Gender differences in mathematics achievement—here today and gone tomorrow?

A number of studies have attempted to assess more systematically the effect of the nature and format of the assessment instrument on performance in mathematics. Anderson (1989) compared the achievement of first-year female and male mathematics students enroled in specialized mathematics degree courses at several universities in the UK on a range of objective tests-multiple-choice items, analysis questions, and two sets of true-false questions, one with and one without a penalty for incorrect answers. The female students, he reported, on average scored somewhat lower than did the males on each of the four types of questions asked. Earlier work by Murphy (1981) reported findings consistent with those of Anderson (1989). Using a sample 15-yearold Irish school students, Bolger and Kellaghan (1990) compared students’ school performance on multiple-choice and free-response test items in mathematics, Irish and English tests. Only slight differences, generally in favour of males, were found on both the multiple-choice and free-response language measures. Gender differences, again in favour of males, were more pronounced on the mathematics tests and ‘Males score[d] one third of a standard deviation higher than females on the free-response test and almost one half of a standard deviation higher on the multiple-choice test’ (p. 170). Johnson (1984) used American college students to trace gender differences in problem solving over time. Nine different experiments were carried out. He concluded that …the male advantage seems to extend broadly through the domain of word problems…and to be independent of prior exposure to the particular problem used… However, when the same subjects are given more formally deductive problem-solving tasks to perform, the male advantage evaporates into insignificance…suggesting that the male advantage may have something to do with translating a verbally expressed situation into a representation that can be attacked analytically or mathematically. A facility for bringing relevant real-world knowledge to bear on the problem may also be involved (pp. 1368-9).