Thursday, February 09, 2012

Citation Conspiracy

Someone recently told me about this, and I was wondering if anyone has participated in something similar:

A group of colleagues makes a specific effort to cite each other's papers -- those paper not involving the author/s doing the citing, so no self-citation is involved -- to help each other get their citation numbers up. They don't gratuitously cite a paper that is irrelevant to the topic at hand, but they proactively seek opportunities to cite each other's papers, and, given a range of options for citation of a particular point, they will choose to cite a paper by someone from this group.

If you have not participated in something like this, does it bother you that some people do this?

I have not participated in a citation-circle like this, and the fact that others do does not bother me. These people are not inappropriately citing their friends -- the citations are all relevant -- and it is likely that most of us do something like this anyway, even without making a concerted effort. We tend to cite papers with which we are familiar, no matter how diligent we try to be in surveying the vast literature in each of our sub/fields.

Does a citation-circle have any measurable positive effect on the career of a particpant? If it is effective, involving a sufficient number of productive (in terms of publications) researchers, it can make the difference in the citation numbers (h-index and so on). Increasingly, career advancement relies on having good citation numbers, so being in a citation-circle might be quite helpful, even if it doesn't result in a dramatic jump in citations.

Does a citation-circle harm those not in it? I suppose one instance in which a citation-circle, even one conducted in an ethical way, might have an unintended negative effect on someone not in the circle would be if one of the "proactively cited" papers becomes one of those papers that is commonly cited in introductions. And then, because it is cited prominently in some papers, it gets picked up as the go-to cite for introductions in papers on similar topics. The citation numbers can then snowball, and other papers might languish in undeserved obscurity.

I think that happens to some papers anyway, with or without a few citation-circles in action. Therefore, I was intrigued by the existence of such citation-circles, but not disturbed. But that's just me, perhaps reflecting my secure position as a mid-career professor who obsesses about citation data mostly out of curiosity rather than out of necessity. I am curious if others feel otherwise, but I need to note again that comment moderation will continue to be sporadic for a few more days (sorry).


23 comments:

Anonymous said...

As a junior professor who may or may not be affected by citation counts, this does not bother me. It's exactly what you expect people to do when you rely on a number with an easy-to-understand formula that leads to easily-comprehensible ways to make that number go up.

What bothers me is the reliance on that number.

I remember commenting on a previous post here that I don't pay much attention to my citation counts, because if I did, I would start thinking about how to game them. This is exactly what I meant.

Anonymous said...

I always assumed that everyone does this... subconsciously. That is of course when we are thinking of citations for a particular statement / result our minds will immediately jump to that paper our buddy wrote on the topic last year (in addition to any earlier, seminal work, of course). We simply may not be aware that a paper which would be just as relevant exists.

I don't really know a way to avoid this other than for more scientists to read more papers... which of course would be good but given the increased competitiveness and focus on research output, ends up being lower on the priority list than it perhaps should be.

As for overtly or specifically doing this... meh. I kind of think that since we all have a subconscious bias towards citing our friends'/collegues' papers anyway, we ought to make a concerted effort to cite papers outside of our circle, not make the bias worse. I don't know if I'd call it immoral but it's a little shady.

Anonymous said...

I know of a large group within one scientific community that does this, It does indeed harm the science because many of the most popular scientific journals have page limitations, or even a limit to number of references. Thus, other papers that are relevant and even crucial to the manuscript study are not cited at all so that the friends' papers can be squeezed in. It is a bad practice!

Mrinny said...

I do not really belong to such a citation-circle as such... but i do know that it would be easy for people who know me or read my work to feel that way.

I work in a large institute where-in the faculty often give several public lectures in a month. It just so happens that many of my ideas are sparked on such events. Also, the brainstorming sessions or the questions asked lead to new insights as well as links to other peoples research work.

