set-size of four), in which the tasks were conjunction search and supposed to have significant set-size effect, patterns of interactions of two independent substitutions of distracters supported parallel processing rather than serial processing of the stimuli. And serial processing was only supported in experiment III in which the task produced a huge set-size effect although the same number of stimuli (four) as in experiments I and II were presented. Parallel processing of stimuli was also confirmed even with moderate set size (8 and 12 stimuli). Overall, the results support Bundesen's (1990) claim that an extreme set-size effect reflects serial processing of stimuli. However, previous parallel models do require some modification to be able to explain the current results. A possible alternative based on previous model (Bundesen,1990) is proposed.




Schweickert, Purdue University.

Subject Area

Psychology|Experiments|Cognitive therapy

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server