The Relationship between Fear for Mathematical Stimuli and Math Anxiety: Validation of Mathematical Pictures
Lieke Kok
Institution: University of Amsterdam Student number: 10488693
Supervisor: Eva Schmitz
Amount of words in abstract: 162
Abstract
In this study, the relationship between fear for mathematical stimuli and two types of measured math anxiety was examined, in order to validate a picture-set for future use in priming tasks. This was tested for 519 secondary school students in the first three years of secondary school. Math anxiety was measured through questions during a performance-task and through the use of the Abbreviated Math Anxiety Scale (AMAS). Participants also had to complete a picture-task, in which they were asked to report how anxious they felt. The
different anxiety scores on all three parts of the task, were compared to each other. Fear for mathematical stimuli was significantly related to both measures of math anxiety. The two measures of math anxiety were also significantly related to each other. However, the pictures used, did not meet all the criteria for validation. In conclusion, there is a relation between fear for mathematical stimuli and measured math anxiety, but the pictures used cannot yet be validated.
The Relationship between Fear for Mathematical Stimuli and Math Anxiety Relevance to Society
A lot of different definitions of mathematical (math) anxiety exist (Suárez-Pellicioni, Núñez-Peña & Colomè, 2016). What these definitions have in common is that dealing with numbers or math-related situations can induce an emotional response in people with math anxiety, which in turn can negatively influence their performance. Therefore, if anxiety is reduced, math performance might improve. Moreover, math anxious individuals often show avoidant behaviour towards math which can result in choosing different career paths
(Chipman, Krantz and Silver, 1992).
Western societies in particular appreciate good education and stimulate people to achieve the best they can in school. Thus, math anxiety is not only a burden on the person itself, but on the educational system as well (Ashcraft, 2002). New ways of reducing math anxiety might be discovered by doing research on this phenomenon. For these reasons, research is very relevant to both math anxious individuals and society itself. However, before research on ways to reduce math anxiety can be done, research has to be done on what the aspects of math anxiety are and how they relate to each other.
Theoretical Framework
In order to examine anxiety as a whole, one must be aware of the different aspects of anxiety. Information, in this case an emotion, is processed in two different ways, which can be explained by the dual process theory. According to the dual process theory, cognitive
processes consist of two different components, which are encoded in different ways. That is, cognitive processes are encoded both implicitly and explicitly (Sun, Slusarz & Terry, 2005). The implicit way is based on intuition and is therefore fast, automatic and difficult to change (Kahneman, 2011). Processes like these are measured with priming techniques or interference
tasks, such as the Stroop task. The explicit way is based on reasoning and is therefore slower, more conscious and easier to change (Kahneman, 2011). Questionnaires are often used to measure these processes.
Anxiety might also be encoded in two different ways. This idea is supported by an experiment carried out by de Jong, van den Hout, Rietbroek and Huijding (2003). They found that explicit fear for spiders differed between fearful people and non-fearful people. However, implicit fear hardly differed between the two groups. Therefore, examining only explicit fear might not be consistent with actual fear. Math anxiety is probably encoded in two different ways as well. Thus, explicit and implicit math anxiety might differ (Rubinsten, Bialik & Solar, 2012).
In most studies, only questionnaires are used to assess math anxiety. However, the dual process theory suggests that implicit and explicit math anxiety might differ and might therefore also have different influences on performance. Thus, it is useful to find a way to measure math anxiety implicitly and to compare explicit and implicit measures of math anxiety.
The Aim of this Study
Before a priming task can be used in an experiment that measures implicit anxiety, the stimuli needed for this task have to be validated. Therefore, the aim of this study was to validate pictures for future use in implicit tasks. This was done through comparing fear for mathematical stimuli with two types of measured math anxiety, math anxiety during a performance-task and self-reported math anxiety on a questionnaire.
