Monday, March 22, 2010

The Few, The Proud, The Underestimated

A report, just released by the American Association of University Women, “Why So Few?”, found that although there have been gains, stereotypes and cultural biases continue to plague women in the sciences.

A few highlights from this report:

1. A study of postdoctoral applicants found that women had to publish 3 more papers in prestigious journals or 20 more in lesser journals to be judged equally productive as male applicants. [Edit added 3/26/2010: note that these extra papers represent what a female applicant would need to achieve the same "competence" score as a male--and is not what was typical of the productivity of applicants (for fellowships in the Swedish Medical Association). In other words, this is what it would take for female applicants to overcome the prejudice in this competition.  See Comments for more information.]

2. Another study found systematic differences in letters of recommendation for academic faculty positions for female and male applicants. More often mentioned attributes for males: achievements, research, abilities; for females: compassion, teaching, effort. These were judged to be unconscious stereotyping as opposed to conscious bias.

3. Being discouraged from entering a science field was cited by female and minority chemists and chemical engineers as the leading contributor to underrepresentation in those fields. Many in this same survey specifically said that they had been actively deterred in college, most often by a professor.

4. The “double-bind”: being competent and well-liked. Studies indicate that likability and competence both matter for workplace success. One study found that when information about performance was not provided, the woman was rated much less competent than the man, whereas when prior success was made explicit, both men and women were rated equally competent. The opposite pattern held for likability. When success was unclear, men and women were judged similarly likable, but when success was clear, the man was judged more likable. This outcome was only true for “male-type” jobs (successful women in “female-type” jobs did not experience the same negativity). See the figure at left.

5. An experiment that tested “contrast sensitivity ability”, a skill made up by the investigators, showed that when men and women in a group were told there was no difference between the sexes in ability, rated their own ability equally. Not surprisingly, when the group was told that men were better, men rated themselves much higher than did women. They were tested by being asked to detect the proportion of white and black on a screen, which actually were equal or nearly so—so there was no real correct answer. However, they apparently did not test the outcome for the situation in which the group was told that women were more skilled. It would have been interesting to see if men still rated themselves higher. In any case, the upshot is that individual aspirations for a career are influenced by one’s perceived ability or potential ability in that field.

6. Some research indicates that even when individuals consciously reject gender stereotypes, they can still be biased at an unconscious level.

7. An interesting part of the report dealt with differences in male and female spatial skills. When administered the Purdue Spatial Visualization Test: Rotations (see example below), women were more than three times as likely to fail as were male peers. However, when failures (male and female) were given a 10-week spatial visualization course, their average scores increased from 52% to 82%. This might be a case in which male-female differences are due to previous exposure to activities requiring spatial skills (playing with erector sets, etc.).

Some positive observations:

1. Small things can make a big difference: a course to acquire a special skill (such as the spatial skills course mentioned above); assurance that ability (in a particular field) is not fixed, but grows with experience.

2. Teaching girls about how stereotypes affect performance can diminish such effects.

3. Mentors are very influential.

4. Where there are clear criteria for judging individual performances, women are more likely to be judged competent and have a higher probability for success.

5. A highly successful, competent woman can improve her likeability by increasing her “communal” nature, e.g., being seen as a nurturing “mother” type. I don’t think this observation should alter one’s behavior, but it does provide some insight into sources of negativity (seeing women as competitors rather than nurturers).

In any case, the report is fascinating reading.

5 comments:

  1. I have seen some press about the report but have not had a chance to read it. Thanks for the great summary. It certainly seems to address some points that are often implicated, but for which there was not supporting evidence.

    ReplyDelete
  2. Thank you for writing about AAUW's new research report! Please join us online on Thursday, March 25, 2010 for a live Webcast about the report. For more information and to register for the Webcast, please visit: http://whysofewwebcast.eventbrite.com/

    ReplyDelete
  3. 20! 3 more high-impact papers feels like a lot but that may depend on what is high impact (e.g. does it need to be Science or is Am Nat sufficient?) but 20 more low impact papers?! How many postdocs have 20+ papers? Not many in my field in any case. Wow, that's depressing.

    ReplyDelete
  4. Anon. Thanks for pointing this out.

    I went to the original paper (Wenneras & Wold, 1997, Nature) that was quoted in the AAUW report. The report did not explain completely what was meant by 3 high impact and 20 low impact papers.

    The original study examined a number of metrics by which post-docs (applying for fellowships from the Swedish Medical Association) were assessed. Each applicant was assessed by five reviewers who assigned a score of 1 to 4 for each category (publications, citation rate, proposal, and "competence"). On average, female applicants received an additional 0.0033 competence points per impact point they accumulated, whereas male applicants received an extra 0.21 points for competence. The authors found two factors (in addition to scientific productivity) that determined competence scores: gender of the applicant and affiliation of the applicant with one of the evaluators. Thus, being female and lacking personal contacts was a "double handicap" in this competition.

    What this meant was that for a female post-doc to be awarded the same competence score as males, she needed to exceed his scientific productivity by 64 points. They translated this difference as equaling 3 extra papers in Nature or Science (impact factors 25 and 22), or 20 extra papers in a journal with impact factor of about 3. This is clearly unlikely to be the case with most applicants, female and male.

    So those figures of 3 and 20 papers are not related to the number of papers actually published by applicants, but just represents the equivalent amount of extra productivity females would need to match up with the scores awarded to males. The paper by Wenneras and Wold did not provide data on the actual number of papers published by post-docs.

    The basic conclusion of this study is the same: that female applicants had to be incredibly outstanding and greatly exceed their male peers in terms of productivity to achieve the same score.

    ReplyDelete
  5. Oh, and one more thing. The paper reported that a female (-.21 competence points) lacking personal connections in the committee (-.22 points) had to present an additional 131 points to MRC reviewers to receive the same score as males.

    The above level of productivity was attained by only 3 of the 114 applicants (one male and two females).

    ReplyDelete