Investigation on a variation of the Stroop test

The role of automatic and controlled processing was examined in an experiment looking at the allocation of processing resources between the two. Previous research found that when a word is written in the same ink as the colour the word denotes, it takes longer to state the colour of these words than if the words are colour neutral i.e. not the name of a colour. In the present experiment the type of word used was manipulated, between colour associated and colour neutral words. The results showed that there was a significant difference between the completion times of the two conditions which provided further support for the idea that automatic processes can interfere with controlled ones.


Best services for writing your paper according to Trustpilot

Premium Partner
From $18.00 per page
4,8 / 5
Writers Experience
Recommended Service
From $13.90 per page
4,6 / 5
Writers Experience
From $20.00 per page
4,5 / 5
Writers Experience
* All Partners were chosen among 50+ writing services by our Customer Satisfaction Team

The cognitive process of attention is said by psychologists to be how an individual processes the information that they receive from their day to day environment and how it is selected to be stored. That is to say that not all information that we obtain from daily life is seen as important for storage, it is essentially a selection process in which the most relevant information is chosen to be retained. William James(1890) believed attention to be dependent upon an allocation of cognitive processing resources(James,(1890),p9). The idea that all actions implore the use of processing resources to some extent, and is reliant upon the amount of resources available to an individual is known as a controlled process. Up until more recent years it was believe that no demand was placed upon these resources when the processes were automatic i.e. occurring without conscious awareness. These automatic responses were thought to use up less processing resources due to the lack of conscious effort involved, thus leaving resources free to attend to other stimuli. Thus essentially attentional processing has to be divided between the two. This was termed the two-process theory by Shiffrin and Schneider((1977),p21).

Researchers felt that this appeared to be a very useful concept as essentially it would leave more resources free but were interested to see at what cost. J.R. Stroop(1935) was one of these researchers(Stroop(1935),p21). He carried out an experiment into the relationship between automatic and controlled processes. This consisted of two conditions, in one participants were given a set of colour name words written in coloured ink but never matched to the colour ink they were written in. e.g. the word yellow written in blue ink. This is known as the Stroop condition. In the other the participants were given a set of neutral words e.g. the word rat all written in the same colour inks as in the previous condition. The participants were required to name the colour ink in which the words were written as fast as possible. It was found that participants could state the colour of the words in the neutral condition more rapidly than in the Stroop condition, leading Stroop to conclude that people are inclined to read the colour the word is portraying rather than the one it is written in. This is said by Stroop to suggest that the automatic process of reading can hinder the controlled process of naming the ink. It also shows that automatic processes are not necessarily ‘free’ as researchers once believe and can on fact be influenced by conscious strategies such as naming the colour in which the word is written.

Many variations of the Stroop test have been carried out since the original to further test individual’s attention and the allocation of resources in both controlled and automatic processing. One such variation is the reverse Stroop test.(Wikipedia) In this experiment there were still two conditions; one where the participant was given a list of colour words and had to point to a block of colour which matched the colour word stated. The second condition was the same as in the original Stroop test where the colour ink the word was written in had to be stated. It was found that it took on average 7 seconds longer to state the colours of the words aloud than it did to simply point to a block of colour that matched it. It was concluded that when the colour of the word stated differed from the colour ink it was written it the automatic process of reading the word again interferes with the controlled process of naming the ink colour. It was also discovered that this can be controlled but that then a time delay is incurred.

This idea that automatic processes do rely on resources just like controlled process, but that they occur without conscious awareness is still open to debate as despite the vast amounts of research done on this topic it still remains to be seen if the ‘Stroop’ effect only occurs with colour words. The idea that it is not only colour words that cause an integration between automatic and controlled processes is explored in the present experiment, by testing a set of colour- related words against a set of non-colour related words rather than a collection of coloured words printed in either black ink or opposing colours. The hypothesis was that people take longer to state the colour ink in which 30 words are written when the words are colour- associated words rather than colour-neutral words. This is a one-tailed hypothesis. The null hypothesis is that there will be no difference in the time taken to state the words in each condition, and if there is it is due to chance.


