Effects of Attentional Interference on Reaction Times

To investigate the effects of attentional interference on reaction times in a replication of Stroop’s (1935) study.

Abstract

The Stroop effect found by Stroop (1935) is considered an extremely robust finding, but modern research has found a reverse effect. The aim of this experiment was to replicate this study to investigate whether the same results are found, and to test the statistical power of the findings. 493 undergraduate psychology students took part in a within-subject experiment using a computer simulated version of the Stroop task. Each participant took part in 5 practice trials and 128 experimental trials, whereby they had to select the appropriate colour on the keyboard which corresponded to the colour of the word presented. The experiment was measuring the reaction time to elicit a response in each condition; congruent, incongruent and neutral. The results supported the experimental hypothesis of ‘there will be a significant slowing in reaction time in the incongruent condition than in the congruent condition’, and supported the results found in the Stroop (1935) study. This has provided evidence to support the robust findings already found as a large effect size and power were also calculated. This research could have implications in improving the quality of life of people with schizophrenia by helping understand the cause of the effect and what can be done to therefore maintain the symptoms.

Introduction

Studies investigating attentional interference have been replicated many times following the initial experiment carried out by Stroop (1935). In this study, Stroop investigated the effect that attentional interference had on reaction time in tasks involving different stimuli, in which attentional interference occurred between one stimulus on another within processing. Mackinnon, Geiselman & Woodward (1985) have defined attentional interference as ‘the difference between the performance times between two kinds of stimuli’ due to different perceptual cognitive processes happening simultaneously. In the Stroop (1935) study, attentional interference was measured using congruent and incongruent colour-word pairings. The results concluded that the reaction time in the congruent condition was faster than the incongruent condition. These findings are extremely robust and so replications have been carried out not only to test the statistical power of the results but also to attempt to explain the higher order cognitive processes behind the Stroop effect.

Different explanations of the existence of the Stroop effect have been established. Brown (1915) found that there was a difference in the speed of processing between reading words and naming colours which was concluded as being due to the association process for simply naming objects being different to that of reading printed words. From this finding it can be suggested that naming colours and reading words require different neural pathways. In an incongruent trial, attentional interference between these pathways could have occurred as the pathways were activated simultaneously which led to a weaker response in naming the colour, leading to an increase in the reaction time taken to elicit the response. This is widely regarded as the parallel distribution processing hypothesis which has helped understand more about the cognitive processing involved in attentional interference. However, an alternative explanation of the Stroop effect has been established by Cohen, Dunbar and McClelland (1990) who found that participants are slower at naming colours than they are for reading words, which has also been found by both Stroop (1935) and Brown (1915) but instead of attributing this effect as a difference in neural pathways, Cohen et al (1990) assumed that this effect was due to information about colours arriving at the response stage of processing slower than that of information about words. This explanation has been further defined as the speed processing hypothesis which states that the processing for words is faster than that for colours. This can be linked to another explanation called the automaticity hypothesis which suggests that the brain can easily and quickly identify words as it is an automatic process (Phaf & Kan, 2007), which could be the reason why information about words arrive at the response stage of processing faster than that for colours. However, although the results of the Stroop task seem inconclusive, there is contrasting research which has found a reverse Stroop effect by changing just one factor: the way in which the response is elicited. Durgin (2000) asked participants to point to the appropriate colour shown in a block rather than state aloud the colour and found that no effect exists. Similarly, when McClain (1983) asked participants to press the appropriate coloured button the effect was eliminated also. This suggests that further research is needed to investigate whether a real effect exists and to establish a definitive cognitive explanation for the effect.

The current research is a replication of Stroop (1935) which aims to investigate the effects of attentional interference on reaction times for each condition, and to test the statistical power of the original study to establish whether the same results are found, or whether a reverse Stroop effect is observed. Previous research into this area have used two levels of attentional interference: congruent and incongruent, but the current research used a further neutral level in order to fully investigate the effect. The main hypothesis for this experiment was that there will be a significant slowing in reaction time in the incongruent condition than in the congruent condition. However, as a third level of attentional interference was included (neutral), it was possible to draw two more comparisons and so there were two more hypotheses to test. These were: there will be a difference between the reaction time to elicit response in the congruent condition than in the neutral condition, and, there will be a difference in the reaction time to elicit a response in the incongruent condition than the neutral condition.

