Method
Design and Controls
This report has been based on the repeated measures design. The main reason for this is that individual differences are eliminated and the time period for which to execute the experiment is minimised. This design, however, introduces order effects which include fatigue and practice effects. These effects can be minimised through counterbalancing, this involves splitting the participants into two equal groups. The first group will go through the experiment from CS to CC and the second group from CC to CS. This is known as ABBA counterbalancing.
The data will consist of two sets of times, one set will be the times that participants took to complete test A and the other set will the times to complete test B (explained in the procedure on the next page). A standard set of instructions can be found on the first page of the web site.
The independent variable will be the colour interference and the dependant variable will be the time it takes for the participant to verbally respond.
Sample
The sample type that will be used for this report will be opportunity sampling. This type will be used because random, stratified and quota samples would take too long to organise and it would be almost impossible for the researches to carry an investigation of such great proportions. Systematic and self-selected samples could be used but the sample, which in this case are the Notre Dame Sixth Form students, is not really large enough. The sample consists of males and females aged between 16 and 19.
Materials
A web site will be designed by two researchers for ease of use. The web site will consist of both JavaScript™ and HTML programming techniques and a print-out of the web site will be attached to the appendix of this report. This web site will also contain the lists of words in both conditions and the tests will be timed by the use of a code-generated stop-watch.
Procedure
Two male researchers from the Notre Dame Sixth Form will carry out the experiments. One researcher will carry out the experiments in the order Cs to Cc whilst the other Cc to Cs. Participants will be approached in the Open Learning Centre (OLC) building located at Notre Dame High School. A student close by will be politely asked whether they would like to take part in a psychological experiment. If the answer in no, then they will be thanked for their time and another student will be asked to participate. If they answer yes, it is then explained that the experiment will be carried out on a computer in the OLC but they need not worry whether they are competent with computers or not because they researcher will operate it himself. The participant will then be escorted to a reserved computer in the OLC.
The participant’s gender and age are noted and the web page is brought up on screen. The standard instructions, displayed on-screen, will then be read aloud politely and the researcher will ask if they are understood fully. If not, the instructions are explained again in more detail until the participant understands. When this point is reached, the researcher waits until the participant is ready and the first test begins. The time it took for the participant to read out all the words will displayed (see Figure 2) and be noted and so will any errors.
Figure 2
The second test will then commence when the participant is ready and again the time and any errors will be noted. The participant will then be thanked for their time and a brief explanation of what is being investigated will be given to the participant. The researcher will then search for his next participant.
This experiment will be carried out according to the British Psychological Society (BPS) Guidelines. As a result, the researchers will not force students of Notre Dame High School to participate; the free choice of the students is understood and respected. Participants will not be deceived or given any misleading information and it will be made known to them that they have the right to withdraw from the experiment at any time. Only the age and gender of the participants will be noted in order to respect the privacy of the participants.
Results
Raw Data
The data from the web page task were collected by two researchers and put into a table (appendix) and also a graph (appendix). Table 1 shows a summary of the raw data in table form:
Raw Data Summary Table
Table 1
Where ‘N’ is the number of participants, ‘Σx’ is the sum of all the data for that condition, ‘x’ is the mean for that condition and ‘S.D.’ is the standard deviation.
Treatment of Results
The data was statistically analysed using the ‘t-Test’. This statistic was used because it is a parametric test for related data. Parametric tests are more powerful, of a higher efficiency, more sensitive to the features of the data and more robust than non-parametric tests. Although non-parametric tests are easier to calculate and there are no needs to meet data requirements, the researchers chose the ‘t-Test’ for related data because of it’s many advantages mentioned above.
The conditions for using the t-Test for related data are shown in Table 2:
Conditions for a Related t-Test
Table 2
The data from the experiments are ‘differences’ because there are two separate, discrete times for each participant i.e. the data is not being correlated. The data for this report is of the ratio level because it is measured in fixed units with equal distances between them (time, in seconds) and zero does provide a ‘baseline’ for proportion calculations in this case. Data that is of a ratio level are the most precise and therefore most reliable type to use. Zero provides a baseline in this case because, unlike oC and oF, you can not have minus seconds. The data is related because the experiment design is repeated measures so therefore a value in one group is related to a value in the other group. The parametric assumptions can be found below:
- The samples were drawn at random for the populations under consideration,
- The variances in the populations are homogeneous,
-
Scores are normally distributed in the population.
The samples were, admittedly, not drawn at random from the population but it is safe to assume that, like memory, recall times would be very similar throughout the population. Variance calculations (appendix) prove that the populations are homogeneous. Scores are normally distributed in the population, as found by the standard deviation, so therefore the criteria are met.
