Unlike boys, when girls perform a math test at school, they may be at risk of experiencing an additional pressure, namely the threat of confirming or being judged in terms of the negative stereotype of women’s poor abilities in math. For instance, the activation of such a negative stereotype causes a decrement in girls’ performance in math (Huguet & Régner, 2007).
This stereotype threat phenomenon (Steele & Aronson, 1995; Steele, 2007) has been generalized to many social groups (i.e., Afro-American, men, the elderly...) and to a wide variety of fields (i.e., academic skills, memory, sports...). It is mediated by situational reductions in working memory (Schmader, Johns, & Forbes, 2008). There has also been many attempts to understand how to reduce stereotype threat. One way to counteract stereotype threat effects is to reduce the negative propositional relation between the concept of group (i.e., women) and the concept of ability domain (i.e., mathematics, Schmader et al., 2008). This may be done by providing girls with a positive counterstereotypic model (i.e., a girl who succeeds in mathematics) when performing a math test. For instance, Marx and Roman (2002) showed that girls performed significantly better a math test when it was administered by a competent woman in mathematics rather than by a competent man (see also McGlone, Aronson, & Kobrynowicz, 2006; Marx & Goff, 2005). These results were generalized to role models other than the experimenter (McIntyre, Lord, Gresky, Ten Eyck, Frye, & Bond, 2005; McIntyre, Paulson, & Lord, 2003). Within the class, we identify two different types of role models for girls: a woman teacher who is competent in mathematics and a girl who is top of the class. The influence of the first kind of positive model has already been studied (see Marx & Roman, 2002), we focus particularly on the second. Thus, we hypothesize that girls do better in math in classes where the top of the class is a girl than in classes where the top of the class is a boy.
To test this hypothesis, we conducted a longitudinal study using questionnaires. A sample of 1082 students was followed during their middle high school years (49.8% were girls and 50.2% were boys). In the present study, we focus on the 7th grade, that is a sample of 811 students (392 girls and 405 boys) nested in 44 classes from 15 colleges. Among teachers surveyed, 35.71% (n =10) were men and 64.29% (n =18) women.
For these students, we collected their initial level of math performance (a common test they performed at the end of 6th grade) and their average score in math at the end of 7th grade. For each class we identified the teacher’s and the top of class’ gender (that was determined on the basis of the student whose mathematics average score at the end of the 7th grade year was the best).