In priming tasks, shortly presented stimuli are used to trigger a certain behavioural or emotional response (Henson, Eckstein, Waszak, Frings, & Horner, 2014). Therefore, stimuli in priming tasks can only be used when they trigger such a response. To make sure pictures
can trigger a certain response, they have to meet a few criteria. The criteria used in this study were based on a study carried out by Pronk, van Deursen, Beraha, Larsen and Wiers (2015). The most important criterion is that variation in stimuli have to cause variation in the outcome variable. That is, stimuli that vary in how much they relate to math should cause variation in reported anxiety. In this study, mathematical pictures, pictures with aspects of math and neutral pictures were used to cause variation in reported math anxiety. Another criterion is that all of the pictures need to have the same features, such as size and complexity. This will make sure that the response is triggered by the variation in how much the picture relates to math, not other aspects of the picture. The last criterion is that participants need to be able to recognise whether a picture contains math or not, in order to make sure that the anxiety participants feel are indeed related to math. In this study, the criteria are tested using a recognition task and a rating task (Pronk et al., 2015).
Earlier Research
An important note to make is that arithmetic and mathematics are two different
subjects in the Netherlands. The Dutch definition of mathematics does not include arithmetic. The study in this paper was carried out in the Netherlands and therefore fear for arithmetic was not measured. Therefore, when discussing the study in this paper, math anxiety will be defined as fear for all aspects of math relevant to young secondary school students, excluding arithmetic. In all the other studies about math anxiety referred to in this paper, math did include arithmetic.
Pictures have been used several times for studies which are not related to math anxiety. Simon, Kischel, Spielberg and Kathman (2012) attempted to validate the use of a priming technique with pictures inducing OCD symptoms. They found that OCD-patients reported higher anxiety for pictures with OCD triggers than participants without OCD.
However, the test could not be validated, because their sample size was too small. Pronk et al. (2015) were more successful. They used a priming technique with pictures to measure
cognitive bias in alcohol usage. Using the criteria mentioned earlier, the picture-set was validated. Studies like these suggest that pictures do trigger an emotional response.
Math-related stimuli have been used in earlier research as well. For instance, Hopko, McNeil, Gleason and Rabalais (2002) found that math anxious individuals experienced more difficulty in an emotional Stroop-task with mathematical words, compared to individuals without math anxiety. A more recent experiment with mathematical words was carried out by Rubinsten, Bialik, and Solar (2012). They found that women responded faster to a math equation after seeing a negative word than after seeing a positive one in an affective priming task. This suggests that they associate math with negative feelings. Studies like these show that math representational stimuli can induce negative emotions such as anxiety and therefore there is reason to believe that math pictures can induce anxiety as well.
A study that did use math pictures was a study from Rubinsten, Eidlin, Wohl and Akibli (2015). They measured attentional bias in math anxiety. Instead of only using mathematical words, they also used equations as primes. Using these words and equations, they found that attentional bias exists in math anxious individuals. Aarnos and Perkkilä (2012) used a more varied picture-set and asked young children about their feeling towards them. Math-pictures were associated with sadness. However, no distinction was made between math anxious children and children without math anxiety and mathematics in this age-group consists mainly of arithmetic. Studies like these do however implicate that mathematical pictures induce certain emotions, like anxiety.
Math Anxiety during a Performance-Task
The research mentioned above implicates that mathematical pictures might trigger an emotional response, in this case anxiety. However, not enough research has been done on math pictures to create a task that could measure implicit math anxiety. It is not yet clear if there is a relation between fear for mathematical pictures and measured math anxiety and this is necessary in order to make sure that the emotional response triggered by the pictures
resembles actual math anxiety during the task. If the results of the different tasks are related to each other, it is more likely that all tasks measure math anxiety.
As stated before, according to Suárez-Pellicioni et al. (2016) there is a relationship between math anxiety and math performance. Also, Mammarella, Devine, Caviola and Szűcs (2015) found that children with dyscalculia and children with math anxiety were impaired in different ways, which suggests that math anxiety influences performance. Therefore, math anxiety could be measured using a math performance-task. However, Ashcraft (2002) suggests that math anxious individuals might not necessarily perform poorly on math tests. Therefore, comparing math performance and fear for mathematical stimuli might not be a reliable method to validate a picture-set. However, math anxious individuals do experience math anxiety when doing a test (Suárez-Pellicioni et al., 2016). Therefore, comparing math anxiety during a math performance-task to self-reported math anxiety on a picture-task, might be a good method to validate a picture-set. The performance-task itself will be assessed in terms of anxiety and performance.
