Over seven percent of all citations on ResearchGate are self-citations

Which scientific disciplines do this most? We looked into it.

A "self-citation" is a citation in which the author has cited one of their own papers. Despite criticism of self-advertising or vanity, self-citations can also be useful, demonstrating how the current article builds on previous findings. To get a sense of how self-citations are used in practice, we looked at publications across disciplines on ResearchGate. We also asked information specialist Hadas Shema of The Inter-University Computation Center to weigh in on what we found.

First, we looked at what percentage of a scientific discipline's total citations are self-citations. Engineering leads the way with 11.06 percent of all its citations being self-citations.

Second, we took specific study areas from within each scientific discipline and looked at what percentage of their total citations are self-citations. Plasma Physics and Algebra lead the way with 14.9 and 14.33 percent of all their citations being self-citations respectively.

This graph is interactive, mouse over each data point for the corresponding study area.

Third, to see the distribution of self-citations within a scientific discipline we have slotted researchers into categories based on what percentage of their total citations are self-citations.

By hovering your mouse over the line you will see that for 13.47 percent of researchers in Medicine (pink line), self-citations make up 1 percent of their total citations. Or that for 0.83 percent of researchers in Medicine, self-citations make up 50 percent of their total citations.

This graph is interactive, you can click on the legend to remove and add scientific disciplines. Note that spikes at percentages like 100 and 50 reflect values from researchers with few publications.

RG: How do you think self-citations are generally perceived in the academic community?

Hadas Shema: That is an interesting question. In theory, self-citations are frowned upon, because they go against the ethos of science, which demands impartiality and forbids the promotion of self-interests. In practice, self-citations are accepted, as long as they're not excessive. Metrics based on citations normally include self-citations, showing they are recognized as valued scientific communication.

RG: Do researchers benefit from self-citations? If so, how?

Shema: Self-citation definitely has its benefits. Fowler and Aksnes found that in ten years, every self-citation generates 3.65 additional citations from other articles. However, this effect diminishes as the number of self-citations increases. There is a limit to the benefit one can receive from self-citations. Fowler and Aksnes estimated that at 40-50 self-citations a year, additional self-citing will decrease citations from others.

ResearchGate: What are your personal feelings towards self-citations?

Shema: I see self-citations as part of the scientific process. Sooner or later, one has to build on previous works and therefore cite herself.

RG: After looking at the graphs, why do you think certain scientific disciplines and study areas self-cite more than others?

Shema: The results are similar to those of previous studies, which found that the mathematical, engineering, and natural sciences have high levels of self-citation. Broadly, I would suggest that the natural sciences tend to have more authors per paper than the social sciences and the humanities, and therefore more opportunities for every paper to be self-cited. This is especially true for certain sub-disciplines of physics, where there can be hundreds of authors per paper.

RG: Does anything here surprise you?

Shema: I would have expected the medicine discipline to have higher levels of self-citation, given that they tend to have a relatively large number of authors per paper.

RG: What do you make of the paper that argues that men cite themselves more than women? Could this have played a role in the data?

Shema: Statistically, men tend to publish more than women, which gives them more papers to self-cite. It is possible that partly because of that, in male-dominated disciplines such as engineering and physics, there will be more self-citing.

RG: What should people take away from the information depicted here?

Shema: Self-citations are part of the academic life and are perfectly acceptable, as long as you don't overdo them.