Gender Stereotypes about Mathematics and Science and Self-Perceptions of Ability in Late Childhood and Early Adolescence

MERRILL-PALMER QUARTERLY, VOL. 54, No. 3

Gender Stereotypes about Mathematicsand Science and Self-Perceptions of Abilityin Late Childhood and Early Adolescence

Beth Kurtz-Costes, University of North Carolina, Chapel HillStephanie J. Rowley, University of MichiganApril Harris-Britt and Taniesha A. Woods, University of North Carolina,

Chapel Hill

A model linking children's perceptions of adults' gender stereotypes about math-ematics and science ability, children's stereotypes, and children's perceptions oftheir own mathematics and science competence was tested in 302 fourth, sixth,and eighth graders. When boys believed that adults hold more traditional stereo-types, they tended to hold corresponding beliefs that girls are relatively lesscapable or that boys are more capable in mathematics and science. These groupcompetence ratings, in turn, were related to self-perceptions of ability for sixth-and eighth-grade boys. In contrast, most paths were nonsignificant for girls. Theresults provided support for both social status theory and experiential theory. Wediscuss implications regarding the influence of stereotypes on motivation andidentity development.

National studies have shown few gender differences in the mathematicsand science skills of children and young adolescents over the last threedecades (Campbell, Hombo, & Mazzeo, 2000). Nevertheless, as youthsenter later adolescence and young adulthood, girls are less likely than boysto take higher-level mathematics and science courses and to pursue careers

Beth Kurtz-Costes, Department of Psychology; Stephanie J. Rowley, Department of Psychol-ogy; April Harris-Britt and Taniesha A. Woods, Department of Psychology.

Data collection for this study was supported by National Institute of Mental Health grant P50MH52429 to Robert Cairns. Preparation of this report was supported by the National ScienceFoundation Grant No. 0335221 and No. 0335308 to Beth Kurtz-Costes and Stephanie Rowley.

Correspondence should be addressed to the first author at Department of Psychology, CB#3270, University of North Carolina, Chapel Hill, NC 27599-3270. Phone: (919) 962-4137. Fax:(919) 962-2537. E-mail: [email protected].

Merrill-Palmer Quarterly, July 2008, Vol. 54, No. 3, pp. 386-409. Copyright © 2008 by WayneState University Press, Detroit, Ml 48201.

386

in domains such as engineering, computer science, and the physical sci-ences (Leahy & Guo, 2001; National Center for Education Statistics, 2003;National Science Foundation, 2001). Many theorists have argued that thesepersistent gender differences in educational and career choices will best beunderstood through investigations of motivational beliefs such as compe-tence perceptions, values, and stereotypes. Indeed, girls report lower per-ceptions of mathematics and science competence than boys, and byadolescence girls report less interest in and valuing of these academic sub-jects (Fredricks & Eccles, 2002; Leahy & Guo, 2001; Marsh, Trautwein,LUdtke, Kbller, & Baumert, 2005; Stetsenko, Little, Gordeeva, Grasshof, &Oettingen, 2000).

The purpose of this study was to examine the role of gender stereotypesin the development of children's self-perceptions of mathematics and sci-ence abilities. We tested two alternative theories-social status theory andcultural/experiential theory--that led to different predictions about age andgender differences in a model linking children's perceptions of adults' gen-der stereotypes, children's own endorsement of those stereotypes, and chil-dren's self-perceptions of their mathematics and science abilities.

Self-Concept, Motivation, and Performance

Motivational theorists have long argued that perceptions of one's own com-petence is a necessary, albeit not sufficient, prerequisite of striving forachievement (e.g., Bandura, 1986; Deci & Ryan, 1985; Eccles & Wigfield,2002). These theorists contend that positive self-competence beliefs pro-mote effort, the selection of adaptive goals, behavioral choices that are con-gruent with those goals, and task persistence. Positive beliefs enable one toapproach the domain, obtaining more feedback and opportunity to practice.

Gender differences in children's and adolescents' perceptions of theirmathematics and science abilities are robust (e.g., Marsh et al., 2005; Stet-senko et al., 2000). These gender differences in self-perceptions of skill andvalues related to mathematics and science are parallel to traditional aca-demic stereotypes: Girls report greater self-competence in verbal domains,whereas boys report greater self-competence in and valuing of mathematicsand science (Bomholt, Goodnow, & Cooney, 1994; Eccles, Wigfield,Harold, & Blumenfeld, 1993; Jacobs, Lanza, Osgood, Eccles, & Wigfield,2002; Marsh et al., 2005; Marsh & Yeung, 1997; Stetsenko et al., 2000;Wigfield et al., 1997). A starting point of the current research was theassumption that these differences in self-concept lead to interests, values,and behaviors (e.g., course selection) that put girls and women at a disad-vantage in mathematics and physical science domains (Eccles & Wigfield,

Gender Stereotypes 387

Merrill-Palmer Quarterly

2002). Therefore, understanding the factors that lead to gender differencesin mathematics and science self-concept will inform our theories aboutgender differences in engagement and success in these fields.

Although it is clear that prior performance influences perceptions ofcompetence, these beliefs are also shaped by other factors such as the typeand extent of feedback that children receive about performance, their attri-butional beliefs about reasons underlying success and failure, and mes-sages, both explicit and subtle, from teachers, parents, and other adultsabout the child's ability in a particular domain. It is this last influence that isthe focus of the current study: If children perceive that adults view particu-lar social groups (in this case, gender groups) as relatively more or lesscapable in the domain, do these perceptions enhance or undermine thechild's beliefs about his or her own ability? Gender stereotypes might leadto gender differences in mathematics and science ability either by enhanc-ing the self-perceptions of boys or by inhibiting mathematics interest, self-competence perceptions, and identification in girls.

The Power of Stereotypes and Adults As Socializing Agentsof Children

Stereotypes are judgments about the abilities or attributes of individualsbased on their membership in a social group (Ruble, Cohen, & Ruble,2001). Research examining gender stereotypes in children has shown thatchildren already have an awareness of some traditional gender stereotypesduring the preschool years and that across middle childhood this knowl-edge becomes more detailed and more differentiated (Huston, 1983; Ruble,Martin, & Berenbaum, 2006; Weinraub et al., 1984).

