In the most expansive analysis ever carried out to assess the gender differences in scholarly article publishing, Carl Bergstrom’s laboratory and collaborators from NYU analyzed over 2 million journal articles that listed 2.7 million authors, spanning across 345 years and hundreds of disciplines. Half of the articles in the study were published between 1665 and 1989, and the other half within 1990-2010. The study found that across the entire 345 years, 22% of all authors and 19% of first authors were female, and women tend to publish more heavily in certain disciplines, such as Cognitive Science (30.2%), Pollution and Occupational Health (35.3%), Demography (32.5%), and Education (37.2%), and are less represented in Mathematics (6.6%), Economics (9.7%), and Philosophy (9.4%). Though, from 1990-2010, the overall proportion of women authorship went up to 27%, and in 2010 alone the figure crept up to 30%.
The prestigious position of first author on a scholarly paper is often used as a metric for granting tenure and promotions in the world of academia, and the position of last author signifies a senior scholar that might not have carried out a majority of the research, but directs the lab in which the research originated. Women were less likely to hold the esteemed position of last author. For example, from 1990 to 2010, women authored 30% of the papers published in Cell and Molecular Biology, but only 16.5% were last authors.
The researchers ask the question "why is it not 50-50?", and this Chronicle of Higher Education article sheds lights on why this might be.
Also, check out the interactive map that is the product of the analyses performed by the Bergstrom lab and their collaborators.