The apparent trend has been on the rise over the past two decades.
Men cite their own papers 56% more than women on average, according to an analysis of 1.5 million studies published between 1779 and 2011.
The analysis looked at papers across disciplines in the digital library JSTOR and found that men’s self-citation rate had risen to 70% more than women’s over the past two decades, despite an increase of women in academia in recent years. Around 10% of a given paper’s references are likely to be self-citations by the paper’s authors regardless of their gender.
What the analysis, posted on arXiv on TK June, cannot clarify is whether this trend is a by-product of the under-representation of women in senior academic positions or some separate effect.
According to the paper, academics working in ecology and evolution, sociology and molecular biology are the most likely to cite themselves, whereas historians and classical scholars are the least likely.
Tooting their own horns
Citations define scholars, for better or worse. They are the currency of widely debated metrics such as the h-index that measure productivity and the impact of researchers’ publications. These metrics are often the basis of decisions made on tenure and grant applications.
Most self-citations are “appropriate,” where scholars are legitimately citing past work, says Molly King, the study's lead author and a sociologist at Stanford University in California. They aren’t necessarily a conscious attempt to game the system.
However, a small fraction could be deliberate attempts to boost an academic’s own citation count. “Every citation a paper receives will attract additional future attention from other scholars,” says King. So if men are citing their own papers more, they will receive extra attention from their peers, she notes.
It’s something that hiring and tenure committees should take into account when assessing the impact of researchers and their work, King says.
But wait …
There are some limitations to the study.
King and her colleagues deciphered the gender of all authors listed on a publication based on first names and their associated sex in US Social Security Administration records. They discarded gender-neutral names. But they also dropped authors listed with only a first initial, which the researchers acknowledge may have excluded women disproportionately.
In fact, the paper suggests that female authors may in some cases be using only their first initials to obscure the fact that they are women.
“Only 56.4% of all the authors in their database were able to be assigned a gender,” says Adrian Letchford, a data scientist at the University of Warwick in Coventry, UK. “That's a very big portion left out, especially considering that women may be actively hiding their gender.”
There was also no way to control for an academic’s productivity — or the fact that men generally have more senior positions in academia, publish more papers and therefore have more work to self-cite, says Cassidy Sugimoto, an information scientist at Indiana University Bloomington.
“Self-referencing is highly dependent upon productivity: one cannot reference a work one has not written,” she says. Around 78% of the authors in the study’s sample population are men and roughly 22% are women.
There are other possible explanations for the trend, according to the study authors, including the idea that men view their abilities more positively than women do or that men face fewer societal penalties for self-promotion than do women. Another possible contributing factor may be that men tend to specialize more. That means that they would have fewer peers working in the same field, and so they might end up citing themselves more.
Nevertheless, says Sugimoto, the topic is an important one to look into. “The degree to which there is gendered behaviour in self-referencing is an important observation and one that can open new lines of inquiry and has potential implications for science policy.”