Relatively little research in gender stereotypes has focused on chil-dren's beliefs about group differences in mathematics and science abili-ties. One study with college women showed that a full 41% reported thatthere was at least "some truth" to the stereotype that men are better inmathematics and spatial skills than women (Blanton, Christie, & Dye,2002). In a college sample of women who were mathematics majors, 24%endorsed the traditional stereotype (Schmader, Johns, & Barquissau,2004). However, in an earlier study with some of the same children whocomprise the current sample, middle school and elementary school girlsand middle school boys viewed girls as better than boys in mathematicsand science (Rowley, Kurtz-Costes, Mistry, & Feagans, 2007). This result(i.e., that children of both genders had positive views of girls) may reflectthe strong emphasis in policy, practice, and research in recent decadesaimed at reducing stereotypes about gender differences in mathematics

388

and science performance and at making these domains more welcoming togirls (Weaver-Hightower, 2003).

Regardless of the extent to which children explicitly endorse genderstereotypes, it is clear that many parents and teachers believe that boys aremore capable in mathematics and science than girls, and some evidenceindicates that adult stereotypes influence children's self-perceptions of abil-ity and decisions about mathematics-related education and careers. Forexample, parents' endorsement of traditional mathematics and sciencestereotypes is related to their perceptions of their own child's ability, whichin turn predicts the child's concurrent self-concept and later career choices(Bleeker & Jacobs, 2004; Jacobs & Eccles, 1992; Tiedemann, 2000). Link-ages among parental stereotype endorsement, parents' beliefs about theirchildren's mathematics abilities, and children's own ability perceptions havebeen demonstrated in both middle childhood and adolescence. However, toour knowledge, no prior research has examined whether parent stereotypesact through children's developing beliefs about group differences in skill.Thus, prior studies suggest that traditional mathematics and science stereo-types are still prevalent even if not universally or explicitly endorsed andthat parents' beliefs are related to children's own self-competency beliefs.Prior research has not established whether children are explicitly aware ofadults' gender stereotypes, whether the connection between adults' stereo-types and children's mathematics and science self-concept is mediated bychildren's beliefs about the competence of their gender, and how these asso-ciations might vary across age and gender groups.

We chose to examine children's perceptions of adult stereotypes ratherthan a direct measure of parents' stereotypes for three reasons. First, giventhe focus in recent decades on the importance of affirming girls' mathe-matics competence, it is likely that adults would underreport their genderstereotypes about mathematics and science ability if asked directly. Sec-ond, children's competence beliefs are shaped by many adults, not justtheir own parents (e.g., the parent of a friend who expresses surprise whenlearning that a sixth-grade girl is in a high-track mathematics class). Thus,children's beliefs about the stereotypes of adults-not just their own par-ents-were assessed in this study. Third, adult stereotypes may influencechildren's mathematics and science self-concept in many ways with orwithout the child's explicit awareness of those stereotypes (e.g., differen-tial feedback, differential provision of educational experiences). In thecurrent study, we chose to examine children's explicit awareness of adults'stereotypes as one avenue by which stereotypes may influence children'sdeveloping self-concept and thus lead to gender differences in values andpursuit of careers in mathematics and the physical sciences.


Gender and Self-Evaluations: Social Status Theory versusCultural-Experiential Theory

We tested two competing hypotheses about possible age and gender differ-ences in the association between mathematics and science stereotypeawareness and perceptions of mathematics and science competence. Thefirst hypothesis, social status theory, grows out of social identity theory.Tajfel (1970) theorized that individual identity is based on our membershipin social groups and that individuals have a natural tendency to showfavoritism to their own group in terms of their attitudes, affiliations, andbehaviors. To maintain positive self-regard, individuals view out-groupmembers more negatively than in-group members. However, members oflow-status groups sometimes endorse socially sanctioned status differencesrather than expressing an in-group bias (Aboud, 1988; Bigler, Brown, &Markell, 2001; Signorella, Bigler, & Liben, 1993; Spencer & Markstrom-Adams, 1990). In contrast to high-status group members, low-status groupmembers may be egalitarian in domains where their group is negativelystereotyped (e.g., Bigler et al., 2001; Rowley et al., 2007). The failure toacknowledge stereotypes that are linked to one's social group is adaptive ifthose stereotypes are negative because the denial of such stereotypes pro-tects self-esteem. Thus, according to social status theory, girls should beless likely than boys to endorse traditional gender stereotypes about mathe-matics and science. A previous study showed status differences in endorse-ment of both race and gender stereotypes: In contrast to EuropeanAmericans, African American students (a low-status group) reported norace differences in academic abilities; girls not only failed to report the tra-ditional mathematics and science stereotype but actually reported that girlsare better than boys in mathematics and science (Rowley et al., 2007).

In this essay we extend these ideas with the hypothesis that stereotypeendorsement will be more strongly related to ability perceptions for high-status groups than for low-status groups. Social psychological studies haveconfirmed that beliefs that others hold negative views of relevant socialgroups tend to be weakly related to self-concept (Rowley, Sellers, Chavous,& Smith, 1998). Individuals in low-status groups may use a variety of cop-ing mechanisms to protect themselves from negative stereotypes such asassuming that negative stereotypes about the group do not apply to oneselfas an individual. In contrast, high-status group members-in this caseboys--should benefit from stereotypes. Therefore, a positive association isanticipated between stereotype endorsement and self-perceptions of abili-ties for boys but not for girls. It should be noted that these hypotheses arebased on our use of explicit measures of perceptions of stereotypes and thusare distinct from research using experimental methods, such as Steele's

Merrill-Palmer Quarterly390

(1997) stereotype threat paradigm. We are not arguing that stereotype threatis not at work but rather that it is not measured in our study.

