The framework of dimensional interaction was used to test the hypothesis that the Stroop effect is partially rooted in mismatches in baseline discriminability, with stimulus differences along the word dimension typically exceeding stimulus differences along the color dimension. Subjects made speeded classifications, with either keypresses or vocalizations, of either words or colors. Stroop congruity and Garner interference were measured under conditions in which discriminabilities were (l) matched (Experiments 1 and 4), (2) mismatched in favor of colors (Experiment 2), or (3) mismatched in favor of words (Experiment 3). When matched, colors and words appeared separable, with small interactive effects being reduced or eliminated through practice. When mismatched, asymmetric Stroop and Garner effects emerged, with the more discriminable dimension disrupting classification of the less discriminable dimension. Asymmetric effects were obtained in both response modes, and were not alleviated by practice. We conclude that (1) the Stroop effect is an optional effect, and (2) unequal discriminability causes a mandatory failure of selective attention.In his comprehensive review, MacLeod (1991) noted that, since the seminal paper by Stroop in 1935, over 700 articles have explored some aspect of the phenomenon that Stroop discovered. Stroop's study is especially intriguing because of the peculiar asymmetry that characterized his subjects' performance: When they read color words printed in various colors, irrelevant color had no effect on performance; however, when they named the colors of these words, the irrelevant words hindered performance significantly. This asymmetry is known universally as the Stroop effect. Much is now known about how this effect relates to a host of experimental variables. Some ofthese relations are discussed below. Despite an enormous amount of empirical study, however, the wider theoretical implications of Stroop's work have largely eluded researchers. In the present study, we view the Stroop effect in a different light-namely, that of research and theory on dimensional interaction. This approach permits us to fit the Stroop effect comfortably into the broader framework of what factors cause selective attention to succeed or fail. Our efforts yield new insights into the nature of the Stroop effect, on the one hand, and the nature of selective attention, on the other.
Selective Attention and Dimensional InteractionThe questions of whether and how humans integrate the separate sources of information in their environment have enjoyed considerable research focus within cognitive psychology. One reason is that these issues have great bearing on the nature of many fundamental cognitive processes, including those in perception (see, e.g., Ashby