Research Question and Hypotheses
This paper will provide insight in the relationship between fear for mathematical stimuli and math anxiety. Based upon all the research above, the following was hypothesised: there is a relationship between fear for mathematical stimuli and math anxiety. Two tasks will
be used to measure math anxiety: the AMAS and a performance-task. These two tasks will be used to compare the two types of measured math anxiety with self-reported math anxiety on a picture-task. In the performance-task, anxiety will be increased by increasing pressure on the participants in order to observe the effect of anxiety on performance. Both performance and anxiety will be measured. According to Hopko, Mahadevan, Bare and Hunt (2003) The Abbreviated Math Anxiety Scale (AMAS) is a parsimonious and valid method for measuring math anxiety. Therefore, construct validity of the picture-task can be assessed by comparing outcomes of the AMAS and the picture-task.
It is expected that scores of self-reported anxiety for mathematical pictures will be positively correlated to scores on math anxiety during the performance-task. Because the pictures need to cause variation in the outcome variable and have to be recognised (Pronk et al., 2015), it is expected that scores on the AMAS will be positively correlated to scores of self-reported anxiety for mathematical pictures and scores on the AMAS will not be related to self-reported anxiety for neutral pictures.
It is also hypothesised that fear has a negative effect on performance. Therefore, a positive correlation between anxiety during the performance task and differences in performance is expected, as well as a higher score on performance and a lower score on anxiety in part 1 of the task, compared to part 2.
Furthermore, it is hypothesised that math pictures will trigger a different response than math-aspect pictures and neutral pictures. Scores of fear for mathematical pictures are
expected to be higher than scores of fear for math-aspect pictures and neutral pictures. Scores of fear for math-aspect pictures should also be higher than scores of fear for neutral pictures.
Neutral pictures should not be recognised as math pictures and math-aspect pictures should be recognised less often as math pictures than actual math pictures. Therefore, it is expected that math pictures get a higher score on recognition than math-aspect pictures and
neutral pictures and math-aspect pictures get a higher score on recognition than neutral pictures. Math pictures have to be recognised by at least 70% of the participants. Neutral pictures should be recognised as non-mathematical pictures by at least 70% of the participants.
The final hypothesis is that the performance-task is a valid method to measure math anxiety. Therefore, the scores on the AMAS and anxiety scores on the performance-task should correlate positively as well.
Method Participants
In total, 519 secondary school students participated in this study. The children were between twelve and sixteen years old (M = 14.37, SD = 0.96). 239 of the participants were boys and 280 girls. The participants were recruited through contacting 64 Dutch secondary schools of all levels of education. Four of these schools agreed to participate. Because of time restrictions only two of the schools could be included for analysis. 579 parents, of children in the first three years of secondary school, were asked to let their children participate through passive consent. Only children in the first three years of secondary school could participate, because all of these children have math classes. Seventeen of the parents sent back a letter stating that they refused to let their children participate and 43 students were absent during testing. The participants did not receive a reward for participation.
Materials
Abbreviated Math Anxiety Scale. A Dutch translation of the Abbreviated Math Anxiety Scale (AMAS) was used to assess math anxiety (Hopko et al., 2003). The AMAS contains nine items. An answer to these items could be given on a five-point Likert-type scale,
ranging from 1 (‘hardly/not anxious’) to 5 (very anxious). This results in a mean score ranging from 1 to 5. A higher mean score on the AMAS indicates higher levels of math anxiety. This test has an excellent internal consistency (α = .90) and test-retest reliability (r = .85) (Hopko et al., 2003). In this study reliability of the AMAS was high as well, Cronbach’s α = .88. A Cronbach’s alpha of .7 or higher is found acceptable for psychological constructs (Field, 2013). The different items are shown in Table 1.