Method

Participants:

The participants were undergraduate psychology students studying at the University of Leeds who were recruited through an opportunistic volunteer sample. There were 493 participants (mean age = 18.88, standard deviation = 2.04; 369 females, 36 males and 88 who did not disclose their gender). The study received ethical approval from the School of Psychology Ethics Board.

Design:

The study used a within-subject design whereby all participants took part in all 128 trials of the task. The independent variable was defined as attentional interference and had three levels (congruent, incongruent and neutral). It was measured on a nominal level of measurement. The dependent variable was the reaction time to elicit a response to the stimuli measured in milliseconds (ms). The dependent variable was measured on a ratio level of measurement.

Materials:

This study used a computer simulated version of the Stroop task whereby there were 5 practice trials and 128 experimental trials for each participant. The task was run on E-prime which randomly generated the congruent, incongruent and neutral trials, and there was an average of 42.6 trials in each condition for each participant. Each trial flashed up on the screen for 10 seconds in which the participants had to respond by selecting the appropriate key from the keyboard which represented the colour of the word. The colours used were red, blue, green and purple. In the congruent condition, the colour of the word matched the word written. In the incongruent condition the colour of the word did not match the colour written. In the neutral condition, the word was shown as ‘XXXX’ in one of the four colours previously stated.

Procedure: 

At the start of the task, the participants created a unique participant code which would be used when entering their data. They also disclosed their age and gender prior to starting the experiment. The participants would then carry out the Stroop task. E-prime automatically recorded the responses of each participant. The participants then received a debrief about the aim of the study and background as to why it had been carried out.

Results

The data for all 493 participants were collated and analysed to find the mean and the standard deviation (SD) for each condition. The results from this analysis are shown in Figure 1.

Figure 1. The mean reaction time (ms) for each level of attentional interference.

A paired-sample t-test was carried out to establish whether the differences in reaction time between each condition were statistically significant. The first paired-sample t-test found that incongruent scores (M= 1047.48, SD= 244.21) were significantly slower than congruent scores (M=911.39, SD = 195.71), t(492) = -22.88, p < 0.001; d=1.10. The paired-sample t-test for the neutral-congruent comparison found that reaction time for the neutral condition (M=911.84, SD=195.66) were slower than congruent scores (M=942.23, SD =216.53), t(491) =4.37, p <0.001. The final paired-sample t-test found that the reaction time in the incongruent condition (M= 1048.47, SD= 243.46) was significantly slower than that for the neutral condition (M= 942.23, SD = 216.53), t(491) = 12.92, p <0.001.

Discussion

The aim of this study was to test the results of the previous Stroop tasks to investigate whether an effect exists. The results show that there was a statistically significant difference between all conditions, and that reaction times in the congruent condition were faster than that of the incongruent condition. This supports the main hypothesis that was that there will be a significant slowing in reaction time in the incongruent condition than in the congruent condition. This suggests that attentional interference can slow down reaction time to tasks involving certain stimuli. The results show that there was also a difference between the reaction times between the incongruent and neutral condition, and between the congruent and neutral condition. Figure 1 shows that the reaction time for the incongruent condition was substantially longer than for the neutral condition. The difference between the neutral and congruent conditions was not as substantial but the result still showed that reaction time in the congruent condition was faster than that of the neutral condition.

The initial Stroop test carried out by Stroop (1935) has been considered robust as it can be replicated to find the same statistically significant results, and in this experiment the same results were found. However, statistically significant results only prove that an effect exists, but not actually how strong and powerful that effect is. In the current study, a large effect size was found suggesting that there is a very significant effect. The power of the study was found to be 0.99 suggesting that the results found are extremely robust and so statistical support for the effect is strong. The result found by the neutral-congruent trial could suggest that there is another factor involved in cognitive processing that has not yet been considered as with both conditions eliminating attentional interference, the reaction time for the neutral trial was slower. This means that although the study was beneficial in providing strong experimental support for the findings that already existed, more research is needed to establish what other factors could also be having an effect in order to explain the effect with improved comprehension.

