Psychonomic Bulletin & Review 2000, 7 (3), 499-503
Stroop interference and negative priming: Problems with inferences from null results PALOMA MARI-BEFFA University of Wales, Bangor, Wales ANGELES F. ESTEVEZ University ofAlmeria, Almeria, Spain
and SHAIDANZIGER University of Wales, Bangor, Wales
The Stroop color-naming effect has often been taken as evidence for the automaticity of word processing (MacLeod, 1991). Recently, Besner, Stolz, and Boutilier (1997) reported that coloring a single letter instead of the whole word eliminated the Stroop effect. From this finding, they concluded that word processing could not be purely automatic, since it can be prevented. Weasked whether the elimination of the Stroop effect is sufficient evidence for concluding that the word is not processed. Combining Besner et al.'s manipulation with a negative-primingprocedure, we found intact negative priming from the prime color word in the absence of a Stroop effect. This result clearly indicates that the meaning of the prime word was processed. The findings highlight the importance of using converging methods to evaluate lack of processing. In one of the most cited studies in experimental psychology, Stroop (1935) reported that participants take longer to name the color of an incongruent stimulus, such as the word BLUE printed in green ink, than to name the color of a neutral stimulus, such as a green square. The Stroop phenomenon is cited as strong evidence that word reading is automatic and cannot be prevented even when it is irrelevant for the task at hand (MacLeod, 1991). Recently, Besner, Stolz, and Boutilier (1997) compared the basic Stroop task, in which the whole word is colored, to a Stroop task in which only one of the letters is colored. They found that in the single-letter condition, the Stroop effect was either reduced (Experiment I) or eliminated (Experiment 2), and they concluded that this result demonstrates the nonautomatic nature of word reading, stating that Nonetheless, as the present experiments show, at least some of these putatively automatic reading processes can be interfered with so that the Stroop effect is reduced ... , and even eliminated .... The ability to produce this outcome would seem to fly in the face of any automaticity ac-
P.M.-B. was supported by a postdoctoral grant from Ministerio de Educacion y Cultura (Spain). We also thank Steve Tipper, Derek Besner, and Bruce Milliken for their helpful comments and suggestions. This work represents equal and shared contributions by the authors. Correspondence concerning this article should be sent to P. Mari-Beffa or to S. Danziger, School of Psychology, University of Wales, Bangor, Gwynedd LL57 2DG, Wales (e-mail: pbeffa@bangor. ac.uk,
[email protected], or
[email protected]).
count in which specified processes cannot be prevented from being initiated. (Besner et al., 1997, p. 224)
In the present study, we asked whether the absence of a Stroop effect necessarily implies a lack of Stroop word processing. The logic behind our concern with Besner et al.'s (1997) conclusion was that although one can argue for word processing if a Stroop effect is found, one cannot validly infer a lack of word processing from the absence of a Stroop effect. In essence Besner et al. base their argument on a lack of an effect, and one must always be cautious when making such inferences. A similar point was made by Tipper and colleagues (Allport, Tipper, & Chmiel, 1985; Driver & Tipper, 1989; Tipper, 1984) with respect to claims that an absence of distractor interference in visual selection tasks necessarily means that distractors are not processed (Francolini & Egeth, 1980). Driver and Tipper, for example, showed that distractors that did not interfere with responses to semantically related targets (and hence might be considered not to have been processed) nevertheless produced subsequent negative priming (see Fox, 1995, for a review ofthe negative-priming literature). That is, on a subsequent probe trials, reaction times (RTs) were slower when a target was related to the previous distractor than when it was not related.
THE PRESENT STUDY We combined Besner et al.'s (1997) version of the Stroop task with a negative-priming procedure. Negative priming served as a second convergent measure of word
499
Copyright 2000 Psychonomic Society, Inc.
500
MARf-BEFFA, ESTEVEZ, AND DANZIGER Table 1 Data for Prime Responses in Experiments 1 and 2: Means of Median Reaction Times (M, in Milliseconds) and Error Rates (%E) as a Function of Type of Printing (All vs, Single Letter Colored) and Lexicality (Word vs. Nonword) and Stroop Effects From Words and Nonwords Conditions Word Nonword XXX Stroop Effect Typeof Printing M %E M %E M %E Word-Nonword Nonword-XXX Experiment I -3 All letters 750 1.7 753 1.7 10 Single letter 809 3.5 799 2.0 Experiment 2 742 1.1 734 1.5 717 2.1 17 8 All letters 790 2.1 780 2.3 762 2.3 IO 18 Single letter
processing. The probe ink color could either have the same meaning as the prime word (related condition) or differ in meaning from the prime word (unrelated condition). If the elimination of Stroop interference reflects the prevention of word processing, there should be no negative priming. If, however, the elimination of the Stroop effect is not due to lack of word processing, but to some other factor, negative priming should be observed.
EXPERIMENT 1 Method
Participants. Twenty-nine students at the University of Almeria participated in the experiment in partial fulfilment of a course requirement. All the participants were right-handed and reported normal or corrected-to-normal visual acuity. The participants were not aware of the experiment's purpose. Stimuli and Materials. The stimuli were the four Spanish color words, raja, azul, verde, and blanco (red, blue, green, and white), and their derived nonwords, ronu, azej, velti, and blasde. Following Besner et al. (1997), the nonwords were chosen to match the words on length and on the first two letters. The stimuli were equally likely to be colored in red (ASCII code 4), blue (ASCII code 9), green (ASCII code 2), or white (ASCII code 15) and appeared on a black background. In the all-letter-colored condition, all of the letters appeared in the same color. In the single-letter-colored condition, one letter was colored, and the remaining letters appeared in gray (ASCII code 8). The colored letter occurred with equallikeIihood at any position within the letter string. The stimuli always appeared in a color different from that indicated by the words meaning (for the word) or by the two first letters (for the nonwords). The words served as the incongruent stimuli, and the nonwords as neutrals. At a viewing distance of 60 ern, each character was 0.480 of visual angle high and 0.380 wide. Stimuli appeared at fixation on a color monitor controlled by a Pentium-compatible PC with a VGA card. The computer controlled stimulus presentation and response acquisition. Design and Procedure. On each trial, a prime stimulus was followed by a probe stimulus. The participants were instructed to report the color of the letter strings as quickly and as accurately as possible while ignoring the letter string. They responded to the four target colors by pressing one of four keys on the keyboard (>,
