The citations found at the end of research papers serve several purposes, like providing background on the current work and giving proper credit where it’s due. They can also, according to a new study, reveal decades’ worth of trends in whole fields of science.

A trio of researchers have waded though more than half a century of research published in astronomy journals and found that studies authored by women receive 10 percent fewer citations than similar studies written by men.

Neven Caplar of the Swiss university ETH Zurich and his colleagues analyzed more than 149,700 papers published between 1950 and 2015 in five journals: Astronomy & Astrophysics, The Astrophysical Journal, Monthly Notices of the Royal Astronomical Society, Nature and Science. They made sure that the papers being cited matched up in variables unrelated to gender, like the lead authors’ seniority in the field, the institutions they wrote from, the total number of authors on the paper, the number of references, the year and journal in which it was published, and the specific field of study. They say their findings, published Friday in Nature Astronomy, quantify the effect of gender bias in citations within astronomy research.

If there were no gender bias in astronomy research and only these factors mattered, the researchers’ analysis predicts that men would actually receive 4 percent fewer citations than women would. So their actual results were surprising—to the algorithms, at least. In the context of history, their findings are not surprising at all.

Since the late 1990s, women in the United States have earned nearly 60 percent of all bachelor’s degrees, but about half of all degrees in science and engineering fields, according to the National Science Foundation. The number of women receiving degrees in science is on the rise, but women remain outnumbered in many of these fields, particularly in physics, engineering, and computer science. In 2013, an analysis of more than 8 million papers in the fields of natural sciences, social sciences, and humanities showed that men are more likely to be listed as lead authors. So it follows that with fewer women getting degrees, becoming researchers and professors, contributing to papers, and then leading papers, there are fewer women to cite.

Some of the gender disparity can be attributed to the nature of the workforce. Most science professionals got their degrees in the last 40 years, and those people tend to be disproportionately male and white, National Science Foundation statistics show. A 2014 report on an annual meeting of the American Astronomical Society found that although the gender ratio of speakers matched that of the audience, more men than women asked questions of the participants. The researchers in this study interpreted this observation to be a product of the workforce. More senior scientists may be more likely to ask questions, they wrote, and senior scientists are usually men. Another survey of participants at a National Astronomy Meeting, organized by Britain’s Royal Astronomical Society, made similar observations about question-askers. A 2016 survey of more than 13,000 requests for use of the European Southern Observatory over eight years found that female applicants had significantly lower chances of getting telescope time. The study attributed this result to the effects of seniority; only 34 percent of the women applying were professionally employed astronomers, compared to 53 percent of the men.

Critics of the effect described in the Nature Astronomy study could argue that researchers seek to use the best sources in their work, regardless of gender. Any perceived preference for male-led work surely must be unintentional. But research has shown that when gender is taken out of consideration, potential implicit biases fade away and the scales balance. In 2001, the journal Behavioral Ecology started using a double-blind review that masked the genders of the applicants being evaluated. This led to “a significant increase in female first-authored papers, a pattern not observed in a very similar journal that provides reviewers with author information,” according to a paper that examined the policy. “No negative effects could be identified.” A similar effect has been found in hiring. In a 2012 study, researchers simulated an application process for a laboratory manager job, randomly assigning applicants either a male or female name. The applicants, members of faculty at a research university, were given identical credentials for the applicants. Yet the participants—both male and female faculty—rated the male applicant as significantly more competent than his female counterpart. Even scientists, some of the loudest advocates for objectivity, are not immune to deeply rooted differences in the perception of men and women.

The Nature Astronomy study does have some encouraging findings. The number of astronomy papers authored by women has increased over the last 50 years, and the difference between the number of female-led and male-led papers in citations has shrunk, the researchers write. They found that back in the 1950s and 1960s, men received between 50 percent and 100 percent more citations than women did.

The average number of citations in a paper has also increased, from about 10 in the 1960s to about 60 today, providing room for more authors to be recognized and credited, male or female. But the disparity persists, in astronomy and likely elsewhere, and even in the very study that examined it. Of the 19 authors cited in the paper, just six are women.