Unlike the previous experiment carried out by Stroop (1935), the current study adopted a large sample size of 493 participants compared to 70 participants in the original study. This large sample size means that the results that were found can generalise to a larger proportion of the population. However, there could potentially be gender bias within the results as research carried out by Van der Elst, Van Boxtel, Van Breukelen and Jolles (2006) found that women perform better than men on the Stroop task. As the current study used an uneven male-female ratio, this effect was not able to be recorded. This means that the current study may have been affected by gender and so an even proportion of males-females should have been used to investigate this potential effect.

The findings from this study could have huge implications within clinical practice as the findings can help understand cognitive functioning. Various clinical studies have used the results from the Stroop tasks to investigate frontal lobe function and the ‘imposition of higher-order control to inhibit or suppress the processing of irrelevant stimuli’ (Bruchmann, Herper, Konrad, Pantev, & Huster, 2010). One area of clinical practice in which the results could have implications is in the diagnosis and treatment of schizophrenia. Westerhausen, Kompus and Hugdahl (2011) found that schizophrenic participants showed an increased level of interference compared to neurotypical participants. In recognising this effect, and considering the results found by earlier and current studies, the research could be used to highlight the brain areas, or the explanations associated with higher order cognitive processing involved in the Stroop task in order to help understand what causes an increased level of interference in schizophrenic participants and potentially help maintain or control the symptoms of schizophrenia. Future research would be beneficial as a follow-on from this study, perhaps involving the use of brain scans like fMRI or EEG to detect the brain areas responsible for the interference so that more is known about the effects and symptoms of schizophrenia, and so more can be done to maintain those symptoms and improve the quality of life of those with the condition.

In conclusion, the results found proved that the findings of the Stroop (1935) study are as robust as they have previously been considered. All three comparisons were found to be statistically significant in the same direction as the Stroop (1935) task which provides strong statistical support for the Stroop effect.

References

• Brown, W. (1915). Practice in associating color-names with colors. Psychological Review, 22(1), 45-55. doi: 10.1037/h0073322
• Bruchmann, M., Herper, K., Konrad, C., Pantev, C., & Huster, R. (2010). Individualized EEG source reconstruction of Stroop interference with masked color words. Neuroimage, 49(2), 1800-1809. doi: 10.1016/j.neuroimage.2009.09.032
• Cohen, J., Dunbar, K., & McClelland, J. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97(3), 332-361. doi: 10.1037//0033-295x.97.3.332
• Durgin, F. (2000). The reverse Stroop effect. Psychonomic Bulletin & Review, 7(1), 121-125. doi: 10.3758/bf03210730
• MacKinnon, D., Geiselman, R., & Woodward, J. (1985). The effects of effort on Stroop interference. Acta Psychologica, 58(3), 225-235. doi: 10.1016/0001-6918(85)90022-8
• McClain, L. (1983). Effects of Response Type and Set Size on Stroop Color-Word Performance. Perceptual And Motor Skills, 56(3), 735-743. doi: 10.2466/pms.1983.56.3.735
• Phaf, R., & Kan, K. (2007). The automaticity of emotional Stroop: A meta-analysis. Journal Of Behavior Therapy And Experimental Psychiatry, 38(2), 184-199. doi: 10.1016/j.jbtep.2006.10.008
• Stroop, J. (1935). Studies of interference in serial verbal reactions. Journal Of Experimental Psychology, 18(6), 643-662. doi: 10.1037/h0054651
• Van der Elst, W., Van Boxtel, M., Van Breukelen, G., & Jolles, J. (2006). The Stroop Color-Word Test. Assessment, 13(1), 62-79. doi: 10.1177/1073191105283427
• Westerhausen, R., Kompus, K., & Hugdahl, K. (2011). Impaired cognitive inhibition in schizophrenia: A meta-analysis of the Stroop interference effect. Schizophrenia Research, 133(1-3), 172-181. doi: 10.1016/j.schres.2011.08.025

Appendix 1

SPSS output

Appendix 2

Effect size calculation

Pooled standard deviation= 195.713-244.209/2= 219.961

Average standard deviation= -24.248 $\surd 2\times (1–0.842)$

= 123.649

Effect size= $\frac{911.391–1047.48}{123.649}=–1.10$