A few months ago, we wrote a letter to the editor at Science, in response to a research article entitled “Global sex differences in test score variability” that was published in Science by economists Stephen Machin and Tuomas Pekkarinen.

These two male researchers analyzed standardized tests given to children around the globe and analyzed them by gender. They found that boys showed greater “variance” on math test sections, meaning that boys scored more frequently than girls in both the highest and lowest extremes on math sections.

This type of “variance” has been reported elsewhere and was the crux of Larry Summers’ argument that women are less genetically predisposed to excellence in math and science than men. In fact, this variance argument was also used by The Economist as a reason for why there are so few female nobel laureates.

The argument is: boys tend to populate the extremes of intelligence, creating more male geniuses than female geniuses.

We were disgusted to see that this kind of “research” is getting funded and published in high places like Science. So, we wrote them a letter to express a scientific counter-point to the article. However, Science did not publish our letter, so we are publishing it for the first time, here, on feministchemists.com.

One final point to ponder is whether Science would have published this article if it had been titled  ”Global Race Differences in Test Score Variability.”

Here is the letter.

A recent article entitled “Global Sex Differences in Test Score
Variability”(1) poses the politically loaded question, “Do boys and girls differ in their intellectual and cognitive abilities?” The ensuing discussion focuses on the authors’ findings that boys exhibit greater variance than girls on standardized tests, particularly in mathematics where boys dominate the highest percentiles. The authors reference, but fail to criticize, that this type of gender-based variance has been used to explain the lower frequency of women at the highest levels of intellectual achievement, and in doing so perpetuate the flimsy hypothesis that there is a causal relationship between gender and intellectual achievement.

The fundamentally flawed assumption of this article is that standardized tests are relevant data sets for predicting achievement and excellence. On the contrary, a recent collaborative publication on expert performance (2) found that measures of basic mental capacities, such as IQ tests, failed to predict the achievements of successful scientists, chess masters, and artists. More accurate predictors of individual achievement included extraordinary familial support and access to excellent mentors, as well as personality traits including interest and self-discipline, factors that cannot be quantified by a standardized test. As it has been well-documented that there is a dearth of female mentors for women in math and science (3), such a lack of mentorship is one valid explanation for women’s underrepresentation in those fields.

The underrepresentation of women at the highest levels of educational and professional success cannot be justified by the variance on tests which have no real value in predicting achievement. The paucity of highly successful women is a result of systemic and institutional biases against women (3). Let’s spend our resources on trying to solve the real problem, rather than reinforcing the damaging stereotype that “girls aren’t good at math.”

References:

1. S. Machin,T. Pekkarinen, Science 322,1331 (2008).
2. K. A. Ericsson, N. Charness, P. J. Feltovich, R. R. Hoffman, Eds., The Cambridge Handbook of Expertise and Expert Performance (Cambridge University Press, 2006).
3. Beyond Bias and Barriers: Fulfilling the Potential of Women in Academic Science and Engineering(Committee on Maximizing the Potential of Women in Academic Science and Engineering, National Academy of Sciences, National Academy of Engineering, and Institute of Medicine, National Academies Press, 2007).

We encourage you to also contact Science regarding this article.