The design of this experiment was within-participants as the same set of participants were used to acquire the data in both conditions. The independent variable was the colour associated words. This consisted of two conditions, a colour associated word condition (condition 1) and a non-colour associated condition (condition 2). In the colour associated condition (condition 1), participants were timed to see how long it took them to state the colour ink of 30 colour associated words when written in opposing colours. E.g. plum written in orange. In the second condition the same procedure was followed but with the participants stating the colour ink of non-colour associated words. E.g. plan written in blue. The dependant variable was the time taken to complete each task, measured with a stopwatch to the nearest second. Several things were done to try and control for any confounding variables. Firstly, all participants were given the same set of instructions in order to prevent them gaining a different interpretation of what was being expected of them. Another thing that was done was to ensure that none of the participants could confer during the time in which the experiment was taking place, thus helping to prevent them from incurring demand characteristics. Participants also took the two conditions in different orders to help control for practice effects. Finally, all participants were shown the exact same copies of the stimuli which made it easier to ensure that this is not a reason for any differences that may have occurred.


Twenty five participants took part in this experiment. 16 of these were recruited by a researcher at the Open University and consisted of their friends and family. The other 9 were recruited through asking friends and family of the researcher. The age of the participants ranged from 17 to 69 and was made up of 12 males and 13 females.


A stopwatch accurate to the nearest second was used to time how long it took each participant to state the colour the words were written in. The visual stimuli presented in each condition consisted of 30 words placed into two columns on an A4 piece of paper. For condition 1 these consisted of 6 different words (blood,plum,carrot,sky,grass and lemon) written in 6 different colours (red,orange,blue,green,yellow,and purple). The word was never written in the correct colour i.e. sky would never be blue. In the second condition a different set of 6 words was used (plan,sty,ledge,grade,career and blame). These were written in the same 6 colours as the words in condition 1. In each condition each word was used 5 times and presented in a random order. A set of each stimuli is shown in the appendicies (appendix 1). A set of standardized instructions was also used to explain what was required of each participant (appendix 2). Finally each participant was required to complete a consent form before carrying out the tasks (appendix 3).


Each participant was asked if they would be willing to take part in a psychology experiment looking at cognitive processes, that would take about 5 minutes of their time. The ones who agreed to were given a consent form to fill in stating that they were willing to take part. They were then tested one at a time with their age and gender being recorded before being read the instructions as to how the experiment would work. i.e. that they would be shown a list of words and were required to state the colour the word was written in whilst being timed to state all 30 colours. They were also told that there would be two conditions and that the same procedure was required for each. (instructions appendix 2). Once all of this was understood and the participant had asked anything they wanted to they were given the words for condition 1. The stopwatch was started as they stated the first colour and stopped after they had stated the 30th. This time was recorded in the same table as used to record the participants age and gender earlier in the experiment.. The same was then repeated for condition 2 and the time recorded. The order in which the participants carried out the two conditions varied. Once all of this was completed the participant was given a debrief about the nature of the experiment and asked if they wanted to ask anything or needed anything. They were also made aware again of their right to withdraw at any point after the experiment should they decide they are uncomfortable with their data being use, and of the fact that their data would remain anonymous should they allow it to be used.


The one-tailed research hypothesis was that people take longer to state the colour ink in which 30 words are written, when the words are colour- associated words rather than colour-neutral words. This was measured with a stopwatch accurate to the nearest second (appendix 4 raw data).

Table 1-Paired -Samples T-test data



Std. Deviation

Std. Error Mean

Pair 1

Colour associated words





Colour-neutral words





As can be seen from table 1 above, the mean response time for the two conditions was fairly different with the mean for the colour- associated words being 2.7 seconds slower than the speed for the colour-neutral words. This difference can further be seen from the error bar chart below(graph 1).

