In relation to the current study we might expect participants to be slower to name the quantity of incongruent characters, as a result of the characters being read faster than the quantity can be named. Thus when numerical characters are presented that are incongruent with the quantity of characters shown interference occurs and we may expect the reaction time of participants to be slower. (NEED A STUDY?? – Cohen & Dunbar pg.344).
The automaticity model states that reading is an automatic process. People can see words without much effort or consciousness; on the other hand, naming colours is not automatic and requires more effort than reading, thus creating interference in the stroop task (WEBSITE REFERENCE). In our study we would thus expect the reading of the characters presented to be more automatic than naming the quantities of the characters presented, and that when required to name the quantities when presented with incongruent numerical characters, as participants attempt to overcome their automatic response their reaction time may be slower, we may also expect more errors for the numerical characters as interference occurs. Evidence for the automaticity approach comes from studies that modify the original stroop task by only colouring one letter in the presented word, thus naming the colour becomes easier as interference is reduced – ‘colouring a single letter eliminates automatic reading processes’ (REFERENCE – cited???) therefore giving the participant time to process the ink colour.
This study aims to try and look at whether the stroop phenomenon can be extended to looking at interference between the naming of characters and quantities, and also whether the theories presented can correctly predict and explain the outcome of our study. One study similar to this, was carried out by Miya., M, MacLeod., C and Colin., M (2004) , they found a significant difference between the length of responses (i.e. interference level) when participants enumerated asterisks while attempting to ignore a single digit (inconsistent condition) compared to when participants attempted the same task whilst trying to ignore a letter (REFERENCE?).
The results of this study as well as those described before it lead to the development of two hypotheses; firstly that reaction time will be slower for numerical characters than non-numerical characters, and second, that numerical characters will produce more errors than non-numerical characters.
In the discussion, if numerical characters produce slower reaction times and produce more errors, it may be necessary to discuss which theory best explains this result, and/or whether a revision of theories is needed?
Method
Design
This experiment used a repeated measures design; participants took part in both the numerical and non-numerical character conditions.
Thus the independent variables were whether numerical or non-numerical characters were presented to the participant. The dependent variables were the reaction time (measured in milliseconds) and the number of errors made by each participant after each presentation.
Participants
The experiment used an opportunity sample of 1st year psychology students (around 180 students). From this cohort 20 participants were randomly chosen for data analysis. No controls for visual acuity or blindness were applied.
Materials
Stimulus presentation and response was controlled by a computer program pre-installed on the computers prior to carrying out the practical. Responses were made using a key-press technique (keys 1,2,3 and 4) on the numerical keypad.
Procedure
Participants were each seated in front of a computer screen in the computer suite of King Henry Building and then given standardised instructions (see appendix OR SHOULD BE HERE?) in how to carry out the task
.
During the task either 1,2,3 or 4 of the same characters were presented on the screen (e.g. 44, ****). The characters (2 types of stimuli) were presented randomly, being either numeric (1,2,3,4) or non-numeric (%, +, =, @). Participants had to press either key 1,2,3 or 4 on their keyboard depending on the number of characters presented. If they had not responded within 2.5 seconds of the stimuli being presented, the programme moved onto the next trial. There were 8 practice trials, followed by 256 test trials given in 4 blocks of 64 with a short break in between each block.
Results
The main findings of the study provide support for both of the hypotheses presented in the introduction. The average reaction time for the naming of quantities was slower when participants were presented with the numerical characters; in addition on average more errors were made in the numerical character condition.
Research Hypothesis 1: Figure 1 shows the means and standard deviations for the reaction times for the two conditions.
It can be seen from the above that reaction time was highest on average when participants were presented with numerical characters and asked to name the quantity.
A related t-test was performed on the data. A significant difference was found between the time taken to respond when presented with numerical characters (mean = 0.4855 ms) and the time taken to respond when presented with non-numeric characters (mean = 0.472 ms), (t (19) = 3.00, p<0.005). Mean reaction time was significantly greater for numerical characters than reaction time for non-numerical characters.
Research Hypothesis 2: Figure 2 shows the means and standard deviations for the reaction times for the two conditions.
It can be seen from the above that the average number of response errors was greater when participants were presented with numerical characters, and asked to name the quantity, than with non-numerical characters.
A related t-test was performed on the data. A highly significant difference was found between the number of errors made when participants were presented with the numerical characters (mean = 5.65 errors) and when participants were presented with the non-numeric characters (mean = 2.3 errors), t (19) = 4.30, p<0.0005). Mean number of errors was significantly greater for numerical characters than for non-numerical characters.
Discussion
The main findings from the study showed that significantly more errors were made for numerical characters than non-numerical characters, and additionally that numerical characters took significantly longer to react to than non-numerical characters. These results confirm our hypotheses and suggest that attempting to process numerical characters visually and incongruently with quantity gives significantly more interference, than in a neutral condition such as the presentation of non-numerical characters. Thus our study can be considered to replicate the basic idea of the original stroop effect.
In relation to previous research, the results of the study agree with that of Mya, McLeod and Colin (CHECK) – another counting interference task, and additionally to that of the original stroop task, where incongruent stimuli produced interference in much the same way. This further supports the assumption that we can transfer the ideas of the stroop phenomenon to numerical characters. Validating both this research study and the many stroop studies that have preceded it by demonstrating replication of ideas, theory and premonitions.
However, we must also consider, whether we can apply the theories discussed to the results obtained in this study, and whether any one can provide the best explanation.
