Shiffrin and Schneiders research was based on visual rather than auditory tasks, so Poltrock et al. (1982) carried out an auditory detection task, which also had the same results.
Healey (1976) carried out research into how we automatically process frequently occurring words such as “of” in sentences and so find it harder to focus on component letters. She presented participants with a piece of English prose and asked them to read it and circle all the t’s in the passage. Participants frequently missed out the t’s in common words such as “the” and more easily identified the t’s in more uncommon words.
This shows that we recognise high frequency words such as “the” as whole units rather than by their individual letters, so automatically process them. This powerfully autonomic nature of reading words is also evident in the following research carried out by “Stroop”.
Stroop (1935) carried out an investigation into autonomic processing, by inventing the stroop effect. In this, he instructed participants to read a list of colour words written in black ink. This, evidently a very simple task was easy for the participants to carry out.
Following this, participants were asked to read a list of colour words written in conflicting coloured inks, (e.g., the word “red” written in blue colour ink) and to call out the colour ink the words were written in. Although this task seems very simple at first and is only matter of simple colour recognition, Stroop found that it took the participants considerably longer to complete this task then the previous. This is because, the powerful autonomic (unconscious) nature of reading words meant that participants automatically wanted to read the words rather than the colour ink they were written in. So, even though the participants didn’t often read the colour word out loud, there was a time delay whilst the participants thought of the correct response (the colour ink).
AIM: In this investigation I aim to carry out a modification on the experiment on automatic processes carried out by Stroop and to discover whether the results correspond.
EXPERIMENTAL HYPOTHESIS: This is a directional hypothesis which states that the length of time taken for participants to name a set of incongruent words will be greater than the length of time taken to name a set of congruent words.
NULL HYPOTHESIS: There will be no difference in the length of time taken for participants to name a set of incongruent words and the length of time taken for participants to name a set of congruent words.
Method
Design:
A laboratory experiment was chosen, with a repeated measures design. This is because, the participants needed to be exposed to each of the experimental conditions (congruent and non congruent words).
The independent variable was the congruent and incongruent colour lists and the dependant variable was the time taken to name the colour of the word.
In the case of this particular experiment, lack of informed consent (presumptive consent), so in a sense, deception and psychological harm to participants were all ethical breaches. So in order to minimise any potential harm, the participants were debriefed (see appendix) to prevent psychological harm, as stress may have been caused due to the pressure of having to carry out the task as quickly as they could.
Also, Confidentiality of participants had to be ensured, by anonymity.
Participants:
The participants included the same researcher, who was a female psychology student aged 17 and a target population of student’s aged 16-19 from Richmond upon Thames College. The sampling method was availability and the actual sample consisted of 20 participants (10 girls and 10 boys). As the design used was repeated measures, the participants were exposed to two conditions (one a list of congruent words and the other a list of incongruent words). The participants were exposed, by being shown a piece of paper with the words on.
Materials:
The materials used in this experiment were six lists of words: 3 congruent and 3 incongruent, a stopwatch and a results table (see appendix)
Standardised procedures:
In order to standardise the procedures, the following exact steps were taken in conducting the research:
- A hypothesis was formulated, which was that participants would take longer to name the colour of a set of congruent words than of a set of incongruent words.
- All the materials were prepared in a quiet room in which the experiment would take place.
- The sample was chosen and the participants were then chosen on opportunity from Richmond College.
- Once chosen, the participants were briefed as much as they could be without jeopardising the whole point of the experiment and presumptive consent to carry out and use the results of the experiment was gained.
- Each participant was taken into the same, well lit quiet room individually each time and was read out a sheet of instructions (see appendix)
- The participant was presented with a list of words and was then instructed to “read through the list of words as fast as you can”.
- At the beginning of the recitation of words the stopwatch was started and at the end, it was stopped and the results were recorded.
- This was repeated with following 5 lists.
- Results were collected and the participants were debriefed.
Controls:
There were a range of possible confounding variables including the quality of the participants’ eye sight and whether any were colour blind or not. In this case, obviously they would not be able to carry out the experiment. In addition, the clarity of incongruent and congruent words, including the choice of colour size and font could also bias the results. For instance, some light colours such as yellow may have been too difficult to read clearly. Other confounding variables could include timing bias and order effects.
