New trends in gender and mathematics performance: a meta-analysis.
Level 1 - systematic review of randomized trials
Meta-analysis and large probability dataset synthesis (design analogy for non-clinical research)
PubMed 21038941 · doi:10.1037/a0021276
What was done
The authors conducted a meta-analysis of 242 studies published between 1990 and 2007 examining gender differences in mathematics performance among 1,286,350 participants. They also analyzed multiple large probability datasets of U.S. adolescents over a 20-year span, including the National Longitudinal Surveys of Youth, the National Education Longitudinal Study of 1988, the Longitudinal Study of American Youth, and the National Assessment of Educational Progress.
What was found
Across the 242 meta-analyzed studies, the overall effect size was d = 0.05 with a variance ratio of 1.08. In the U.S. adolescent probability datasets, effect sizes for gender differences ranged from -0.15 to +0.22, and variance ratios ranged from 0.88 to 1.34.
Why it matters
These findings show that male and female performance in mathematics is essentially equivalent in contemporary cohorts, contradicting persistent assumptions of substantial gender disparity and unequal variance in math ability.
Limits
The abstract does not report subdomain-specific performance, age-stratified breakdowns, or formal assessments of publication bias. The secondary longitudinal datasets are restricted to U.S. adolescent populations, which may not generalize globally.
Cited by
- supports There are no learning science papers showing that middle school girls are incapable of learning middle school math and science.