The level of significance that was chosen was p=0.05 and this was because it is conventional in directional (one-tailed) psychological research. Type one errors might occur if the significant population actually lies within that 5%. Type two errors occur if a false null hypothesis fails to be rejected and the probability of which is designated as β (beta). The probability of a type one error is designated as ‘p’.
A summary of the statistics (see appendix for more) found using the related t-Test can be found in Table 3:
Summary of Results
Table 3
Decision Rule
The obtained t is larger than the critical t therefore the alternative hypothesis can be accepted.
Discussion
In the study by Brown (1915) mentioned in the introduction, it was concluded that “…the association process in naming simple objects like colours is radically different from the association process in reading printed words” and therefore the interference occurs. The results show that, according to the statistic, it is very unlikely that the null hypothesis should be accepted (thus reducing the likelihood of a type 1 error). This can be said because the obtained t-value is much higher than the critical t-value. There was one atypical result where the participant was actually quicker in the CS. The participant was interviewed after the experiment had taken place and it was explained that the result was uncommon. The participant explained that her main interest was art and she also studied it at A-Level. She also explained how she saw the colour first, regardless of the text or whether text was involved at all. This atypical result cannot be generalised to the population of artists because there were two more participants that studied A-Level art in this experiment. However, it would be interesting to do a study on the differences in results between artists and non-artists.
Stroop (1935) showed how it took an average of an extra 2.3 seconds longer to read words in the conflicting condition, the results from this report show how it took and average of and extra 5.3 seconds. Theses two results can be considered as similar because the sample taken for this report is three and a half times smaller than that of Stroop’s seventy college undergraduates. This difference can also have occurred because Stroop’s participants had to read lists of 100 words, whereas only lists of 25 words were used for this report. Telford (1930) also found that, in slightly different research, it takes more time to name colours than to read the names of colours.
When considering methodological limitations, a few factors can be taken into account in terms of this report. Since the experiment was based around computers then participants may suffer slight eye-strain or frustration due to the flickering of the screen. This could have been minimised by using TFT monitors which do not flicker and reduce eye-strain. Some participants leaned closer to the screen when they came to the orange colour and some errors occurred when they generalised this colour to red (see Figure 2):
Figure 2
This could be overcome by modifying the stimulus material to use less ambiguous colours, as in Figure 3:
Figure 3
It could be said that this report lacks ecological validity because the sample selected from Notre Dame High School may not be representative of the entire population of the world. However, as discussed previously, the results in this report do not differ massively from those taken by Ph.D. students in the professional field of psychological research. The 1998 Data Protection Act states that when conducting research of this nature, researchers should not need to disclose personal information of the participants and any needed for the experiment should be destroyed after use. The results taken in this experiment were kept secret from the participants and they were recorded only in terms of age and gender. British Psychological Society Guidelines were also followed (as explained in the procedure) so therefore ethical issues were dealt with as much as was possible.
The relationship between the anomalous result and the fact that it was an artist is interesting and could be further investigated. Using two conditions (one with artists and one without) and a repeated measures design, differences could be found in recognition times between the groups. Other factors could also be investigated further, such as the practice effect. A group that has been practising and another group that hasn’t could be compared to see if there are any significant differences. Another modification could be the comparison of results from this report with the results of a similar study conducted using the suggested, less ambiguous colours.
Conclusion
Since it was found that participants did take longer in the Cc, then it can be concluded that the visual interference of colour does effect process times.
References
Brown, W. (1915) ‘Practice in Associating Colour Names with Colours’ Psychol. Rev.
Cardwell, M., Clark, L, Meldrum, C, ‘Psychology for AS Level’, Collins (2000).
Cardwell, M., Clark, L, Meldrum, C, ‘Psychology for A2 Level’, Collins (2001).
Kahneman, D. (1973) ‘Attention and Effort’ Englewood Cliffs, NJ: Prentice-Hall.
Shaffer, L. H. (1975) ‘Multiple Attention in Continuous Verbal Tasks’, in P.M.A. Rabbitt and S. Dornic (eds) Attention and Performance, London: Academic Press.
Shiffrin, R.M., Schneider, W. (1977) ‘Controlled and Automatic Human Information Processing: Perceptual Learning, Automatic Attending and a General Theory’, Psychological Review.
Stroop, J.R. (1935) ‘Studies of Interference on Serial Verbal Reactions’, Journal of Experimental Psychology.
Telford, C.W. (1930) ‘Differences in responses to colours and their names.’ J. Genet. Psychol.
An Experiment on the Stroop effect and hearing,
British Psychological Society Code of Conduct for Psychologists,
Cognitive Psychology, Wadsworth CogLab online laboratory,
Neuropsychological Model of the Stroop Effect,
Neuroscience for Kids – The Stroop Effect,
Parametric Assumptions,
The Stroop Effect – Attention and Memory,
The t-test,