The essay below was written in 2003. On Monday, 24 January, 2005, the New York Times in a story about remarks on gender by Harvard president, Larry Summers, wrote the following (a correction appeared Saturday, 29 Jan.):
"C. Megan Urry, a professor of physics and astronomy at Yale who led the American delegation to an international conference on women in physics in 2002, said there was clear evidence that societal and cultural factors still hindered women in science.
Dr. Urry cited a 1983 study in which 360 people--half men, half women--rated mathematics papers on a five-point scale. On average, the men rated them a full point higher when the author was "John T. McKay" than when the author was "Joan T. McKay". There was a similar, but smaller disparity in the scores the women gave."
Here's the truth about this study.
In 1983, in an article in the journal, Sex Roles, vol.9, Paludi and Bauer conducted a study in which 360 college students (half male, half female), were asked to evaluate an article (abridged to 1500 words) in the field of politics, or psychology of women, or education, that was written (supposedly) by Joan T. McKay, John T. McKay, or J. T. McKay. The study found that the evaluations depended quite significantly on whether the name on the student's paper was female or male or inconclusive (J. T.), with, not surprisingly, higher ratings going to the male author, whether or not the student was male or female.
The first time I know of in which this study was mentioned in math circles is in an article in the Notices of the American Math Society, vol 38, Sept. 1991, 707-714, by Lynne Billard, titled "The Past, Present, and Future of Academic Women in the Mathematical Sciences". Here is the relevant paragraph under a section titled "5. Publications and Performance".
"Although women have performed at least as well in graduate school, they publish less than do men. A study on citations showed that in mathematics 2.9% of articles classifiable by gender were written by women although 13% of the membership of professional associations were women and 7.6% of those doctorates in the mathematical sciences employed in educational institutions in 1979 were women (from Vetter (1981)). A study by Goldberg (1968) was repeated by Paludi and Bauer (1983) in which it was shown that publications perceived to have been written by women were considerably less favorably reviewed than those thought to be written by men (see Table 6). In this study, 180 men and 180 women were asked to rate comparable articles, one-third of which were ``authored'' by John T. McKay, a Joan T. McKay, and a J. T. McKay. The most favorable rating was one, with five being the least favorable. Upon questioning later, the majority of the raters believed the articles by J. T. McKay were in fact authored by a woman and this belief is reflected in the mean scores shown in the table. It is interesting to observe that women also rated male-authored papers more highly than they did for Joan T. and J. T. papers, but not quite as favorably as did men rate men authors.
Article Article Authored By reviewed by John T. McKay Joan T. McKay J. T. McKay Men 1.9 3.0 2.7 Women 2.3 3.0 2.6Source: Paludi and Bauer (1983)"
Note: 1 is high, 5 is low.
Also note that Billard does not identify either the evaluators or the subject of the articles by the McKays. Since the article is about the mathematical sciences, appears in the Notices of the AMS, and the paragraph quoted begins with data on citations in math, it is easy to assume that the Paludi and Bauer article is about a study involving mathematicians.
In fact, this is the most likely explanation for an article in Science, 255, 13 March, 1992, page 1382, by Paul Selvin (then a grad student in biology at Berkeley). The article is about Jenny Harrison's suit against the Berkeley math department, and in a discussion about bias against women in the sciences, Selvin presents a bar graph with a caption stating: "In a 1983 study both male and female mathematicians rated math papers they thought were written by a man ("John T. McKay") more highly than those they thought were written by a woman ("Joan T. McKay")."
Science printed a correction (256, 19 June 1992, page 1615) which I partially quote: ". . a caption . .. was incorrect. The data did not show how male and female mathematicians evaluated mathematical articles. In fact the articles in the study that was discussed [M. A. Palardi and W. D. Bauer, Sex Roles 9, 787 (1983)] were about politics, the psychology of women or education; the subjects were not mathematicians but male and female college students." Note that Science misspelled Paludi's name, so this is not found in web searches.
Later the Economist (June 1996, Science and Technology section) ran the same mistake, but did not print a correction when I informed them of their error.
What brought this to my attention yet again was a recent article by Meg Urry, physics professor at Yale, which appears several places (including bulletin boards) but began in the fall 2002 issue of The Gazette, the newsletter of the APS Committee on the Status of Women in Physics.
Here is Urry's relevant paragraph:
"3. Sociology holds some of the answers, if physicists would only listen. Studies have shown that referees judge gender of author, not quality of work. In 1983, Paludi and Bauer published a revealing study about the influence of gender on perception of excellence. Three-hundred-sixty referees, half men and half women, were each sent a mathematics paper to rate, with the author's name given variously as John T. McKay, Joan T. McKay, or J. T. McKay. The reviewers found that the man's paper was considerably better than the woman's. The neutral, initials-only designation was also rated rather lower than the man's paper, apparently because many referees believed the initials to represent a woman. Both men and women found the paper written by the woman to be markedly less good than the man's paper. So it isn't just men undervaluing women's work, it is all of us."
If one "Googles" Paludi and Bauer, one find a few dozen sites, a number of which refer to Urry's article. Oddly enough, because Science magazine misspelled Paludi (as Palardi), one does not find their correction to Selvin's article. (On my web site, I make the same mistake http://math.berkeley.edu/~kirby/sexism.html .) Thus unless one reads carefully and finds Allyn Jackson's correct statement in a Notices AMS article, there is apparently no correct statement to be found.
It is interesting that so many articles mention Paludi and Bauer but do not mention any other such studies. Are there none? It is not surprising that authors who believe there is discrimination against women make such a point of citing the one good piece of "evidence" of bias in rating papers. The fact that the journal, Sex Roles, is somewhat hard to find may explain why no one searched further to see what else the paper had to say.
Still, any mathematician would have been suspicious. If I were asked to rate an article, it would presumably be in my subfield, topology, of mathematics, and then I would immediately know that there are no John or Joan McKays. Moreover, for most papers in mathematics, there do not exist 360 people who are competent to evaluate the paper. In fact, I can't quite imagine how such a study could be conducted in mathematics. There were suggestions decades ago of instituting blind refereeing in math, with the purpose of avoiding discrimination, but no one could see how to make it work, because if we are competent to referee a paper, we are likely to have seen the preprint already, or to have heard the author lecture, or at least to have heard of the result.