10 October 2012

Retraction classic: Physics and feminism

When I want to talk to students about how the scientific publishing process can go awry, I almost always end up telling the story of how the Canadian Journal of Physics published a screed against feminism.

I was in graduate school in the early 1990s when the story broke. The Canadian Journal of Physics had published a set of papers on chaos. Chaos theory was big at the time, following James Gleick’s best-seller Chaos (1987). Among the regular sort of papers one would expect to find in a journal of physics was one very unusual short paper.

First, it was labelled “Sociology.”

Second, it contained exactly two citations. And one reference was to a dictionary for a definition.

Third, the paper was outlandish. It was an almost unimaginable broadside attack on feminism, blaming it for most all of society’s ills. Almost as an aside, it incidentally insulted almost all social science at the same time. To give you a little sampling, this paper claimed:

  • Women should not be in the work force, because they are “nurturers.”
  • Half the children of working mothers suffered “serious psychological damage.”
  • Surveys and controlled experiments in social science distort findings, and “wisdom” is a superior method of obtaining knowledge in the social sciences.
  • Abstinence until marriage would improve things, which should be vigorously promoted through television advertising campaigns.

It’s rare to find such blunt “women belong in the home” arguments in any print medium, let alone a scientific journal. It was astonishing.

The article was penned by Gordon Freeman (pictured), who was the guest editor of this one issue of the journal. It was pretty obvious what had happened, in broad strokes: he abused his editorial power to get his poisonous opinion piece into the pages of the journal.

The details of exactly how this happened were a little more complicated. Freeman organized a conference on chaos theory, and was assembling papers that had been presented at a conference for publication in the Canadian Journal of Physics. Apparently, the deal was that the journal would publish all the papers Freeman compiled, provided that they were presented at the conference, and that they were peer-reviewed.

Freeman lied about presenting the paper at the conference.

Somehow, Freeman managed to get back a positive review. The mind boggles, but he did.

The paper was published, and the excrement collided with the rotary cooling device.

Once this story blew up, as was inevitable, the regular editor of Canadian Journal of Physics, Ralph Nicholls, refused to reveal who the reviewers of Freeman’s article were, following the standard practice of anonymous peer-review that most journals follow. Nicholls was removed as editor for refusing to show the review to the editor in chief for all the National Research Council journals.

It also came out that Freeman’s institution, the University of Alberta, considered this to be human subjects research, and that Freeman had not obtained approval for it.

The journal apologized for the paper nine months after it was published (Dancik 1991a), but did it in a very uninspired way. The did say Freeman’s work “had no place in a scientific journal,” but it didn’t specifically disavow Freeman’s opinions. And there was no explanation of how Freeman’s paper managed to get into the journal in the first place. Although Science magazine later described the notice as a “retraction,” the word didn’t appear in the three sentence editorial.

The Canadian Association of Physicists certainly considered it a retraction, and wrote to Freeman asking him to shut up about his paper. (Freeman, by all accounts, loved the attention this article was giving him.)

The Editor-in-Chief of the NRC Research Journals has published an editorial note in the Canadian Journal of Physics stating that your article has no place in a scientific journal and expressing regret that it was published. This clearly retracts any approval of the article by the CJP and, in effect, retroactively disavows its publication. ... I would ask you, therefore, in view of the harm engendered to the scientific reputation of the Canadian Journal of Physics and, by extension, to the whole physics community, to refrain from making any further references to this article in any public forum, and not to distribute reprints with the name of the journal attached. I believe that scientific ethics demand that this article effectively be struck from the public record(.)

Moreover, the retraction / apology was printed on an unnumbered page, which made it hard to cite, and to link to the original paper. This led the journal to print the apology a second time a few moths later (pictured; Dancik 1991b).

The outrage over this paper was so great that many people were advocating reprinting the entire issue without Freeman’s article, and sending it to every library, so there would be no trace of it left. But a petition to reprint the issue didn’t get any response until the story appeared in Science magazine in 1992 (Crease, 1992).

The Nation Research Council, the publisher of the journal, at one point considered publishing a special issue about the controversy around Freeman’s paper as a mea culpa. But it was not appearing. At one point, the plan not to publish this came out held a conference on the ethics of publishing (reported in Crease 1993; Huston 1993). The uninspired response for why there was no supplement? “It seems a little late now.” This seems to support a claim by newspaper columnist Morris Wolffe on how the controversy was handled:

The (National Research Council), it was clear, was hoping the Freeman matter would go away.

Ultimately, some additional commentary did appear in the March/April 1993 issue of the Canadian Journal of Physics. Proceedings of the ethics symposium appeared in the journal Scholarly Publishing.

If you go to the Canadian Journal of Physics website today to look up the article, this is what you find:

It’s still available, with no indication that this wretched paper has been retracted, or warranted an apology, not once, but twice.

(Update, 11 October 2012: This part of the story, at least, gets a happier ending: there is now a note and link out to follow-ups that were published in the journal. Read more about how this happened here.)

This story was one of the first cases where I became aware of the concept of retraction. It definitely shaped by view of what retraction was meant to do. The notion of reprinting the issue set it in my mind that the goal of retraction was to, as near as possible, expunge an article from the scientific record. I’ve since realized that in practice, retraction is a much more complicated beast.

But if ever there was a paper that deserved an unambiguous retraction, this would be it.

Update, 8 October 2020: Retraction Watch reports that this paper is officially being retracted, 30 years on.


Freeman GR. 1990. Kinetics of nonhomogeneous processes in human society: Unethical behaviour and societal chaos. Canadian Journal of Physics 68(9): 798. DOI: 10.1139/p90-116

Crease R. 1992. Canadian chemist takes on working women. Science 255(5048): 1066. DOI: 10.1126/science.1546307. JSTOR: http://www.jstor.org/stable/2876664

Crease R. 1993. Fallout from paper on working mothers. Science 259(5101): 1531. DOI: 10.1126/science.8456280. JSTOR: http://www.jstor.org/stable/2880644

Dancik BP. 1991a. Editorial / Éditorial. Canadian Journal of Physics 69(6): iii. DOI: 10.1139/Editors_comments_on_KNP_issue

Dancik BP. 1991b. Editors Comments on KNP Issue. Canadian Journal of Physics 69(12): 1403. DOI: 10.1139/Editors_cjp6912

Spurgeon D. 1992. Physics and family values. Nature 360(6404): 504. DOI: 10.1038/360504a0

Huston P. 1993. Article in Canadian physics journal put academic integrity on trial. Canadian Medical Association Journal 148(10):1771-1773. http://www.cmaj.ca/content/148/10/1771.citation. Full text on PubMed: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1485550/

External links

The sexist science of Gordon Freeman

No comments: