Lab Report Sample: The Stroop Effect: A Classroom Replication

Model Answer

Abstract

This report describes a classroom replication of the classic Stroop (1935) colour-word interference effect in a sample of UK undergraduate psychology students. Twenty-eight participants completed a computerised Stroop task in which they named the ink colour of words presented in congruent, incongruent and neutral conditions. Mean reaction time was slowest in the incongruent condition (M = 789 ms, SD = 103) compared with the neutral (M = 648 ms, SD = 81) and congruent (M = 612 ms, SD = 74) conditions. A one-way repeated-measures ANOVA confirmed a significant effect of condition on reaction time, F(2, 54) = 54.62, p < .001, partial η² = .67, and error rates showed the same pattern. These findings closely replicate the original Stroop effect and support automaticity accounts of word reading.

Introduction

The Stroop effect is one of the most robust and widely replicated phenomena in cognitive psychology (MacLeod, 1991). In the original demonstration, Stroop (1935) found that participants were slower to name the ink colour of a word when that word itself named a different colour (for example, the word ‘RED’ printed in blue ink) than when the word and ink colour matched, or when a neutral, non-colour stimulus was presented. This interference effect has since been replicated many hundreds of times across languages, age groups and testing methods, and is frequently used as a marker of selective attention and cognitive control (MacLeod and MacDonald, 2000).

Several theoretical accounts have been proposed to explain why word reading interferes with colour naming rather than the reverse. The most widely accepted, automaticity theory, holds that word reading is a highly practised, automatic process that occurs rapidly and without intention, whereas colour naming is a slower, more controlled process; when the two processes generate conflicting responses, the automatic word-reading response intrudes on the slower colour-naming response and must be actively suppressed, producing a measurable delay (MacLeod, 1991). An alternative, speed-of-processing account suggests that word identification is simply completed faster than colour identification, so that by the time colour information is available, word information has already begun to influence the response (Cohen, Dunbar and McClelland, 1990). Connectionist models have since simulated this interference using networks in which word-reading pathways are more strongly weighted than colour-naming pathways as a result of extensive practice (Cohen, Dunbar and McClelland, 1990).

Attentional accounts add a further dimension, proposing that Stroop interference reflects a failure to fully suppress irrelevant information, implicating executive control and working memory capacity; individuals with lower working memory capacity tend to show larger interference effects (Kane and Engle, 2003). Despite the depth of this literature, the Stroop effect remains a valuable and accessible task for classroom replication, since it produces a large, reliable effect with a relatively small sample and simple apparatus. The present study aimed to replicate the classic Stroop interference effect in a sample of UK undergraduate students using a computerised, within-subjects design, comparing reaction times across congruent, incongruent and neutral conditions. It was hypothesised that reaction times would be significantly slower in the incongruent condition than in either the congruent or neutral conditions.

Method

Design. This study used a within-subjects (repeated-measures) design. The independent variable was word-colour congruency, with three levels: congruent (word and ink colour matched), incongruent (word and ink colour conflicted) and neutral (a string of X characters presented in colour, with no word content). The dependent variables were mean reaction time, in milliseconds, and percentage of naming errors.

Participants. Twenty-eight undergraduate psychology students (21 female, 7 male; M age = 20.1 years, SD = 1.4) took part as part of a practical research methods class. Participants were recruited via an opportunity sample from a single UK university cohort. All reported normal or corrected-to-normal vision and no colour vision deficiency. Participants gave informed consent and were free to withdraw at any point; the study received departmental ethical approval.

Apparatus. The Stroop task was programmed and presented using PsychoPy (Peirce et al., 2019) on identical laboratory computers with 24-inch monitors. Stimuli were the words RED, GREEN, BLUE and YELLOW, presented in one of four ink colours (red, green, blue, yellow) on a white background. In the congruent condition, word and ink colour always matched; in the incongruent condition, word and ink colour always conflicted; in the neutral condition, the string ‘XXXX’ was presented in one of the four colours. Participants responded by pressing one of four labelled keys (R, G, B, Y) corresponding to the ink colour, using their dominant hand. Each condition comprised 40 trials, giving 120 experimental trials in total, preceded by 10 practice trials.

