See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/247782442

Mathematics Anxiety and the Affective Drop in Performance

Article in Journal of Psychoeducational Assessment · June 2009

DOI: 10.1177/0734282908330580

CITATIONS READS

482 10,402

2 authors:

Mark H Ashcraft Alex M Moore

University of Nevada, Las Vegas Illinois College
73 PUBLICATIONS 10,493 CITATIONS 14 PUBLICATIONS 890 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mark H Ashcraft on 09 February 2015.

The user has requested enhancement of the downloaded file.

 Journal of Psychoeducational
Assessment
http://jpa.sagepub.com

Mathematics Anxiety and the Affective Drop in Performance

Mark H. Ashcraft and Alex M. Moore
Journal of Psychoeducational Assessment 2009; 27; 197 originally published online Apr 13, 2009;

DOI: 10.1177/0734282908330580

The online version of this article can be found at:

http://jpa.sagepub.com/cgi/content/abstract/27/3/197

Published by:

http://www.sagepublications.com

Additional services and information for Journal of Psychoeducational Assessment can be found at:

Email Alerts: http://jpa.sagepub.com/cgi/alerts

Subscriptions: http://jpa.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations http://jpa.sagepub.com/cgi/content/refs/27/3/197

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

 Journal of Psychoeducational
Assessment
Volume 27 Number 3

Mathematics Anxiety and the

June 2009 197-205
© 2009 Sage Publications
10.1177/0734282908330580

Affective Drop in Performance http://jpa.sagepub.com

hosted at
http://online.sagepub.com
Mark H. Ashcraft
Alex M. Moore
University of Nevada, Las Vegas

The authors provide a brief review of the history and assessment of math anxiety, its relationship
to personal and educational consequences, and its important impact on measures of perform-
ance. Overall, math anxiety causes an “affective drop,” a decline in performance when math
is performed under timed, high-stakes conditions, both in laboratory tests as well as in educa-
tional settings. This means that math achievement and proficiency scores for math-anxious
individuals are underestimates of true ability. The primary cognitive impact of math anxiety is
on working memory, particularly problematic given the important role working memory plays
in math performance. The authors conclude with a discussion of risk factors for math anxiety
and some factors to be kept in mind when working with math-anxious students.

Keywords:   math anxiety; math achievement; working memory; affect

M athematics anxiety is a person’s negative affective reaction to situations involving

numbers, math, and mathematics calculations, “a feeling of tension and anxiety that
interferes with the manipulation of numbers and the solving of mathematical problems in a
wide variety of ordinary life and academic situations” (Richardson & Suinn, 1972, p. 551).
Math anxiety reactions can range from mild to severe, from seemingly minor frustration to
overwhelming emotional (and physiological) disruption; for example, we had a college stu-
dent burst into tears during a lab experiment that tested simple subtraction facts (Ashcraft,
2002). People can experience math anxiety in formal settings, in a math classroom or when
taking a high-stakes standardized math test, or in more everyday settings, for example when
trying to balance a checkbook or figure a tip on a restaurant bill when others are watching.
Research on math anxiety has traditionally examined its relationships with other affec-
tive constructs, such as test or generalized anxiety, with various academic outcomes such
as school grades or with motivational factors such as intent to take additional math courses
or pursue math-related college majors (e.g., Hembree, 1990). Our research on the topic has
taken a different approach, investigating the possible cognitive consequences of math
anxiety when people perform math-related tasks in the laboratory. Taken together, these
lines of research suggest strongly that math anxiety is a significant impediment to math
achievement, one that affects a considerable portion of the population and one that merits
serious attention both in terms of assessment and intervention.

Authors’ Note: Please address correspondence to Mark H. Ashcraft, Psychology – University of Nevada, Las
Vegas, Box 455030, 4505 S. Maryland Parkway, Las Vegas, NV 89154.

