Wednesday, February 09, 2011

Orders of Magnitude

Thanks to all who sent me links to the recent PNAS article by Ceci & Williams on "Understanding current causes of women's underrepresentation in science". I had read reports of the results, and now have read the article itself.

The authors of the article focused on some of the most commonly cited reasons for the underrepresentation of women in "math-intensive fields of science": discrimination in reviewing of proposals and manuscripts; being selected for interviews in faculty positions; and in hiring. They propose that current data do not support these as primary causes for the underrepresentation of women today.

Their main conclusion is that "differential gendered outcomes in the real world result from differences in resources attributable to choices, whether free or constrained", and that underrepresentation would be best alleviated through changes that take into account "differing biological realities of the sexes."

I think they make some important points with their study, and I believe that the current situation for women being evaluated for jobs, grants, and publications is better than it has ever been. However, I continue to see and hear examples of discrimination in reviews and hiring committees -- faculty who doubt that women have their own ideas (but have no trouble believing this about male candidates), or who don't like "aggressive" women (but think this is a fine trait in a man), are alive and well. These issues are not as obvious and widespread as in the past, but neither are they isolated, rare incidents that can safely be ignored as irrelevant to current practices.

I was surprised that the study focused so much on databases related to the life sciences, a realm of science in which, as the authors note, women now make up the majority of PhD recipients. I realize that some biological sciences are quite "math-intensive", but the authors seem to use this term to refer specifically to the non-biological sciences. It would therefore make sense to base conclusions primarily on studies other than those involving the life sciences.

(A quibble: In the sentence about how the number of PhDs awarded to women in the life sciences has increased, 13% is described as "orders of magnitude less" than 52%. It is not.)

In the section on Discrimination Against Women in Journal Reviewing, the authors rely in part on a study of acceptance rates for the journals Behavioral Ecology, the Journal of Biogeography, and Nature Neuroscience. Those all seem kind of life sciencey to me.

Similarly, in the section on Discrimination Against Women in Grant Funding, the authors rely on studies of databases of Medical Research Councils and similar organizations of various countries, including the NIH. There is also mention of NSF and the Australian Research Council, both of which cover a wide range of fields in science, engineering, education, and beyond. If possible, it would have been interesting to see a separate analysis of recent data for the "math-intensive" sciences.

Having seen such data for my own field, I believe that NSF works -- with some success -- to provide a "gender-fair grant review process", but I don't think the authors of this particular study have demonstrated that with their chosen databases. [I chose the phrase "works to provide" rather than a simple "provides" based on experiences such as this (which referred to, but did not specify, that the problem was offensive sexist comments by a program officer) and this and this and this, all of which the excellent and enlightened staff at NSF can and do filter so that there are no deleterious effects on female PIs].

The authors note that there are "more women in teaching-intensive, part-time posts where research resources are scarce", and attribute this to life-style choices or constraints. When discussing the situation at major research universities, the authors cite an NRC task force report that concluded that women were not at a disadvantage for interviews and offers (and may have a slight advantage) in a study of 6 fields of "natural science". That's great, but I think it is premature to propose, based on these data, that universities should discontinue efforts to train hiring committees to avoid bias (explicit or implicit).

Although ultimately not as convincing as it could be (owing to the datasets used), this is a useful study in that the authors try to focus our collective attention on additional factors that affect the underrepresentation of women in math-intensive fields of science, and suggest that universities explore new options for career tracks ("The linear career path of the modal [sic?] male scientist of the past may not be the only route to success.."). I agree; just don't throw out the methods that seem to be working so far. That would make the situation orders of magnitude worse than it already is for women in math-intensive fields of science.

32 comments:

  1. In my field, an interesting dynamic has occurred where a few successful women have chosen (whether free or constrained) not to pursue tenure track positions. They have decided to work on soft money grants--often as research faculty but sometimes more like contractors. This is beginning to look like a trend now, I think, because many of the younger women in the field naturally look up to those successful woman as role models. It's an interesting situation, and many of us wonder what its consequences might be.

    ReplyDelete
  2. "(A quibble: In the sentence about how the number of PhDs awarded to women in the life sciences has increased, 13% is described as "orders of magnitude less" than 52%. It is not.)"

    It is if you're a computer scientist. ;^)

    ReplyDelete
  3. How many women avoid "dating down"? Just the fact that this expression exists shows how prevalent the phenomenon is. When the female chief surgeon, say, doesn't want to marry a male nurse, it's usually not because the nurse doesn't want to. As long as this mentality exists, men will have to, on average, earn more than women, otherwise they can never get married.

