Cognitive Report PDF

Preview

Summary

A report based upon the numerical stroop effect, research conducted in workshops during university. Title: "The Numerical Stroop Effect"....

Description

NUMERICAL STROOP EFFECT PSYC2092 Cognitive Psychology Assignment

P2437307 De Montfort University

2

Abstract Cognition is a complex phenomenon, especially when numbers are involved. Therefore, a study was conducted on one British, female, psychology student to find a difference in mean congruency response times using the numerical Stroop task. Using PsychoPy (version 3.2.3), the numerical Stroop task ran 95 trials, ranging from congruent, incongruent and neutral. The number pairs would range in sizes between 0.2 and 0.5, to see if the participant can control processing when finalising which number has higher numeric value. No significant difference was found amongst congruency response times to the different conditions, where both congruent and incongruent mean averages were 0.72. This suggests that there is some area of control in numerical processing, and that once a mind is set on a task of finalising which number is numerically higher, it can conquer any throw offs.

3

Introduction Our brains are capable of automatic processing when we have the accurate amount of specific knowledge to execute a task, for example, riding a bike. However, sometimes automatic processing needs to be inhibited, for example, when a child walks in front of your bike, you would need to stop riding. The Stroop Task is upon some of the highest researched tasks that have been used in cognitive psychology to examine individual’s ability to inhibit responses that are irrelevant during tasks and thus heighten our control over automatic processing (Dadon & Henik, 2017). An extension to the common colour-word Stroop task was created by Henik and Tzelgov (1982), known as the numerical Stroop task, using the comparison of physical sizes or numerical values of two numbers and to ignore irrelevant dimension values.

Being one of the most common ways to assess numerical processing, there has been a high volume of research previously conducted. Heine et al. (2010) studied the numerical Stroop effect in sixty-six primary school children, correlating them with level of achievement in maths. The Stroop effect was found across all levels of children in the physical comparison, but only when the distance between the numbers was substantial. The numerical distance effect relied on the condition, with the usual effect of distance in the congruent condition, but a conflicting effect in the incongruent condition. However, the research holds sample bias as all participants originated from Berlin, as well as the sample being small considering the high availability and accessibility of primary school children (Heine et al., 2010). On the opposition, the researcher took consent from the children’s parents and made sure that each child understood the task by asking questions and running many practice trials.

4

Furman and Rubinsten (2012) studied numerical processing in 31 adults with and without developmental dyscalculia (inability to process numerical information), where the sample population consisted of 6 males and 25 females. It was found that the DD group there was no significant results between the incongruent and congruent conditions, whereas the control group possessed a mean response time difference by 26.62ms. This suggested that those who suffer with DD struggle with differentiating between numerical value and physical size, as well as numerical processing. Although the sample size was small, the research can be accredited for possessing face validity due to finding the results they initially set out to find regarding the processing of numerical values in those suffering with developmental dyscalculia (Furman & Rubinsten, 2012). However, due to the sample including those who already struggle with the processing of numerical information, running tests based upon numeracy on those who struggle can become quite distressing and form unethical research (Von Aster and Shalver, 2007).

Based upon previous research and the main focus of samples being those with learning difficulties or children in general (Furman & Rubinsten, 2012; Heine et al., 2010), the current study will focus on a Psychology student with no background of any cognitive impairments. This will allow to study the effect of the numerical Stroop task in a neutralized context. As well as this, it will ensure that the participant understands the task, and will not have to reach out to parents for consent. Due to the current study’s neutralised context, it means there will be no struggles faced in terms of distress due to numerical learning difficulties as found in Furman and Rubinsten (2012).

5

Being able to research into numerical processing it’s a great step towards understanding how individuals’ cognitions work. This can be beneficial towards society; advancing education to children learning maths, increased knowledge on cognition and learning difficulties and neuroanatomy. Therefore, studying the numerical Stroop task in a neutral context, with participants struggling from no impairments can give us a good grasp on how to alter learning for those who do.

Regarding what has been explored above, the hypothesis for each condition has been drawn due to the result that previous research has reached, where incongruent conditions have found to increase reaction times in both the studies (Furman & Rubinsten, 2012; Heine et al., 2010). The hypotheses are as follows: The participant’s reaction time during congruent conditions will be smaller in value. The participant’s reaction time during incongruent conditions will be larger in value. The participant’s reaction time will have no significant difference during the neutral conditions.

Method Design The numerical Stroop task contains three different conditions, making these the independent variables: neutral - the numerical value of the two separate numbers is continuous; congruent - the number sizes and values correlate to each other; and incongruent - the numbers value and the physical sizes do not match. The dependant variable was the time taken to press the corresponding key to the largest number value (reaction time). The study was a within-subjects experiment as the

6

participant took part in every condition. As well as this, the participant undertook a practice trial to ensure complete understanding of the numerical Stroop task and how to complete it.

Materials The Numerical Stroop Task was developed by Henik and Tzelgov (1982), which the participant carried out on a late 2013, thirteen-inch MacBook Pro laptop (MacOS Mojave) with the keyboard attached. The Stroop task was made through 3 complete files (script, trial types and practice trial) all saved in the same folder and run via the application PsychoPy, version 3.2.3 (Peirce et al., 2019). The text presented on screen altered between sizes 0.2 and 0.5, with a frame rate of 59.996. Once the trials were completed, the data transferred onto a Microsoft excel document which was then collated and analysed on the spreadsheet.