Firstly with regard to hemispheric differences, the predictions the theory made in relation to this study have been supported – the use of both hemispheres in processing does seem to lengthen the time required to process information, especially when the stimuli are incongruent, thus explaining the significantly longer reaction time found for numerical characters. Additionally this theory can explain why more errors where made when asked to name the quantity of incongruent stimuli as it is the left hemisphere that seems to suffer from most interference and it is this hemisphere that is involved in verbal processing.
However, we must consider possible flaws with suggesting this theory provides a reliable account of why we got the results obtained. Firstly, perhaps we cannot assume that typing the quantity of characters into the keyboard is the same as naming them verbally; surely this brings visual processing into a condition we are considering to be purely verbal. Secondly, not all research has agreed even with regards to using this theory to explain even the basic stroop effect, therefore it is perhaps doubtful that we can apply it to an extension of the stroop effect. We would possibly need to carry out a further study in which characters were presented to the different visual fields (i.e. different hemispheres) to assess the contribution of this theory to explaining the results of the current study.
Next we can consider the results of this study in terms of the speed of processing model. As this model predicted reaction time was slower for the incongruent condition (when participants were asked to give the quantity of a presented number of numerical characters) compared to the consistent condition (naming the quantity when presented with non-numerical characters. We could therefore state that this was in fact a result of the processing of the characters being quicker than the processing of the quantity, thus when these responses competed in the inconsistent condition, reaction time was slower and more errors were made. This does seem a plausible explanation, however some argue that this approach must be rejected. The argument of the model is that ‘there is a parallel processing of the two dimensions of the stimulus at different speeds’. (REFERENCE MACLEOD) Therefore it assumes that the faster dimension should interfere with the slower one and not the other way round. In fact, many types of interference can occur, ‘ (REFERENCE) ‘interference is not an orderly function of relative speed of processing’ (REFERENCE). These points possibly make the hypotheses of this approach invalid, not only in explaining the stroop effect, but also in explaining the results of the current study. Interference may be the result of quantity interfering with the characters, but the characters may not interfere with naming the quantity. This makes the theory unfalsifiable in many aspects. (CHECK/UNSURE).
Lastly we consider interpreting the results in terms of the automaticity approach. The predictions of this model were quite accurate for the study. Participant’s reaction times were slower whilst attempting to overcome their automatic responses to read the characters in the inconsistent condition; this interference also produced significantly more errors. This approach seem perhaps the more feasible, as it makes sense that reading is a more practiced and thus automatic skill in comparison to naming quantities, and this can not only be applied to words and colours, but characters as well. However, others say that there are also flaws in this approach arguing that what is taking place is more to do with attention than automaticity. Taking evidence to support this approach these critics state that results that have shown increased processing time when only one letter is coloured is due to increased attention on the participant’s part rather than a aid in reducing an automatic response. (REFERENCE? /THOUGH USED OWN INTERPRETATION OF WHAT WAS SAID).
In conclusion it seems that although all these theories gave predictions supported by the results obtained, and can also provide reasonable explanations of such results, they all have weaknesses that prevent domination of one theory. This suggests that further research needs to be carried out on the stroop effect and its extension to other mediums. A focus on new theories perhaps needs to be adapted. One such theory is the parallel distributed processing model (COHEN???) described in (MACLEOD REFERENCE).’It incorporates many of the virtues of automaticity and relative speed of processing but few of the liabilities’. This model’s basic idea is focussed on an incorporation of a clear role for attention, stating that ‘it moderates the operation of processing units in a pathway’. (REFERENCE). Processing is performed in a system comprised of interconnected modules, containing units for inputs and outputs, information flows in one direction through these modules.. When the model is instructed to perform a task, it selects a pathway that includes various units, ‘the set of connections in this pathway specifies its strength, and the choice of pathway therefore determines both the speed and accuracy of the processing’ (REFERENCE). Interference is a result of two pathways being activated simultaneously and producing conflicting activation at their intersection. (RE_WORD? SHORTEN? INCLUDE? ONLY TALK ABOUT WITH REGARDS TO STROOP EFFECT?).
In discussing this study, not only must we consider theoretical approaches to explaining it, but also methodological issues with the actual study itself. We can then ascertain whether its results are valid and reliable enough to be included in relation to theory and further discussion. The sample of participants is perhaps one area that should be addressed, the study only looked at first year psychology students, and generalisation to the wider population is therefore perhaps limited. The participants were also aware of the research background prior to participation in the study and although unlikely as the task was quite complex and defined, their expectations may have affected their performance. A further study could perhaps include a more diverse range of participants, maybe those who are completely unaware of the basic stroop phenomenon.
IN HERE COULD INCLUDE OUTLIER< BUT ONLY IF CAN REDUCE WORD COUNT
We should also question the practical use of these results and in fact that of the original stroop task. After all it does seem a rather artificial task that seems to lack application to the real world. In reply to this criticism however, there have been cases in which the stroop phenomenon has been used to identify brain damage, perhaps our numerical characters study could dot the same. Further research should also consider the applicability of research findings of the stroop effect and extensions of it in an applied context.
Validity of the data is also an important concept; we must question whether the study measured what it was supposed to measure. Practice effects may have decreased the validity of the data obtained, there were a large number of trials and participants may have got used to having to distinguish between quantity and visual incongruency. However this criticism has to be weighed up against the reduction of reliability that would occur if the number of trials was reduced, or in fact the addition of individual differences in attention if an independent groups design was preferred.
In conclusion, this study showed that participants took significantly longer to react to the stimuli presented when they were incongruent (i.e. numerical characters presented and then asked to name the number of characters present) and made more errors in this condition in comparison to when non-numerical characters were presented. This replicates research done into the stroop effect, which suggests interference occurs when verbal and visual processing conflicts. However, the theories proposed to explain these results and that of previous stroop research is somewhat weak and further discussion and theoretical consideration is definitely needed.