Table 1.
Items on the Abbreviated Math Anxiety Scale (AMAS)
Item
1. Having to use tables in the back of a math book 2. Thinking about an upcoming math test 1 day before
3. Watching a teacher work an algebraic equation on the blackboard 4. Taking an examination in a math course
5. Being given homework assignment of many difficult problems that is due the next class Meeting
6. Listening to a lecture in math class
7. Listening to another student explain a math formula 8. Being given a ‘pop’ quiz in math class
9. Starting a new chapter in a math book
Performance-task. Both math anxiety and math performance were measured with a performance-task. The aim of this task was to measure the effect of math anxiety on
two parts. The first part consists of six screens, each showing twelve math problems. In each screen only four of the twelve problems have to be solved. In order to measure math anxiety, participants are asked how anxious they feel at the moment, before and after every set of two screens with math problems.
The second part of the performance-task also consists of six screens. Only this time, four problems are shown per screen and all of the problems have to be solved. Again before and after every set of two screens participants are asked how anxious they feel. Also, the participants are told their answers will be compared to that of others and a timer is shown on the screen. These aspects increase the pressure on the participants.
An example was given before the actual task. In the first part of the task the
participants could choose problems of three different levels of difficulty. The problems that were given in the second part of the task were problems of the same level as the problems that had been chosen before.
The math problems were multiple-choice questions. A correct answer is worth 1 point and a wrong answer 0 points, resulting in a mean score ranging from 0 to 1. A higher mean score indicates higher math performance. Answers on the questions about anxiety could be given through sliding an arrow to one of five faces showing more or less anxious expressions. The least anxious face was assigned 1 point and the most anxious face 5. Therefore, the mean score on these questions range from 1 to 5. A higher mean score indicates higher math
anxiety.
The first anxiety question on part 1 of the task was asked before the participants had to solve math problems. The aim of this task was to examine anxiety during the performance-task. Therefore, the first question in part 1 of the task will not be used in analysis.
Picture-task. Self-reported anxiety for mathematical stimuli was observed through the use of a picture-task. The picture-task consists of three sets of pictures: 21 mathematical pictures, ten math-aspect pictures and five neutral pictures. As mentioned earlier, the different pictures are used to cause variation in anxiety. The task consists of three parts.
The first part is a recognition task (Pronk, et al. 2015), in which participants are asked whether the pictures are considered mathematical or not. Answers are given through pressing either the ‘e’ button (not mathematical) on the keyboard or the ‘i’ button (mathematical). The ‘i’ response is assigned one point and the ‘e’ response zero. For the mathematical set of pictures, the mean score of each picture-set ranges from 0 to 1. A higher score indicates more pictures that are recognized as mathematical.
The second part is a rating task (Pronk et al., 2015) in which participants are asked how anxious they feel after seeing the picture. Answers are given through clicking on any point of a line ranging from ‘not scared at al’ to ‘very scared’. The beginning of the line was assigned 1 point and the end of the line 100. This results into a mean score ranging from 1 to 100 for each picture-set. A higher total score on the mathematical picture-set indicates higher anxiety for mathematical stimuli.
The third part is a familiarity task in which participants are asked whether they have seen the picture in math class or not. Answers are given through pressing either the ‘e’ button (not familiar) on the keyboard or the ‘i’ button (familiar). The ‘i’ response is assigned one point and the ‘e’ response zero. This part was not used for analysis in this paper.
All of the tasks listed above were presented on a computer screen to each participant individually. The pictures in the picture-task were shown separately and in random order. Each picture was shown on the screen for 500 milliseconds. First, all the pictures appeared in the recognition task, then again in the rating task and in the familiarity task only neutral pictures were not shown.