An alternative theory is a socialization perspective, which is based onthe assumption that stereotypes are like any other kind of world knowledgein that they result from learning about and observing the social environment(Hall & Jose, 1983). As children encounter more world knowledge, they arelikely to encode that information, to retain it, and to evaluate subsequentinformation according to that knowledge. Thus, children's increasing cul-tural knowledge over time would be evidenced through a developmentalincrease in stereotype awareness (Leahy, 1983). By this reasoning, as youthsenter adolescence we would expect a greater recognition among both boysand girls that some adults hold traditional gender stereotypes, a greater per-sonal endorsement of these stereotypes, and a stronger association betweenstereotype endorsement and self-perceptidns. Whereas social status theorypredicts gender differences in the endorsement of stereotypes, cultural-experiential theory predicts age differences in stereotypes and in the relationbetween stereotype endorsement and self-concept regarding skills in mathe-matics and science.

The Current Study

In the current study, children reported their beliefs about adults' genderstereotypes regarding mathematics and science abilities, children's own per-ceptions of group (i.e., boys, girls) competence, and their perceptions oftheir own mathematics and science abilities. We predicted that children'sbeliefs about adult stereotypes would influence their perceptions of theirown abilities both directly and indirectly through the association of thesevariables to children's beliefs about the competence of their own socialgroup (i.e., boys, girls). In other words, one purpose of the study was to testhypothesized linkages among children's perceptions of adult stereotypes,children's own stereotypes, and their perceived competence in mathematicsand science (Figure 1).

An additional purpose of the study was to test the competing hypothesesfrom social status theory and cultural-experiential theory. As outlined above,cultural-experiential theory predicts that stereotype-perceived competencelinkages would hold for both boys and girls and would be more likely amongolder children than younger children. Social status theory, in contrast, pre-dicts that girls might be resilient to the potential negative effects of stereo-types and thus would either be less likely to endorse stereotypes that reflectednegatively on their social group or that their competence perceptions wouldnot be related to their endorsement of traditional stereotypes.


392 Merrill-Palmer Quarterly

PercetionsMathematicsOf Adult and Science

Stereotypes Self-Concept

Figure 1. Theoretical model showing hypothesized relationships among children'sperceptions of adults' gender stereotypes about mathematics and science, children's beliefsabout the abilities of boys and girls in mathematics and science, and children's ownmathematics and science self-concept. End-of-year mathematics and science grades are

entered as a control variable.

We examined these associations in three age groups spanning the tran-sition to middle school, a time when perceptions of competence drop forboth boys and girls. Furthermore, middle school--compared to elementaryschool-has more tracking of mathematics classes, a higher percentage ofmale teachers, increased choice of which classes a child takes and will takein high school, and a stronger focus on competition and grades (Eccles,Midgley, et al., 1993; Feldlaufer, Midgley, & Eccles, 1988). Fourth, sixth,and eighth graders participated in the study. We predicted that if girlsshowed lower stereotype endorsement or a weaker relation between stereo-type endorsement and perceived competence than boys, these gender dif-ferences were more likely to be found in middle school than in elementaryschool. That is, we expected that status effects would be stronger in the con-text of increased competition and greater salience of stereotypes. Con-versely, if age differences appeared for boys, both cultural-experientialtheory and status theory predict that stronger associations were more likelyto be found in middle school than in elementary school.

Method

Participants

Data for the study were gathered as part of the larger Children's Beliefsabout Stereotypes Study. Protocol changes that occurred during the studymeant that not all participants completed the same measures. Therefore,data reported here were from only those participants who completed the

Merrill-Palmer Quarterly392

Gender Stereotypes

two central measures of beliefs about adults' gender academic stereotypesand students' own assessments of the performance of girls and boys. A totalof 302 children completed these central measures, 93 of whom were infourth grade (49 boys and 44 girls), 91 in sixth grade (37 boys and 54 girls),and 118 in eighth grade (38 boys and 80 girls). Mean ages were 9.4 (SD =0.54), 11.6 (SD = 0.65), and 13.5 (SD = 0.61) years, respectively. Therewere no gender differences in age at any of the three grade levels. Abouthalf of participants (162) were European American, 89 were African Amer-ican, 30 were Hispanic/Latino, 7 were Native American, 1 was AsianAmerican, and 13 were of mixed race/ethnicity. Most fourth graders werestudents in elementary schools that served first through fifth grades, andolder children were in middle schools that served sixth through eighthgrades. One participating school served children in kindergarten througheighth grade. All students were from a rural school district in the southeast-ern region of the United States.

Measures

Children's perceptions of adult stereotypes were measured with 5-pointLikert-type items. For each of several domains (e.g., mathematics, music,writing, sports), children were asked to circle a number indicating howadults believed that boys compared to girls (I = girls are much better thanboys, 2 = girls are somewhat better than boys, 3 = boys and girls are equallyskilled, 4 = boys are somewhat better than girls, 5 = boys are much betterthan girls). Children's responses to two items about mathematics and sci-ence were averaged to yield a single score indicating their beliefs aboutadult endorsement of traditional gender stereotypes regarding mathematicsand science abilities.

Children's beliefs about gender group competence in mathematics andscience were measured using 100-millimeter scales with anchors labeled"not very good" and "very good." One set of items corresponded to boys'abilities and another to girls' abilities. Children were also asked to rate thecompetence of other social groups; those results are not included in thisreport. Separate items were used to assess mathematics, science, and otherdomains. Children made a mark on each line to show the competence ofeach social group in each domain. The distance from the left endpoint of thescale to the child's mark was measured for each item to create scores rangingfrom 0 (not very good) to 100 (very good). The scores for the mathematicsand science items (two items rating boys' competence and two items ratinggirls' competence) were averaged, yielding a composite score for children'sperceptions of girls' competence' and a single score for perceptions of boys'

393


competence in mathematics and science. In the analyses reported below,scores are analyzed within the categories of same-gender and other-gendercompetence; thus, for example, girls' rating of girls' abilities corresponds tosame-gender group competence, whereas girls' ratings of boys' abilities cor-responds to girls' other-gender group competence ratings.

Children's mathematics and science self-concept was assessed with amodified version of Nicholls (1978) measure. Histograms that each had acolumn of 25 faces were used, with "best in the class" printed at the top and"worst in the class" at the bottom of each item. Children circled the face ineach column that best represented their own standing in comparison toclassmates. Competence in several domains was assessed (e.g., reading,sports, grades). For this essay, the two mathematics and science items wereaveraged to create a composite score with a possible range of 1 to 25.

