Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

Applied Cognitive Psychology, Appl. Cognit. Psychol.

(2014)
Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/acp.3058

A Quick and Easy Strategy to Reduce Test Anxiety and Enhance Test Performance

MYRTO-FOTEINI MAVILIDI1, VINCENT HOOGERHEIDE1 and FRED PAAS1,2*


1
Institute of Psychology, Erasmus University Rotterdam, Rotterdam, The Netherlands
2
Early Start Research Institute, University of Wollongong, Wollongong, Australia

Summary: The negative thoughts that anxious children experience while sitting for an exam consume working memory resources
at the cost of resources for performing on the exam. In a randomized field experiment (N = 117) with primary school students, we
investigated the hypothesis that stimulating students to look through the problems of a math test before they start solving them
would reduce anxiety, release these anxiety-related working memory resources, and lead to higher test performance than not
allowing students to look ahead in the test. The results confirmed the hypothesis, indicating that the positive effects of looking
ahead applied to all students, regardless of their anxiety level (low, medium, or high). The results suggest that by looking ahead in
a test, less working memory resources are consumed by intrusive thoughts, and consequently, more resources can be used for
performing on the test. Theoretical and practical implications of the results are discussed. Copyright © 2014 John Wiley & Sons, Ltd.

People’s educational and professional careers are to a large perform up to their maximum potential. This article presents
extent determined by the outcomes of examinations that an experimental study into the effects of an instructional
evaluate the quality and quantity of their learning. Past strategy that allows students to look ahead in a test at the start
research has shown that test anxiety represents a major threat of it, on anxiety, working memory load, and test performance.
to students’ performance on examinations (e.g., Ashcraft, According to cognitive load theory (CLT; Paas, Renkl, &
2002; Ashcraft & Krause, 2007; Dutke & Stober, 2001; Sweller, 2003; Sweller, 1988; Sweller, Ayres, & Kalyuga,
Eysenck, Derakshan, Santos, & Calvo, 2007). Test anxiety 2011), human cognitive architecture consists of a working
may lead to so-called choking under pressure (e.g., Baumeister, memory that is limited in capacity to 4 ± 1 elements of infor-
1984), which has been described by Ramirez and Beilock mation (Baddeley, 1986; Cowan, 2001; Miller, 1956) and
(2011) as a situation in which people perform more poorly than duration to about 20 seconds (Peterson & Peterson, 1959)
expected given their skill level, especially when there are and that interacts with a comparatively unlimited long-term
positive consequences for good performance and negative memory. Central to CLT is the notion that working memory
consequences for poor performance. Such a situation will architecture and its limitations should be a major conside-
typically result in intrusive thoughts, which incorporate nega- ration when designing instruction. These limitations are
tive perceptions related to expectancies, past experiences ameliorated by changes in long-term memory associated with
(Meece, Wigfield, & Eccles, 1990), and value judgments such learning (Paas & Sweller, 2012; Sweller, Van Merriënboer, &
as worries, thoughts of low self-esteem and self-confidence, Paas, 1998). During learning, new information is temporarily
feelings of having to write a difficult and unfamiliar test, prob- stored in working memory, and, if rehearsed, coded, or
ability of failure, sense of inefficacy and of being unprepared for imagined, it can be permanently stored in cognitive schemas
a test, negative social comparisons with fellow students, and a in long-term memory (Atkinson & Shiffrin, 1971).
wish to elude from the test situation (Zatz & Chassin, 1985). To perform well on a test, information that was previously
It is assumed that these anxiety-related intrusive thoughts stored in long-term memory needs to be activated and
or worries about the situation and its consequences consume retrieved, after which it can be manipulated in working memory.
some of the limited working memory resources, which con- Complex tasks, such as equations in mathematics or chemistry,
sequently cannot be used for activities that are relevant for typically consist of a high number of interacting information
performing successfully on the test (Beilock & Carr, 2005; elements and consequently require more working memory
Grimley & Banner, 2008). During tests, high anxious people capacity (i.e., impose a higher cognitive load) for the information
need significantly more response time and greater effort to manipulation process than easy tasks (Sweller et al., 2011).
achieve the same performance level than low anxious Every other process that simultaneously competes for work-
people. This happens because, in accordance with selective ing memory capacity will interfere with performance on the
attentional mechanisms, high anxious children tend to focus test, especially for complex problems. These competing pro-
more on emotionally threatening stimuli such as failure and cesses can vary from monitoring information in the environ-
mistake and less on the required task (Ramirez & Beilock, ment (e.g., Vredeveldt, Hitch, & Baddeley, 2011) to monitor-
2011; Vasey, El-Hag, & Daleiden, 1996). Because anxious ing distracting information that is generated by oneself, such as
students often perform below their ability, it is considered thoughts about failure (e.g., Ramirez & Beilock, 2011).
important to discover instructional strategies that can Cognitive load theory focuses on an optimal use of
manage the student’s anxiety level so that they can use all working memory resources and the prevention of cognitive
working memory resources for task-relevant processes and overload in learning, especially in complex tasks (Paas &
Van Merriënboer, 1994a). It is concerned with the develop-
ment of instructional methods that stimulate people’s learning
*Correspondence to: Fred Paas, Institute of Psychology, Erasmus University
Rotterdam, PO Box 1738, 3000 DR Rotterdam, The Netherlands.
and transfer of acquired knowledge and skills by efficiently
E-mail: paas@fsw.eur.nl using their limited cognitive processing capacity (Paas, Renkl,

Copyright © 2014 John Wiley & Sons, Ltd.


M.-F. Mavilidi et al.

et al., 2003; Paas, Renkl, & Sweller, 2004; Sweller, 1988, could reduce anxiety and its negative effects on test perfor-
1999). In cognitive load research, the perceived amount of mance in sixth-grade primary school children. Similar strate-
invested mental effort is considered a reliable estimator of gies, such as skimming a text before reading, have been
the working memory capacity that is consumed, or the found to facilitate the construction of mental representations
cognitive load that is imposed by the learning task (Paas, through the mapping between the structure and content of
Tuovinen, Tabbers, & Van Gerven, 2003; Paas & Van information (Duggan & Payne, 2009; Payne & Reader,
Merriënboer, 1994b). 2006). With regard to the looking ahead strategy, it was
Cognitive load theory categorizes two types of cognitive expected that skimming the test problems would activate
load, that is, intrinsic and extrinsic, which together determine the cognitive schemas that are necessary to solve the prob-
the total cognitive load imposed by the task on the learner’s lems and reduce the students’ anxiety. Consequently, less
cognitive system (Sweller, 2010; Sweller et al., 2011). Intrin- working memory resources would be occupied by worries
sic cognitive load is generated by the materials to be learned or intrusive thoughts about failure and more working mem-
and is imposed by the extent to which essential elements of ory resources would be available for task-relevant processes.
information interact. If there are many interacting elements Therefore, our main hypothesis was that, for low, medium,
that are intrinsic to the task, intrinsic cognitive load is high, and high anxiety students, those who were allowed to look
and subsequently, that study material is difficult to be learned. ahead in a test at the start of it would experience less anxiety
Extrinsic or extraneous cognitive load is generated by the and achieve higher test performance than those who were not
instructional procedures that accompany the to-be-learned allowed to look ahead in the test. Although no differences in
materials, the physical learning environment (Choi, Van overall working memory load were predicted between the
Merriënboer, & Paas, 2014), and the learners’ affective low, medium, and high anxious students, it was expected
responses, such as anxiety, to these procedures and the that the students, who would have the opportunity to look
learning environment. An essential assumption of CLT is that ahead in the test, would show a more favorable relationship
while working on complex tasks, the total pool of working between working memory resources used for performing
memory resources is divided among activities that are relevant on the test (i.e., intrinsic cognitive load) and working mem-
to task performance (i.e., intrinsic load) and activities that are ory resources used for anxiety-related processes (i.e., extra-
not relevant to task performance (i.e., extraneous load). This neous cognitive load) than the students who would not have
means that resources that are used for activities that impose the opportunity to look ahead in the test.
an extraneous load are not available for activities that impose
an intrinsic load. Therefore, an important goal of instructional
design according to CLT is to optimize the relationship METHOD
between intrinsic and extraneous load by minimizing extrinsic
cognitive load so that as much as possible working memory Participants
resources can be devoted to intrinsic cognitive load (Sweller The experiment was carried out in three Greek primary
et al., 2011). schools in the city of Athens. Participants were 117 sixth-
The intrusive thoughts that result from anxiety on a test grade children (51 boys and 66 girls) 11–12 years of age
can be considered an important source of extraneous cogni- (M = 11.59, SD = 0.44). Parents gave their consent after
tive load that should be reduced to release resources that being informed about the experiment by the teachers. The
can be used to deal with intrinsic load, thereby enhancing children were divided into low, medium, and high anxiety
test performance. In a recent study, Ramirez and Beilock groups on the basis of their scores on the cognitive anxiety
(2011) examined whether the strategy of writing down test questionnaire (Cassady & Johnson, 2002). The cognitive
anxiety-related intrusive thoughts before a high-pressure test anxiety questionnaire was administered 1 week before the
could be used as a strategy to improve performance. In four start of the experiment during a usual school day in a
experiments with ninth-grade students, they compared three classroom-based setting. In each school, the child that
conditions, in which students were asked either to write achieved the highest test score received a gift card of €10,
about their test-related thoughts for 10 minutes immediately as a prize. Participants were matched for anxiety level and
before a math exam or to write about an irrelevant unemo- randomly assigned to either the control or the experimental
tional subject or to sit down just quietly. Their hypothesis condition. This resulted in 20 low anxious, 18 medium
that students’ exam performance would improve most when anxious, and 20 high anxious students in the control
they expressed their intrusive thoughts and worries before condition (nonlooking ahead strategy; N = 58), and 19 low
the test about the forthcoming exam was confirmed by the anxious, 21 medium anxious students, and 19 high anxious
results. The expressive writing about worries and not just students in the experimental condition (looking ahead
writing itself was a beneficial strategy for reducing anxiety strategy, N = 59).
in high-pressure situations, especially for high anxious
students. The results of this study suggest that reducing the
Design
working memory load that is imposed by intrusive thoughts
improves students’ performance. The dependent variables math test performance, invested
In this study, a randomized controlled classroom study mental effort to achieve that performance, and total test task
was conducted to examine whether a ‘looking ahead’ complexity were used in a 2 (test strategy; looking ahead vs.
strategy, in which students were allowed to look shortly nonlooking ahead) × 3 (anxiety level; low vs. medium vs.
through a math test before starting to solve the test problems, high) between-subjects design. Moreover, in a mixed