Procedure
The Ethics Committee (CE) has given permission to do this research and ask for participation through passive consent. Each individual school scheduled 50 minutes in a computer room for each participating class separately. Participants were asked to take place behind a computer, if possible, each time leaving one empty seat in between them. All the participants were given a participation number with which they could log into the
questionnaires and other tasks. Participants were asked to carefully read each question and click on the answer that resembled their own feelings most. They were told there were no right or wrong answers and that the answers they gave were confidential. If they had a
question or when they had completed one of the two parts they had to raise their hand. One of the experimenters helped them switch from computer to laptop or the other way around. The picture-task was made on a laptop and the other tasks were made on a computer. A few participants started with the picture-task and then switched to the computer, others finished the tasks in reversed order.
On the computer, the first screen showed a few questions about the sex, age and education level of the participant. After answering these questions, the AMAS, performance-task and another questionnaire, not relevant to this paper, were presented in random order. All of the tasks were explained on the computer screen. The nine questions of the AMAS were presented simultaneously on the screen.
The participants and their teacher were thanked for their participation and either left school or went to the next class.
Results
Two of the participants did not complete the questionnaires and were therefore not included for analysis. The remaining 517 were included for analysis of the AMAS. However,
only 463 of those 517 were included for analysis of the picture-task, because others were not able to finish the task due to time restrictions. Also, due to an error in the performance-task, 120 participants have answered more than 24 questions in total. The extra answers, however, did not appear on the screen. Therefore, some participants got frustrated and started clicking on more different answer options. This could have resulted in more wrong answers, even though the participant might have known the correct answer. It was not possible to retrieve which questions had been answered first and therefore, these 120 participants have been excluded from analysis of the performance-task. In total 397 participants were analysed in the performance-task.
Performance-task
In order to test whether there is a difference in performance for different parts of the task, a paired sample t-test was used on the mean performances of both parts of the test. The assumptions of normal distribution and equal variances were met. The difference in
performance, .012, was not significant t(396) = -1.94, p = .053. Therefore, the mean
performance did not differ. However, it was expected that performance would decrease in part 2 of the task.
To test whether there is a difference in anxiety for different parts of the task, a paired samples t-test was used on the mean anxiety of both parts of the test. The assumptions were met and the difference in anxiety, .15, was significant t(516) = -6.6, p < .001. The mean anxiety was higher in part 2 of the task than in part 1. This was expected. For each part of the task, the mean and standard deviation of performance and anxiety (FAS) were calculated, see Table 2.
Table 2.
Mean Scores and Standard Deviations (between Brackets) of Performance and Anxiety on Part 1 and 2 of the Performance-Task
Part 1 Part 2
Performance .77 (.23) .78 (.22)
FAS 1.42 (.72) 1.56 (.83)
To determine whether anxiety has a negative effect on performance, a few correlations were calculated. There was no significant relationship between FAS1 and the difference in performance, r = -.08, p = .12 and neither for FAS2 and the difference in performance, r = .02, p = .67. It was expected that higher anxiety in both parts of the task would be related to a bigger difference in performance. The difference in math performance was calculated by subtracting the score on part 2 from the score on part 1.
Two explorative analyses were done on the mean performance and the mean FAS for both parts of the task. This was done because a non-significant correlation between the difference in performance and anxiety does not necessarily mean that there is no relationship between anxiety and performance. There was a significant relationship between FAS1 and performance on part 1, r = .32, p < .001, as well as for FAS2 and performance on part 2, r = -.26, p < .001. This means that higher anxiety is related to lower performance. All correlations and p-values are shown in Table 3.
Table 3.
Correlations and P-Values (between Brackets) for the Performance Task
Difference in performance
Performance part 1 Performance part 2
FAS1 -.08 (.12) -.32 (< .001) -
FAS2 .02 (.67) - -.26 (< .001
Picture-task
In order to test whether the different picture-sets differed in mean recognition, a paired samples t-test was used. The assumptions were met for all three of the analyses. The
difference between math pictures and math-aspect pictures, .16, turned out to be significant
t(462) = 8.95, p < .001. These findings are consistent with the expectations. The mean
recognition score on math pictures was higher than the mean recognition score on math-aspect pictures. The difference between math pictures and neutral pictures, -.01, was not significant
t(462) = -.97, p = .332. This was not expected. The recognition score should be higher for
mathematical pictures than for neutral pictures. The difference between math-aspect pictures and neutral pictures, -.17, was significant t(462) = -8.85, p < .001. It was expected to be significant, but then the other way around. Recognition scores for math-aspect pictures should be higher than scores for neutral pictures.