    Role models? Successful, rich women who marry a man less successful and less rich and don't hide the fact.

    ReplyDelete
  4. regarding Phillip Helbig comment...I think I'm missing your point and it's connection to this post? And I guess I'm one of those rare women who "married down" - I'm a PhD engineer, while my husband only has a MS. Thank you for considering me a role model, never thought of it that way.

    ReplyDelete
  5. I haven't rad the actual report, but as a bio sci person I'd like to point out that even though more than half of PhDs are earned by women, there is still a low percentage of bio sci TT women.

    Not entirely sure where they go--some think industry, since there are often more family-friendly companies than universities. Anony@5:47 observations may also partially explain it, as would my own observations that bio sci women PhDs are more likely than men to take positions such as study coordinators, etc and not feel like they have to land that TT job.

    ReplyDelete
  6. The Lesser Half2/09/2011 02:42:00 PM

    So . . . it is all about resources . . . which of course you get fewer of if you are discriminated against and have to take a less-prestigious job.

    So . . . what was their point again?

    ReplyDelete
  7. In my department two women have just been hired on in part time tenure track appointments. (I am one of them.) Both of us have small children and that is in part what makes the part time appointment appealing though it is still unclear exactly what it will take to earn tenure in a part time position. I think it is a growing trend, and a good thing regarding some of the issues highlighted in this study.

    ReplyDelete
  8. Phillip Helbig's comment: I have no problem with so called "dating down", but many men I meet sure seem to be uncomfortable with my education level! And being successful doesn't mean you're rich.

    ReplyDelete
  9. I have a PhD and my last boyfriend topped out at a college degree; he mentioned that he felt inadequate every couple of months, despite the fact that he was clearly much better educated than me on all things not covered in my PhD field and I never told (or, I think, implied/ suggested/etc) him that he was anything other than super smart. Doesn't make for a very comfortable relationship.

    ReplyDelete
  10. I didn't have time to read the original paper in detail, but I felt like it had this feeling of, "These empirical studies are all bogus, therefore, the reason there are no women in higher levels of academia is because they [dropped out to have babies, feel like math fields don't involve enough people, etc]"

    I'm all for overturning poorly designed studies (if they were indeed poorly designed), but proposing an alternative hypothesis seemingly randomly without testing it makes me feel uncomfortable.

    I also dislike that they ignored the social aspects to working in a math-laden field, for example, being surrounded by people with abrupt and sometimes abrasive interaction styles, alpha geek-ism, etc.

    ReplyDelete
  11. I also dislike that they ignored the social aspects to working in a math-laden field, for example, being surrounded by people with abrupt and sometimes abrasive interaction styles, alpha geek-ism, etc.

    FCS,

    What do you think of this essay on how academia is inhospitable to people who don't do well around people?

    ReplyDelete
  12. Doh! Forgot the link!

    http://www.insidehighered.com/advice/2011/02/07/margaret_price_on_the_search_process_for_those_with_mental_disabilities

    ReplyDelete
  13. Thanks for the insightful comments, FSP. I read the full article in PNAS, and I was continually confused by the authors' lack of clarity about "math-intensive fields." As you point out, the data they used to support their arguments were drawn from the life sciences or from all sciences, so I am not sure why they felt qualified to comment on "math-intensive fields."

    I, for one, hope their basic conclusion is correct - that discrimination against publishing, funding, and hiring woman scientists is (mostly) a thing of the past. But I agree with Female Computer Scientist that the authors didn't put much effort into supporting their alternative hypotheses of choice and resource constraint - especially with regard to "math-intensive" fields as opposed to academia in general. They seemed to treat it as an foregone conclusion.

    And their glaringly incorrect use of the phrase "several orders of magnitude" irked me immensely, all the more so considering that the topic of the paper is quantitative careers!

    It's a shame this paper wasn't more carefully reviewed, because it asks some really important questions. Hopefully it will catalyze more rigorous future efforts.

    ReplyDelete
  14. It seems to me that there's still quite active discrimination in hiring committees, at least in Chemistry, at least based on my experience (I served on 6 so far). Comments like"she's not serious about science" because she has a baby and is active in the women in science association.