Copyright © 2014 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. (2014)
A strategy to reduce anxiety

design, time of measurement of children’s anxiety during


the test (at the start, during, and at the end) was used as a
within-subjects factor and anxiety level and test strategy
as between-subjects factors.

Materials
Overview of materials
Booklets were prepared for the cognitive anxiety test
questionnaire and for the math test. A leaflet with general
instructions about the content and process of the test was
created, which included two examples of test problems, as
well as rating scales of task complexity and mental effort,
and a draft paper.

Measurement of anxiety
A Greek translation of the cognitive anxiety test question-
naire (Cassady & Johnson, 2002) was used as an instrument
to measure the students’ level of test anxiety. The test was Figure 1. Two examples of test problems
carefully introduced and explained to the children by the
researcher until she was assured that all the children under-
stood the questionnaire. Measurement of task complexity
Cassady and Johnson (2002) have shown that the ques- After each test problem, the children had to rate the level of
tionnaire has a high internal consistency (α = .86). The ques- complexity on a 9-point scale ranging from 1, very very low
tionnaire consisted of 27 items that measured trait anxiety on task complexity, to 9, very very high task complexity (Sweller
a 4-point item scale ranging from 1, not typical of me, to 4, et al., 2011).
very typical of me. They were asked statements such as ‘I
lose sleep over worrying about examinations’, or ‘I have less Measurement of mental effort
difficulty than the average college student in learning After each test problem, the mental effort that children
assigned chapters in textbooks’. Also, statements about the invested for each test problem was assessed through the
relationship between cognitive test anxiety and performance mental effort scale of Paas (1992; see also Paas, Tuovinen,
were included, that is, ‘finding unexpected questions on a et al., 2003; Paas & Van Merriënboer, 1994a), which
test causes me to feel challenged rather than panicky’, or consists of a 9-point symmetrical category scale ranging
‘when I first get my copy of a test, it takes me a while to calm from 1, very very low mental effort, to 9, very very high men-
down to the point where I can begin to think straight’. The tal effort. The amount of mental effort invested by children
minimum score that a child could achieve was 27 and the per problem was decoded into a numerical value, and for
maximum 108. the whole test, the average amount of mental effort invested
was calculated.

Math outcome measure


Procedure
Previous math grades assessed by the children’s teachers in
the current semesters were taken into consideration for The study was conducted in the schools’ classrooms by one
estimating students’ prior knowledge and mathematical experimenter (i.e., first author). The teachers handed out the
background. For both conditions, the math test consisted of anxiety questionnaire to the students. On the basis of the
10 open-ended problems, selected from a sample of the students’ responses in the questionnaire, they were categorized
California standard tests (STAR, nd; http://www.cde.ca. as low, medium, and high anxious students and randomly
gov/ta/tg/sr/documents/cstrtqmath6.pdf) and test problems assigned to either the control or experimental condition. One
created by the Dutch National Institute for Educational week after the anxiety questionnaire had been administered
Measurement (CITO, nd) for the end of primary school test to students, they sat the math test. Before the start of the test,
(www.eindtoets.nl). It incorporated problems that covered the experimenter explained the procedure of the test in
the following topics from a syllabus that children had 15 minutes. It was explained that no further clarifications
studied throughout the year: simple algebraic equations, or explanations would be given during the test. A countdown
geometry, ratios, and combined problems. Examples of timer was positioned in a central place in the classroom in
the test problem are shown in Figure 1. The overall order to be easily accessible to all students and induce anxiety.
complexity of the test was considered high because of the Participants worked on the 10 test problems one by one. When
number of interacting information elements in each prob- the time was up for a problem, the experimenter indicated
lem (Sweller & Chandler, 1994) and the time pressure that that the students had to move on to the next problem. The
was imposed on the participants. For each correct answer, control group had 4 minutes for solving the first problem
students achieved one point. and 3 minutes for each of the nine remaining problems.

Copyright © 2014 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. (2014)
M.-F. Mavilidi et al.