197

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

198   Journal of Psychoeducational Assessment

The purpose of this article is to outline some of the important findings about math anxiety
and to indicate how math anxiety relates to other important performance measures, both
achievement tests and cognitive tests of math processing. We explain current thinking about
the mental mechanism thought to underlie the affective drop in performance mentioned in
the title, and then, we explore a model that suggests some of the factors that may place stu-
dents at risk for math anxiety. We do not claim that math anxiety is a genuine case of math-
ematics learning disability, in any traditional sense of the term disability (e.g., Mazzocco,
2007). We do, however, suggest that math anxiety functions as a disability in the sense that
there are well investigated—and negative—personal, educational, and cognitive conse-
quences of math anxiety. Unfortunately, these negative consequences affect a substantial
percentage of the population (e.g., Ashcraft, 2002; Ashcraft, Krause, & Hopko, 2007).

Historical Background and Assessments

Two reports in the 1950s were instrumental in triggering modern interest in the topic of
math anxiety: first, an anecdotal report by a teacher who commented on her students’ emo-
tional difficulties with math (Gough, 1954) and second, a surprisingly insightful paper on
“numerical anxiety” by Dreger and Aiken (1957). Dreger and Aiken added three math-re-
lated items to the Taylor Manifest Anxiety Scale and renamed it the Numerical Anxiety
Scale, thus providing the first objective test of math anxiety. They also made three general
predictions: (a) that math anxiety would be different from general anxiety, although the two
would overlap to a degree; (b) that math anxiety would not be systematically related to
general intelligence, although it would probably correlate negatively with quantitative
scores on IQ tests; and (c) that there would be an inverse relationship between math anxiety
and academic performance in math coursework. All three of these predictions have been
confirmed repeatedly by subsequent research (see Hembree, 1990).
A major advance in the research occurred when Richardson and Suinn (1972) published
a general purpose assessment instrument for math anxiety, the Mathematics Anxiety Rating
Scale, commonly called the MARS. The MARS is a 98-item rating scale on which respond-
ents rate, on a 1 to 5 Likert-type scale, how anxious they would feel in situations ranging
from formal math settings (e.g., taking a pop quiz in math class) to informal, everyday
situations (totaling a dinner bill you think has overcharged you). Given its objectivity, avail-
ability, and accompanying psychometric data (e.g., the 2-week test-retest reliability was
.85), the MARS became the test of choice for those investigating math anxiety.
For research or assessment purposes, however, the 98-item MARS is burdensome. It is
time consuming to administer, cumbersome to score (scores range from 98 to 490), and
given that shorter instruments are now available, far too long. Several descendants have
emerged more recently that maintain good reliability and reduce administration time. For
example, in our laboratory work with college students, we have used a 25-item version of
the MARS, which we call the sMARS (shortened MARS; e.g., Ashcraft & Kirk, 2001,
based on Alexander & Martray, 1989). It is a convenient and quick way of assessing a par-
ticipant’s level of math anxiety, available at no cost in the Alexander and Martray (1989)
publication, which correlates very strongly (.97) with the original MARS. Hopko,
Mahadevan, Bare, and Hunt (2003) have recently constructed the nine-item Abbreviated