Only pictures that would be either considered mathematical or non-mathematical by 70% of the participants should have been included for analysis. However, none of the neutral stimuli passed that criterion, and only a few mathematical stimuli did. More than 60% of the participants recognised the neutral pictures as math pictures. Therefore, no pictures could be deleted and all pictures have been used for analysis.
To test whether participants were more fearful of math pictures than math-aspect pictures and neutral pictures, paired samples t-tests were used on the mean rating score. The assumptions for all tests were met. The difference in mean rating between math pictures and math-aspect pictures, 8.94, was significant t(462) = 15.18, p < .001. The mean rating score was higher for math pictures than math-aspect picture. This was expected.
The difference in mean rating between math pictures and neutral pictures, 17.32, was significant t(462) = 16.93, p < .001. The mean rating score was higher for math pictures than neutral pictures. This was also expected.
The difference in mean rating between math-aspect pictures and neutral pictures, 8.38, was also significant t(462) = 13.56, p < .001. The mean rating score was higher for math-aspect pictures than neutral pictures. This was expected.
The mean and standard deviation of the recognition task and rating task were calculated for each picture-set, see Table 4.
Table 4.
Mean Scores and Standard Deviations (between Brackets) of Recognition and Rating task for each picture-set
Recognition Rating
Math pictures .67 (.36) 25.48 (22.34)
Math- aspect Pictures .51 (.18) 16.54 (14.94)
Comparisons
In order to examine the relationship between fear for mathematical stimuli and math anxiety, correlations were calculated between the different tasks. All correlations and p-values are given in Table 5.
Table 5.
Correlations and P-Values (between Brackets) for the Different Tasks
AMAS FAS1 FAS2 Rating
math Rating aspects Rating neutral AMAS - .47 (< .001) .46 (< .001) .63 (< .001) .56 (< .001) .22 (< .001) FAS1 .47 (< .001) - - .43 (< .001) .38 (< .001) .21 (< .001) FAS2 .46 (< .001) - - .46 (< .001) .42 (< .001) .22 (< .001)
Scores on the AMAS were significantly related to FAS1, r = .47, p < .001. There was a significant relation between scores on the AMAS and FAS2 as well, r = .46, p < .001. This means that high scores on the AMAS are related to high scores on anxiety during both parts of the performance-task. This was expected.
Higher scores on the AMAS were, as expected, significantly related to higher rating scores on mathematical stimuli, r = .63, p < .001. However, higher scores on the AMAS were also significantly related to higher rating scores on neutral stimuli, r = .22, p = < .001. This was not expected.
Higher rating scores for mathematical stimuli were significantly related to higher FAS1 scores, r = .43, p < .001. Higher rating scores for mathematical stimuli were also significantly related to higher FAS2 scores, r = .46, p < .001. These findings are consistent with the expectations. However, fear for neutral stimuli was also significantly related to both FAS1, r = .21, p < .001, and FAS2, r = .22, p < .001. This suggests that higher rating scores on neutral pictures are related to higher anxiety scores on both parts of the performance-task. This was not expected.
Discussion
The aim of this research, was to examine the relationship between fear for
mathematical stimuli and two types of measured math anxiety, in order to validate a set of pictures for future use in priming techniques. The results show that there is a positive
relationship between the AMAS and anxiety during a performance-task, as well as the AMAS and fear for mathematical stimuli. Fear for mathematical stimuli is also positively related to anxiety during a performance-task. However, fear for neutral stimuli was also related to both types of measured math anxiety. This could have happened because a lot of participants recognized neutral stimuli as mathematical stimuli. Still, this problem is not very serious, for math pictures had a much higher correlation with measured math anxiety than neutral
pictures, or math-aspect pictures.