The experimental group had 1 minute to look through all the was not significant, F(1, 111) < 1, p = .739. Tukey’s post-
test problems at the start of the test and 3 minutes for solving hoc tests revealed that low anxious students (M = 6.00,
each test problem. At the end of each problem, participants SD = 1.69) performed significantly better than high anxious
indicated on the rating scales how complex they perceived students (M = 4.85, SD = 1.82), t(2) = 2.84, p = .015, r = .07.
the problem and how much mental effort they invested in Moreover, medium anxious students (M = 5.89, SD = 2.06)
it. At the end of the test, participants rated on the anxiety performed significantly better than high anxious students
scale how anxious they were at the beginning of the test (M = 4.85, SD = 1.82), t(2) = 2.56, p = .032, r = .05. Finally,
(when starting to solve the first test problem), during the test math performance of low anxious students (M = 6.00,
(while solving the problems), and at the end of the test (after SD = 1.69) and medium anxious students (M = 5.89, SD = 2.06)
solving the last test problem). The whole process lasted did not differ significantly, t(2) < 1, p = .956, r = .00 The main
45 minutes and was conducted first in one class and subse- effect of test strategy indicated that the experimental condition
quently in the other class. The delegation of conditions in (M = 6.14, SD = 1.71), which allowed children to look ahead in
the two classrooms was counterbalanced among different the test, performed significantly better on the math test than the
schools. It was also ensured that communication between control condition (M = 5.00, SD = 1.96), which did not allow
students of different classrooms was not possible while children to look ahead in the test. The effect sizes observed
conducting the experiment. for the test strategy were moderate to large (d = 0.62).
Finally, in addition to the anxiety induced by the limited
time, anxiety was induced by telling the students that they Mental effort
were matched with a fellow student from the other class
who had already sat the test, creating a group of two With regard to the mental effort that children invested when
students. They were also told that their partner had received completing the test, a two-way ANOVA was conducted with
a satisfactory grade in the test and subsequently the possibility test strategy and anxiety level as between-subject factors.
of earning the prize would depend on them. This procedure Table 1 portrays the means of invested mental effort as a
was adopted from the study of Ramirez and Beilock (2011). function of test strategy and anxiety level. Results showed that
Ecological validity was ensured by testing participants under neither the main effects of test strategy, F(1, 111) = 3.45,
normal classroom conditions, in which the math test was p = .066, η2p = 0.03, also explained by the small effect sizes
administered as a regular examination. observed (d = 0.02), and anxiety level, F(1, 111) = 1.03,
p = .359, η2p = 0.01, nor the interaction between test strategy
and anxiety level, F (1, 111) < 1, p = .692, was significant.
RESULTS

The assumptions of analysis of variance (ANOVA) were Task complexity


tested, and no violations were found in the sample. An For analyzing children’s ratings regarding the task complex-
ANOVA performed on the children’s previous math grades ity of the test, a two-way ANOVA was conducted with test
revealed that there was no difference as a function of strategy and anxiety level as between-subject factors. Table 1
experimental condition (looking ahead vs. nonlooking portrays the means in task complexity as a function of test
ahead), F(1, 115) < 1, p = .541. In addition, the ANOVA strategy and anxiety level. There was no significant main
performed on children’s anxiety level, as evaluated by the effect of test strategy, F(1, 111) < 1, p = .353, while the
cognitive anxiety test questionnaire before the test, revealed effect sizes observed were small (d = 0.17), nor a signifi-
no differences between the looking ahead and nonlooking cant interaction between test strategy and anxiety level,
ahead conditions, F(1, 115) < 1, p = 1.0. F(1, 111) < 1, p = .162. There was, however, a significant main
effect of anxiety level, F(1, 111) = 4.35, p = .015, η2p = 0.07.
Math test performance Tukey’s post-hoc tests revealed that low anxious students
Children’s anxiety level was negatively related to children’s (M = 4.20, SD = 1.34) experienced the test as less complex than
previous math grades, r = .20, p = .028, and their grades on high anxious students (M = 5.03, SD = 1.27), t(2) = 2.92,
the test in this study, r = .25, p = .007. This means that the p = .012, r = .07. At the same time, medium anxious students
higher the anxiety of children, the lower their mathematical (M = 4.51, SD = 1.13) experienced the test just as complex as
performance on previous tests and on the current test. Also, high anxious students (M = 5.03, SD = 1.27), t(2) = 1.84,
there was a positive relationship between children’s previous p = .162, r = .02. Finally, low anxious students (M = 4.20,
math grades and their performance on the test in this study, SD = 1.34) and medium anxious students (M = 4.51, SD = 1.13)
r = .53, p < .001. did not differ in their perception of the complexity of the prob-
Data on math performance were analyzed with a two-way lems, t(2) = 1.08, p = .527, r = .01.
ANOVA, with test strategy and anxiety level as between-
subjects factors. Table 1 shows the mean scores on the math Anxiety at the beginning, during, and at the end of the test
test as a function of test strategy and children’s anxiety level.
Table 2 shows the mean differences in children’s anxiety at
The main effects of test strategy, F(1, 111) = 11.50, p = .001,
the beginning, during, and at the end of the test.
η2p = 0.09, and children’s anxiety level, F(1, 111) = 4.73, A mixed ANOVA was performed with time of measuring
p = .011, η2p = 0.08, were found significant. However, the anxiety during the test (at the start, during, and at the end) as
interaction between test strategy and children’s anxiety level within-subjects factor, and test strategy and anxiety level as