Children completed other survey measures that are not included in thisreport. Reports of children's mathematics and science grades were obtainedfrom school records at the end of the school year. Some schools reportedgrades on a 100-point scale, whereas others provided letter grades. We usedeach school's grading scale midpoint (e.g., B = 85%) for schools that pro-vided letter grades. Each child's end-of-year mathematics and sciencegrades were averaged to create a score representing the child's mathematicsand science achievement. Grades were not available for 50 children. Forthese children, we imputed the sample mean score to use in analyses thatcontrolled for achievement.

Procedure

Informed parental consent was obtained for all participants. Fewer than 3%of families who received consent forms declined to participate. Childrencompleted questionnaires in small groups at their school during schoolhours and received a small gift (e.g., stress ball, flashlight) for their partici-pation. Data collection occurred throughout the school year but mostlybetween the months of October and February. Because we administered theabove measures as part of a larger study about stereotype awareness and inorder to reduce response bias, measures of adult stereotypes, group compe-tence, and self-concept were not adjacent in the study questionnaire.

Results

Descriptive statistics appear in Table 1. Nonsystematic missing data (e.g.,when a child failed to answer a few items on the questionnaire) led to slightvariations in cell sizes. Unless otherwise specified, all significant results

394

Gender Stereotypes

were evaluated with an alpha level of .05, and Newman-Keuls analyses wereused for post hoc comparisons. Before testing the model, we conducted aseries of 2 (gender) x 3 (grade) analyses of variance tests (ANOVAs) toassess group differences in scores. For perceptions of adult stereotypesscores, the main effects of gender and grade were significant: F(l, 289) =7.07 and F(2, 289) = 6.30, respectively; Cohen's d = 0.29 (boys versus girls),0.50 (fourth versus sixth grade), 0.27 (fourth versus eighth grade), and 0.29(sixth versus eighth grade). The main effects were qualified by the signifi-cant gender x grade interaction (F[2, 289] = 11.05). Newman-Keuls con-trasts showed that fourth-grade boys believed that adults hold traditionalstereotypes about mathematics and science (i.e., boys are better than girls),whereas the other groups reported on average that adults hold egalitarianviews (p < .01; Cohen's d ranged from 0.56 to 1.03).

Same- and other-gender competence scores were examined with a 2(gender) x 3 (grade) x 2 (group) repeated measures ANOVA in which therepeated (within subject) factor was same-gender group competence versusother-gender group competence. A repeated measures analysis was usedbecause each child ranked the competence of same-gender (e.g., girls' rank-ing of girls) and other-gender (e.g., girls' ranking of boys) groups. The maineffects of group and gender were significant (F[l, 511] = 106.8 and 12.3;Cohen's d = 0.45 and 0.29, respectively). On average, each group rated itsown gender more positively than the other gender, and boys' ratings weremore positive overall than the ratings of girls. However, these main effectswere qualified by significant group by grade and group by gender interac-tions (F[2, 511 ] = 11.4 and F[ 1, 511 ] = 41.6, respectively). Eighth gradersrated the other-gender group more positively than did fourth (d = 0.26) orsixth graders (d = 0.21), and girls rated boys more negatively than boys ratedgirls (d = 0.52). The main effect of grade, the grade x gender interaction, andthe grade x gender x group interaction were nonsignificant (p > 0.10).

The ANOVA on self-concept scores yielded a significant main effect ofgender (F[I, 289] = 8.34; Cohen's d = 0.38). Boys reported more positiveself-concepts regarding their mathematics and science abilities than girls.The main effect of grade and the grade x gender interaction were nonsignif-icant for self-concept scores. In addition, the main effect of gender in anANOVA on mathematics and science grades was significant (F[1, 252] =4.72). Girls received higher mathematics and science grades (M = 88.96)than boys (M = 86.75) (d = 0.26). The effect of grade and the interaction ofgender and grade were nonsignificant.

In summary, only fourth-grade boys reported on average that adultshold traditional mathematics and science stereotypes, and mean same-gender group competence scores, which were in the 70s on the 100-point

395

0 o M- 01 W') 0 NN- '0-Z - 0

al - 9 -o 9cl

M" C6 C") O0DCO 0N- N-

C0

"-o2

0

CON00

c0'O1

II

z

41)Wa-

E0U

a)

Uc0

-)

U

E

a)

N

0 II

z

0 U1

'5ZIIz

o ii

E

0a)

U

"CE)UD

C


L5

LoCq)

C.4

c") oN LOC'J

'o00 "0

CO)

04 0

0 I.

00

i7,

O0 '0C,O

'aN COO0',0 'at CO '0- O 6

0

0

2

CL

CL

V)

*0

r

.0C

*0

U)

Ca

U

CaUca1)

E0

c

a)0V)

0

0_

C" N 00 0O

C') CC) -

'0- N' -N

-n -1 -0 0 0

N -O 0 '

Lt O '0 COD 00 'N-Z: '0-Z -

scale, suggest that all groups viewed their gender group as highly compe-tent in mathematics and science. If we interpret mean differences betweensame-gender and other-gender competence ratings as stereotype scores, itwould suggest that girls have strong nontraditional stereotypes about gen-der differences in mathematics and science and that the effect is driven bytheir relatively low ratings of boys' competence. This in-group advantage,however, is not coupled with better mathematics and science self-conceptfor girls. Their self-concept scores were significantly lower than those ofboys for all three age groups in spite of strong performance as indicated byschool grades.

