Explaining Gender Differences on SAT-Math Items

Egyptian Journal of Psychological Studies, 2009

The present study investigates gender differences in Academic achievement (school grades in sciences and languages) as mediated by reasoning ability in a large sample with a clustered data structure from an educational context. Whereas girls outperformed boys in languages, boys excelled in sciences and reasoning. Multilevel analyses indicated a small indirect effect of gender on school grades mediated by reasoning ability. Gender differences in sciences, but not in languages were largely explained by reasoning ability, but not by factors such as gender-ratio in the classroom or mental speed. The predictive power of reasoning ability for languages, but not sciences was larger in classrooms with higher mean reasoning ability, whereas gender differences in languages, but not sciences varied significantly across classrooms. Reasoning ability appears to be important for predicting Academic achievement in sciences, whereas languages are more affected by gender.

ETS Research Report Series, 1992

Twenty-four students who scored 650 or more on the Scholastic Aptitude Test Mathematics test (SAT-M) were asked to think aloud while solving 13 mathematics items in either multiple-choice or free-response format. Strategies students used to solve the items were classified as either algorithmic or insightful. Data analyses indicated that items in the free-response format were significantly more difficult for females than multiple choice items. No significant difference was found for males. Females were more likely than males to use algorithmic strategies in both the multiple-choice and free-response formats, with no significant difference in types of strategies used in either format. Males were more likely to use insightful strategies in the multiple-choice format than in the free-response format. Finally, on multiple-choice items, both males and females used options to detect calculation errors, with females using this strategy somewhat more than males. One table gives score distribution by sex. Appendix A contains multiple-choice and free-response versions of items used in the protocols. Appendix B provides samples of solution strategies. (Contains 31 references.) (Author/SLD)

Frontiers in Psychology, 2015

Learning strategies and general cognitive ability as predictors of gender-specific academic achievement.

Contexts in Mathematics Education Proceedings of the 16th Annual Conference of the Mathematics Education Research Group of Australasia Brisbane July 9 13 1993 Brisbane, 1993

Educational Research and Evaluation, 2009

This study was undertaken to assess gender differences in mathematics achievement and retention by using Problem-Based Learning (PBL). The design of the study was pre-posttest quasi-experimental. Four hundred and twenty eight senior secondary one (SS I) students using multistage sampling from ten grant-aided and government schools were involved in the study. Two hundred and sixty one male students and one hundred and sixty seven female students were taught algebra using PBL method of instruction. Algebra Achievement Test (AAT) constructed by the researchers was the main instrument used for data collection. Two hypotheses were raised for the study and tested using t-test at .05 level of significance. The study revealed that male and female students taught algebra using PBL did not significantly differ in achievement and retention scores, thereby revealing that male and female students are capable of competing and collaborating in mathematics. In addition, this finding showed that performance is a function of orientation, not gender. The studies recommend the use of PBL by mathematics teachers to overcome the male image of mathematics and enhance students' (male and female) achievement and retention.

The Development of Sex Differences and Similarities in Behavior, 1993

Gender differences in cognitive abilities and achievement are not new. They have been reported for several decades. Yet what is new is that recent reports seem to indicate that these differences are steadily diminishing in normative samples. That is, males and females apparently are converging toward a common mean on a variety of abilities, including mathematics (Feingold, 1988; Hyde et al., 1990; Rosenthal and Rubin, 1982). Hyde et al. (1990). for example, in their meta-analytic review of 100 studies, reported that the average effect size for the gender difference in mathematics was only .14 for studies published in 1974 or later compared to .31 for studies published earlier. Feingold (1988) studied scores on two test batteries over a 30-year period and also concluded that females have been catching up with males. Such findings have led several individuals to conclude that research on gender differences is better referred to as research on sex similarities (Connell, 1987; Riger, 1992). Some investigators have been more flippant. asserting that gender differences in cognitive functioning are decreasing "faster than the gene can travel". Although such conclusions are encouraging and consonant with the current Zeitgeist (Halpern, 1992), they are perhaps misleading. Stanley et al. (1992) have noted that, for at least the past 20 years, some test publishers have attempted to minimize what some would call"gender bias" by discarding, from one revision to the next, items that show the greatest gender disparities. So meta-analytic reviews in this area are difficult to interpret; and further, not all studies have documented a decline in gender differences (Benbow, 1988; Lubinski and Benbow, 1992). Firm conclusions are still unavailable. What we do know, however, is that gender differences vary as a function of a variety of variables, including age, ability-level of sample surveyed, ethnicity, and the ability itself. For example, Hyde et al. (1990) revealed that girls showed a slight 87 M. Haug et al. (eds.). The Developl7Ulnt o/Sex Differences and Similarities in Behavior. 87-109.