When i do feel that someone has contributed substantially to the work i am doing, they end up getting cited. The more time i spend with them, the more our work links up. So i guess, to someone 'outside' of my citation circle, it would seem like what you are saying.

It is just that in this particular case, emails, lectures and friendships can often lead to a citation bias that is a logical outcome of the influences certain people have in your work, ideas and how often you discuss work with them at the cafeteria.

I DO know of some citation circles that do not fall under this category and that is when i will double-check other papers where the work has been cited to assess the persons credibility.

Anonymous said...

Unless these group are producing 20+ articles per year, I don't think this is going to impact their career in any significant way. It is unethical however to do it deliberately. No, it doesn't disturb me.

Phillip Helbig said...

It's just a small step between this and bribing people to cite you. It makes an unjust system worse. It demonstrates that judging scientists on bibliometry is almost as bad as judging music by the singles charts.

EscapedWestOfTheBigMuddy said...

On a related note, there is a paper floating around that examines the citation metrics of some journals that seemed to encourage citing of papers previously published in the same rag.

As I recall the argument was that the usual metrics of "impact" could be gamed by this method.

arXiv:1010.0278

Find it at http://arXiv.org for those not in fields that use arXiv.

barbylon said...

Interestingly, I just read about this study of journals pressuring authors to cite that journal's articles to bring up the journal's impact factor. I have many papers in different journals and I have never experienced this firsthand. http://blog.al.com/huntsville-times-business/2012/02/university_of_alabama_in_hunts_3.html

John Vidale said...

It's disturbing. However, as there are numerous ways to bump ones citation count without organization, and there are also inherent biases towards inflating citation counts of influential people, it is not uniquely disturbing.

Suggestions as a reviewer, as an editor, or as a colleague can also promote one's own papers. Even without a nudge, authors try to placate potential reviewers with juicy citations. And once a paper is well-cited, others often just copy the citations they find in the literature (even copying the typos).

There would be a price to pay for the citation circle strategy, as one would gain a reputation for being parochial and in-bred, on the other hand one might get friends as reviewers if lazy editors just scan the reference list for reviewers.

Citations may be easiest way to get a handle on the influence of a scientist in an unfamiliar specialty, but the numbers are not completely objective.

Anonymous said...

I'm curious: Does the "someone" you spoke with participate in such a circle? I agree that coordinating efforts to cite each other's work makes sense on some levels, even though it seems like those levels could become sinister rather quickly. Hence I would not be particularly surprising if something like this existed. But I have no actual knowledge of one. Does anyone? Or are these just one more urban legend designed to freak out assistant professors (and a healthy swath of the rest of us) about how much we're being cited??

Anonymous said...

My impression (it may be incorrect) is that this is more of a problem in Europe, where citation counts are very much used in evaluations (e.g.list the PI's 5 most relevant papers and their citation counts).

So there can be a lot of incentive to game the system and so a lot of people do. A cynical person (like me, for example) might even suggest that this seems to be the purpose of the EU "network of excellence" programs...

I think one part of the problem is that when you see that others are gaming the system, it's easy to feel pressured to game the system yourself.

This is reinforced by the fact that it's really hard to catch this sort of behavior - partly because reviewers look for important missing citations rather than 'extra' ones - partly because the literature for any topic is so large that no one feels like they master it - and partly because, as many have pointed out, there is some natural tendency to cite within one's collaborative circle because it's the work one is most familiar with.

Anonymous said...

the question is not whether it "bothers" me or you or anyone else. That may be a personal matter. The real question is whether it is hurting the scientific process, and I think the answer to that is a definite YES.

I see the main argument in the other direction to be: we already do this at some level, we already cite the work of our buddy's who work in the area, ...


Really? This is not the case for me at all. I have never met the authors for the majority of papers I cite. In fact, I always do an independent literature review for each paper, and try to choose the ones that are most relevant to the current work. Aren't we all supposed to be doing this? Isn't this what we teach our students to do?

Anonymous said...