Criterion of Recognition
It was hypothesised that neutral pictures should not be recognised as math pictures and math-aspect pictures should be recognised less often as math pictures than actual math
pictures. However, neutral pictures were just as often recognised as math pictures, compared to mathematical pictures. This means that the criterion that participants need to be able to
recognise whether a picture contains math or not, is not met (Pronk, et al. 2015). Therefore, the picture-sets cannot yet be validated. First, a distinction of pictures must be developed, which most participants agree on. In this study, symbols like a smiley and a heart were used as neutral stimuli. In future research other types of neutral pictures could be used, for instance landscapes instead of symbols.
Criterion of Variation
It was also hypothesised that math pictures will trigger a different response than math-aspect pictures and neutral pictures. This hypothesis is supported by the results. Participants were more fearful of math pictures than math-aspect pictures or neutral pictures. Therefore, the variation in stimuli cause variation in the outcome variable (Pronk et al., 2015). This is supported by the theory behind priming tasks, where shortly presented stimuli are used to trigger a certain behavioural or emotional response (Henson, Eckstein, Waszak, Frings, & Horner, 2014).
Effect of Fear on Performance
The hypothesis that fear has a negative influence on performance could not be supported by the results. Even though anxiety differed in both parts of the task, it seems like anxiety did not affect performance. However, anxiety and performance were related in the specific parts of the task. According to Beilock (2008), stressful situations do influence math performance. However, the first part of the task could already have been quite stressful. The difference in stress and anxiety between the first part of the task and the second part might therefore not be big enough to cause an effect. Also, the results could have been influenced by a practice effect. A practice effect influences response time and response caution, which could
result in better answers (Dutilh, Vandekerckhove, Tuerlinckx, & Wagenmakers, 2009). Maybe this effect caused the results of the second part of the task to be higher than expected.
Performance Task to Measure Anxiety
The hypothesis that the performance-task is a reliable method to measure math anxiety is supported by the results. The task correlated positively with both the AMAS and fear for mathematical stimuli. This is supported by Suárez-Pellicioni et al. (2016), who stated that math anxious individuals experience math anxiety when doing a test.
Shortcomings
A lot of participant solved too many math problems during the performance-task and therefore had to be excluded from analysis. This could have affected the results. When more than four answers were given, the extra answers were not shown on the screen. Therefore, some participants started clicking on other options, because they thought the task had crashed. It is not possible to verify that the extra answers given, were answers the participants thought to be correct. Therefore, excluded participants might have made more mistakes than included participants, which could have resulted in different conclusions. It is very important that in future research a browser is used in which this problem does not occur. During testing we found out that Google Chrome did not have this problem, however the lay-out in Google Chrome was very different from that of Internet Explorer. If the lay-out could be changed, it would be encouraged to use Google Chrome as browser.
Another problem, as mentioned earlier, had to do with the picture-sets. The
participants were not able to discriminate between math pictures and neutral pictures. This has probably resulted in higher scores on anxiety for neutral pictures, which can also explain the positive correlation between fear for neutral pictures and two types of measured math anxiety. This could have happened due to the question. Participants were asked whether a picture was
mathematical or not. Because all the tasks had to do with math, they could have been primed into thinking that the pictures were all supposed to resemble math. To avoid this in future research, a scale could be used on which participants indicate how much pictures resemble math. Also tasks about other subjects could be included to avoid a priming effect.
However, his shortcoming does not cause a very serious problem, because stimuli did differ in how much anxiety was induced. This means that the pictures are probably valuable to use in implicit tasks.
Conclusion
Despite a few shortcomings and unsuspected results this study is very valuable. Normally, participants aged between eleven and sixteen years old are very difficult to test. In most research settings they do not pay attention during instructions or discuss answers with their peers. However, most participants in this study were very motivated and during testing the classrooms were very quiet. This was thanks to a few teachers who explained to the students what the importance of this study is. This was observed in the comments participants made after the tasks. It can therefore be concluded that the results in this study are very reliable.
The most important conclusion of this study, is that there is indeed a relationship between fear for mathematical stimuli and two types of measured math anxiety. This means that pictures can be used in priming techniques in order to measure math anxiety implicitly.