    An other problem is the cumulative effect of certain choices, which may or may not be obligated but are certainly gender dependent. I recently canceled my participation to a week-long conference, where I had an invited talk, because I didn't feel comfortable leaving hubby alone with two small children, the youngest a few months old, and it was too far for a quick trip. Based on our experiences with the first kid we thought it would be OK, but this one is more challenging. Yes, I am aware that a man might have chosen otherwise.
    Anyways the organizer found (with some input from me) a replacement, a little more junior than I am. He must have done well, because the organizer has invited him to an other conference on the same topic he's putting together for the summer. This time I was not asked. Sigh.

    ReplyDelete
  15. Alex - I do agree that people with invisible disabilities, including mental health disorders, are greatly discriminated against in the workplace and in society in general. Academia is no exception to this. We as a community should work harder to be more accommodating and inclusive.

    That being said, I think if one has a mental health disorder that puts them in a state where they really dislike talking to people and would rather not interact with them, a service-oriented profession is probably not their best job choice. Teaching, advising, and increasingly researching are all activities that involve interacting with other people, usually on a daily basis. It is no longer sufficient that one can get paid for sitting in a dark corner being brilliant - I think those days are long gone.

    ReplyDelete
  16. I don't think it is quibbling to point out that a difference between a number of the order of 50% and one of the order of 10% don't differ by several orders of magnitude, as they asserted.

    This is because it is part of the scaffolding of their argument, which is roughly: "Sure, maybe there was overt discrimination before, and maybe that's why women's representation was several orders of magnitude less, but we aren't too committal on that. But nowadays, when the difference is minuscule, there's something else going on."

    If in fact the differences between yesteryear and today are not several orders of magnitude, it is several orders of magnitude less convincing that women face qualitatively different barriers to success than they did in the distant past.

    ReplyDelete
  17. @ another Anonymous commenter: If it wasn't tragic, the thought that being part of a women in science organization means one is not serious about science would be absolutely hilarious.

    ReplyDelete
  18. We are currently in the process of hiring for a tenure track chemistry position. I don't know if it's all in my head or what but I definitely have noticed that my colleagues are much harder on the women (including me when I was hired) then they are on the men during the proposed research presentation. It's really been pissing me off.

    So article or no, gender bias does play a role in hiring.

    I'm the only female in my department and the 4th female to ever be on the TT here.

    ReplyDelete
  19. Oh, I can't help myself- @Phillip Helbig, you're just wrong. Back when I was single, my number one method for getting a guy I wasn't interested in to leave me alone was to casually mention that I have a PhD. Worked like a charm.

    (I am now happily married to a man with "just" and MS, and I make more than he does... so clearly some men don't mind. But some really, really do.)

    On topic: I'm amused by the prevalent assumption that industry is more family friendly than academia. The (admittedly sparse) data I've seen indicates that we work similar numbers of hours. Maybe the fact that there isn't a time-sensitive gate like a tenure clock in industry makes it more amenable to people who want to slow down a bit when they have kids. But perhaps there is also an aspect of self-fulfilling prophecy in it. We hear when we're in grad school that industry is more family-friendly, so ambitious women who want kids might be more likely to choose industry. They get there, work hard, have kids, succeed, and as the examples of people successfully combining career and parenthood pile up, it does indeed get easier to have a family and work in industry, because it is seen as the normal thing to do.

    ReplyDelete
  20. I have really appreciated the flexibility of my academic job in terms of scheduling. I guess it's hard to separate out the effects of individual issues, including how helpful one's spouse/partner is, but I seem to have a much easier time than my friends in industry, even though we work about the same amount.

    ReplyDelete
  21. The reason why it's a quibble is because the point they were making is crystal clear based on the numbers they gave. they just don't appear to know what o.o.m. means.

    I thought the article was interesting and reasonably convincing, but really unnecessarily verbose. If they stopped repeating themselves in five-line sentences which say one simple idea over and over, maybe they wouldn't have had to stick all the actual data into the SI.

    ReplyDelete
  22. It's not that industry is more family friendly per se, but there are several reasons why it can be easier with a family in industry.

    First, most people in industry have permanent jobs. Thus, one can be more aggressive and assert one's rights. It takes a lot to fire someone, especially for the wrong reason; it takes little not to hire someone after a short-term soft-money contract expires.

    Second, most people in industry earn more money, so problems can be solved more easily (money for the babysitter).

    Third, the role of the individual is less in industry. This is why it isn't as fulfilling. On the other hand, one is usually part of a team. This has to be the case, since the fortune of the company can't depend on an individual who might become ill, die, leave the company etc. So, there is already a mechanism built in so that slack is taken up when someone takes off. An academic who takes time off creates a space which cannot be so easily filled.