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

 Ashcraft, Moore / Mathematics Anxiety   199

Math Anxiety Scale (AMAS) test. Two-week test-retest reliability was .85 for this test, and
factor analysis yielded a two-factor structure, with factors of learning math anxiety and
math evaluation/math test anxiety accounting for 70% of the overall variance in scores.
Both the sMARS and the AMAS can be used with high school students, although enough
of the items refer to advanced topics (e.g., algebra, graduation requirements) that younger
respondents probably could not respond meaningfully. In passing, we note that in our work
with college students, we also ask a one-item question: “On a scale from 1 to 10, with 10
being the most anxious, how math anxious are you?” Our results from this question have cor-
related anywhere from .48 to .85 with sMARS scores. Although we do not advocate using
this one question instead of a validated survey like the AMAS, it does suggest that a very
quick, informal direct question can often be used as a preliminary screen for math anxiety,
with some degree of validity.
Younger respondents can be given one of the age-appropriate descendants of the original
MARS—for example, the MARS-A (for adolescents) or the MARS-E (for elementary
school students, starting with fourth graders), although these tests are now rather dated. In
our preliminary work with grade school children, we are exploring alternatives to our one-
question statement, as very young children are likely unfamiliar with adult-like connota-
tions for terms such as anxious (or possibly math as opposed to arithmetic). Based on a
similar attitude scale about children’s reading, which used drawings of Garfield with facial
expressions ranging from happy to sad, we simply provide children with a row of schematic
faces, a classic “smiley” face down through a “frowney” face, and ask “How do you feel
when you do arithmetic in school?” It is not yet clear whether this alternative elicits
anxiety-related responding or responding based on some other factor.
A more formal definition of math anxiety is necessary for research, however, much as
formal definitions of mathematics learning disability are needed for assessment and inter-
vention (see Mazzocco, 2007). Using the sMARS as our example, respondents answer the
25 items by responding on a 5-level Likert-type scale from no anxiety at all up through
extreme anxiety. We then award 0 through 4 points, respectively, for their responses and
then sum the points, yielding a possible range of 0 to 100. Across many college samples,
the overall grand mean has been 36.0, with a standard deviation of 16. For purposes of
forming low, medium, and high math-anxious groups, we have simply established statisti-
cal cutoffs based on these values. For the low math anxiety group, we count scores at or
below 1 SD below the grand mean, 20 or lower, and for the high-anxious group, scores at
or above 1 SD above the mean, 52 or higher. For the medium-anxiety group, we center the
16-point SD on the grand mean, such that this group’s scores range from plus or minus 0.5
SD around the mean, from 28 to 42. Based on the normal curve, roughly 17% of the entire
population would be expected to be low in math anxiety, 17% would be high in math
anxiety, and 38% of the population should fall within 0.5 SD of the mean.
This is a purely statistical definition, however, used for purposes of forming groups in a
research design. As such, it is clearly mistaken to draw conclusions such as “17% of the
population is high in math anxiety.” Instead, it is important to use this statistical criterion as
an initial guide and to then refine those definitions by examining actual performance data to
see how performance depends on an individual’s actual level of math anxiety. We turn to that
topic shortly.

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

200   Journal of Psychoeducational Assessment

Relationships to Other Factors

One possible reaction to a student’s math anxiety—whether the student confesses to
having math anxiety or the teacher “diagnoses” it—is to merely acknowledge it and then
say something like “Well, you’ll just have to deal with that” or “You should be able to
overcome that,” or even “So what?” These reactions fail to appreciate the pervasive rela-
tionships that can exist between math anxiety and other aspects of personal, educational,
and cognitive functioning. A large literature now exists on how math anxiety relates to
aspects of personality and educational attainment, and we review that literature briefly here.
We then turn to our research work on some cognitive consequences of math anxiety, which
are relevant to the actual doing of math in the moment—in the classroom during a test, in
the lab during a cognitive test of math performance. Given society’s increasing reliance on
standardized testing, knowing how math anxiety affects performance as students actually
perform math, or as they are assessed in math, is extremely important.
Hembree’s (1990) and Ma’s (1999) articles provide meta-analyses of the existing litera-
ture on math anxiety, with Ma’s article focusing exclusively on precollege samples. The
articles summarize what is known about the relationship between math anxiety and a variety
of other personal and educational factors. To begin with, math anxiety correlates only –.17
with overall IQ and a nonsignificant –.06 when only the verbal aptitude sections of standard-
ized tests are used (except as noted, all correlations reported in this section are from the
Hembree and Ma meta-analyses). Clearly, math anxiety is not a correlate of intelligence.
Math anxiety does correlate with several other anxiety measures, most strongly with test
anxiety (.52) but moderately with general, trait, and state anxiety (respectively, .35, .38,
and .42). Most researchers interpret this to mean that test anxiety is a strong correlate of
math anxiety and that math anxiety is separate from, although related to, other anxiety
constructs.
Correlations between math anxiety and several math attitude measures are uniformly
negative and frequently quite strong—for example, with enjoyment of math in precollege
samples (–.75), with self-confidence in math in precollege samples (–.82), and with motiva-
tion in math (–.64). The correlation between math anxiety and ratings of the usefulness of
math is also negative (–.37), as is the correlation with ratings of math teachers (–.46).
Negative conversational remarks about math are relatively common in our culture, of
course, but persistent, negative remarks about oneself and math, especially those that dis-
play dislike and low self-confidence, are probably diagnostic.
Turning to measures that examine educational effects, math anxiety correlates –.34 with
math achievement scores in precollege samples (and –.31 in college samples), –.30 with
high school math grades, and –.27 with college math grades (Ashcraft & Kirk, 2001). And
it correlates –.31 with the extent of high school math courses taken (electives) and –.32
with intent to enroll in more math courses (college; see Ashcraft et al., 2007, for a more
complete discussion of these correlational results). We return in a moment to the negative
correlation with math achievement, given its centrality to educational assessment.
In sum, the higher one’s level of math anxiety, the lower one’s score is on math achieve-
ment tests, the fewer math courses one takes, and the lower one’s grades are in the math
courses that are taken. Furthermore, with higher math anxiety, one’s attitudes about math are
poorer, including one’s intent to take further math courses. This of course influences deci-
sions in college concerning majors and career paths. Predictably, those majoring in physical

