Brad DeLong congratulates Steven Levitt on winning the John Bates Clark Medal and gives us a list of previous winners. Awarded every other year since 1947, all of the winners are men. Brian Weatherson picks up on this point and notes that analytic philosophy “has done a very poor job over the years in attracting and keeping bright women,” though he thinks the situation is not quite as bad as economics. Matt Yglesias also notes the problem, and wonders whether anything will be done about it.

This is a complex issue with many nuances but, to be blunt about it, there’s a lot of good sociological research showing that essentialist explanations for occupational segregation are very weak. I don’t even have to read the comments section on Matt’s weblog to predict that, soon, people will be making confident assertions that women don’t go into analytic philosophy because it’s so tough and logic-oriented and they don’t like that, and by nature they don’t like all the arguing and nit-picking and so on. In fact (having just looked anyway), there’s one there already:

The problem is the nature of current analytic philosophy. It is the kind of purely abstract exercise that appeals primarily to men. … Women, bless them, care about people, and they care that their work have some relationship, even if tenuous, to the concerns and interactions of human beings in the world.

Ack. Explanations grounded in stereotypes like this are no good for any number of reasons. To pick one at random: there are a lot of women in the world. Don’t you think some of them might be interested in abstract analysis? It’s not as if philosophy is a very big field. But the key problem with stereotypes is that they are too flexible. Aren’t women also supposed to be endless talkers, complainers, nit-pickers and more detail-oriented than men? Sounds like a perfect background for philosophy to me. Or to put it another way, is making pizza a high-status man’s job or a low-status woman’s job? It depends who you ask.

Debate along these lines tends to be unproductive, because proponents of the essentialist line routinely respond by saying “Aha! So are you really saying you think there are no genetic differences between men and women?” which is very irritating. A good example of this has also popped up in Matt’s comments:

So, fellas, I ask a simple question:

Acknowledging that there will be exceptions to any rule, and acknowledging that any institutional discrimination is anathema in its own right, and acknowledging that differences in temperament or inclinations do not imply any particular moral conclusions, are there differences between the genders about which we can make true generalizations?

The poster here is using the word “acknowledging” as if it meant “ignoring”. “And what have the Romans ever done for us?” he did not add.

Rather than flog that horse any further, let me quickly point to some other aspects of this issue in the vain hope that it will prevent a pointless argument on the narrow topic of natural differences between men and women.

  1. Outright discrimination is real, but probably not the main mechanism. Explicitly-held beliefs about women’s innate inability to do economics, philosophy, algebra or whatever do persist, but are not terribly prevalent. Systems of occupational segregation do not need individual-level prejudices to sustain themselves, though of course it helps.
  2. Small but persistent hits are more common than outright bias. Although old-fashioned “girls can’t do it” discrimination is less common than it was, many women report subtler forms. This can involve things like having a slightly higher standard applied, not getting a second chance, not being listened to properly, not getting the same work-distribution as male colleagues, and any number of other similar problems. All of this, naturally, affects productivity. For instance, the MIT Gender Equity Report found that “Marginalization increases as women progress through their careers at MIT, and was often accompanied by differences in salary, space, awards, resources, and response to outside offers; women receiving less despite professional accomplishments equal to those of their male colleagues.” In nearly all cases, Department chairs or other decision-makers could give plausible particularistic reasons why a decision was made in a specific case, but the outcome was gender-specific.
  3. Occupational segregation is very fine-grained. There are women in economics and philosophy, of course. But these fields are also internally differentiated. Thus, many women in philosophy are in fields like ethics and ancient philosophy rather than self-styled “core” areas like metaphysics and epistemology. I’m less familiar with the internal structure of economics, but I’d predict something similar holds. Insofar as women are directed (e.g., by advisers) into lower-status subfields they won’t win things like the Clark Medal.
  4. Newer Fields are a Good Test Bed. In fields where occupational segregation is very well institutionalised, it is unreasonable to expect women to beat the odds, though they may sometimes do so. (When Dorothy Hodgkin won the Nobel Prize for Chemistry in 1964 the New York Times ran the headline “Nobel Prizewinner is a Grandmother” and the Observer described her as an “affable-looking housewife,” despite the fact that she was Wolfson Professor of Crystallography at Oxford.) Newer fields—like molecular biology, for instance—might be expected to have a more equitable gender distribution because women can participate from the beginning.
  5. Homophily is a powerful process. Homophily is the tendency of groups to be composed of members with similar attributes. Almost all of your social interaction will happen with your neighbors in “Blau space,” [pdf] the social space defined by the universe of sociodemographic attributes. This fundamental aspect of social interaction has strong implications for the kinds of patterns we expect to see when groups control resources, allocate rewards and so on.
  6. Introspection is not a good guide to outcomes. Few people think of themselves as unfairly biased against any particular group. Instead, they tend to think that outcomes are explained either (a) by some aspect of the group’s nature, or (b) by the simple aggregation of choices and preferences of group members. Situational and structural factors influencing choice tend to be ignored. (This is a variety of the old fundamental attribution error.)