Unz/Gelman Exchange

March 13, 2013

Dear Prof. Gelman,

I wonder if you could assist me on a matter that draws on your considerable statistical expertise…

In recent weeks, you have heavily promoted the research of Prof. Janet Mertz, publishing four or five separate columns quoting or citing her serious doubts about the accuracy of my ethnic analysis.  In fact, you found her work so persuasive that you were led to characterize me as merely a “political activist” whose research methods were “sloppy.”

All her data was drawn from her November 2008 AMS paper, which you cited and linked to on multiple occasions:

Given your strong endorsement of her analytical skills and the objectivity of her research methods, I assume you reviewed her paper in close detail, and am aware that her central focus was a sweeping refutation of the widespread belief that men tend to have greater math ability than women, especially at the high end.  Indeed, a few months after her paper appeared, she received media coverage for her paper’s findings, being quoted as saying “It’s not an innate difference in math ability between males and females.”

The massive verbiage of her 10,000 word research paper was a bit eye-glazing for me to read, but her Math Olympiad data constituted the central core of her analysis, and those figures were conveniently summarized in Table 6 on p. 1252.  This table provides the number of top male and female math students for the 34 leading countries over a couple of decades.

Obviously, tiny datasets may have statistical fluctuations, but I noticed that averaged across all the countries, about 95% of the top math students were male and 5% female.  For almost every country, the top math students for that twenty year period were close to 95% male, with the figure for the vast majority of the countries being in the range 93% to 97% male.

Now you are an award-winning Ivy League statistics professor while I may indeed be a “sloppy” researcher.  So I wonder if you could tell me whether you agree with Mertz that her data constitutes strong evidence that males and females have very similar math ability at the high end, and if so, your reasoning.  I’d appreciate your using simple words that I can easily understand.

Yours Sincerely,

Ron Unz

=======

March 14, 2013

Hi, Ron.  Just to be clear, I characterized you as a political activist, not as “merely” a political activist.  I could have followed Wikipedia and characterized you as a “former businessman and political activist” but the “former businessman” part didn’t seem so relevant.  As I wrote on the blog, I do not think of the term “political activist” as a criticism!  Also, if you do not like the term “sloppy” you can feel free to replace it with “cursory,” which is the term that you used.  If you look carefully, I did not describe everything you do as sloppy; rather I referred to “sloppy counting.”  But, again, it would be fine to do a global search and replace and change “sloppy” to “cursory” or “causal.”

I have skimmed Mertz’s articles but did not read them in detail.  I have no particular judgment on her analytical skills; I merely take her word that she interviewed various Olympiad team members and others, which is my impression of where her ethnicity information came from.  This is not something I have followed in detail, and I have tried throughout the discussions to make it clear that I am referring to the work of others.  My impression is that you put a lot of work into writing your article but that in certain particular points, others such as Mertz happen to have additional information that is relevant.  This is one of the benefits of having your article widely circulated, that various experts will learn of it and provide additional relevant information.

Finally, I have not studied the question of male and female math abilities but of course I’ve noticed a preponderance of males in various math events.  Ideally it is possible for people to agree on the numbers and then to argue about various explanations.  It’s hard for me to say more than this, not for any political reasons, but just because this is something I haven’t studied.  Regarding the numbers that you, Mertz, and others have provided, I have been careful to emphasize that to note the mistakes you have made is not to dismiss your ideas.  There certainly have been large demographic changes in the United States in recent decades, and the result is increasing academic competition. These things are worth studying.

Andrew

Comments are closed.