Wow. This strikes me as one of those situations in which no one think s/he's really doing "the wrong thing"--just favoring friends a bit more than others-- but the consequences at the population level are terrible. Citations matter enormously to employment and thus indirectly (and maybe directly) to funding. Deferring to cliques instead of sampling (reading) fairly and citing based on merit increases the power of cliques and diminishes academic mobility and diversity, which can't be good for science as a whole. It's also damn annoying for people who aren't members of in-groups or who do good work but can't afford to fly to all the conferences.

I've made an active effort for years not to cite friends just because I'm familiar with their stuff and it's easy. I always try to look more broadly for work I don't know. I'm a postdoc and I find it deeply distressing that most commenters, presumably including senior researchers, don't find "citation circles" troubling.

Anonymous said...

This sort of thing must go on on all different size scales. In my field of physics there's a well known citation circle. We call it France.

EliRabett said...

More troubling is the refusal of some to cite others who did important relevant work, although it is sometimes fun to watch them waltz around the obvious

Anonymous said...

For those that are saying this is bad for science.. isn't the reliance on citation counts, h-indexes, and impact factors even worse for science? If that reliance went away, so would this.

Anonymous said...

I have never heard of any such formal arrangement. However I do know of a couple of close knit, rather irrelevant fields which for all practical purposes work as citation circles, but again this is not an open agreement.

It goes as follows: select uninteresting problem with room for many variations. Create a journal/conference/society on the subject and start publishing epsilon increments on said problem. No one else cares about the problem, so all the citations come from within this group of colleagues.

Anonymous said...

8:18 Anon again. Another issue here is to what extent participation in formal citation circles affects bias in reviews. It's already hard enough to review dispassionately the manuscript of someone I know and like, and it's probably a bit harder when that manuscript (accurately) cites my own work. It seems that participation in citation circles would greatly compound these problems.

I've no problem with friendships shaping the formation and development of collaborations, but keep them out of the major metric of scientific merit. We can't be perfect, but we can try.

Anonymous said...

This is a morally problematic but perfectly rational response to the weight of citations on people's livelihoods. Any system can be gained. Any ranking system for academics will undoubtedly be able to be gamed. The players are, typically, smarter than the game designers and there are definitely some with far fewer scruples.

David S said...

It bothers me in so far as those scientists are spending precious time actively organizing a way to game the system, rather than doing Science.

Anonymous said...

isn't the reliance on citation counts, h-indexes, and impact factors even worse for science? If that reliance went away, so would this.

Depends what it is replaced with. In many countries the use of indexes and bibliometrics implied a marked increase in the correlation between quality of research and funding decisions.

Also do you really believe that your institution will replace the use of indexes with careful reading of your records and papers? If so then by all means go ahead and never use them again.

If on the other hand you have an overworked hiring/promotion committee who can use bibliometrics as a rough guide as to which candidates deserve more attention (be it positive or negative) then we are better of using them.

Anonymous said...

Wow - Anon 2:05 - France was my answer too! I'm in biology and there's a well known french group in my field that really just cites themselves and each other. I'm sure there are other wonderful biologists in France not doing this (in fact I know some) but it does make people snipe about the issue and get a little anti-euro at times.

Personally I think it's a crappy thing to do in an organized fashion but I admit I cite my friends a lot. This is not a conspiracy but they are my friends from grad school - we worked on similar topics and they do good work so should I not cite them? I have avoided citing more peripheral papers from my very productive advisor because I don't want to seem to be riding his coattails but in the core stuff his work is the best so what can I do? I think there is a line between citing papers because they're the right ones as well as a friend's and setting out to find ways to cite your buddies.

Martin said...

So if there is a clustering of citations why not do a weight measure of citations? Papers with lots of citations with a lot of spread in terms of domains would get higher ranking than clustered citations. The weighing could be based upon university, proximity etc.

(I've just started my PhD so maybe I am missing some point here.)