Friday, October 31, 2014

The flawed and offensive logic of "Academic Science Isn’t Sexist" in the @nytimes

OK.  It is Halloween night and I am tired and need to get my kids to sleep.  But someone on Twitter just pointed me to an opinion piece just out in the New York Times: Academic Science Isn’t Sexist - NYTimes.com and after reading it I felt I had to write a quick post.

The opinion piece is by Wendy M. Williams and Stephen J. Ceci and discusses work by them (and coauthors).  In particular they discuss findings in a massive report "Women in Academic Science: A Changing Landscape" by Stephen J. Ceci, Donna K. Ginther, Shulamit Kahn, and Wendy M. Williams in Psychological Science in the Public Interest.  I note - kudos to the authors for making this available freely and under what may be an open license and also apparently for making much of their data available behind their analyses.

The opinion piece and the associated article have a ton of things to discuss and ponder and analyze for anyone interested in the general issue of women in academic science.  I am not in any position at this time to comment on any of the specific claims made by the authors on this topic.  But certainly I have a ton of reading to do and am looking forward to it.

However, I do want to write about one thing - really just one single thing -  that really bothers me about their New York Times article.  I do not know if this was intentional on their part, but regardless I think there is a major flaw in their piece.

First, to set the stage -- their article starts off with the following sentences:
Academic science has a gender problem: specifically, the almost daily reports about hostile workplaces, low pay, delayed promotion and even physical aggression against women.  Particularly in math-intensive fields like the physical sciences, computer science and engineering, women make up only 25 to 30 percent of junior faculty, and 7 to 15 percent of senior faculty, leading many to claim that the inhospitable work environment is to blame.
This then sets the stage for the authors to discuss their analyses which leads them to conclude that in recent times, there are not biases against women in hiring, publishing, tenure, and other areas.  Again, I am not in any position to examine or dispute their claims about these analyses - to either support them or refute them.

But the piece makes what to me appears to be a dangerous and unsupported connection.  They lump together what one could call "career progression" topics (such as pay, promotion, publishing, citation, etc) with workplace topics (hostility and physical aggression against women).  And yet, they only present or discuss data on the career progression issues.  Yet once they claim to find that career progression for women in math heavy fields seems to be going well recently, they imply that the other workplace issues must not be a problem.  This is seen in statements like "While no career is without setbacks and challenges" and "As we found, when the evidence of mistreatment goes beyond the anecdotal" and "leading many to claim that the inhospitable work environment is to blame."

Whether one agrees with any or all of their analyses (which again, I am not addressing here) I see no justification for their inclusion of any mention of hostile workplaces and physical agression against women.  So - does this mean that a woman who does well in her career cannot experience physical aggression of any kind?  Also - I note - I am unclear I guess in some of their terminology usage - is their use of the term "physical aggression" here meant to discount reports of sexual violence?   This reminds me of the "Why I stayed" stories of domestic violence.  Just because a women's career is doing OK does not mean that she did not experience workplace hostility or physical or sexual violence.  I hope - I truly hope - that the authors did not intend to imply this.  But whether they did or not, their logic appears to be both flawed and offensive.

UPDATE 1. November 1, 8:30 AM

Building a Storify about this.

UPDATE 2: Nov 3, 2014. Some other posts also criticizing the NY Times piece
UPDATE 3: Nov. 4, 2014.  More posts about the NY Times piece

14 comments:

  1. Thanks, Jonathan. More reading.

    ReplyDelete
  2. Dear Jonathan, I speed-read Ceci et al. 2014 Psychological Science in the Public Interest. 15(3) 75 –141. I urge others to read it.
    The op ed in the New York Times doesn't do the review justice.
    In fact, it glosses over the nuanced analysis that the authors have done.
    all the best, dawn

    ReplyDelete
  3. Ugh, sorry. I left out the review's title: Women in Academic Science: A Changing
    Landscape.
    It appears to be open access and can be found at: http://bit.ly/1tmhi8V

    The one thing I would say about the article is that it ignores the sociological approach of Thick Analysis, which takes a blended approach to qualitative information (some might call this anectodal) and quantitative data.
    http://bit.ly/1zoczbC

    cheers, dawn bazely

    ReplyDelete
  4. Are there any (minimum) estimates for sexual harassment of academic women? What used to be my group at UCL has been run for the last 10 years by a (very bright) woman, She says she hasn't suffered any physical aggression. I have certainly heard of it happening, but how common is it? That's not easy to answer I guess.

    ReplyDelete
    Replies
    1. You could start with the Clancy et al. paper - Survey of Academic Field Experiences (SAFE): Trainees Report Harassment and Assault

      Delete
    2. This comment has been removed by the author.