Analysis Plan

We had two primary goals: (1) to test a model linking children's percep-tions of adult stereotypes, children's own stereotypes, and children's self-concepts, and (2) to determine if these associations differed by grade andgender. Path analysis was chosen so that we could model the variablescleanly in a single analysis and so that comparisons across groups could bemade. The model (see Figure 1) tested direct paths from children's reportsof adults' gender stereotypes about mathematics and science abilities to thechildren's own ratings of girls' and boys' mathematics and science compe-tence as well as their mathematics and science self-concepts. We usedsame-gender and other-gender competence ratings in the model in lieu of adirect measure of children's stereotypes. Thus, by examining paths fromadult stereotypes to group competence ratings, we were able to observe towhat extent children's perceptions of adult stereotypes are related to chil-dren's own beliefs about gender differences in mathematics and scienceabilities and whether beliefs about adult stereotypes served to enhance per-ceptions of boys' abilities versus diminished perceptions of girls' abilities.In addition, the path between ratings of girls' and boys' competence andmathematics and science self-concepts was assessed. These models wereestimated separately for gender (boys, girls) and grade (fourth, sixth,eighth) groups. Because our hypotheses suggested that many paths shouldbe nonsignificant (e.g., that older girls should be less sensitive to adultstereotypes than younger girls), we will not report fit statistics for the mod-els. Instead we report significant paths and measures of effect sizes. Mathe-matics and science achievement (i.e., children's end-of-year grades in thosesubjects) were controlled in these analyses.

In line with status and experiential theories, we found greater supportfor the model in data from sixth- and eighth-grade boys than among fourth-grade boys (Figure 2). In fourth grade, adult stereotypes were related to



-0.40

-0.32

0.32

"0.33 0.31

Percepton s ofAdult

Stereotypes

Metheematicsand ScienceSelf-Concept

Figure 2. Perceptions of adult stereotypes, perceptions of group competence, andmathematics and science self-concept in fourth-, sixth-, and eighth-grade boys. Dashed linesindicate nonsignificant paths. End-of-year mathematics and science grades are controlledin all analyses.

398

boys' assessments of girls' abilities. The more strongly boys reported thatadults view boys as more competent than girls in mathematics and science,the lower their ratings of girls' abilities (P3 = -0.40, p < .01). No other pathswere significant. The model accounted for 33% of the variance in mathe-matics and science self-concept. In sixth grade, boys' beliefs about adultstereotypes were associated with their own assessments of boys, which inturn were related to their mathematics and science self-concepts. Here,beliefs that adults hold more traditional stereotypes were associated withmore favorable ratings of boys' abilities (P3 = 0.36, p < .05), and assess-ments of boys' abilities were associated with mathematics and science self-concepts (P3 = 0. 3 2, p < .05). Assessments of girls' abilities in mathematicsand science were also related to self-concepts (P = -0.32,p < .05) such thatas their assessments of girls increased, their beliefs about their own per-formance decreased. This model accounted for 36% of the variance inmathematics and science self-concept. Data for eighth-grade boys similarlyshowed mediation of reports of adult stereotypes through assessments ofboys' competence to self-concept. That is, the stronger boys' beliefs abouttraditional adult stereotypes, the better these older boys rated boys' compe-tence (13 = 0.33, p < .05), and the more positive their views about boys' per-formance, the better they felt about their own performance (13 = 0.31 ,p < .05).Thirty-seven percent of the variance in eighth-grade boys' self-concept wasexplained by this model.

The analyses on girls' data showed fewer significant paths, particularlyfor older girls (Figure 3). In fourth grade, girls' beliefs about adult stereo-types were related to girls' assessments of girls' competence (f = -0.4l ,p <.01). The better girls perceived that adults viewed girls (i.e., the lower thescore on the 5-point scale), the higher their own ratings of girls' abilities.Thirty-nine percent of the variance in self-concept was explained by thismodel. In sixth grade, adult stereotypes were related to mathematics andscience self-concept (P3 = -0.42, p < .05). Girls who believed that adultsviewed boys as better than girls in mathematics and science tended to havepoorer mathematics and science self-concepts. Twenty-one percent of thevariance in sixth-grade girls' self-concept was explained by this model. Nopaths were significant for eighth-grade girls, and only 2% of the variance inself-concept was explained by the model.

Summary

Results from the path analyses showed that whereas fourth-grade boys'self-views did not reflect their own beliefs about the mathematics and sci-ence abilities of boys and girls, there was greater correspondence between



-0.41

-0.42

iComptece/4 ofGirls ' %

/q

i#

/S

/

/ Competence.v of Boys

%S

%. %

Perceptions0f fM ttiemacsAdult - --- ---------------------------- andScience

Stereotypes Selt-Concept

Figure 3. Perceptions of adult stereotypes, perceptions of group competence, andmathematics and science self-concept in fourth-, sixth-, and eighth-grade girls. Dashed linesindicate nonsignificant paths. End-of-year mathematics and science grades are controlledin all analyses.

400

stereotypes and self-concept for middle school boys, with the relationbetween adult stereotypes and self-concept being mediated by boys' assess-ments of boys' mathematics and science ability. These results differed fromthe results of girls in that beliefs about adults' stereotypes were only relatedto self-concept for sixth-grade girls, and girls' beliefs about girls' and boys'abilities in mathematics and science were never related to correspondingself-concept. None of these paths were significant for eighth-grade girls.

Discussion

In spite of strong performance in mathematics and science as measured bytheir school grades, girls in the current study reported lower self-competencein those domains than boys. The self-concepts of boys appeared to beenhanced by their perceptions of adult stereotypes favoring boys over girls,stereotypes that were unrelated to self-concept among girls. Thus, results ofthe study were consistent with social status theory, which posits that endorse-ment and effects of stereotypes differ for high- versus low-status groups.Some support was also provided for cultural-experiential theory in thatstereotypes were more consistently related to self-concepts among boys inmiddle school than among fourth-grade boys.

Stereotypes and Self-Concept Development

Stereotype threat theory posits that stereotypes can harm performancewhether or not they are explicitly endorsed (Schmader & Johns, 2003;Spencer, Steele, & Quinn, 1999; Steele, 1997). The current study showedthat in addition to the implicit threat of stereotypes, an explicit endorsementis linked to enhanced self-concept among boys. Furthermore, for sixth- andeighth-grade boys, our results supported a mediational model linking per-ceptions of adult stereotypes, students' stereotypes, and self-concept. Priorresearch had shown that parents' stereotypes are related both to their chil-dren's concurrent perceptions of competence in mathematics and to thesechildren's later career choices (Bleeker & Jacobs, 2004; Jacobs, 1991;Jacobs & Eccles, 1992, Tiedemann, 2000). To our knowledge, researchershad not examined whether children perceive that adults hold traditionalgender stereotypes about mathematics and science or whether such percep-tions are related to children's own beliefs about gender differences in math-ematics and science abilities.

