Math Anxiety: Personal, Educational, and Cognitive ...

1 Copyright 2002 American Psychological SocietyCURRENT DIRECTIONS IN PSYCHOLOGICAL SCIENCE181 Abstract Highly math -anxious indi-viduals are characterized by astrong tendency to avoid math ,which ultimately undercutstheir math competence andforecloses important careerpaths. But timed, on-line testsreveal math - anxiety effects onwhole-number arithmeticproblems ( , 46 27),whereas achievement testsshow no competence differ-ences. math anxiety disruptscognitive processing by com-promising ongoing activity inworking memory. Althoughthe causes of math anxiety areundetermined, some teachingstyles are implicated as riskfactors.

2 We need research onthe origins of math anxiety andon its signature in brain ac-tivity, to examine both its emo-tional and its cognitivecomponents. Keywords anxiety ; mental arithmetic; math competence; workingmemory; problem solvingMy graduate assistant recentlytold me about a participant he hadtested in the lab. She exhibited in-creasing discomfort and nervous-ness as the testing session pro-gressed, eventually becoming sodistraught that she burst into assistant remarked that manyof our participants show some un-ease or apprehension during test-ing trembling hands, nervouslaughter, and so forth.

3 Many ask,defensively, if their performancesays anything about their overallintelligence. These occasionally ex-treme emotional reactions are nottriggered by deliberately provoca-tive procedures there are no per-sonally sensitive questions or in-tentional manipulations of , we merely ask collegeadults to solve elementary-schoolarithmetic problems, such as 46 18 ? and 34 19 ?The reactions are obvious symp-toms of anxiety , in this case mathanxiety induced by ordinary arith-metic problems presented in timedtasks. On the one hand, it is almostunbelievable that tests on such fun-damental topics can be so upset-ting; knowing that 15 8 7ought to be as basic as knowinghow to spell cat.

4 On the otherhand, culture abounds withattitudes that foster math Anxiety: math is thought to be inherentlydifficult (as Barbie dolls used tosay, math class is hard ), aptitudeis considered far more importantthan effort (Geary, 1994, chap. 7),and being good at math is consid-ered relatively unimportant, oreven this article, I discuss what hasbeen learned about math anxietyacross the past 30 years or so, andsuggest some pressing issues to bepursued in this area. An importantbackdrop for this discussion is thefact that modern society is increas-ingly data and technology ori-ented, but the formal educationalsystem seems increasingly unsuc-cessful at educating students to anadequate level of numeracy, themathematical equivalent of literacy(Paulos, 1988).

5 math anxiety DEFINED AND MEASURED math anxiety is commonly de-fined as a feeling of tension, appre-hension, or fear that interferes withmath performance. The first sys-tematic instrument for assessingmath anxiety was the MathematicsAnxiety Rating Scale (MARS), pub-lished by Richardson and Suinn(1972). In this test, participants ratethemselves on the level of anxietythey would feel in various every-day situations, such as trying to re-figure a restaurant bill when theythink they have been overchargedor taking a math test. My co-work-ers and I use a shortened version ofthe test, which yields scores thatcorrelate well with scores obtainedusing the original test and also hasvery acceptable test-retest reliabil-ity ( , an individual who takesthe test on different occasions gen-erally receives similar scores).

6 Wehave also found that for a quick de-termination, one can merely ask, On a scale from 1 to 10, how mathanxious are you? Across at least ahalf-dozen samples, responses tothis one question have correlatedanywhere from .49 to .85 withscores on the shortened is a rather extensive litera-ture on the personal and educa-tional consequences of math anxi-ety, summarized thoroughly inHembree (1990). Perhaps the mostpervasive and unfortunate ten-dency is avoidance. Highly math -anxious individuals avoid take fewer elective mathcourses, both in high school and incollege, than people with low mathanxiety.

7 And when they take math ,they receive lower grades. Highlymath-anxious people also espousenegative attitudes toward math ,and hold negative self-perceptions math anxiety : Personal, Educational, and Cognitive Consequences Mark H. Ashcraft 1 Department of Psychology, Cleveland State University, Cleveland, Ohio182 VOLUME 11, NUMBER 5, OCTOBER 2002 Published by Blackwell Publishing Inc. about their math abilities. The cor-relations between math anxietyand variables such as motivationand self-confidence in math arestrongly negative, ranging between.

8 47 and .82. It is therefore nosurprise that people with mathanxiety tend to avoid college ma-jors and career paths that dependheavily on math or quantitativeskills, with obvious and unfortu-nate , math anxiety isonly weakly related to overall in-telligence. Moreover, the small cor-relation of .17 between math anx-iety and intelligence is probablyinflated because IQ tests includequantitative items, on which indi-viduals with math anxiety performmore poorly than those withoutmath anxiety . The small correlation( .06) between math anxiety andverbal aptitude supports this inter-pretation.

9 However, math anxietyis related to several other impor-tant characteristics. As conventionalwisdom suggests, it is somewhathigher among women than gender difference is rather small,may be particularly apparent inhighly selected groups ( , col-lege students), and may be partlyattributable to a greater willingnesson the part of women to disclosepersonal attitudes. Nonetheless,when we recruited participants forresearch on math anxiety , wefound fewer men than women athigh anxiety levels, but just the re-verse at low levels (Ashcraft &Faust, 1994).

10 Individuals who are high inmath anxiety also tend to scorehigh on other anxiety tests. Thestrongest interrelationship is withtest anxiety , a .52 correlation. De-spite the overlap among kinds ofanxiety, however, the evidence isconvincing that math anxiety is aseparate phenomenon. For in-stance, intercorrelations betweenalternative assessments of mathanxiety range from .50 to .70, butintercorrelations of math anxietywith other forms of anxiety rangefrom .30 to .50. In a particularlyclear display of the specificity ofmath anxiety , Faust (1992) foundphysiological evidence of increas-ing reactivity ( , changes in heartrate) when a highly math -anxiousgroup performed math tasks of in-creasing difficulty.