Higher Superstition, chapter 5: Gender

Chapter 5 is about feminist critiques of science.  To the extent that feminists critique the culture of scientific workplaces I freely agree that there is much that is worthy of criticism, much to work on, and much more to be done.

To the extent that feminist critics have gone after scientific knowledge, I would say the following:
1) When people like Sandra Harding offer up essentialist arguments for why some particular idea in the basic sciences is more "masculine" or more "feminine", I think they are not only wrong but dangerous.  There's no distinct female perspective on chemical bonding or thermodynamics or electromagnetic waves.  To argue that there is risks bringing in the idea that women and men have different comparative advantages in the basic sciences.  That idea is not only devoid of empirical support, it is also an open door to justifying gaps and discrimination.  Many feminists have thus rejected essentialist arguments, and justifiably so.

2) When feminist critics raise concerns about the topics chosen by applied scientists, I think they have a better argument, at least in some fields.  There's no distinctly feminine viewpoint on quality control in chemical synthesis or optimizing the design of fiber-optic networks, but perhaps if there had been more women in biomedical research sooner then it wouldn't have taken so long for the medical community to recognize that heart attack symptoms in women are often (not always) different from those that are most common in men.  It's not about whether women or men are more qualified to analyze the data or perform the medical procedures (it's obvious that women and men are equally qualified to work in medical research), it's that much research begins with an anecdote (since every hypothesis lacks proper support before it's tested) and female clinical researchers might notice certain patterns in anecdotes from patients.  (Just as male clinical researchers might notice certain patterns in anecdotes.)

Similar things could be said about other areas of medical and behavioral research.  Female engineers working on consumer products might pay a bit more attention to, say, differences in average body size, differences in typical user experiences, etc.

A harder issue is the tendency to favor the use of male mice in biomedical research.  I've heard many female biomedical researchers defend this practice, on the grounds that the reproductive cycle in female mice lasts only 4 days, so there's much more variability in their physiology over the course of a study.  If one takes seriously the notion that female physiology matters, then it matters that female mice are more variable so the data will be noisier.  With resources being finite (raising and handling mice takes time and money, as does tracking their reproductive cycles so the data can be properly analyzed) it makes sense that many studies should be done first with male mice to get some preliminary data.  But if you really want to generalize to humans, and you aren't studying a male-specific question, then at some point you need to study female mice.  There's a difference between justifying greater use of male mice and justifying exclusive use of male mice.

But, of course, this is something where the funding agencies need to get more blame than the people working in the trenches with limited budgets.

So my take on this chapter is that Gross and Levitt start off strong but go too far in rejecting feminist critiques.  They need to keep in mind the distinction between pure and applied research.  There's no distinct feminine perspective on arterial plaque, or even a female perspective on molecular mechanisms of cervical cancer, but life experiences will matter when evaluating clinical anecdotes that might lead to the formulation of working hypotheses, and certainly the technology used in cervical cancer treatment should be designed with input from women who actually undergo such procedures.

Similar things can be said about race.  There's no distinct ethnic/racial perspective on statistical analysis in a clinical study, but one's experiences might affect whether one notices certain lifestyle patterns in different racial/ethnic/economic groups, and that matters when formulating hypotheses.  Ethnic diversity surely matters on an engineering team working on facial recognition software, as shown by some unfortunate examples with consumer products.

Mind you, there's a difference between research and practice.  I've been examined by competent female health professionals for some male-specific problems, and if I had a skin disease I'd be happy to go to a dermatologist of any ethnic/racial background.  Race and gender need not affect a conscientious professional's competence to apply existing knowledge in practice, but life experiences might affect the hypotheses that one frames in research. The testing of hypotheses is or ought to be an objective matter, but the choice of a hypothesis is highly subjective, and perspective matters.