The current project extended prior findings by showing that boys whobelieve that adults favor boys in these academic domains hold similar beliefsthemselves. Adults' gender stereotypes may influence the child's mathemat-ics and science self-concept through many mechanisms, including parents'



and teachers' recommendations for course selection, attitudes toward home-work, and expectations of the child's performance conveyed in both subtleand blatant ways (e.g., attributing girls' mathematics successes to effort butnot to ability). Results of our study indicate that one way in which adults'gender stereotypes influence children's competence perceptions is throughthe formation of the children's own beliefs about gender differences in abil-ity. For all age and gender groups, whenever a perception that adults holdtraditional stereotypes was related to group competence or self-concept,those relations were parallel: Traditional adult stereotypes were related tochildren's own positive beliefs about boys and to negative beliefs aboutgirls. Among boys, our results indicate that one mechanism by which stereo-types lead to gender differences in mathematics and science outcomes isthrough the individual's personal endorsement of the stereotype.

An intriguing and somewhat surprising finding was the extent to whichall students viewed girls' mathematics and science performance positivelywhen they were evaluated as a gender group. Although status theory wouldpredict that girls as the low-status group would be less likely than boys toendorse traditional stereotypes, only fourth-grade boys ranked boys betterthan girls in their group competency ratings. On average, older boys wereegalitarian in beliefs, and girls in all three age groups showed a pronouncedin-group bias. These positive ratings of girls as a group, even among olderboys, may be due to the fact that throughout the elementary school grades,girls' classroom achievement is often superior to that of boys, or it may bedue to attempts in recent years to make mathematics and science educa-tional settings more gender inclusive. However, girls' positive views oftheir gender group's performance were not translated into more positiveself-perceptions: Girls' self-perceptions of mathematics and science abilitywere lower than those of boys, and gender group competence ratings wereunrelated to self-perceptions.

Social Status Theory, Stereotypes, and Identity Development

We tested two alternative theories about age and gender differences instereotype endorsement and in the associations among perceptions of adultstereotypes, children's stereotypes, and children's self-concept. Social sta-tus theory predicts that because of their lower status in this domain, girlswould be less likely than boys to endorse traditional stereotypes. Meangroup competence scores supported this hypothesis: Girls' estimates ofgirls' competence were strong, and boys' reports of girls' competence weresignificantly greater on average than girls' reports of boys' competence.Reports of one's own group's mathematics and science competence did not

402

differ by gender: All same-gender competence means were approximately70 to 75 on the 100-point scale. Most research examining the tendency ofindividuals to show in-group bias has involved experimental manipulationof group membership. In this study, social status theory was supportedwith a social category- gender--that is biologically based and is one ofthe most salient social categories for humans. In a prior study we foundsupport for social status theory in children's endorsement of race and gen-der stereotypes (Rowley et al., 2007). The current study extends thatresearch by showing that group competence perceptions are self-enhanc-ing for boys, are somewhat detrimental for young girls, and are unrelatedto self-perceptions for eighth-grade girls.

Even though middle school boys on average reported egalitarian beliefsboth for adults and themselves, their own self-perceptions of ability werehigher than those of girls, and the positive relation between stereotypes andself-concept showed that among boys who endorsed traditional stereotypes,endorsement was self-enhancing. In contrast, ratings of gender group compe-tence were unrelated to self-competence for middle school girls, who alsoreported lower self-perceptions of ability. Middle school girls who excel inmathematics and science and who have positive self views may downplaytheir group membership (e.g., "girls are not good in mathematics and science,but I'm not like other girls"), or they may hold more flexible views that allowthem to reconcile the seemingly inconsistent beliefs (e.g., "boys are betterthan girls in mathematics and science on average, but I'm not an averagegirl"). With a college-age sample, Blanton and his colleagues (2002) foundthat women who endorsed mathematics gender stereotypes evaluated theirown mathematics ability by comparing themselves only to other women.Like the college women in Blanton et al.'s sample, the girls in the currentstudy who endorsed traditional stereotypes may have used other girls as aframe of reference in evaluating their own abilities rather than basing self-evaluations on their social group membership.

Although our results might seem to imply that gender stereotypes havelittle negative effect on girls' striving for achievement, we believe that thecumulative evidence in this field strongly argues to the contrary. First, thepositive relations between adult stereotypes and self-concept for middleschool boys and the more positive mathematics and science self-concept ofboys found in our data and in other studies indicate that boys benefit fromthese stereotypes, whereas girls do not. Second, where perceptions of adultstereotypes were related to competence perceptions among girls, thoseassociations were negative (i.e., negative views of girls among fourthgraders and negative self-concept among sixth graders). Results of thisstudy imply that for those girls who are consistently exposed to gendered



views about mathematics and science, the harm is already done to girls'self-views by the time they reach middle school. Research with collegewomen who were majoring in mathematics or a mathematics-related fieldshowed that whereas stereotype endorsement among these women wasunrelated to their perceptions of their own ability (like the adolescent girlsin our sample), stereotype endorsement was related to their general per-formance self-esteem (e.g., "I feel just as smart as others") (Schmader et al.,2004). Efforts in recent years to encourage mathematics and science excel-lence in girls may be having an impact on explicit endorsement of stereo-types but not on girls' perceptions of competence, leaving them vulnerableto the implicit threat of stereotypes.

Finally, substantial gender differences continue to exist in secondaryschool educational tracks and career selection, and women are thus grosslyunderrepresented in fields such as computer science, engineering, and thephysical sciences (NCES, 2003; National Science Foundation, 2001). Con-sistent with the views of Eccles, Jacobs, and their colleagues (2002), webelieve that by the time girls reach middle school, a significant number ofthem have already formed identity beliefs that denigrate the importance ofmathematics and physical science for their own educational and career tra-jectories. Thus, a higher percentage of girls, compared to boys, havedecided that these domains are not interesting, are not of value to them per-sonally, and do not play a role in defining their identities. Therefore, ifstereotypes are not related to self-concept among girls, it is because manyof these girls may have already chosen other areas to value and in which toexcel. Our results are consistent with other research that shows that by thetime they reach high school, many girls have turned away from mathemat-ics and the physical sciences as areas that are unimportant to their sense ofself. This process is probably gradual and continuous throughout adoles-cence and young adulthood: Even among mathematically talented womenin college, gender stereotypes about mathematics are related to women'srejection of mathematically related careers and graduate studies (Nosek,Banaji, & Greenwald, 2002; Schmader et al., 2004).