Copyright © 2014 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. (2014)
A strategy to reduce anxiety

Table 1. Means (SD) of math performance, mental effort, and task complexity as a function of children’s anxiety level and test strategy
Looking ahead condition Nonlooking ahead condition

Low anxious Medium anxious High anxious Total Low anxious Medium anxious High anxious Total
Math performance 6.68 (1.48) 6.48 (1.70) 5.24 (1.67) 6.14 (1.71) 5.35 (1.65) 5.19 (2.28) 4.48 (1.92) 5.00 (1.96)
Mental effort 4.43 (1.37) 4.27 (1.34) 4.87 (1.19) 4.52 (1.31) 4.80 (1.70) 5.02 (1.11) 5.13 (1.09) 4.98 (1.32)
Task complexity 4.11 (1.34) 4.33 (1.25) 5.03 (1.27) 4.47 (1.34) 4.28 (1.39) 4.72 (0.96) 5.07 (1.21) 4.69 (1.23)

Table 2. Means (SD) of children’s anxiety at the beginning, during, and at the end of the test as a function of test strategy and their general
anxiety level
Looking ahead condition Nonlooking ahead condition

Low anxious Medium anxious High anxious Total Low anxious Medium anxious High anxious Total
Start of test 2.79 (2.59) 3.33 (2.13) 3.63 (2.14) 3.25 (2.28) 2.60 (2.35) 4.89 (2.56) 5.95 (2.48) 4.47 (2.81)
During test 3.84 (2.54) 3.90 (2.19) 4.05 (2.24) 3.93 (2.29) 2.55 (2.58) 4.17 (1.79) 6.00 (2.29) 4.24 (2.65)
End of test 2.26 (2.00) 2.48 (1.99) 2.42 (2.38) 2.39 (2.09) 1.75 (1.83) 2.56 (2.28) 3.85 (2.78) 2.72 (2.46)
Total 2.96 (2.38) 3.24 (2.10) 3.37 (2.25) 3.19 (2.22) 2.30 (2.25) 3.87 (2.21) 5.27 (2.52) 3.81 (2.64)

between-subjects factors. Mauchly’s test indicated that the effective for reducing anxiety and enhancing performance.
assumption of sphericity was not violated, χ(2) = 5.04, It is assumed that the working memory resources that were
p = .081. Results revealed significant main effects of time of released through the reduction of anxiety could be used for
measuring anxiety, F(2, 222) = 30.13, p < .001, η2p = 0.21, task-relevant processes. It is believed that the short (1 minute)
and anxiety level, F(2, 111) = 7.87, p = .001, η2p = 0.12. How- looking ahead intervention had both cognitive and affective
effects. From a cognitive point of view, the results suggest
ever, the main effect of test strategy, F(1, 111) = 3.20,
that skimming the test problems before starting to work on
p = .076, η2p = 0.02, was not significant. The interactions
them activated the relevant cognitive schemas for solving
between time of measurement and anxiety level, F(4, the test problems. From an affective point of view, the results
222) = 1.22, p = .304, η2p = 0.02, time of measurement and test suggest that the students who skimmed the problems knew
strategy, F(2, 222) = 3.02, p = .051, η2p = 0.03, and time of mea- what to expect on the test and might have felt more confident
surement, anxiety level, and test strategy, F(4, 222) < 1, (see also Jansen et al., 2013; Protheroe & Perkins-Gough,
p = .647, η2p = 0.01, were not significant. The main effect of 2000). The use of strategies and students’ self-efficacy beliefs
time of measurement indicated that children’s anxiety differed are determining factors of problem-solving accuracy and effi-
as a function of the time of measurement. Follow-up analyses ciency and consequently math performance, especially in
revealed that anxiety at the beginning of the test was not signi- high complex tasks (Hoffman, 2010; Hoffman & Schraw,
ficantly different from anxiety during the test, F(1, 111) = 1.34, 2009; Hoffman & Spatariu, 2008; Pajares & Kranzler,
p = .249, η2p = 0.01. However, anxiety at the beginning, 1995). Social cognitive theory advocates that self-efficacy is
related to personal beliefs about performing and achieving
F(1, 111) = 33.19, p < .001, η2p = 0.23, as well as anxiety
the expected gains (Bandura, 1986). Students demonstrating
during the test, F(1, 111) = 48.18, p < .001, η2p = 0.30, high self-efficacy are more likely to invest more effort and
was significantly different from anxiety at the end of the show more persistence in solving complex math tasks
test. The main effect of anxiety level indicated that anxiety (Pajares & David, 1994; Pajares & Graham, 1999; Pajares
scores were differentially affected for low, medium, and & Kranzler, 1995). Future research could unravel the effects
high anxious students. Follow-up analyses indicated that of anxiety reduction and schema activation, by adding an ex-
whereas low anxious students (M = 2.63, SD = 2.31) were perimental condition in which anxiety is reduced, but
less anxious than high anxious students (M = 4.32, schemas are not activated. An example of such a condition
SD = 2.39), t(2) = 4.04, p < .001, r = .13, there was no is the writing intervention that was used by Ramirez and
difference between low anxious and medium anxious Beilock (2011), who asked students to write down their
students (M = 3.56, SD = 2.16), t(2) = 2.13, p = .088, worries about the test before it started.
r = .04, and between medium anxious and high anxious Although the overall mental effort invested in the test did
students, t(2) = 1.90, p = .141, r = .03. not differ between both conditions, the results on perfor-
mance indicate that the students in the looking ahead condi-
tion were more efficient (Paas & Van Merrienboer, 1993;
DISCUSSION Van Gog & Paas, 2008) in the sense that they achieved
higher performance with the same mental effort investment.
We demonstrated that an instructional strategy that allowed This suggests that there was a more favorable relationship
low, medium, and high anxious students to look ahead in a between extraneous and intrinsic cognitive load in the
math test before starting to solve the test problems was looking ahead condition than in the nonlooking ahead