Procedure The trials and scripts were downloaded into the same folder onto a late 2013, thirteen-inch MacBook Pro (MacOS Mojave) laptop from Blackboard. The trials were opened into PsychoPy (version 3.2.3) on the MacBook and started to run. To begin with, the participant had to put their participant ID into the pop-up box in the form of their student number, as well as which number participant the trial was running for. The participant first completed a practice trial, after being explained which keys to press on the keyboard to indicate which number was higher in numerical value. The keys were: left arrow for the number on the left, and the right arrow for the number on the right. After the practice trial, which contained 12 rounds, the participant was bought to a screen which they could press any key to proceed onto the actual trial of

7

95 rounds, where their reaction times would be recorded for analysis. In each round, the numbers would appear either the same size or different sizes, as well as being different distances away from each other. The participant had to identify which number was larger in numerical value, instead of physical size. Once the trials were completed, they were taken to a thank you screen. The data was then automatically transferred into a Microsoft excel spreadsheet ready for collation and analysis.

Results Once the data was transferred into a Microsoft excel spreadsheet, the raw data (see Appendix 1) was collated to ensure it was easy to read, and only contained the information needed. The raw data was saved as a separate file to any newly analysed data in case of mistakes or emergencies. To begin with, all practice trials were deleted, as these are only a run through to ensure the participant understands the task – they therefore are invalid towards any means and standard deviations. After this, one trial was incorrect and therefore this had to be deleted. This was done by sorting the cells alphabetically by the ‘resp.corr’ column, and then selecting the incorrect rows and deleting them. Once the incorrect cells were deleted, the trials needed to be sorted by congruency. This was done in a similar way, but instead sorting the cells alphabetically by the ‘resp.rt’ column.

To analyse the results into means and standard deviations, the subtotals tool was utilised by analysing the average of ‘resp.rt’ at each change in congruency. The same was done for standard deviation, but changed average to SD. The results show no significant difference between the congruent and incongruent trials (see Appendix 2), with the congruent mean score being 0.72 (SD = 0.18) and the

8

incongruent mean score being 0.72 (SD = 0.11). The neutral mean score was 0.77 (SD = 0.26). The standard deviation scores highlight a clear overlap of scores, cancelling out any difference between the neutral mean scores against the congruency scores and shows no significant results.

Discussion The aim of this research was to find a difference in mean response times between congruency conditions when numerically processing information, such as value and size. The results did not find any difference across any of the congruency conditions and thus suggests that the participant’s numerical processing was perhaps different to those who have partaken in previous studies explored. In order to be able to conclude any findings from this research, the sample would need to change. Both genders and different ethnicities need to be involved on a larger scale. As well as this, a level of professionalism would need to partake, due to the researcher also being the participant. This creates intense researcher bias, and perhaps no difference was found due to knowing the hypothesis and wanting to get the results to accept them.

However, if these findings were similar in a study with a much larger sample, it would bring some new evidence to light, suggesting that individuals have control over their processing and can quickly identify the numerical value of a number, no matter its physical size. Both studies previously explored (Furman & Rubinsten, 2012; Heine et al., 2010) found that incongruence increased response time, however this was not the case in the current experiment. This could however be due to the sample differentiation, where kids nor individuals with learning difficulties were used. Not

9

only would it create new evidence, it would be able to help individuals understand the individual differences within processing, and could perhaps advance teacher training throughout education, allowing them to adapt to different styles of processing types. De Smedt, Noël, Gilmore and Ansari (2013) found a clear link between different types of numerical processing and their mathematical achievement level, suggesting that those who are perhaps smarter and require harder work, process numbers in a different way to those who struggle.

Altering this study for the future, a larger sample size would be required to ensure the validity of results and a more in-depth overview of numerical processing. As well as this, no psychology students should be used as this can create a practice bias, allowing participants to have great understanding of the task and responding to fit the hypothesis. Both females and males will be used, to ensure all areas of individual differences of numerical processing are covered. This also includes cultures, ranging from worldwide ethnicities to cancel out any sampling bias, making results more reliable and ecologically valid.

For future research, I would suggest looking into how many different levels of numerical processing there are, ranging in ages, genders and cultures in order to figure out whether we are all the same or can control our processing when it comes to tasks like the numerical Stroop effect.

10

References Dadon, G., & Henik, A. (2017). Adjustment of control in the numerical stroop task. Memory & Cognition, 45(6), 891-902. doi:http://dx.doi.org.proxy.library.dmu.ac.uk/10.3758/s13421-017-0703-6 De Smedt, B., Noël, M. P., Gilmore, C., & Ansari, D. (2013). How do symbolic and non-symbolic numerical magnitude processing skills relate to individual differences in children's mathematical skills? A review of evidence from brain and behavior. Trends in Neuroscience and Education, 2(2), 48-55. Furman, T., & Rubinsten, O. (2012). Symbolic and non symbolic numerical representation in adults with and without developmental dyscalculia. Behavioral and Brain Functions, 8(1), 55. Heine, A., Tamm, S., De Smedt, B., Schneider, M., Thaler, V., Torbeyns, J., . . .Jacobs, A. (2010). The numerical stroop effect in primary school children: A comparison of low, normal, and high achievers. Child Neuropsychology, 16(5), 461-477. doi:10.1080/09297041003689780 Henik, A., & Tzelgov, J. (1982). Is three greater than five: The relation between physical and semantic size in comparison tasks. Memory & cognition, 10(4), 389-395. Peirce, J., Gray, J. R., Simpson, S., MacAskill, M., Höchenberger, R., Sogo, H., ... & Lindeløv, J. K. (2019). PsychoPy2: Experiments in behavior made easy. Behavior research methods, 51(1), 195-203. Von Aster, M. G., & Shalev, R. S. (2007). Number development and developmental dyscalculia. Developmental Medicine & Child Neurology, 49(11), 868-873.

11

Appendices

Appendix 1

Figure 1: Screenshot of the raw data saved straight after trials were completed.

12

Appendix 2

Figure 2: Screenshot of the collated and analysed data, showing the congruent and incongruent means and standard deviations in red....