précis by Rosslyn Ives
This article is a summary of a study conducted by Christine Wennerås and Agnes Wold from the Department of Medical Microbiology and Immunology & Department of Clinical Immunology, Göteborg University, Sweden. It was published in Nature (Vol. 387, p. 341) and given prominent exposure in the editorial, "Naming the game", which appeared in New Scientist May 24. The findings give hard evidence of the "old boy networks" underlying who rises to the top echelons of science.
In 1995, 52 women and 62 men applied for postdoctoral fellowships from one of the major biomedical research agencies, the Swedish Medical Research Council (MRC). Out of the 114 applicants, 4 women and 16 men obtained fellowships, i.e. women were only half as successful as fellow male applicants.
The evaluations leading to this outcome were the task of different specialised committees each consisting of five reviewers. The procedure involves each reviewer scoring the applicants on a scale of 0–4 for three qualities, "scientific competence", "quality of proposed methodology" and "relevance of research project". Final scores were obtained by multiplying the three variables giving a range of 0–64 for each applicant.
The evaluations of the MRC peer review system are not normally made available, supposedly to ensure objectivity. This secrecy was challenged through the courts, using Sweden’s Freedom of the Press Act, with the result that the 1995 scores were released, making the study by Christine Wennerås and Agnes Wold possible. See results in Table 1.
The consistently lower score for females was of significant interest, leading the researchers to investigate the factor which appeared to be most lacking in women, i.e. "scientific competence". They did this by assuming that the quality of a researcher would be reflected in her/his scientific productivity, which itself could be gauged from a more objective measure of productivity based on original scientific publications. Using information from fellowship applications, such as number of papers, number of first authored papers, prestige of journal in which papers appeared, and citation rates, the researchers derived a "total impact" score. This showed a good correlation with "competence" as perceived by reviewers, and once again revealed women as consistently lower.
To exclude the possibility that the lower scores for women were due to confounding factors, multiple regression analyses were performed. In summary this model indicated that female applicants were being given 0.21 points lower on "competence" than male applicants having the same productivity level. Based on the use of publications as an evaluation of productivity it was concluded that a woman would have to be 250% more productive to receive the same "competence" score as a male.
Another finding of this study showed that affiliation between reviewers and applicants resulted in 0.22 points higher for "competence". Although the MRC has a rule that biased reviewers are not allowed to score an applicant with whom they have personal connections, remaining committee reviewers still tended to give higher competence scores than were warranted by scientific productivity. This "affiliation bonus" seemed to accrue irrespective of the applicant's gender. However, females who lacked personal connections were doubly disadvantaged, requiring such levels of compensating productivity that were attained by only 3 (one male and two females) of the 114 applicants for 1995.
Overall the two key variables that enhanced "competence" scores were being male and having personal connections. Variables like basic education, field of research, university and national affiliation were largely irrelevant.
As the MRC reviewers consisted of 50 men and 5 women it was not possible to make a useful comparison between the two genders. However, despite lack of data other studies suggest that while more female reviewers is desirable it may not necessarily of itself contribute to the abolition of discrimination against female scientists.
In conclusion the study showed that the peer review system used by the MRC was incapable of evaluating women and men applicants equally. It was also the same system used to review grant applications, where once again women scientists were turned down more often than men.
This small yet significant study supports the work of Jonathan R. Cole and Burton Singer who proposed "a theory of limited differences" to explain the gender based discrimination in the sciences. Their theory refers to the small disparities at every stage of a woman scientist’s career which combine to create a subtle yet virtually unassailable barrier to success.1
Rosslyn Ives is a Melbourne based, independent science consultant, researching science as a social practice. She can be contacted by email at rosslyn@netspace.net.au, telephone (03) 9857 9717 or fax (03) 9857 9466.
1. "A Theoretical Explanation" in The Outer Circle: Women in the Scientific Community, edited by Harriet Zuckerman, Jonathan R. Cole & John T. Bruer, 1991. W.W. Norton & Company, New York, London.
Table 1. Averaged scores for all applicants
| Females | Males | Difference | |
| Scientific competence | 2.21 | 2.46 | 0.25 |
| Quality of proposed methodology | 2.37 | 2.54 | 0.17 |
| Relevance of research project | 2.49 | 2.62 | 0.13 |
| Total score | 13.8 | 17.0 | 3.2 |