20

If one looks superficially at the publications of author X, he seems to be one of the top scientists in his field. But looking more closely his citations are 60% from himself 35% from his co-author, and 5% from others. (Some people might say: “It doesn’t matter, probably he is the only one working in such a field and he has to cite himself.” However, let’s say that it is not the case.)

This author is only one example of many people who do self-citations. Many of us have grown up to see citation as a measure and started to mishandle it. As it is easy to do so. My questions are:

  1. Isn’t there a system which monitors unethical self-citations apart from the peer-review process, where such authors pass such tests and still extensively cite themselves?
  2. Isn’t there a system that blacklists authors who involve in similar unethical activities as “fake scientists” and never allows them to publish anywhere else?
  3. Do institutions not realize that such professors are a potential threat to science and ask them to resign?
  4. Don’t such people demean real science people who spend years of hard work to publish their work?
  • 49
    I strongly disagree with the notion of self-citations being in general bad/inappropriate/unethical. – Walter Dec 3 '17 at 19:05
  • 22
    Most comments and answers discuss self-citation per se, while the case here is about excessive self-citation. The original version included a link to a Google Scholar profile where the citation count went from 500 to 5000 in a single year, mostly based on self-citations. – lighthouse keeper Dec 3 '17 at 20:02
  • 9
    The specific case covered is quite notorious, and seems to be a case of an institution trying to outright cheat the system in order to increase their University’s rank: see e.g. the following article, or this quora Q&A. – dfri Dec 3 '17 at 20:42
  • 12
    Self-citations are not unethical, but if you do any bibliometry, they should be ignored. – yo' Dec 3 '17 at 22:59
  • 10
    Goodhart's law states that ""When a measure becomes a target, it ceases to be a good measure." At some point people decided that citation count was a good measure of an academic productivity (and it probably was). As soon as important decisions (promotion, tenure, etc) started being made based on citation count, some academics started gaming it. If we "fix it", then some academics will game the new fixes. – emory Dec 4 '17 at 13:33
43

I don't think there is a good way around self-citations, since most of the time, your current research builds upon your previous research, and therefore, your previous research (or that of your collaborators) is relevant to your current research. However, from my experience, self-citations form a small part of the overall number of citations.

Assuming the author's group is not the only one working on a topic, there seem to be two other problems here:

  1. other people don't seem to cite his work, which could mean that he is ignored, his work isn't visible or his work is just not good and/or relevant enough.

  2. basing quality on number of citations is flawed. The number of citations is closely related to an authors H-index, which is criticized regularly. The same criticism should be valid for number of citations.

To tackle the "problem" of self-citations (I would not say this is a problem in the majority of cases), some citation-based measures are based on citations excluding self-citations. This does of course get tricky when collaborators are citing someone's work, but the large bulk of self-citations will be removed this way.

  • 12
    I agree with everything, but I think the key part is that self-citation is normally not a problem. – TimothyAWiseman Dec 3 '17 at 23:04
17

To answer the questions directly:

  1. There is no such "system" that I know of. There are a variety of independent "watchdog" groups, bloggers, etc, who sometimes draw attention to what they see as abuses in academic publishing, and occasionally notorious cases will be discussed in academic or popular media. But I'm not aware that these people or organizations monitor these issues in a systematic way, beyond individual cases. There is certainly no single global body overseeing the issue.

  2. Again, there is no system. Academic publishing is decentralized and every publisher makes their own decisions about how to run their journals. Individual publishers might have their own internal blacklists of authors who have committed abuses in that publisher's journals, and whose work will no longer be considered by that publisher. But there is no global blacklist, nor any central organization that would have the power to enforce it. Many people would consider such a blacklist to be unethical - such systems have in the past been abused to exclude people for inappropriate reasons - and in some jurisdictions the idea of an industry-wide blacklist might be illegal.

  3. Institutions make their own decisions. They might not know about such abuses until they are brought to their attention by someone else. They might or might not consider them significant enough to fire a researcher. They might be swayed by the researcher's success in other areas. There might be disagreement within the institution about what to do. In many places there are tenure policies that can make it difficult to fire a researcher; it might require a broad consensus among faculty and administration that the researcher's actions are unacceptable, as well as requiring a long and costly legal or quasi-legal process. Institutions might not feel that self-citation abuse is worth the effort.

