16 September 2011

Sex, hormones, STEM, and the chronic abuse of correlation

Get ready for a small tirade on gender, science and the abuse of statistics.


This month, a scientific study was released to the press about how overexposure to a certain hormone while in the womb is correlated with more interest in "thing" related careers (farmer, scientist), and less interest in "people" related careers  (social worker, teacher).  In other words, yet another study was released claiming to have found evidence for what Harvard ex-president Lawrence Summers called women's "different availability of aptitude" in STEM fields.  The Penn State researchers gave surveys about job preferences to young women with congenital adrenal hyperplasia (CAH), an endocrine disorder caused by too much male sex hormone in the womb.  Indeed, the women with CAH were more likely to choose the "thing" jobs than women without CAH, implying a biological reason why women are underrepresented in science.  Suggested a graduate student on the project: "[m]aybe we could show females ways in which an interest in people is compatible with STEM careers."



What first got my hackles up was the lack of any discussion about how else CAH could have affected the lives of these women.  CAH is the most common cause of sexual ambiguity, or people who identify as intersex. In people without a Y chromosome (biology refresher: biologically born women are typically XX, and biologically born men are typically XY), this means anything from "partial masculinisation that produces a large clitoris, to virilisation and male appearance. The latter applies in particular to Congenital adrenal hyperplasia due to 21-hydroxylase deficiency, which is the most common form of CAH" (Wikipedia - yeah, I'll cite Wikipedia; I have nothing to be ashamed of).  In other words, the women in this study most likely have their bodies marked by CAH in a way that probably impacted their sociocultural experience, especially in the formative years of puberty.  According to the website intersexualite.org, "some CAH individuals have been identified as males at birth and are reared as boys despite the presence of XX chromosomes and ovaries. In other cases, the masculinization of prenatal life is interrupted at birth, and the child is surgically and hormonally treated and reared as a girl. These girls often have characteristics that are popularly stereotyped as masculine. In addition, many CAH individuals identify themselves as lesbians."  While all the participants in the subject had female genitalia and were raised as female (whatever that means), there is no mention of other CAH-related variables that may have marked their experiences in society.   I'm not saying its not an interesting result, but clearly there may have been additional post-natal variables impacting the choices of these women.  I would love to interview the participants and find out more about their lives, their communities, their educational experiences, and what they believe shaped their career decisions.
While clearly biology has an impact on who we are, researchers who study identity have shown that who we are is a more complicated mix of biological and sociocultural factors than these kinds of studies can possibly sort out.  Just to be clear here, my problem is not with the science: it's interesting, and I'm sure the researchers were thorough.  My problem is with the way in which the results have been presented, both by the researchers, and the science reporters, and the conclusions that have been drawn from them, without reference to other variables.

How many times do I have to say this?  CORRELATION DOES NOT IMPLY CAUSALITY PEOPLE!  Haven't these journalists ever taken statistics?  It's this kind of reckless conclusion "suggesting" that fuels the continuing debates on scientific realities like climate change and evolution.  I know, I know, I worked for a newspaper, I know what it's like to have the A1 editor breathing down your neck about having a compelling lede, and getting people to think this is something totally new, and its all good if you get one (preferably two) dissenting quotes down near the bottom.  But stories like this can be hurtful to our collective idea of what a scientist is, and what a woman is, and what a woman scientist is.  My advisor always says that being a woman scientist is paradox in our society, and when an article opens by suggesting that "[i]f you are a woman who wanted to be a physicist, you may have been fighting your own sex hormones to do it," I know she's got a point.

To me this study makes a strong case for the feminist notion of corporeality, of our identities being marked on our bodies, especially the bodies of women.  Traditionally science, rationality, reason and masculinity have been associated with a mind-body distinction unique to the biological male, while shallowness, sensuality, irrationality and femininity have been associated with a preoccupation with the physical self that makes rational, scientific work unsuitable for ladies.  This is the same classic duality that shows up explicitly in the study, as an interest in "careers related to things" (heady, male careers) as opposed to "careers related to people" (natural, female careers).  (As a side note, just by the choice of those categories we already see evidence of science being socially constructed, even as it is constructing and reconstructing social reality.)   Feminist theory confronts this duality of male and female, rationality and corporeality, by making explicit the importance of the body in our construction of gender and identity.  "The body prompts memory and language, builds community and coalition. The body is a pedagogical devise, a location of recentering and recontextualizing the self and the stories that emanate from that self" (Cindy Cruz,  "Toward an Epistemology of a Brown Body").  I would love to know how the women in this study have experienced CAH, and how it has affected their corporeality, and sociocultural reality.

Here are several links to articles based on this study:

1. Science 2.0 story
2. Time Healthland story
3. Science Daily story

And the original citation:

Beltz, A.M., Swanson, J.L., and Berenbaum, S.A (2011).Gendered occupational interests: Prenatal androgen effects on psychological orientation to Things versus PeopleHormones and Behavior, 60(4).

And the citation for Cindy Cruz's brilliant article, Toward an Epistemology of a Brown Body, which explores the importance of the body and the queer Chicana identity:

Cruz, C. (2001). Toward an epistemology of a brown body. Qualitative Studies in Education,14(5), 657–669.

As a final treat if you didn't get riled up enough by my tirade, here's a cartoon that was a popular email forward for a while, a modern reification of the traditional male-female duality.



2 comments:

Anonymous said...

Hey Zoe - great post! This kind of stuff really bothers me too. These studies are doing no one a service. Females and males may have differences but so do people born in large families vs. small families, people raised in cities vs. country, etc. Choosing two random women and comparing them will come up with as many differences as comparing a random man and a random women. So by starting off the study with the assumption that women and men have different scientific aptitudes, the researchers have already decided that there is a significant difference. Research is always done within a certain framework and expecting a certain hypothesis - and I have had it with these male/female studies.

Craig Faustus Buck said...

This sort of science reminds me of the randomness illustrated in The Nietzsche Family Circus: http://www.losanjealous.com/nfc/