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

 Ashcraft, Moore / Mathematics Anxiety   201

sciences, engineering, and math in college have low math anxiety, and those in humanities
and non-math-related disciplines have higher math anxiety. Particularly worrisome, the col-
lege major with the highest level of math anxiety is elementary education (Hembree,
1990).
A convenient way of summarizing much of this correlational research is to appeal to the
concept of “avoidance,” saying that avoidance of math is an overriding characteristic of
math-anxious individuals. Math-anxious individuals avoid taking math courses whenever
possible, avoid selecting courses of study in college that involve math, and of course avoid
career paths that involve math. It seems likely, although this has not been tested yet, that
such avoidance would be observable in the math classroom as well, in terms of student
engagement in the lesson, participation in class, time spent on studying, submission of
homework, and so on. Turner et al. (2002) reported that avoidance was the typical student
behavioral pattern in response to teachers whose teaching style was cold and unsupportive;
students avoided eye contact during class, avoided out-of-class make-up and help sessions,
and the like when they had such teachers. Such avoidance seems only one step removed
from math anxiety, given some of the risk factors to be discussed below.

Relationships to Performance
In the published literature, no other relationship is as troublesome as the negative cor-
relation between math anxiety and math achievement; in college samples, the correlation is
–.31 (and –.34 in precollege samples; Hembree, 1990). From the standpoint of doing
research on math anxiety, the relationship means there is a confound between the two—any
time a high math-anxious group performs more poorly than a low-anxious group, the inter-
pretive question is whether their poor performance was due to high anxiety or to their lower
level of mastery of the math. Our solution to this vexing problem in the lab has been to test
college students on fairly simple arithmetic (e.g., two-column addition problems) on which
the anxiety groups have been shown not to have differences in mastery when they are tested
in an untimed, low-pressure setting (Faust, Ashcraft, & Fleck, 1996; see also Ashcraft
et al., 2007, who showed no math anxiety differences on accuracy using the first three lines
of the Wide Range Achievement Test, the three lines that test whole number arithmetic).
But the relationship between math anxiety and math achievement is more knotty than
this. We typically measure not only a student’s aptitude and achievement with standardized
tests but also the effectiveness of instruction and indeed the quality of education provided
by the school. As such, it is important to know how math anxiety affects performance, how
it alters how people actually perform math in the moment of doing it. This is the line of
research we have pursued in our lab for several years, examining how cognitive processing
is affected by math anxiety. To foreshadow our results, we find evidence for what might be
termed an affective drop, a drop in performance that can be attributed to math anxiety inde-
pendent of the individual’s competence or achievement in math.
In our early research (Ashcraft & Faust, 1994; Faust et al., 1996), we found that math
anxiety had very little effect on performance to the simplest kinds of arithmetic: the basic
facts of addition and multiplication; for the most part, college students perform such prob-
lems in a fairly automatic fashion, by means of memory retrieval. More complex arithmetic
problems, however, showed either slower performance or considerably less accurate