In order to prevent any confounding variables from affecting the validity of the results, the participants’ eyesight was checked beforehand, the ink and font of the words was made clear, easy to read and kept the same for every participant. Also the colours used were chosen carefully to make sure the words could all be read clearly.
In addition, as the experiment design was repeated measures, in order to prevent order effect, counterbalancing was carried out. So, each participant was exposed to a different sequence of the lists of words.
Table of results
Descriptive statistics
Mean
From a table containing the congruent and incongruent results (see appendix), the overall means for both was calculated. Using the results the following bar chart was created showing the time it took for participants to read the congruent and incongruent lists.
From looking at this bar chart, it is evident that the incongruent mean is considerably higher than the congruent mean, which shows that it took considerably longer for participants to read the lists of congruent words than to read the lists of incongruent words. In fact the exact value for the congruent mean is 7.45 whereas the value for the incongruent mean is 13.30. This shows that participants on average took a greater time of 5.85 to recall the incongruent lists of words.
Median
I then calculated the median of my data. (See appendix) The median is the middle of a distribution, and is less sensitive to extreme scores than the mean, which makes it a useful calculation to carry out.
For the congruent lists of words, the median is significantly lower at only 7 seconds than the median, at 12.50seconds. Again, comparing the medians has given me an idea as to how widely the values in this data set are spread apart. Here it is evident that there is a very large difference in the time taken to recall the congruent lists, than the incongruent lists.
Range
The range also will tell me about the spread of the data and by calculating this (appendix), it is evident that the range for the incongruent lists of words is greater, by 3 seconds than the congruent lists of words.
Standard deviation
Now, in order to measure the spread of data around the mean, the standard deviation of the time taken to read the congruent and incongruent lists of words was calculated (see appendix). The standard deviation for the incongruent lists of words is almost double that of the congruent lists. This really does highlight the difference in the length of time taken to read both lists of words.
Inferential statistics
The Wilcoxons matched pairs signed ranks test was used in order to test the median difference in the paired data. This statistical test is suitable for my data as it is a test of difference, suitable for data gained from a repeated measures design. It can be used with data on an interval level of measurement, which is then converted to an ordinal level for the purpose of the test (see appendix).
Using the critical value table, using a significance level of 0.05, the critical value for 20 pieces of data is 60.
As the observed value of T (0) is less than 60, the null hypothesis can be rejected in favour of the alternative hypothesis.
This essentially means that the alternate hypothesis, that the length of time taken for participants to name a set of incongruent words will be greater than the length of time taken to name a set of congruent words, can be supported.
Discussion
The aim of this research was to replicate the original experiment carried out by strop and gain the same results. The aim was achieved, as the results gained supported the hypothesis that it would take considerably longer for participants to name the colour of the words on the incongruent list than those on the congruent list. This means there was interference between the autonomic nature of reading words and the colour, first suggested by Stroop and other researchers.
Limitations
There are a number of methodological limitations that could have affected my results.
- Biased sample- the sample consisted only of students of a particular age range which meant that the results gained are only valid for this population. The results are not generalisable to the whole population.
- Timing- the experimenter could never be fully accurate with the timing, as using a stop watch is bound to cause human error. If carried out in future, perhaps the Stroop effect could be done on a computer screen with the colours flashing on a screen before the participants’ eyes and once completed, the computer could itself stop the timer.
- The experiment itself was very artificial and carried out in an artificial environment. People would not usually find themselves in situations where they have to read out lists of words from a sheet of paper in real life. The experiment therefore lacks ecological validity and the findings can not be applied to real life situations.
The findings from the stroop effect do have greater implications. The discovery that the autonomic nature of reading tasks can interfere with other controlled processes is very significant, in for example texting whilst driving. This experiment shows that the automatic process of reading may draw attention away from the task of driving and could cause interference. This could result in a serious car accident.
Future studies into the Stroop effect could include using shapes and objects rather than just colours to see whether the autonomic nature of reading words is just as strong when paired with using objects or other shapes.
References
- Cardwell, Clark and Melldrum, Collins psychology for A2 level
- David Rice and Mike Haralambos, psychology in focus A2 level