Copyright © 2014 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. (2014)
M.-F. Mavilidi et al.

condition. In the latter condition, mental effort scores imply their answers on the trait anxiety questionnaire and after-
that more working memory resources were allocated to wards linked to their state anxiety through their rating scale
anxiety-related worries about failure in high anxious scores. This might explain why anxiety level was not an
students. Unfortunately, the mental effort rating scale does important contributor to the interactions between test
not allow us to differentiate between different types of strategy in performance and mental effort invested. In future
cognitive load. Future studies could use a newly developed research, using a real measure of children’s state anxiety
instrument for measuring different types of cognitive load could be more indicative for the relationship between
(Leppink, Paas, Van der Vleuten, Van Gog, & Van anxiety, performance, and test strategy. The current study
Merriënboer, 2013). was conducted on examining students’ math performance.
Interestingly, recent clinical anxiety studies have used Other STEM domains such as science, technology, and engi-
hypotheses and found results that are consistent with the neering in older students, as well as language, vocabulary,
theoretical framework of CLT used in this study. For and text comprehension of first and second language in
example, Amir and Bomyea (2011) showed that nonanxious younger students, could be further investigated in terms of
individuals demonstrated better working memory capacity the effectiveness of the looking ahead intervention. In addi-
than individuals with generalized social phobia for neutral tion, it would be interesting to determine whether the effects
words but not for social threat words. Individuals with of the looking ahead strategy generalize to older students.
generalized social phobia demonstrated better working From a practical perspective, it is clear that the short
memory capacity performance for threat words relative to duration and simplicity of the looking ahead strategy
neutral words. These results suggest that it is not the total make it very useful for educational practice. Although
working memory capacity that differs between nonanxious further studies need to be conducted to show whether
individuals and individuals with generalized social phobia the strategy generalizes to other topics, such as language,
but the ratio between working memory capacity for neutral or that a longer period to look ahead will have a greater
words and social threat words. In addition, the study of impact on anxiety and performance, the strategy seems
Judah, Grant, Lechner, and Mills (2013) showed that work- very promising in enabling students to perform up to their
ing memory load moderates late attentional bias in social maximum potential.
anxiety such that avoidance occurs if working memory load
is low and difficulty disengaging attention occurs if working REFERENCES
memory load is high. This result suggests that the effects of
anxiety on test performance might differ as a function of the Amir, N., & Bomyea, J. (2011). Working memory capacity in generalized
working memory load imposed by the test problems. social phobia. Journal of Abnormal Psychology, 120, 504–509.
Relatively simple problems might have different effects on Ashcraft, M. H. (2002). Math anxiety: Personal, educational, and cognitive
consequences. Current Directions in Psychological Science, 11, 181–185.
anxiety and test performance than relatively complex problems.
doi: 10.1111/1467-8721.00196
The analyses on task complexity showed that high Ashcraft, M. H., & Krause, J. A. (2007). Working memory, math performance
anxious students rated the test as more complex than low and math anxiety. Psychological Bulletin & Review, 14, 243–248. doi:
anxious students. Because the highly anxious students were 10.3758/BF03194059
likely, on average, less skilled, as suggested by the negative Atkinson, R. C., & Shiffrin, R. M. (1971). The control processes of short-
term memory. Scientific American, 224, 82–90.
correlation between participants’ anxiety level and previous
Baddeley, A. D. (1986). Working memory. Oxford, England: Oxford
math grade, it is no surprise that they experienced the test University Press.
as more complex. In high anxious students, more resources Bandura, A. (1986). Social foundations of thought and action. Englewood
are allocated to anxiety-related aspects, leaving less Cliffs, NJ: Prentice-Hall.
resources for task-relevant aspects, which are needed to Baumeister, R. F. (1984). Choking under pressure: Self-consciousness and
perform well on the task. This is also reflected by their per- paradoxical effects of incentives on skillful performance. Journal of
Personality and Social Psychology, 46, 610–620. doi: 10.1037/0022-
formance on the math test, with the low and medium anxious 3514.46.3.610
students outperforming the high anxious students. Children’s Beilock, S. L., & Carr, T. H. (2005). When high-powered people fail:
perception about the complexity of the test and their perfor- Working memory and “choking under pressure” in math. Psychological
mance may be mediated by the confidence they have in their Science, 16, 101–105. doi: 10.1111/j.0956-7976.2005.00789.x
California standard tests (STAR). (nd). Released test questions. Retrieved
own mathematical capabilities (i.e., self-efficacy; Maynard
from: http://www.cde.ca.gov/ta/tg/sr/documents/cstrtqmath6.pdf
& Hakel, 1997). Cassady, J. C., & Johnson, R. E. (2002). Cognitive test anxiety and
From a theoretical perspective, the current study is rele- academic performance. Contemporary Educational Psychology, 27,
vant for CLT (Sweller et al., 2011). Although the available 270–295. doi: 10.1006/ceps.2001.1094
working memory capacity plays a central element of the Choi, H. H., Van Merriënboer, J. J. G., & Paas, F. (2014). Effects of the
physical environment on cognitive load and learning: Towards a new
theory, cognitive load researchers have only been interested
model of cognitive load. Educational Psychology Review, 26, 225–244.
in the allocation of working memory resources to aspects doi: 10.1007/s10648-014-9262-6
of the learning task. This study shows that the amount of CITO problems. (nd). Retrieved from: http://www.eindtoets.nl/rekenen-
available working memory resources is also determined by oefentoets-eindtoets-basisonderwijs-citotoets/
affective factors, such as anxiety-related worries about Cowan, N. (2001). The magical number 4 in short-term memory: A recon-
failure. These worries occupy working memory capacity, sideration of mental storage capacity. Behavioral and Brain Sciences,
24, 87–114. doi: 10.1017/S0140525X01003922
which reduces the available capacity, and make it relevant Duggan, G. B., & Payne, S. J. (2009). Text skimming: The process and
for cognitive load theorists and researchers to consider. In effectiveness of foraging through text under time pressure. Journal of
this study, children’s anxiety level was measured through Experimental Psychology: Applied, 15, 228–242. doi: 10.1037/a0016995