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

202   Journal of Psychoeducational Assessment

p­ erformance for the higher anxious participants. In one condition, which tested difficult
problems in all four arithmetic operations, our highest anxiety group performed as rapidly
as the low-anxiety group but made twice as many errors. This tends to be the typical pattern;
participants of medium and high math anxiety respond quite slowly, compared to ­low-anxious
participants, and/or perform quite inaccurately whenever we ask them to perform beyond the
level of single-digit arithmetic. This is the affective drop in performance that we observe in
the lab, the kind of drop that will detract from performance on typically administered tests
of math proficiency and achievement. Importantly, we find this drop for both medium- and
high-anxious groups, composing somewhat more than half of our participants.
In the most theoretically meaningful result we have obtained (Ashcraft & Kirk, 2001),
we tested college adults on two-column additions, showing them problems with and
without carrying and having them furnish the answers out loud in a timed setting. On a
portion of the trials, they simultaneously had to hold either two or six unrelated letters of
the alphabet in working memory while doing the addition problem and then report the let-
ters back in order. This is referred to as the “dual task” method in cognitive psychology,
taxing working memory by performing two tasks at once.
In brief, our results showed that two-column addition, especially when it involved carrying,
placed a heavy load on working memory, as shown by higher errors in the letter recall task.
Most important, the increase in letter recall errors was especially pronounced among the high
math-anxious participants. Our interpretation was relatively straightforward. High math-
anxious individuals are already using some of their limited working memory resources
worrying about their anxiety whenever they perform a math task. When given an especially
taxing math task, say one involving carrying, the load on working memory becomes even
more intense. Coupling this with the letter recall task led to serious difficulties. It was as if
the high math-anxious participants were participating in a three-way competition for their
limited working memory resources: difficult math, letter retention and recall, and their own
math anxiety. The load on working memory became so pronounced that their performance
deteriorated markedly—affective drop.
Extrapolate this situation to high-stakes testing, say a standardized math test needed for col-
lege entrance or, quite common these days, a math proficiency test needed to obtain a high
school diploma. The pressure to perform well is intense, the math itself is challenging, and the
student must still grapple with the internal worries and fears associated with the math anxiety.
Just as found in the lab, the student’s math anxiety compromises the essential working memory
resources necessary for successful completion of the math problems. In testing situations, we
typically interpret a student’s score on the standardized test as an indicator of the student’s mas-
tery of math, literally of his or her math achievement (or of the quality of instruction). Based on
our research findings, this is an interpretive mistake. In the pressure-inducing, high-stakes set-
ting, some portion of the math-anxious student’s low score is more appropriately attributable to
math anxiety, to the same kind of affective drop we see in the lab.
Consistent with this interpretation, Hembree (1990) reported a fascinating finding in his
meta-analysis. High math-anxious individuals who had undergone cognitive behavioral
interventions for their math anxiety subsequently showed math achievement scores “in the
normal range.” On the assumption that the therapeutic intervention did not involve teaching
any math, this clearly suggests that their pretherapeutic math achievement scores were

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

 Ashcraft, Moore / Mathematics Anxiety   203

underestimates of their true math achievement. Reducing math anxiety thus removed an
impediment to adequate performance on the achievement tests. The message for assess-
ment seems very clear here: Math anxiety reduces a person’s assessed math achievement
regardless of the person’s true level of mastery of math. This is the same kind of affective
drop that our dual task study of college students showed in two-column addition.
We wish to repeat an important point here regarding the cognitive component that seems
especially vulnerable to the effects of math anxiety. It was working memory that was compro-
mised in our study of college students’ performance and math anxiety; working memory suf-
fered the brunt of the math anxiety effect because of the inner-worries and self-doubts that are
reported by math-anxious individuals. This is the same cognitive mechanism at work when
females’ math scores drop because they have been exposed to a negative stereotype concerning
women’s allegedly poorer performance in math—the stereotype threat effect (e.g., Beilock,
Rydell, & McConnell, 2007; see also Eysenck & Calvo, 1992). The involvement of working
memory is important to note because math becomes especially abstract and difficult when it
involves multistep computations, sequencing of mental procedures, mental look up of formulas
and equations, and the like. All of these factors place a heavy load on working memory, as
documented in the literature (e.g., LeFevre, DeStefano, Coleman, & Shanahan, 2005).