    Fourth, many people are in industry because academia didn't work out, and even those who didn't come from academia realise that the interests of the workers and the management are often different, so taking time off is just one more aspect of that.

    ReplyDelete
  23. In my own experience, and that of most of the women I know who haven't been able to crack the TT ceiling, the major hurdle to get there is managing dual careers. This can appear to take the form of 'dropping out' and not applying for jobs - you can't apply places where you can't both find work. I was shocked to know in a recent faculty search that only 10% or so of the applicants were female - and this is in a life science department with lots of women faculty already. If your spouse is ahead of you (likely since most women marry men older than themselves, or have their career slowed down by kids), you are often not taken seriously as a candidate. As in, 'she's not serious because she won't leave the university where her husband has tenure, she's just fishing for an outside offer to finagle a position at her home institution'. Or because they think you put your husband's career ahead of your own. So at some point you stop trying, or don't try so hard. Most of the women I've come across who've succeeded are single, unattached and without kids.

    ReplyDelete
  24. @Philip Helbig: plus one to the number of female commentors whose ex had really, really big insecurity issues around the fact that he was less "successful" than her when it came to educational achievement (in her chosen, math-intensive natural science field). This contributed to major relationship problems (including abusive behaviors), with the the result that I will be very hesitant to ever again date a man who does not (a) have a PhD, and/or (b) earn more money than me and (c) feel very very secure about his level of success in his chosen career. It's not that I would look down on a man who is, for instance, a nurse. It's that I don't want to risk putting myself through that kind of treatment again.

    Lest you think it is my/our imagination or just a bunch of anecdotes, there is evidence from sociological studies showing that (heterosexual, married) women who earn more than their husbands, especially if they are employed and their husbands are not, are at increased risk for becoming victims of physical and emotional abuse by their husbands, being cheated on by their husbands, or facing divorce. For example, check out Macmillan and Gartner (1999), who found that a woman being employed while her partner was unemployed increased the risk that she would become a victim of abuse by almost a factor of three. Atkinson, Greenstein and Lang (2005) find that "Wives’ share of relative incomes is positively related to likelihood of abuse only for traditional husbands."

    ReplyDelete
  25. "Most of the women I've come across who've succeeded are single, unattached and without kids."

    I've seen a rather different result. Almost all the tenured women I know are married and have kids. I don't have any solid statistics though, so this may be observer bias.

    ReplyDelete
  26. same here. The tenured women in my department all have kids. The one female professor who did not get tenure in recent years was single and childless. I know there are studies about this, but it is a dangerous exaggeration to say that a woman has to be single and/or childless to succeed in academia.

    ReplyDelete
  27. Interesting point in the comments re: industry being more appealing due to lack of tenure or similar "gating" and self-selection based on the pre-existence of a critical mass.

    I am very tired of "alpha geek-ism".

    This PNAS article pissed me off since my experiences are at odds with the claims made by the authors. I'm glad to see that other people also found it somewhat less than convincing for multiple reasons.

    But, one could argue that my field is fairly math and chemistry intensive, so that's probably (?) why... Assuming we're going to believe their claims that "biology" is oh so much better now than it used to be.

    ReplyDelete
  28. The PNAS article says:

    "Forty years ago, women’s presence in most of these fields was several orders of magnitude less; e.g., in 1970 only 13% of PhDs in life sciences went to women (1). In the most math-intensive fields, however, women’s growth has been less pronounced (2–4)."

    If you check statistics for PhDs between 1970 and 2006, you see that “orders of magnitude” is wrong if it’s meant to characterize growth in “math-intensive fields” as opposed to other scientific fields.

    For example, the percentage of engineering PhDs that went to women was 0.5% in 1970 but 20.2% in 2006.

    In contrast, for psychology, it was 23.5% in 1970 and 71.3% in 2006.

    If “order of magnitude” is meant to be a multiplicative comparison (which seems to be the usual meaning), then it’s pretty hard to interpret this situation as saying there was more growth in psychology than engineering. If it’s meant as an additive comparison (e.g., number of deciles), then psychology shows more growth.

    I’ve discussed other problems with work on this subject here.

    ReplyDelete
  29. I was reading this and then suddenly I saw something that trolled everyone off track.

    Orders of magnitude and several.
    I had learned that several was 3 to 7 but then I googled for definitions online and I found "more than 2 but not many" "2 or more" and other such definitions. So I'm going to assume several is 2 or more.

