Within countries, it is common to find larger differences among schools than between girls and boys within any particular school. Looking at differences across four high schools in Shanghai, Jinni Xu and Edwin Farrell found larger school than gender differences. In Australia, the percentage of students from different mainland states in year 8 who went on to complete year 12 math ranged between 16% and 55% for females and 21% and 50% for males (Willis). Similarly, one U.S. study found that calculus course completion varied across schools from 7% to 42% for females and 5% to 41 % for males (Dick & Rallis). Differences among schools often reflect different curricula. Higher SES schools offer more advanced mathematics courses whereas schools in marginalized neighbourhoods tend to offer more remedial courses.

In reality, gender differences often are not found. When they are found, typically they are small, and, if anything, have been getting smaller over time (Hyde et al. 1990a, Linn & Hyde). School, ethnic, class and cultural differences in mathematics achievement are consistently much larger than gender differences.³ So why the fuss about gender?

Gender and the Symbolization of
Mathematics

Despite the fact that girls and boys have similar achievement patterns on many achievement tests and in mathematics classrooms, there remains in Eurocentric cultures a persistent belief that mathematics belongs to the realm of the masculine. I am speaking here of what Evelyn Fox Keller calls "the symbolic work of gender" (p.17) and Sandra Harding describes as the symbolic sex-gender system. In this analysis, gender is an analytical tool that can be applied to all culturally constructed human endeavours including mathematics. Although in certain contexts a cultural analysis may overlap with an analysis of individual gender, the two systems are logically, and to a large extent empirically, independent. Thus it is possible for mathematics to be gendered symbolically even though women and men do equally well at mathematics.

On the most obvious or conscious level is the stereotype of mathematics as a male domain. In a meta-analysis of gender differences in mathematics attitudes, Janet Hyde and her colleagues found the largest gender difference (effect size .90) for math as a male domain with girls rejecting the stereotype of mathematics as masculine much more strongly than boys (Hyde et al. 1990b). Furthermore, this gender difference was more than twice as large as the largest gender differences in mathematics achievement (effect size .40), which are found only with highly select samples. In an area of investigation where gender differences are so small, this difference in stereotyping stands out. What are the girls trying to say in their almost total rejection of the stereotype of math as a male domain?

I would suggest they are expressing a strong objection to their possible exclusion from a culturally masculine endeavour. On the other hand, by not rejecting this stereotype as I strongly as girls, boys may be expressing a reluctance to give up mathematics as a male territory. Given their male peers' significantly greater willingness to endorse the stereotype of math as a male domain, girls will be exposed to a number of subtle or not-so-subtle comments over the years, the cumulative effect of which will be to reinforce their sense of not belonging. In math intervention programs, drawing girls' attention to the sexist nature of math has been shown to increase their anxiety (Fennema 1993).

At a more subtle level, I want to examine the persistent bias in the achievement literature that men's mathematics skills and proficiencies are more valuable than those of women. For example, the disproportionate emphasis on standardized tests exists partly because these are the measures of achievement that reinforce the view that men are better at mathematics. Furthermore, the fascination of researchers and the media with studies of precocious samples, where the gender differences favouring males are the largest, reflects both the masculine and the elitist symbolization of mathematics.

Another assumption of the value of men's mathematics skills and proficiencies ,over those of women is the uncritical association of males with "higher level" problem solving skills and females with "lower level" computational skills. This generalization and the related assumption that when boys do math they are autonomous learners whereas girls are rote learners remain common beliefs even though empirical data raise a number of questions.