This study can be seen as an important step in the research of math anxiety. If in future research picture-sets are used which meet all criteria, math anxiety can be measured implicitly and math anxiety can finally be examined as a whole of two parts.
References
Aarnos, E., & Perkkilä, P. (2012). Early signs of mathematics anxiety? Procedia Social and
Behavioral Sciences, 46, 1495-1499.
Ashcraft, M. H. (2002). Math anxiety: Personal, educational, and cognitive consequences.
Current Directions in Psychological Science, 11, 181-185.
Beilock, S. L. (2008). Math performance in stressful situations. Current Directions in
Psychological Science, 17, 339-343.
Chipman, S., Krantz, D. H., & Silver, R. (1992). Mathematics anxiety and science careers among able college women. Psychological Science, 3, 292-295.
Dutilh, G., Vandekerckhove, J., Tuerlinckx, F., & Wagenmakers, E. J. (2009). A diffusion model decomposition of the practice effect. Psychonomic Bulletin & Review, 16, 1026-1036.
Field, A. (2013) Discovering statistics using IBM SPSS statistics. (4th ed.) London: Sage Henson, R. N., Eckstein, D., Waszak, F., Frings, C., & Horner, A. J. (2014). Stimulus-
response bindings in priming. Trends in Cognitive Sciences, 18, 376-384.
Hopko, D. R., McNeil, D. W., Gleason, P. J., & Rabalais, A. E. (2002). The emotional stroop paradigm: Performance as a function of stimulus properties and self-reported
mathematics anxiety. Cognitive Therapy and Research, 26, 157-166.
Hopko, D. R., Mahadevan, R., Bare, R. L., & Hunt, M. K. (2003). The abbreviated math anxiety scale (AMAS) Construction, validity and reliability, Assessment, 10, 178-182. Jong, P. de, Hout, M. van den, Rietbroek, H., & Huijding, J. (2003). Dissociations between
implicit and explicit attitudes toward phobic stimuli. Cognition and Emotion, 17, 521-545.
Kahneman, D. (2011). Thinking, fast and slow. New York: Farrar, Straus and Giroux. Mammarella, I. C., Hill, F., Devine, A., Caviola, S., & Szűcs, D. (2015). Math anxiety and
developmental dyscalculia: A study on working memory processes. Journal of
Clinical and Experimental Neuropsychology, 37, 878-887.
Pronk, T., Deursen, D. S. van, Beraha, E. M., Larsen, H., & Wiers, R. W. (2015). Validation of the Amsterdam beverage picture-set: A controlled picture-set for cognitive bias measurement and modification paradigms. Alcoholism: Clinical and Experimental
Research, 39, 2047-2055.
Roos, A. L., Bieg, M., Goetz, T., Frenzel, A. C., Taxer, J., & Zeidner, M. (2015).
Experiencing more mathematics anxiety than expected? Contrasting trait and state anxiety in high achieving students. High Ability Studies, 26, 245-258.
Rubinsten, O., Bialik, N., & Solar, Y. (2012). Exploring the relationship between math anxiety and gender through implicit measurement, Frontiers in Human Neuroscience,
6, 1-11.
Rubinsten, O., Eidlin, H., Wohl, H., & Akibli, O. (2015). Attentional bias in math anxiety.
Frontiers in Psychology, 6, 1-9.
Simon, D., Kischel, E., Spielberg, R., & Kathman, N. (2012). A pilot study on the validity of using pictures and videos for individualized symptom provocation in obsessive– compulsive disorder. Psychiatry Research, 198, 81-88.
Suárez-Pellicioni, M., Núñez-Peña, M. I., & Colomè, À. (2016). Math anxiety: A review of its cognitive consequences, psychophysiological correlates, and brain bases. Cognitive
Affective Behavioral Neuroscience, 16, 3-22.
Sun, R., Sluzarz, P., & Terry, C. (2005). The interaction of the explicit and implicit in skill learning: A dual process approach. Psychological Review, 112, 159-192.