    Orders of magnitude. Well it depends on your intepretation of order. Most definitions I can find at least relate it to log scales or exponents. On a computer your exponents are of base 2. If you do information theory your base is 2. If you do computer science your base is 2. If you do many kinds of mathematics your base is often e. Other fields like the fuzzy and warm comfort of 10 and 10 seems to be "common" according to our esteemed wikipedia.

    So here's my setup:
    Orders: [2, e, 10]
    Several_1: |log_order(a) - log_order(b)| >= 1
    Several_2: |log_order(a) - log_order(b)| >= 2
    Several_3: |log_order(a) - log_order(b)| >= 3

    And we're going to test 52% versus 13%, 20.2% versus 0.5% and 71.3% versus 23.5%.

    Orders: 2 2.71828 10
    52 13 2 1.38629 0.60206 Life Science PhDs awarded to Women?
    20.2 0.5 5.33628338786443 3.69883 1.60638 Women PhDs in Engineering
    71.3 23.5 1.60124131987889 1.10990 0.48202 Women PhDs in Psychology

    Excuse the nasty values, but In the more than 1 case of several for order/base 2 and e all 3 measurements of women PhDs differ by several orders of magnitude.

    In the case of 2 or more is several.
    The difference between PhDs in order/base 2 and difference in engineering PhDs in order 2 and e are still "several"

    For 3 or more is several, engineering PhDs in order 2 and e are still "several".

    So while those quibbles are noted and in only 1 case did base 10 order of magnitude matter, for this scientists the claims are accurate. In the case of 2 liberal interpretations of several and 2 different bases there was a difference of multiple orders of magnitudes.

    ReplyDelete
  30. "Most of the women I've come across who've succeeded are single, unattached and without kids."

    I've experienced exactly the opposite and think some of the problem is simply bias against non-conformists, which I certainly am. I constantly am faced with people, including academics and professionals, treating me as "unlucky" or like a second-class citizen because I have NO interest in getting married or in having kids. In fact, I'm really not interested in becoming rich either. I'm very unconventional, love my freedom, and welcome opportunities people with more responsibilities couldn't afford such as part time work and freelance work (science writing, adjunct positions). I love not having to worry about saving for someone's college and being free to go anywhere at moment's notice. And I'm not 21. I started as a writer and am now pursuing a Masters in astronomy, with the goal of doing planetary science research and/or astronomy writing and public outreach. Because I've gotten into this field "later" as in long past college, I also experience discrimination from people who assume I have to take a conventional path and then retire at a certain age when the reality is I don't want to do that. Longevity runs on both sides of my family, and I feel I can look forward to a long career of many years. Women who decide to go into science "later" in life should not face this sort of discrimination either.

    ReplyDelete
  31. I'm a bit late to the party, but I had the 'interesting' experience of hearing Profs Ceci and Williams speak at a departmental lunch a few weeks ago. I almost didn't go, given my negative feelings about some aspects of the study and also about the way the media had been covering it.
    Anyway, I was hoping that I would have my mind changed but alas it was not so. Any presentation that includes a bar graph of the average number of eggs a woman has at various ages is just bound to make me get all ragey.
    They dredged up lots of statistics about SAT scores, bell curves showing that men are more likely to score at the extreme ends of the bell curve on standardized tests and how being in the top 0.1% on the math SAT are so much more likely to become professors (of what, they didn't say.) They then said that didn't matter, and presented their data model without describing what their study pool was, and then distilled all that down into "women are not entering the tenure track just because they want to have babies." They then presented their solution of a more flexible tenure track. I don't mind the flexible tenure track part, and I know that family does influence some people. What I objected to is that they repeatedly discredited all opposition to their study as being anecdotal and not data driven, yet peppered their talk with lots of sexist anecdotes themselves. "Our daughter majored in engineering and is great in math, but by the time she and her female friends got to grad school they were gravitating towards the "softer" subfields such as sustainability" They selectively agreed with anecdotal stories from the audience when it fit their hypothesis, but called it anecdotal when they disagreed with it.
    Sigh.

    ReplyDelete
  32. Old Biddy, thanks for this. One of my friends and I have been wondering about Ceci & Williams's motivation and your comment sheds a little more light on it. I don't suppose they have their slides posted anywhere (I couldn't find any such slides on the web site of their institute, the Cornell Institute for Women in Science or via Google search).

    In some fields, women hold tenure track positions at more or less the rate at which they get PhDs. These fields include math and engineering but not psychology. Statistics are here.

    ReplyDelete