A Model of Math Anxiety Susceptibility—Risk Factors

Given the wealth of information about correlates of math anxiety, it is somewhat surprising
that little if any research has been reported concerning its onset or possible causes. Much is
known developmentally, of course. For example, across Grades 6 through college, females
score approximately 0.3 SD higher on math anxiety scales (Hembree, 1990; Hopko, 2003)
than males, although it is still not clear that they display greater disruption in math perform-
ance as a consequence of this. Math anxiety peaks at the Grade 9 to 10 level and then essen-
tially flattens thereafter (see Figure 1, Hembree, 1990). Hardly any research has examined
children younger than the fourth grade (Hembree, 1990; Ma, 1999). But there appears to be
no empirical information about the onset of math anxiety or firm evidence about why it grows
through the middle school years.
Partly as an attempt to prompt research on this topic, we have recently proposed a model
of math anxiety susceptibility and highlighted several plausible risk factors (Ashcraft et al.,
2007). Our model predicts that a child with adequate skill in math, adequate motivation,
and adequate working memory will exhibit adequate mastery and performance in math and
in particular will not fall prey to a set of cognitive biases that disrupt responding. Risk fac-
tors for math anxiety, on the other hand, would be low skill or ability in math, inadequate
motivation, or insufficient working memory.
Specifically, a child with lower than normal math skill will struggle with math concepts,
possibly in the early math curriculum but certainly once the curriculum advances beyond sim-
ple arithmetic. Such a child may show early deficits in “number sense,” a warning sign given
the growing body of evidence that shows how critical number sense is in early schooling and
how predictive it is of math achievement (e.g., Jordan, Kaplan, Olah, & Locuniak, 2006;
Siegler & Booth, 2004). More generally, we predict that children with lower than average math
ability will fall behind their peers, receive negative feedback from teachers and parents, and be

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

204   Journal of Psychoeducational Assessment

more likely to develop negative attitudes and motivations about math. Likewise, a child with
lower than average working memory capacity is one who will struggle with more complex
arithmetic and math, given the importance of working memory to performance in math and
other educationally relevant cognitive skills (Engle, 2002; LeFevre et al., 2005).
In the model, these risk factors by themselves can lead to performance deficits and avoid-
ance; for instance, low math ability and consequent poor motivation could easily lead a
student to avoid elective math in high school, a pattern that would also detract from perform-
ance on standardized tests that cover advanced topics. But the model also suggests that
several cognitive biases may also result, leading to math anxiety that co-occurs with the
performance deficits and avoidance. In particular, there may be an increase in negative math
attitudes and self-focused attention in some children, internalizing the negative feedback
received from poor performance. In some children, there may also be a heightened vulner-
ability to anxiety, or vulnerability to public embarrassment, that becomes associated with
math; anecdotally, we have had several college students tell us that their anxiety stemmed
from having to work a difficult math problem on the blackboard during math class and the
resultant embarrassment they felt in front of their peers and teachers. The “cold, unsupport-
ive” teacher style mentioned above (Turner et al., 2002) may feed into this vulnerability too,
although there is no way of knowing how common such a style is in math classrooms.

In Practice
We conclude briefly with three points aimed at practice and assessment. First, our pre-
liminary data suggest that children in the first 2 or 3 years of elementary school are not
troubled by math anxiety (or do not report it), although by fourth and fifth grade some start
to indicate a degree of apprehension. Thereafter, and certainly whenever there are other
indicators of math anxiety, it is not possible to interpret a student’s score on a standardized
math test as a clear indicator of math ability or achievement. Our evidence indicates that
every time a math-anxious individual is asked to perform math in a timed, high-stakes set-
ting, the individual’s math anxiety is aroused and causes affective drop, a significant decline
in performance. Indeed, Hopko’s (2003) interpretation of factor analysis results suggests
that any math test arouses anxiety. The score on the test is simply not an accurate reflection
of the respondent’s ability or achievement.
Second, it is likely that a student’s math anxiety is aroused in the math classroom itself,
possibly only to a minor degree during routine class activities but almost certainly when the
student is called on to answer a question or solve a problem. It is certainly aroused when the
student takes a test. Thus, class grades and in-class assessments are also suspect as indicators
of true achievement, at least to the extent that they rely on timed tests completed under stress
or pressure. Overall, it seems more likely than not that the math-anxious student learns some-
what less in the math classroom than the nonanxious student, although testing for this dif-
ferential learning is obviously difficult because of the anxiety drop factor discussed here.
Finally, the existing literature is quite clear on the long-term effects of math anxiety.
Math-anxious individuals avoid elective math coursework, avoid college majors that
require math, and avoid career paths that involve math. Given society’s increasing reliance
on technology and current concerns over STEM (science, technology, engineering, and
math) training, anything that can be done to investigate, understand, and thereby prevent

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

 Ashcraft, Moore / Mathematics Anxiety   205

math anxiety is to be encouraged, beginning with promoting a sensitivity to its effects on

students in the K-12 math classroom.