Procedure. After providing consent, participants received standardised on-screen instructions emphasising that they should respond to the ink colour and ignore the word’s meaning, and completed 10 practice trials with feedback. The three test blocks (congruent, incongruent, neutral) were then presented in a counterbalanced order across participants using a Latin square, to control for order and practice effects. Each trial began with a central fixation cross for 500 ms, followed by the stimulus, which remained on screen until a response was made or 3,000 ms had elapsed. Reaction time and response accuracy were recorded automatically. On completion, participants were debriefed and thanked.

Analysis. Mean reaction time for correct trials only was calculated for each participant in each condition. Data were screened for outliers; responses below 200 ms or above 2,500 ms were excluded as anticipatory or lapsed responses, affecting fewer than 1 per cent of trials. A one-way repeated-measures ANOVA was conducted on mean reaction time, with condition (congruent, neutral, incongruent) as the within-subjects factor, followed by Bonferroni-corrected pairwise comparisons.

Results

Table 1 shows mean reaction time and error rate for each condition. As predicted, reaction times were slowest in the incongruent condition and fastest in the congruent condition, with the neutral condition falling between the two. Error rates followed the same pattern, indicating no evidence of a speed-accuracy trade-off.

Table 1. Mean reaction time and error rate by condition (N = 28).

A one-way repeated-measures ANOVA showed a significant main effect of condition on reaction time, F(2, 54) = 54.62, p < .001, partial η² = .67, indicating that congruency accounted for a substantial proportion of variance in naming speed. Bonferroni-corrected pairwise comparisons showed that the incongruent condition was significantly slower than both the congruent condition, t(27) = 9.96, p < .001, and the neutral condition, t(27) = 8.10, p < .001, and that the neutral condition was significantly slower than the congruent condition, t(27) = 3.45, p = .002.

Table 2. One-way repeated-measures ANOVA summary table for reaction time by condition.

The size of the classic Stroop interference effect can be expressed as the difference between mean incongruent and congruent reaction times: 789 ms − 612 ms = 177 ms. Using the within-subjects difference scores for each participant (SD of difference scores = 94 ms, n = 28), the standard error of the mean difference was calculated as SE = 94 / √28 = 94 / 5.29 = 17.76 ms. Dividing the mean difference by this standard error gives t = 177 / 17.76 = 9.96, matching the paired-samples test reported above and confirming that the interference effect was both large and highly reliable across participants.

Discussion

The results of this classroom replication closely mirror the classic Stroop (1935) effect: participants were substantially slower, and made more errors, when naming the ink colour of colour words whose meaning conflicted with that colour, compared with congruent or neutral stimuli. The 177 ms interference effect obtained here falls within the range typically reported in the literature (MacLeod, 1991), and the accompanying rise in error rate in the incongruent condition indicates that the slowdown reflects genuine interference in processing rather than a simple trade-off between speed and accuracy in the opposite direction.

These findings are consistent with automaticity accounts of the Stroop effect, which propose that word reading is a highly practised, largely automatic process that intrudes upon the slower, more effortful process of colour naming (MacLeod, 1991). Because reading is over-learned to the point of being difficult to suppress voluntarily, participants could not simply ignore the word and attend only to its ink colour; instead, the conflicting word-based response had to be actively inhibited before the correct colour-based response could be produced, adding measurable time to each incongruent trial. The pattern is also compatible with connectionist accounts in which word-reading pathways carry stronger, faster-conducting weights than colour-naming pathways as a consequence of extensive practice with reading relative to colour naming (Cohen, Dunbar and McClelland, 1990), producing systematic interference whenever the two pathways generate competing outputs.

The results also speak to individual differences in cognitive control. Although this study did not measure working memory capacity directly, the sizeable variability in interference scores across participants (SD = 94 ms) is consistent with prior evidence that the magnitude of Stroop interference varies with executive control resources, with individuals lower in working memory capacity typically showing larger interference costs (Kane and Engle, 2003). A useful extension of this classroom study would be to correlate interference scores with an independent measure of working memory or attentional control, to establish whether this relationship holds within a UK undergraduate sample.