Copyright © 2014 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. (2014)
A strategy to reduce anxiety

Dutke, S., & Stober, J. (2001). Test anxiety, working memory, and cognitive Paas, F., Renkl, A., & Sweller, J. (2003). Cognitive load theory and instruc-
performance: Supportive effects of sequential demands. Cognition and tional design: Recent developments. Educational Psychologist, 38, 1–4.
Emotion, 15, 381–389. doi: 10.1080/0269993004200231 doi: 10.1207/S15326985EP3801_1
Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and Paas, F., Renkl, A., & Sweller, J. (2004). Cognitive load theory: Instruc-
cognitive performance: Attentional control theory. Emotion, 7, 336–353. doi: tional implications of the interaction between information structures and
10.1037/1528-3542.7.2.336 cognitive architecture. Instructional Science, 32, 1–8. doi: 10.1023/B:
Grimley, M., & Banner, G. (2008). Working memory, cognitive style and TRUC.0000021806.17516.d0
behavioural predictors of GCSE exam success. Educational Psychology, Paas, F., Tuovinen, J. E., Tabbers, H., & Van Gerven, P. W. M. (2003).
28, 341–351. doi: 10.1080/01443410701635058 Cognitive load measurement as a means to advance cognitive load theory.
Hoffman, B. (2010). “I think I can, but I’m afraid to try”: The role of Educational Psychologist, 38, 63–71. doi: 10.1207/S15326985EP3801_8
self-efficacy beliefs and mathematics in mathematics problem-solving Pajares, F., & David, M. M. (1994). Role of self-efficacy and self-concept
efficiency. Learning and Individual Differences, 20, 276–283. beliefs in mathematical problem solving: A path analysis. Journal of Ed-
Hoffman, B., & Schraw, G. (2009). The influence of self-efficacy and ucational Psychology, 86(2), 193–203.
working memory capacity on problem solving. Learning and Individual Pajares, F., & Graham, L. (1999). Self-efficacy, motivation constructs, and
Differences, 19(1), 91–100. mathematics performance of entering middle school students. Contempo-
Hoffman, B., & Spatariu, A. (2008). The influence of self-efficacy and rary Educational Psychology, 24, 124–139.
metacognitive prompting on math problem-solving efficiency. Contem- Pajares, F., & Kranzler, J. (1995). Self-efficacy beliefs and general mental
porary Educational Psychology, 33(4), 875–893. ability in mathematical problem-solving. Contemporary Educational
Jansen, B. R. J., Louwerse, J., Straatemeier, M., Van der Ven, S. H. G., Psychology, 20, 426–443.
Klinkenberg, S., & Van der Maas, H. L. J. (2013). The influence of Payne S. J., & Reader W. R. (2006). Constructing structure maps of multiple
experiencing success in math on anxiety, perceived competence, and on-line texts. International Journal of Human-Computer Studies, 64,
math performance. Learning and Individual Differences, 24, 190–197. 461–474. doi: 10.1016/j.ijhcs.2005.09.003
doi: 10.1016/j.lindif.2012.12.014 Peterson, L. R., & Peterson, M. J. (1959). Short-term retention of individual
Judah, M. R., Grant, D. M., Lechner, W. V., & Mills, A. C. (2013). Working verbal items. Journal of Experimental Psychology, 58, 193–198.
memory load moderates late attentional bias in social anxiety. Cognition Protheroe, N., & Perkins-Gough, D. (2000). Essentials for principals:
and Emotion, 27, 502–511. Meeting the challenges of high-stakes testing. Alexandria, VA: National
Leppink, J., Paas, F., Van der Vleuten, C. P., Van Gog, T., & Van Association of Elementary School Principals.
Merriënboer, J. J. (2013). Development of an instrument for measuring Ramirez, G., & Beilock, S. L. (2011). Writing about testing worries boosts