Sex differences in cognitive abilities, particularly at the extremes of ability distributions, have important implications for the participation of men and women in highly valued and technical career fields. Although negligible mean differences have been found in many domains, differences in variability and high ratios of males to females in the tails of the ability distribution have been found in a number of studies and across domains. A few studies have also observed trends over time, most noting decreasing ratios of boys to girls in the highest levels of mathematics test performance. In this study, sex differences in means, variances, and ratios were evaluated in four cohorts (1984, 1992, 2000, and 2011) in verbal, quantitative, and nonverbal/figural reasoning domains as measured by the Cognitive Abilities Test. Samples included U.S. students in grades 3-11. Overall, the results were consistent with previous research, showing small mean differences in the three domains, but considerably greater variability for males. The most surprising finding was that, contrary to related research, the ratio of males to females in the upper tail of the quantitative reasoning distribution seemed to increase over time. Explanations for this finding are explored.

Contemporary …, 1999

Males from select populations receive better scores on standardized math achievement tests than females. The research reported in this article evaluates the hypothesis that the reason for these differences is that males are faster at retrieving basic math facts. Studies 1-3 demonstrate that math-fact retrieval predicts performance on math achievement tests with students in grades 5-8 and in college. Studies 4-6 show that males and females in grades 2-8 and in college have different patterns of mathfact retrieval performance and that males at the high positive end of the retrieval distribution are faster than comparable females. Study 5 also demonstrates that math-fact retrieval varies in three populations (Anglo-American, Chinese-American, Hong Kong Chinese) and that speed of retrieval improves with practice. Studies 7-9 tested the hypothesis that males are faster than females on retrieval tasks in general. Study 7 showed that there were no gender differences on simple retrieval tasks, and Studies 8 and 9 showed that females were slightly faster than males on verbal-processing tasks. The General Discussion indicates that the math-fact retrieval hypothesis is consistent with previous research. It also relates the math-fact retrieval hypothesis to theories of cognitive performance and introduces the practice and engagement hypothesis. This hypothesis explains the origin of gender differences in math and reading and relates those differences to the existing literature on gender differences in academic performance. The article concludes with a description of needed future research and a discussion of the educational implications of the math-fact retrieval hypothesis.

The relative influence of gender, coursework, affect, and other hypothesized determinants cf mathematics achievement among high school seniors was examined, using data from the High School and Beyond (HSB) database. The HSB base-year survey used a multi-stage sampling scheme. Complete data on the variables of this study were available for 16,358 students (8,774 females and 7,584 males); all were membLrs of the sophomore cohort at the beginning of the HSB study. The dependent variable was mathematics achievement. The independent variables were: (1) mathematics courses; (2) mathematics attitudes; (3) educational aspirations; (4) academic orientation; (5) parental involvement; (6) prior achievement; (7) verbal ability; (8) socioeconomic status; and (9) gender, the central consideration.

Intelligence, 2010

One factor in the debate surrounding the underrepresentation of women in science technology, engineering and mathematics (STEM) involves male-female mathematical ability differences in the extreme right tail (top 1% in ability). The present study provides male-female ability ratios from over 1.6 million 7th grade students in the right tail (top 5% in ability) across 30 years (1981-2010) using multiple measures of math, verbal, and writing ability and science reasoning from the SAT and ACT. Male-female ratios in mathematical reasoning are substantially lower than 30 years ago, but have been stable over the last 20 years and still favor males. Over the last two decades males showed a stable or slightly increasing advantage in science reasoning. However, more females scored in the extreme right tail of verbal reasoning and writing ability tests. The potential role of sociocultural factors on changes in the male-female ability ratios is discussed and the introduction of science reasoning as a potential new factor in the debate is proposed. The implications of continued sex differences in math and science reasoning is discussed within the context of the many important interlocking factors surrounding the debate on the underrepresentation of women in STEM.