Stereotypes, Group Identity and Competence in Younger Boys

The cultural-experiential theory posits that middle school children wouldbe more likely than fourth graders to be aware of adult stereotypes andwould be more likely to report a self-concept that was consistent with per-ceptions of adult beliefs. Our results with boys supported this hypothesis.Whereas posited stereotype--self-concept links were found for sixth- andeighth-grade boys, these connections were not significant among fourth-

40A

grade boys. However, contrary to the cultural-experiential prediction, thefourth-grade boys in our sample were more likely than all other groups toreport that adults view boys as better than girls in mathematics and science.We believe that this result may be due to two factors.

First, as noted above, social identity theory shows the importance ofthe individual holding positive views about his or her social group and thebias that results from this self-enhancement. Younger boys might be morevulnerable to this sort of distortion of beliefs related to group membershipthan middle school boys (Eccles, Midgley, et al., 1993). In addition, theefforts in recent years to strive toward gender equity in mathematics andscience may have cued older children to the politically correct response.Thus, middle schoolers (and younger girls, a low-status group) reported onaverage that adults hold egalitarian views, whereas younger boys may havebeen less aware of the doctrine of no gender differences. An altemativeexplanation of the findings is that relatively small cell sizes limited statisti-cal power; thus, additional research is warranted to replicate and extendthese findings.

Identity Development, Perceptions of Competence,and the Power of Peers

Research in peer relationships and social development speaks strongly tothe role of group membership in identity development. Adolescents areprofoundly influenced by the interests, values, and behaviors of the socialgroups to which they belong (Cairns & Cairns, 1994; Kandel, 1978). Aslong as ability in mathematics and the physical sciences is viewed as incon-gruous with a feminine identity, it is not surprising that girls are turning toother areas in which to excel as they reach adolescence. The process bywhich social stereotypes lead to differences in competence perceptions,which in turn lead to decreased value and interest in mathematics amonggirls, is similar to the process that leads to academic disengagement formany African American youths (Rowley et al., 2007).

Marsh and his colleagues (2005) demonstrated that prior math self-concept predicted later interest scores as well as school grades and standard-ized test scores in a longitudinal study of young German adolescents.Similar research by Jacobs and colleagues (2002) showed that changes incompetence beliefs for children from first through twelfth grades accountedfor changes in task values. As children enter adolescence and their socialgroup membership becomes increasingly important to them, their beliefsabout their own skills and interests become more aligned with their percep-tions of normative behavior and values for their social group. The energy



they devote to various activities and the choices they make in terms ofcourse selection, extracurricular activities, and future career objectives soonlead to disparate paths for boys and girls. Our results show that late child-hood is an appropriate developmental period for interventions aimed atreducing the gender gaps in mathematics and science educational and careertrajectories. Such interventions should consider children's self-competenceperceptions, interests, and social identities as well as the deeply engrainedstereotypes of adults.

References

Aboud, F. E. (1988). Children and prejudice. Oxford, UK: Blackwell.

B andura, A. (1986). Social foundations of thought and action: A social-cognitivetheory. Englewood Cliffs, NJ: Prentice-Hall.

Bigler, R. S., Brown, C. S., & Markell, M. (2001). When groups are not createdequal: Effects of group status on the formation of intergroup attitudes in chil-dren. Child Development, 72, 1151-1162.

Blanton, H., Christie, C., & Dye, M. (2002). Social identity versus reference framecomparisons: The moderating role of stereotype endorsement. Journal ofExperimental Social Psychology, 38, 253-267.

Bleeker, M. M., & Jacobs, J. E. (2004). Achievement in math and science: Domothers' beliefs matter 12 years later? Journal of Educational Psychology, 96,97-109.

Bornholt, J. J., Goodnow, J. J., & Cooney, G. H. (1994). Influences of genderstereotypes on adolescents' perceptions of their own achievement. AmericanEducational Research Journal, 31, 675-692.

Cairns, R. B., & Cairns, B. D. (1994). Lifelines and risks: Pathways of youth in ourtime. Cambridge: Cambridge University Press.

Campbell, J. R., Hombo, C. M., & Mazzeo, J. (2000). NAEP 1999 trends in aca-demic progress: Three decades of student performance. Education StatisticsQuarterly, 2, 31-36.

Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination inhuman behavior. New York: Plenum.

Eccles, J. S., Midgley, C., Wigfield, A., Buchanan, C. M., Reuman, D., Flanagan,C., & MacIver, D. (1993). Development during adolescence: The impact ofstage-environment fit on young adolescents' experiences in schools and infamilies. American Psychologist, 48, 90-101.

Eccles, J. S., & Wigfield, A. (2002). Motivational beliefs, values, and goals. In S. T.Fiske, D. Schachter, & C. Zahn-Waxler (Eds.), Annual review of psychology:53, 109-132. Palo Alto, CA: Annual Review.

406

Eccles, J. S.,Wigfield, A., Harold, R. D., & Blumenfeld, P. (1993). Age and genderdifferences in children's self and task perceptions during elementary school.Child Development, 64, 830-847.

Feldlaufer, H., Midgley, C., & Eccles, J. S. (1988). Student, teacher, and observerperceptions of the classroom environment before and after the transition tojunior high school. Journal of Early Adolescence, 8(2), 133-156.

Fredricks, J. A., & Eccles, J. S. (2002). Children's competence and value beliefsfrom childhood through adolescence: Growth trajectories in two male-sex-typed domains. Developmental Psychology, 38, 519-533.

Hall, W., & Jose, P. (1983). Cultural effects on the development of equality andinequality. In R. Leahy (Ed.), The child's construction of social inequality (pp.253-285). New York: Academic Press.