References
Alexander, L., & Martray, C. (1989). The development of an abbreviated version of the Mathematics Anxiety
Rating Scale. Measurement and Evaluation in Counseling and Development, 22, 143-150.
Ashcraft, M. H. (2002). Math anxiety: Personal, educational, and cognitive consequences. Current Directions
in Psychological Science, 11, 181-185.
Ashcraft, M. H., & Faust, M. W. (1994). Mathematics anxiety and mental arithmetic performance: An explora-
tory investigation. Cognition and Emotion, 8, 97-125.
Ashcraft, M. H., & Kirk, E. P. (2001). The relationships among working memory, math anxiety, and perform-
ance. Journal of Experimental Psychology: General, 130, 224-237.
Ashcraft, M. H., Krause, J. A., & Hopko, D. R. (2007). Is math anxiety a mathematical learning disability? In
D. B. Berch & M. M. M. Mazzocco (Eds.), Why is math so hard for some children? The nature and origins
of mathematical learning difficulties and disabilities (pp. 329-348). Baltimore: Paul H. Brookes.
Beilock, S. L., Rydell, R. J., & McConnell, A. R. (2007). Stereotype threat and working memory: Mechanisms,
alleviation, and spillover. Journal of Experimental Psychology: General, 136, 256-276.
Dreger, R. M., & Aiken, L. R. (1957). The identification of number anxiety in a college population. Journal of
Educational Psychology, 48, 344-351.
Engle, R. W. (2002). Working memory capacity as executive attention. Current Directions in Psychological
Science, 11, 19-23.
Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing efficiency theory. Cognition
and Emotion, 6, 409-434.
Faust, M. W., Ashcraft, M. H., & Fleck, D. E. (1996). Mathematics anxiety effects in simple and complex addi-
tion. Mathematical Cognition, 2, 25-62.
Gough, M. F. (1954). Mathemaphobia: Causes and treatments. Clearing House, 28, 290-294.
Hembree, R. (1990). The nature, effects, and relief of mathematics anxiety. Journal for Research in Mathematics
Education, 21, 33-46.
Hopko, D. R. (2003). Confirmatory factor analysis of the Math Anxiety Rating Scale—Revised. Educational
and Psychological Measurement, 63, 336-351.
Hopko, D. R., Mahadevan, R., Bare, R. L., & Hunt, M. A. (2003). The Abbreviated Math Anxiety Scale
(AMAS): Construction, validity, and reliability. Assessment, 10, 178-182.
Jordan, N. C., Kaplan, D., Olah, L., & Locuniak, M. N. (2006). Number sense growth in kindergarten: A lon-
gitudinal investigation of children at risk for mathematics difficulties. Child Development, 77, 153-175.
LeFevre, T., DeStefano, D., Coleman, B., & Shanahan, T. (2005). Mathematical cognition and working mem-
ory. In J. I. D. Campbell (Ed.), Handbook of mathematical cognition (pp. 361-377). New York: Psychology
Press.
Ma, X. (1999). A meta-analysis of the relationship between anxiety toward mathematics and achievement in
mathematics. Journal for Research in Mathematics Education, 30, 520-541.
Mazzocco, M. M. M. (2007). Defining and differentiating mathematical learning disabilities and difficulties. In
D. B. Berch & M. M. M. Mazzocco (Eds.), Why is math so hard for some children? The nature and origins
of mathematical learning difficulties and disabilities (pp. 29-47). Baltimore: Paul H. Brookes.
Richardson, F. C., & Suinn, R. M. (1972). The Mathematics Anxiety Rating Scale. Journal of Counseling
Psychology, 19, 551-554.
Siegler, R. S., & Booth, J. L. (2004). Development of numerical estimation in young children. Child
Development, 75, 428-444.
Turner, J. C., Midgley, C., Meyer, D. K., Gheen, M., Anderman, E. M., Kang, Y., et al. (2002). The classroom
environment and students’ reports of avoidance strategies in mathematics: A multimethod study. Journal of
Educational Psychology, 94, 88-106.

Downloaded from http://jpa.sagepub.com at LUND UNIVERSITY LIBRARIES on December 29, 2009

View publication stats