      Delete
    3. You could start there, but it won't get you anywhere. That study, which is about harassment and assault experiences at Anthropology (mainly) field site, is so uncontrolled as to be uninterpretable. The sample is self-selected and thus likely to be overly rich in those who had bad experiences. The time of the incidents was not even collected thus a reported assault could have happened 30 years ago and it still is figured into the percent experiencing assault. There is no evidence to establish that the reported data are current or nearly current. Nothing prevented people from repeatedly taking the survey. There is no verification that any of the respondents actually went to a field site or actually experienced assault. In sum, the study is horribly done and uninterpretable. I think it is no accident that it was published in PLoS 1, which is getting quite a reputation for publishing dreck.

      Your point about it being possible to have a good career and still be harassed/assaulted is a good and important one. The problem as exemplified by the paper you linked, is that there are no data to evaluate your hypothesis. This was not included in the Ceci et al., 2014 paper not because they wanted to leave it out, but because there were no data. I hope that you will read the Ceci, et al. paper as I think you will find it much more complex and fair-minded than its critics (who in most cases appear to have never read the paper) suggest.

      Lastly, people seem to confuse science not being a sexist institution with an absence of sexism within science. Ceci and Williams argue that science is no longer inherently a sexist institution, but they don't argue that sexism doesn't exist. Sexism does exist, it just no longer endemic to the whole enterprise as it was in bygone days.

      Delete
    4. It was Ceci and Williams who included the statements about hostile workplaces and physical aggression against women in their editorial and clearly tried to imply that the data showed that these were not issues. They did not say "there is no data on this". They linked those topics to their claims about careers and thus tried to make people think there was data proving these were not issues. That was a flawed part of their editorial and that was the point of my post. I think that is pretty clear from my post.

      Delete
    5. And Kim, just for the record, I believe the Clancy et al paper is excellent. They describe what they did, they have extensive discussions of possible limitations and they show that their results are consistent with what has been seen in many / most other work environments.

      As for PLOS One, well, I fundamentally disagree with your assessment. Yes, PLOS One has some papers that are dubious. So do all other journals. This comment of yours shows that, contrary to what you are going on and on about all over the web, you do not think that people should read papers but rather that they should be judged by where they are published. Yet at the same time, it seems that, when you disagree with someone's comments you resort with brilliant comments like "Have you read the paper". I would suggest you just stick to the actual comments that people make and not try to guess what they did or did not do leading up to making those comments.

      Delete
    6. I sincerely do not think that people should judge a paper by where it is published. They should read the paper and decide. I would encourage everyone to read the paper you mentioned. I do wonder though, how you can think a paper so poorly controlled is excellent? I listed several threats to validity and even though they discussed some of them they presented no evidence that shows that their data are not seriously skewed. Self-selection of the sample by itself invalidates this study as objective science. If one wants to report that 26% of women at field sites are assaulted one damn well better be able to show what the sample is. They can't. They do not know when it occurred.where it occurred or if it occurred. This is not good science. This isn't even poor science. This isn't science. They can't bolster their argument by saying that their data are consistent with previous studies because THERE AREN'T ANY!

      Of course Ceci and Williams didn't say "there is no data on this" because they know that data is a plural noun and they didn't even comment on harassment in their op-ed. Why not? Because there are no data. Your point that it is possible to have a good career and still experience harassment or assault is well taken and important. Unfortunately, there are no data to evaluate your supposition and we need such data. The Clancy, et al. paper are certainly not those data.

      You are entitled to your opinion about PLoS 1, as am I. To me I see far too much crap in PLoS 1. It is a dumping ground.

      Delete
    7. I do not even know where to begin with responding to your comments but here is a try.

      #1 - I get you do not like the Clancy paper. But to say "This isn't science." just proves you are not being even remotely objective about your assessment. It is unquestionably science. Again, one can debate their methods and approaches and conclusions. But there is no reason to debate details with someone making idiotic statments like you are.

      #2 - thanks for them grammarly policing. That is so relevant-like and helpfully.

      #3 - your statement that "they didn't even comment on harassment in their op-ed." is simply incorrect and shows that either you did not read their article or you simpy wish to wish away the things they wrote. They open their piece with "ACADEMIC science has a gender problem: specifically, the almost daily reports about hostile workplaces, low pay, delayed promotion and even physical aggression against women. " That sentence is pretty clearly discussing forms of harassment.

      Delete
    8. wow, just wow.
      I wish I had old fashioned Ice Skating Judges scorecards to hold up.
      Anyway, Jonathan, I give you perfect for courtesy and patience, and I give the other person (whom I just looked up on web of science) perfect for "I sound like a cranky old white male academic".

      Delete
  5. Oh, hey, I'm doing a paper for my class about the ethos of this article. I think you may have just helped me out a LOT with those links.

    ReplyDelete

Most recent post

My Ode to Yolo Bypass

Gave my 1st ever talk about Yolo Bypass and my 1st ever talk about Nature Photography. Here it is ...