Huston, A. C. (1983). Sex typing. In E. M. Hetherington (Ed.), Handbook of childpsychology: Socialization, personality, and social development (Vol. 4, pp.388-467). New York: Wiley.

Jacobs, J. E. (1991). Influence of gender stereotypes on parent and child mathemat-ics attitudes. Journal of Educational Psychology, 83, 518-527.

Jacobs, J. E., & Eccles, J. S. (1992). The impact of mothers' gender-role stereotypicbeliefs on mothers' and children's ability perceptions. Journal of Personalityand Social Psychology, 63, 932-944.

Jacobs, J. E., Lanza, S., Osgood, D. W., Eccles, J. S., & Wigfield, A. (2002).Changes in children's self-competence and values: Gender and domain differ-ences across grades one through twelve. Child Development, 73, 509-527.

Kandel, D. B. (1978). Homophily, selection, and socialization in adolescent friend-ships. American Journal of Sociology, 84, 427-436.

Leahy, E., & Guo, G. (2001). Gender differences in mathematical trajectories.Social Forces, 80, 713-732.

Leahy, R. L. (1983). Development of the conception of economic inequality:Explanations, justifications, and concepts of social mobility and change.Developmental Psychology, 19(1), 111-125.

Marsh, H. W., Trautwein, U., LUidtke, 0., Kbller, 0., & Baumert, J. (2005). Aca-demic self-concept, interest, grades, and standardized test scores: Reciprocaleffects models of causal ordering. Child Development, 76, 397-416.

Marsh, H. W., & Yeung, A. S. (1997). Causal effects of academic self-concept onacademic achievement: Structural equation models of longitudinal data. Jour-nal of Educational Psychology, 89, 41-54.

National Center for Education Statistics. (2003). "Table 255. Bachelor's, master's,and doctor's degrees conferred by degree-granting. institutions, by sex of stu-dent and field of study: 2000-2001." http://nces.ed.gov/pubs200l/proj0l/tables/table29.asp. Retrieved June 3,2003.


National Science Foundation. (2001). Characteristics of doctoral scientists andengineers in the United States. NSF 03-310. Project Officer, Kelly H. Kang.Arlington, VA: Division of Science Resources Statistics.

Nicholls, J. G. (1978). The development of the concepts of effort and ability, per-ception of academic attainment, and the understanding that difficult tasksrequire more ability. Child Development, 49, 800-814.

Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). Math = male, me =female, therefore math # me. Journal of Personality and Social Psychology,83, 44-59.

Rowley, S. J., Kurtz-Costes, B., Mistry, R., & Feagans, L. (2007). Social status as apredictor of race and gender stereotypes in Black and White youth. SocialDevelopment, 16, 150-168.

Rowley, S. J., Sellers, R. M., Chavous, T. M., & Smith, M. A. (1998). The relation-ship between racial identity and self-esteem in African American high schooland college students. Journal of Personality and Social Psychology, 74,715-724.

Ruble, D. N., Martin, C. L., & Berenbaum, S. A. (2006). Gender development. InW. Damon & R. M. Lerner (Series Eds.) & N. Eisenberg (Vol. Ed.), Handbookof child psychology: Vol. 3, Social, emotional, and personality development(6th ed., pp. 858-932). Hoboken, NJ: Wiley.

Ruble, T. L., Cohen, R., & Ruble, D. N. (2001). Sex stereotypes. American Behav-ioral Scientist, 27, 339-356.

Schmader, T., & Johns, M. (2003). Converging evidence that stereotype threatreduces working memory capacity. Journal of Personality and Social Psychol-ogy, 85, 440-452.

Schmader, T., Johns, M., & Barquissau, M. (2004). The costs of accepting genderdifferences: The role of stereotype endorsement in women's experience in themath domain. Sex Roles, 50, 835-850.

Signorella, M. L., Bigler, R. S., & Liben, L. (1993). Developmental differences inchildren's gender schemata about others: A meta-analytic review. Develop-mental Review, 13, 147-183.

Spencer, M. B., & Markstrom-Adams, C. (1990). Identity processes among racialand ethnic minority children in America. Child Development, 61, 290-310.

Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat andwomen's math performance. Journal of Experimental Social Psychology, 35,4-28.

Steele, C. M. (1997). A threat in the air: How stereotypes shape intellectual identi-Sties and performance. American Psychologist, 52, 613-629.

Stetsenko, A., Little, T. D., Gordeeva, T., Grasshof, M., & Oettingen, G. (2000).Gender effects in children's beliefs about school performance: A cross-culturalstudy. Child Development, 7], 517-527.



Tajfel, H. (1970). Experiments in intergroup discrimination. Scientific American,233, 96-102.

Tiedemann, J. (2000). Parents' gender stereotypes and teachers' beliefs as predic-tors of children's concept of their mathematical ability in elementary school.Journal of Educational Psychology, 92, 144-15 1.

Weaver-Hightower, M. (2003). The "boy turn" in research on gender and educa-tion. Review of Educational Research, 73, 471-498.

Weinraub, M., Clemens, L. P., Sockloff, A., Ethridge, R., Gracely, E., & Myers, B.(1984). The development of sex role stereotypes in the third year: Relation-ships to gender labeling, gender identity, sex-typed toy preferences, and familycharacteristics. Child Development, 55, 1493-1503.

Wigfield, A., Eccles, J. S., Yoon, K. S., Harold, R. D., Arbreton, A., Freedman-Doan, C., & Blumenfeld, P. C. (1997). Change in children's competencebeliefs and subjective task values across the elementary school years: A 3-yearstudy. Journal of Educational Psychology, 89, 451-469.

COPYRIGHT INFORMATION

TITLE: Gender Stereotypes about Mathematics and Science andSelf-Perceptions of Ability in Late Childhood and EarlyAdolescence

SOURCE: Merrill-Palmer Q 54 no3 Jl 2008

The magazine publisher is the copyright holder of this article and itis reproduced with permission. Further reproduction of this article inviolation of the copyright is prohibited. To contact the publisher:http://wsupress.wayne.edu/

Gender Stereotypes about Mathematics and Science and Self-Perceptions of Ability in Late Childhood and Early Adolescence

Documents