different types of cognitive load. Behavior Research Methods, 45, 1–15. exam performance in the classroom. Science, 331, 211–213. doi:
doi: 10.3758/s13428-013-0334-1 10.1126/science.1199427
Maynard, D. C., & Hakel, M. D. (1997). Effects of objective and subjective Sweller, J. (1988). Cognitive load during problem-solving: Effects on learn-
task complexity on performance. Human Performance, 10, 303–330. doi: ing. Cognitive Science, 12, 257–285. doi: 10.1207/s15516709cog1202_4
10.1207/s15327043hup1004_1 Sweller, J. (1999). Instructional design in technical areas. Camberwell,
Meece, J. L., Wigfield, A., & Eccles, J. S. (1990). Predictor of math anxiety Australia: ACER Press.
and its influence on young adolescents’ course enrollment and intentions Sweller, J. (2010). Element interactivity and intrinsic, extraneous and ger-
and performance in mathematics. Journal of Educational Psychology, 82, mane cognitive load. Educational Psychological Review, 22, 123–138.
60–70. doi: 10.1037/a0021276 doi: 10.1007/s10648-010-9128-5
Miller, G. A. (1956). The magical number seven, plus or minus two: Some Sweller, J., & Chandler, P. (1994). Why some material is difficult to learn.
limits on our capacity for processing information. The Psychological Cognition and Instruction, 12, 185–233.
Review, 63, 81–97. doi: 10.1037/0033-295X.101.2.343 Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. New
Paas, F. (1992). Training strategies for attaining transfer of problem solving- York: Springer.
skill in statistics: A cognitive-load approach. Journal of Educational Sweller, J., Van Merriënboer, J. J. G., & Paas, F. (1998). Cognitive
Psychology, 84, 429–434. doi: 10.1037//0022-0663.84.4.429 architecture and instructional design. Educational Psychology Review,
Paas, F., & Sweller, J. (2012). An evolutionary upgrade of cognitive load 10, 251–296. doi: 10.1207/s15516709cog1202_4
theory: Using the human motor system and collaboration to support the Van Gog, T., & Paas, F. (2008). Instructional efficiency: Revisiting the
learning of complex cognitive tasks. Educational Psychology Review, original construct in educational research. Educational Psychologist, 43,
24, 27–45. doi: 10.1007/s10648-011-9179-2 16–26. doi: 10.1080/00461520701756248
Paas, F., & Van Merrienboer, J. J. G. (1993). The efficiency of instructional Vasey, M. W., El-Hag, N., & Daleiden, E. L. (1996). Anxiety and the
conditions: An approach to combine mental effort and performance processing of emotionally threatening stimuli: Distinctive patterns of
measures. Human Factors: The Journal of the Human Factors and selective attention among high- and low-test-anxious children. Child
Ergonomics Society, 35, 737–743. doi: 10.1177/001872089303500412 Development, 67, 1173–1185. doi: 10.2307/1131886
Paas, F., & Van Merriënboer, J. J. G. (1994a). Variability of worked Vredeveldt, A., Hitch, G. J., & Baddeley, A. D. (2011). Eyeclosure
examples and transfer of geometrical problem solving skills: A cognitive- helps memory by reducing cognitive load and enhancing visua-
load approach. Journal of Educational Psychology, 86, 122–133. doi: lisation. Memory & Cognition, 39, 1253–1263. doi: 10.3758/s13421-
10.1037//0022-0663.86.1.122 011-0098-8
Paas, F., & Van Merriënboer, J. (1994b). Instructional control of cognitive Zatz, S., & Chassin, L. (1985). Cognitions of test-anxious children under
load in the training of complex cognitive tasks. Educational Psychology naturalistic test-taking situations. Journal of Consulting and Clinical
Review, 6, 351–371. doi: 10.1007/BF02213420 Psychology, 53, 393–401. doi: 10.1037/0022-006X.53.3.393

Copyright © 2014 John Wiley & Sons, Ltd. Appl. Cognit. Psychol. (2014)

You might also like