Several limitations should be acknowledged. The sample was a relatively small, single-site opportunity sample of psychology students, who may differ from the general population in reading fluency, attentional control or familiarity with laboratory-style tasks, limiting generalisability. The computerised, keyboard-response version of the task used here also differs somewhat from the original card-based, vocal-response Stroop paradigm, and response modality has been shown to influence the size of the interference effect in some studies (Besner, Stolz and Boutilier, 1997); direct comparison with the original 1935 results should therefore be made cautiously. Finally, because the design was purely within-subjects with respect to condition, no causal claims can be made about the specific cognitive mechanisms underlying the interference effect beyond what the experimental manipulation itself supports.

Future replications could usefully incorporate eye-tracking or electrophysiological measures to examine the time-course of interference more precisely, or could manipulate response modality directly to test whether keyboard- and voice-based Stroop tasks produce comparable effects. Extending the design to include a measure of bilingual language experience would also be of interest, given evidence reviewed by MacLeod and MacDonald (2000) that bilingual experience can modulate the size of Stroop interference.

Beyond its theoretical interest, the Stroop task has practical applications that this classroom replication helps to illustrate. Variants of the task are used clinically as brief, low-cost screening measures of selective attention and cognitive control, including in the assessment of attention-deficit conditions and in some neuropsychological batteries following brain injury, precisely because the interference effect is large, reliable and sensitive to disruptions in inhibitory control (Kane and Engle, 2003). The present findings, obtained with a modest undergraduate sample and simple apparatus, reinforce why the task has remained a staple of both research and applied cognitive assessment for close to a century.

Conclusion

This classroom study successfully replicated the classic Stroop interference effect in a sample of 28 UK undergraduate psychology students, with significantly slower and less accurate colour naming in the incongruent condition than in the congruent or neutral conditions. The size and reliability of the effect obtained here, together with the pattern of results across conditions, closely match the wider literature and support automaticity-based explanations of Stroop interference. Beyond its theoretical value, the study demonstrates that a small, undergraduate-run replication can reliably reproduce a foundational finding in cognitive psychology using modest sample sizes and simple computerised apparatus, making the Stroop task a practical and pedagogically valuable exercise for research methods teaching as well as a continuing tool for investigating attention and cognitive control.

References

• Besner, D., Stolz, J.A. and Boutilier, C. (1997) ‘The Stroop effect and the myth of automaticity’, Psychonomic Bulletin & Review, 4(2), pp. 221–225.
• Cohen, J.D., Dunbar, K. and McClelland, J.L. (1990) ‘On the control of automatic processes: A parallel distributed processing account of the Stroop effect’, Psychological Review, 97(3), pp. 332–361.
• Kane, M.J. and Engle, R.W. (2003) ‘Working-memory capacity and the control of attention: The contributions of goal neglect, response competition, and task set to Stroop interference’, Journal of Experimental Psychology: General, 132(1), pp. 47–70.
• MacLeod, C.M. (1991) ‘Half a century of research on the Stroop effect: An integrative review’, Psychological Bulletin, 109(2), pp. 163–203.
• MacLeod, C.M. and MacDonald, P.A. (2000) ‘Interdimensional interference in the Stroop effect: Uncovering the cognitive and neural anatomy of attention’, Trends in Cognitive Sciences, 4(10), pp. 383–391.
• Melara, R.D. and Algom, D. (2003) ‘Driven by information: A tectonic theory of Stroop effects’, Psychological Review, 110(3), pp. 422–471.
• Peirce, J., Gray, J.R., Simpson, S., MacAskill, M., Höchenberger, R., Sogo, H., Kastman, E. and Lindeløv, J.K. (2019) ‘PsychoPy2: Experiments in behavior made easy’, Behavior Research Methods, 51(1), pp. 195–203.
• Stroop, J.R. (1935) ‘Studies of interference in serial verbal reactions’, Journal of Experimental Psychology, 18(6), pp. 643–662.