Legendary Gender Equality Study Falls – Men, Not Women, Are Discriminated Against

For nearly three decades, a Swedish study has been used as proof that women must be significantly more qualified than men to succeed in academia. Now, a thorough scientific re-examination shows that the conclusion rests on a gross mistake. When the material is correctly analyzed, not only does the alleged gender effect disappear – in fact, the result indicates that it is men who are discriminated against through an underestimation of their competence.

In the spring of 1997, an article was published in Nature that quickly achieved iconic status in the debate on gender equality. Researchers Christine Wennerås and Agnes Wold analyzed how the Medical Research Council evaluated applications for postdoctoral positions. Their conclusion was that women needed to be 2.4 times more qualified than men to be perceived as equally competent.

The figure 2.4 quickly spread far beyond the research community. It appeared in international reports, political documents, and Swedish gender equality programs – and became a kind of feminist symbol for alleged systematic discrimination against women in academia.

A Study That Was Never Retested

Despite its enormous impact, the study’s analysis was never properly checked. The original data was inaccessible for a long time, and the study’s status meant that few questioned its results. The story fit the times and aligned neatly with a broader societal discussion shaped by leftist feminist perspectives.

Only now have two researchers done what should be standard in science: returned to the source material and tested if the conclusions hold up.

Researchers Redo the Analysis from Scratch

In a new study published in the journal Quantitative Science Studies, Ulf Sandström and Ulla Riis recreated the entire analysis step by step. Using material from the National Archives and the Medical Research Council’s archives, they reviewed the evaluation committees’ statements and checked all data points.

“We decided to redo the analysis from scratch,” write the researchers. The result was something very different from what had been established as the truth.

Comparing Apples and Oranges

The core of the criticism concerns how competence was measured. In the original study, researchers’ competence was linked to how successfully they published in scientific journals, based on the journals’ expected citations. But different research fields have widely varying publication traditions. “Here, apples and oranges are compared,” note Sandström and Riis.

A preclinical biomedical researcher often publishes many international articles with several co-authors. Clinical researchers publish fewer articles, and in the 1990s, behavioral scientists often published in Swedish journals or report series. These differences reflect academic cultures – not gender.

Gender Bias Disappeared with Correct Analysis

The evaluation committees assessing the applications were specialized and well aware of their respective fields’ norms. When Sandström and Riis took account of differences between disciplines, publication patterns, and the working methods of the eleven separate evaluation committees, the picture changed completely. “What was interpreted as gender bias turned out to be an analytical artifact.”

When the analysis was adjusted for these factors, the gender effect disappeared. Instead, the results pointed in the opposite direction: the competence of highly productive men may have been underestimated in the assessments.

Oversight in the Original Study

An important detail Wennerås and Wold did not consider was that the distribution of grants between the evaluation committees had already been decided. Each group in practice had about two grants to award, guaranteeing distribution between research areas – regardless of the gender distribution among applicants.

Overlooking this structure further contributed to the misleading conclusion about discrimination.

Necessary Correction

Sandström and Riis emphasize that their results should not be interpreted as proof that academia is entirely free from injustices. “Our results do not say the system is equal, but that the evidence in this particular case was weaker than previously believed.”

However, the re-examination shows how important it is that even the most cited and politically useful studies are subjected to critical scrutiny – especially when they confirm what one perhaps wanted to prove in advance. Otherwise, methodological errors risk living on as established truths and undermining trust in research.

Self-Criticism Is Science’s Strength

Returning to an “untouchable” study is not an attack on previous researchers, but part of science’s core mission, Sandström and Riis emphasize. The new analysis confirms that Wennerås and Wold identified differences in the evaluations – but corrects the interpretation by showing that these differences primarily reflect structural and field-specific disparities, not gender discrimination.

The lesson is clear, according to the two researchers: Science’s credibility is not based on always being right, but on the willingness to test, revise, and correct itself.