  4. Arguably. However, that is a matter of opinion and off-topic for this site.

12

1) I suspect social values concerning self-citations may vary field by field, but in my (partly-former) sub-discipline within AI, self-citations are not frowned upon as far as I can see.

2) It's not necessarily that people think it's a good way to increase the h-index because some metrics exclude self-citations. Instead, self-citing helps boost the profile of the paper, which in turn benefits real citations.

3) Even if self-citing is terrible, then the few that do self-cite a lot force others to do so too, that is, some people may simply be doing what they have to do.

People are gaming academia in such cases, but not Science itself. I don't think citations and the h-index are partly-constitutive of Science. So, if it is wrong, then it is violating an academic rather than a scientific norm, in my opinion. The 'real science' remains science regardless of if it is cited a lot or not at all.

  • 3
    Self-citing in itself is not terrible. If I extend my previous research and don't self-cite I'll probably be plagiarising myself, so it's even necessary sometimes. Only excessive self-citing could be 'terrible'. – Hans Janssen Dec 4 '17 at 13:47
8

The problem is that citation counts are nothing better than a gross approximation to a measure of quality. You can see this from the fact that "Smith is a moron who published an extensive series of papers [1-7] claiming that influenza is spread by eating cabbage; Jones [8] has conclusively demonstrated this not to be the case" gives Smith's junk many more citations than Jones's proper science.

  1. Do institutions not realize that such professors are a potential threat to science and ask them to resign?

They're not a threat to science at all. They're a threat to people who believe that science is about counting citations.

  1. Don’t such people demean real science people who spend years of hard work to publish their work?

No. We demean real science people who spend years of hard work to publish their work by reducing that hard work to "I have ten citations and you have six so I'm better than you."

7

Single-author self-citations (in peer reviewed publications) are hardly capable of generating a significant citation count (for peer reviewed publications) or H-index, unless author X publishes at a very high rate in peer reviewed journals. The story is different with large collaborations: if each of ~100 authors cites their paper in their next few publications, this generates several 100 citations.

Edit. Oft course only peer reviewed publications should be considered to assess scholarly impact.

  • 3
    "Single-author self-citations are hardly capable of generating a significant citation count or H-index" It depends on the tool used for counting the citations. For example, Google Scholar indexes every PDF that looks like a paper, including legitimate, but non peer-reviewed work (e.g. undergrad homework), and plain bogus papers. – lighthouse keeper Dec 3 '17 at 20:00
  • 1
    The majority of "citations" in the profile originally linked were form self-published stuff on Researchgate and the like. – Cape Code Dec 3 '17 at 20:57
  • @lighthouse keeper yes it is feasible and many people are doing it. I can repost the link of several profiles I know from Google scholar. – Ramanathan Varadharajan Dec 3 '17 at 23:30
  • @lighthousekeeper Homeworks aren't going to contribute to self-citation counts. – David Richerby Dec 5 '17 at 17:54
  • @DavidRicherby Indeed, thanks for clarifying. An author could still use homework to boost his citation count at GS, in particular if he is the instructor of a well-attended class. The homework could be to write a short review of his recent paper and to upload this review online. – lighthouse keeper Dec 6 '17 at 8:08
5

There is a system that ignores such citations. Ideas RePec (Research Papers in Economics) compiles citation counts for economics research. Author rankings are computed that exclude self-citations and which, when ranking journals or working paper series, exclude citations from that same outlet.

For example:

https://ideas.repec.org/top/top.person.nbcites.html

3

There is nothing, at all, unethical about self-citation. Authors should cite their previous work where it is relevant to their current work, as should co-authors and anyone else who refers to the work.

On the other hand, treating citation counts as a measure of ability in unethical. As with every other metric of this kind, it perverts and distorts science.

Your question therefore is aimed at the wrong problem, and because the question is based on an incorrect premise there is no good answer to it.

1

Self citation is a reasonable thing to do if you want it to be easy for people to follow the history of your work.

Most progression of science is gradual and most progression of each scientists individual work is also gradual.

Therefore it is not very strange if a scientist explains the gradual evolution of their work by citing themselves.

1

This is bound to happen; researchers follow their line of inquiry. And yes, it will help their H-index unless manually corrected. But, most top-tier institutions are wise to this, and a hiring committee will look at this in-depth. At my former institution, the H-index was recomputed less self citations. Hiring committees and grant panels won't fall for this, so don't be too concerned.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.