Naive theories of gender differences in maths

An interesting study (Dar-Nimrod & Heine, 2006) [1] sheds some new light on the “gender and math” controversy, (in)famously reignited last year by Larry Summers.

Ilan Dar-Nimrod and Steven Heine have just published a paper entitled “Exposure to Scientific Theories Affects Women’s Math Performance”. At first I though “oh, more science education means that women are better at math” but wait, isn’t that true of men too? And isn’t that sort of obvious?

In fact, the study bears on the influence of theories explaining gender differences in math performances: can genetic theories of gender differences contribute to create the difference? The authors are inspired by the well know phenomenon of stereotype threat in which members of a stereotyped group perform in a way that conforms with the stereotype when their belonging to the group is primed. For example African-American whose race had been primed will perform less well on intelligence tests and women fare less well in math tests when primed with gender.

Dar-Nimrod and Heine used a clever trick to prime participants with different ‘facts’ related to the alleged gender gap in math skills. The has their participant complete a GRE like task that had three parts: the first and the third were mathematical problems, and the second was reading comprehension. The essay of the second part was used as the prime: in one case it said that there were no gender differences in math skills; in another case, it simply primed gender, without alluding to the gender gap (as in a classic stereotype threat experiment). The last two cases are the most interesting: both mentioned the gender gap, but they offered different explanations for it. One explanation was in terms of genetic differences, and the other in terms of personal experience.

The authors hypothesized that women who read the essay mentioning genetics as a cause for the gender gap would perform less well (as the women primed with gender) but that the essay explaining the gender gap by differences in experience would give similar results to the essay saying that there were no differences whatsoever. And this is exactly the result they obtained (both in this experiment and in a replication).

This shows two interesting facts: first, explaining differences in terms of genetics can have a self-fulfilling effect. Women who think that they are less good in math because of genetic differences will be less good. Second, an explanation in terms of experience can negate the stereotype threat effect. This finding may be even more interesting, since it suggests a way to reduce the gender gap: getting people to understand that experience plays an important role in their performances should improve the performances of people who might be suffering of a stereotype threat (and since this includes woman and minorities like African-American, it means that it is more than half of the population).

I’m not advocating that we engage in wishful thinking: if we had evidence that genetic differences are indeed a cause of the gender gap in math skills, this evidence should not be kept hidden in a vault. However, scientists and journalists should be very careful in their public statements to mention that experience still plays a role, or something to that effect. This might negate the otherwise deleterious effects of the findings on the math performances of women.

Now we should say: “You throw like a girl, but girls throw like girls because of their experience”.

---

[1] Dar-Nimrod, I., & Heine, S. J. (2006). Exposure to scientific theories affects women’s math performance. Science, 314(5798), 435-435.

This post was originally published in 2006 on the Alphapsy blog.

***

Comments

1. On Friday 20 October 2006 by olivier

I happened to be reading Michael Dummett’s “Grammar and Style”, in which you can read, in the entry titled “Gender and sex”:

“The word gender used to be applied exclusively to nouns, pronouns and adjectives in Indo-European and semitic languages to indicate the distinction between masculine and feminine (or between masculine, feminine and neuter. one of the uses of the world sex was to distinguish between male and female human beings and animals.

Anthropologists borrowed the word gender from the grammarians in order to distinguish social from biological classification: a transvestite might be of the female gender, though of the male sex. For a reason obscure to me, feminists have recently promoted the use of gender as a substitute for the word sex in the use mentioned above, even when the distinction is purely biological; they still, however, speak of sexism and not of genderism.

The change seems pointless, and is surely toi be depreated; it make sit impossible or at least clumsy, for example, to explain that in English the gender of the third-person pronoun depends on the sex of the person referred to (…). It also ruins the immortal line spoken by Jack Lemmon in Some Like It Hot, “It’s a whole different sex”.

2. On Thursday 26 October 2006 by Sandy G

This is an important experiment, but the results should be interpreted in a slightly different manner, with more emphasis on the differences in the underlying conceptualisation of Mathematical ability instead of that based on stereotype threat model alone.

I have attempted exactly such an interpretation in my blog post on The Mouse Trap : “Gender bias in Math skills : a case of Traits Vs. Environment/Effort feedback?”; in the absence of the mediating mechanism based on different conceptualisation of Math ability, it seems problematic to me that experiential explanation do not prime the gender stereotype, but genetic explanations do! What could be the rationale behind that?

3. On Thursday 26 October 2006 by hugo

I’ve read your post, and I like your explanation. The explanation based on stereotype and yours are not incompatible (I guess).
Two things I liked: you try to explain how the difference in the way skills are perceived (traits vs. acquired) can cause the difference in results (with people trying harder in the second case). I find this explication intuitively appealing. Moreover, you could link that with a huge literature (in personality psychology) on the ways people attribute their successes and failures to different causes (internal vs. external, etc.). It has been shown that depending on your ‘attribution mode’, the success with which you cope with failure (sic) will be very different.

4. On Monday 30 October 2006 by olivier

There are 2 nice posts at Cognitive Daily about stereotypes.

They say you can smother the cognitive impact of stereotypes by telling people how they work. Or you may expose them to positive stereotypes.