Melvin Konner
Emory University, Atlanta, GA 30322, USA
Respond to this E-Letter:
Re: Assessment of Ability in Mathematics and Science
|
The article, “Gender Similarities in Mathematics and Science,” by J. S. Hyde and M. C. Linn (27 Oct. 2006, p. 599), makes one very important point but obscures another. The sex difference in mathematical ability is small or trivial when we consider means and effect sizes. When assessing applicants for a position requiring average, high, or perhaps even very high mathematical ability, we should expect equal or almost equal numbers of males and females. However, the article neglects the tails of the distribution; worse, it presents a false picture of them in a figure depicting idealized Gaussian curves for males and females, for a hypothetical difference with an effect size of 0.2. Here, the curves are exactly the same distance apart on the x axis except at the tails, where they merge before being truncated. Why? The reality in mathematical ability, for example, is that male variance is higher, so that both tails extend beyond the female distribution. This fact lends a different interpretation of the similarities between the means and, more importantly, it suggests a predominance of males at the very top as well as at the bottom; both have been shown in many studies. For readers of Science, the very top is very important. Even for the Fields Medal, only the work, not the person, should matter, but the expectation of gender balance at this level is unfair and counterproductive. I have three daughters who got A’s in math and I want them to do well; my wife is much better at math than I am. I applaud Shibley Hyde and Linn for smoothing the path for them. But to obscure the very different situation at the tails of the distribution is not going to help us move forward.
Melvin Konner
Emory University, Atlanta, GA 30322, USA. |
E-Letters: ![[Read E-Letter]](/icons/misc/sectionDOWN.gif){kind=icon}
