Relating Search and Stroop Phenomena

Presented with Stroop color-word stimuli, participants were asked to make one response (“Yes”) to the appearance of the word RED, the color red, or both and another response (“No”) when none of these targets (defined by redness) appeared. The same participants also performed in the standard Stroop task, naming the print color of the same set of color words. Because we were careful to (1) construct stimuli equally discriminable on color and word, and (2) combine color and word such that there was zero correlation over the experimental trials between the two components, no Stroop effect was found. The results of the search task, augmented by Townsend’s double factorial paradigm, showed the underlying system to be minimum time parallel, independent, and unlimited capacity. Collectively, the results support the contextual approach to the Stroop phenomenon by which it is neither robust nor inevitable. A taxonomy for elementary cognitive processes developed by Townsend (1974, 1990; Townsend & Ashby, 1983) has been augmented by a mathematical theory and related experimental methodology (Townsend & Nozawa, 1995; Wenger & Townsend, 2000) to permit strong tests and conclusions regarding the properties of the cognitive system under investigation. Four logically distinct but often interrelated characteristics of information processing are architecture (serial vs parallel, and, within the latter, separate vs coactive), capacity (limited vs unlimited vs supercapacity), stopping rule (self-terminating vs exhaustive), and stochastic dependence (dependence vs independence). The prototypical experimental setup to uncover these properties comprises visual search within the “redundant target design.” Of the set of possible stimuli, some are defined as targets. The observer is to make a certain response if one or more targets appear in a display and a different response if no target is presented. Speeding up on trials with multiple (i.e., redundant) targets typically occurs relative to trials with fewer targets – the redundant target effect. The magnitude and distribution features of the redundancy effect then serve to uncover the noted characteristics of the particular cognitive system. Recently, Townsend and Nozawa (1995) have developed a novel paradigm, the double factorial design, which further aids in model diagnosis. In that design, a second variable such as saliency or intensity is manipulated apart from that of the presence of the target. The two factors are crossed with each other; hence the double factorial design. Given the added factor, strong factorial test of architecture and (in)dependence are possible with the subset of data in which all the targets are present. These tests are independent of the capacity analyses routinely performed on the single vs multiple target trials. The bifurcated analysis permits stronger evidence on the underlying properties of the cognitive system under investigation. In the current study, we applied the double factorial design for an examination of the Stroop phenomenon (Stroop, 1935). The single most popular phenomenon within cognitive psychology, the Stroop effect is the classic example of the human failure to attend selectively: Naming the print color of color words is impaired by the meaning of the words, although reading the words is not similarly hindered by irrelevant print color. Selective attention fails for color, but it is perfect for word. The Stroop effect has accrued theoretical interest because, presumably, the effect is the inescapable outcome of pitting the automatic process of word reading against the less automatic or controlled process of color naming. Because reading is more automatic and speedy than naming, the latter suffers intrusions from involuntary word reading, but reading is not similarly hampered by conflicting print color. Recently, we (Algom, Dekel, & Pansky, 1996; Dishon-Berkovits & Algom, 2000; Melara & Mounts, 1993; Pansky & Algom, 1999 ; Sabri, Melara & Algom, 2001; Shalev & Algom, 2000) have challenged both the alleged robustness of the Stroop phenomenon and its traditional explanations. We demonstrated that the effect is malleable experimentally by several factors of context. By judiciously manipulating these factors, one is able to determine the direction, the magnitude, and, in fact, the very presence of the Stroop effect. Hence, the effect is neither robust nor, indeed, inevitable. These results, in turn, cast doubt on theories positing the automatic activation of the meaning of words. To distinguish between the disparate approaches to the Stroop phenomenon and, hopefully, to aid in theoretical resolution, we decided to apply the double factorial design of visual search to the original Stroop stimuli. The stimulus set comprised the factorial combination of the color words, RED, and GREEN, and the print colors, red, and green. As in the classic Stroop task, a single color word printed in color appeared on a trial. Redness was defined as the target: The participant was to press one key (“Yes”) if either the word RED, the print color red or both were presented; otherwise she or he was to press another key (“No”). The added factor was saliency: We degraded the quality of the font, the quality of the color, or both on some of the trials. In a separate session, the participants also performed in the classic Stroop task with the same stimuli. They named, while timed, the print colors of color words. The Stroop effect was calculated as the difference in performance between congruent (the word naming its print color) and incongruent (conflicting word and color) stimuli.