Jun 01 2009
A new review article in Proceedings of the National Academy of Sciences (PNAS) analyzes recent math testing data and concludes that the gender performance gap in mathematics is non-existent in the aggregate level from K-12 in the US (thanks No Child Left Behind for all the wealth of testing data we have for low cognitive level math!). Summers’ now oft-quoted statement poses that the men have higher variability in intrinsic math ability, therefore there are more ‘high performing’ (and also ‘low performing’) men than women in math, and this difference is not a result of cultural/socialization issues: Directly addressing Summers’ hypothesis, the authors’ compare the high performing (99th percentile) population of students for a Minnesota state exam, showing some interesting differences by ethnicity. The ‘expected’ male:female ratio according to this hypothesis is 2.15, which actually correlates well with the White subgroup of this population (2.06). However, the ratio is 0.91 for Asians: ie girls were actually more likely to score higher than boys, albeit by a small margin. Girls score equally and in some cases better than boys from UK, Iceland, and Thailand for an international exam in 2003. Variability analysis of various exams from various countries conclude that the variability in scores for girls and boys is… wait for it… variable. (Some countries had negigible difference in score variability of girls vs boys, others had differences one way or the other, etc…)
Definitely some flaws in the article but interesting nonetheless.
Easily digestible summary on Science Daily.
Review article on PNAS (if you can’t see the full article and would like to read it shoot me an email)