Graph 1 shows that we can have 95% confidence that the population mean for condition 1 (the colour associated words) will lie within :-

1.08117 x 1.96 = 2.12

Therefore the lower bound is 25.62-2.12= 23.50 seconds

Upper bound is 25.62+2.12=27.74 seconds

Or 25.6 2± 2.12 seconds

Graph 1 also shows that we can have 95% confidence that the population mean for condition 2 (the colour neutral words) will lie within :-

.86163 x 1.96 = 1.69

Therefore the lower bound is 22.9 – 1.69 = 21.21 seconds

Upper bound is 22.9 + 1.69 = 24.59 seconds

Or 22.9 ± 1.69seconds

Therefore, although it appears from the experiment that when condition 2 is used the individual’s can read the words more quickly, this cannot be confirmed with any confidence. It is possible that in reality that condition 1 could result in those being test reading the words than under condition 2. For example, if the real mean of the whole population is towards the upper bound for condition 2 and towards the lower bound for condition 1, the individuals will be able to complete the test more quickly under condition 1.if there was no overlap between the bounds for each condition it would have been possible to say with 95% confidence that one condition allowed the individuals to complete the test more quickly.

A paired t-test was conducted on the data which showed that (t(30)=4.214;p=<.001;d=0.6). This analysis showed the difference to be statistically significant. It is therefore possible to reject the null hypothesis and accept the research hypothesis which was that there will be that people take longer to state the colour ink in which 30 words are written, when the words are colour- associated words rather than colour-neutral words. Using this together with the descriptive statistics, this analysis shows that test completion times for the colour-associated words(condition 1) are longer than for the colour-neutral words(condition 2).


The results of the present experiment showed that using colour-neutral words did mean that participants could state the colour ink the words were written in, in a faster time than they could for colour-associated words. Task completion times in condition 1(colour-associated words)were shown to be statistically significantly different from those in condition 2(colour-neutral words). This suggests that the automatic process of reading the words interfered with the controlled process of naming the ink colour.

A similar result was found by J.R.Stroop (1935), in the original Stroop test (Stroop,(1935),p21).He used this result to state that there is substantial interference from the colour name, causing people to find it hard to respond with the colour ink in which the word is written rather than the colour the word describes. Stroop’s research also provided evidence for the concept that resources are limited, and have to be allocated accordingly. The present research goes to support this idea, in the fact that the resources appear to have been allocated to the automatic process rather than the controlled one. This suggests that in this case the limited processing resources have not been allocated to the required controlled processing function required to state the ink colour. Thus supporting Stroop’s ideas. It also provides support for Sciffrin and Schneider’s two-process theory, as it shows that there has to be a division of resources between the two processes(Schneider and Schiffrin,(1977),p21).

However, although the present research supports the previous research carried out by Stroop there were still several flaws with both methodology and procedure. Firstly, although to some extent practice effects were controlled for by the random allocation of the order in which participants carried out the two conditions, it doesn’t allow for the fact that they were still carried out simultaneously. This could mean that the participants had already had a go at one of the tests and although the words were not the same they were similar. They therefore had practice at the task and the difference between their two times could have been down to practice and not due to allocation of processing resources. Another problem was that due to the small scale of the experiment i.e. only using 25 participants, the research is hard to draw conclusions from. If a larger sample of the population was used the difference may have appeared even more significant and the conclusions could have been drawn with more certainty. The sample was also not a representative one as it only used available participants which may have had an effect upon the results. Age may also be a confounding variable as older people may take longer to carry out processing tasks daily anyway. Finally, the experiments were not carried out at a particular time of day which may have had an effect upon the results. E.g. if one participant was tested at 11am they may be more alert and have a quicker completion time than a participant taking the test at 11pm.The ethical risk associated with this experiment was not particularly high and was accounted for by the use of consent forms and debriefing of participants, as well as making them aware of their rights to withdraw at all stages.

If this research was carried out several modifications could be made to gain more accurate results. Firstly, to limit practice effects occurring a time delay could be used between the two conditions taking place, to ensure that the differences between the two conditions are not due to the participants having practice at the experiment. Another way to control for this A much larger sample could be used to gain more accurate data that could more easily be generalized to the whole population, thus making the results more valid. The experiment could be carried out at a set time of day in order to ensure that it is not a confounding variable responsible for the differences in results. Ethical considerations should be taken into account as with the present research and any modifications made should also be given ethical consideration and amended accordingly.

Word count:- 2,109

You Might Also Like