Marginal Revolution: Does self-citation pay?

November
S	M	T	W	Th	F	St
						1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30

« Predictions about 2000 from 1900 | Main | Resale price maintenance »

Does self-citation pay?

in a word, yes:

Self-citations – those where authors cite their own works – account for a significant portion of all citations. These self-references may result from the cumulative nature of individual research, the need for personal gratification, or the value of self-citation as a rhetorical and tactical tool in the struggle for visibility and scientific authority. In this article we examine the incentives that underlie self-citation by studying how authors’ references to their own works affect the citations they receive from others. We report the results of a macro study of more than half a million citations to articles by Norwegian scientists that appeared in the Science Citation Index. We show that the more one cites oneself the more one is cited by other scholars. Controlling for numerous sources of variation in cumulative citations from others, our models suggest that each additional self-citation increases the number of citations from others by about one after one year, and by about three after five years. Moreover, there is no significant penalty for the most frequent self-citers—the effect of self-citation remains positive even for very high rates of self-citation. These results carry important policy implications for the use of citations to evaluate performance and distribute resources in science and they represent new information on the role and impact of self-citations in scientific communication.

Well, that explains some observations. Here is the paper, indirectly the pointer is from Aleks.

Posted by Tyler Cowen on April 20, 2007 at 03:56 AM in Education | Permalink

Comments

Wouldn't self-citation also correlate with total number of published papers?

Posted by: talboito at Apr 20, 2007 4:26:33 AM

Personal gratification tempts Indian economists to list almost all of their works in their references.Self-advertisement and self-promotion!

Posted by: GVV at Apr 20, 2007 8:29:29 AM

How do they know that self-citations CAUSE the other citations? This question would be important in considering "policy implications for the use of citations to evaluate performance and distribute resources in science."

Posted by: mae at Apr 20, 2007 9:46:32 AM

wtf? i agree with mae. the quality of the paper might correlate both with self-citations and 3rd-party citations, i.e. you self-cite your good papers a lot, and other people cite them often , too, but not because of your self-citation, but because it was a good paper.

Posted by: Nonsequiturboy at Apr 20, 2007 9:56:27 AM

This may be the result of the neutral study, but I can also say
that anecdotally people who over-self cite, especially when it is
perceived as not being well-deserved, and especially if they are
doing so while ignoring deserving, especially more deserving
sources to be cited, can get bad reputations. This is not as
straightforward as it seems.

I think this is a matter of a tail on the distribution. I am not
surprised at the general relation. Those who really are on top
deserve to self-cite a lot. But there is a subset who do not and
probably damage themselves somewhat. But they are too few in
number to offset the deserving self-citers.

Posted by: Barkley Rosser at Apr 20, 2007 10:30:55 AM

I recently came across a paper by an economist where out of 34 citations, he self-cited an astonishing (to my modest standards) 16 papers (6 as a sole author). The paper was published in the first issue of a review edited (created?) by the author. By the way I also found this American Professor to be very good at self promotion (with his own domain name and web page): self citations may be only one manifestation of an individual's capacity to promote themselves, thus not necessarily the direct cause of more cites down the line.

By the way, I quoted the guy…

Posted by: Jean-Christophe at Apr 20, 2007 11:00:58 AM

By the way, I cited the guy…

Posted by: Jean-Christophe at Apr 20, 2007 11:02:06 AM

I can think of at least two other effects of self-citation:

a. It serves as an ad for your old papers. Hey, guys, if you think this work is interesting, maybe you'll finally go pound your way through my previous three papers dealing with other aspects of the same problem.

b. Sometimes, this serves as a source of examples from other fields or areas within your field. Here's my new attack on block ciphers, and here's my previous work on protocol design, where this attack would have an interesting impact, and here's some previous work on security proofs, where this attack isn't accounted for in the proof model, and....

In the second case, self-citation may indicate breadth of interest, which may be correlated somehow with success. (I'm not sure about the direction, though.)

Posted by: albatross at Apr 20, 2007 11:17:24 AM

Isn't self-citation highly efficient? If you have already done the work before and don't want to have to re-explain everything, just send the readers to your prior article where everything is explained. Then use that to build your current edifice.

Makes sense?

Posted by: Stuart at Apr 20, 2007 12:56:41 PM

Also, isn't self-citation to some extent a proxy for having written important papers or being one of the primary experts on a particular topic? If many other people are citing your papers, then chances are that you're going to cite them too, for the same reason.

Posted by: Brandon Berg at Apr 20, 2007 2:46:04 PM

Stuart's got it right. In my end of the academic world (computational fluid dynamics, supercomputers, etc.), self-citation is critical. I often need to cite older papers where a method has been explained or the development of a computer program has been described so that I don't have to waste 5 pages redescribing how something works when those details are immaterial to the point of the current work. I can just say "Go see refs 3, 4, 5, and, 6 for details on how this all works," and then move on to the important results in the new paper. This is quite common in mathematics, physics, engineering, etc.

I suppose things are often different in the humanities where one may be referencing an old argument rather than a technique, but nobody should look askance at an author for citing a paper that describes a technique in detail that is being employed in the citing work.

Posted by: billb at Apr 20, 2007 4:20:32 PM

billb,

This is also the case in econ. Again, much of the self-citing is
not only justified, but in fact necessary. If somebody has alreaedy
published a crucial part of an argument in a current paper, it is
highly legitimate and often demanded by editors or referees that they
not waste journal space in the new paper by repeating the already
published and available results, derivations, whatever.

Posted by: Barkley Rosser at Apr 21, 2007 6:42:40 PM

I know the study has 65,000 papers or so, but I can't help but wonder whether Stiglitz is driving the results.

Posted by: Eric Crampton at Apr 22, 2007 11:11:10 PM

I would note one other practice that is more questionable.
It is the referees on submitted papers who demand that the
author(s) cite them, especially them and only them, and maybe
several times to boot. Now, once again, there are times when
this is totally justified. The referee really has written an
important paper in the area that should be cited by the author(s)
of the paper. But a lot of it is sheer blackmail, "cite me if you
want your paper published!!!" I would suggest to readers of this
who wish to do this that your credibility will be heightened if
you suggest citations of some other appropriate sources besides
your own glorious masterpieces. Otherwise, well, editors have
less respect for this sort of overdone thing than they do for
overdone self-citations in papers themselves.

Posted by: Barkley Rosser at Apr 23, 2007 1:08:40 PM

I appreciate very much the comments here, but I strongly recommend readers actually read the paper before judging it. For example, talboito thinks we should control for total number of publications and nonsequiturboy thinks we need to control for paper quality. In the paper we do both! We control for a number of things including: 1) previous citations to the author's work by others (people who write better papers will previously have been cited more), 2) total number of current publications (this increases the number of opportunities to self-cite, but it will also correlate with average paper quality since people who write bad papers do not get published) 3) total number of previous publications (this increases the number of opportunities to be cited by oneself, and again is an indicator of previous paper quality), 4) number of coauthors (more coauthors means more people will cite the author because they are citing themselves), and 5) the expected citation rate for all the author's papers based on the mean citation rate of the journals they were published in. The most important one in this list is number one -- what we are essentially doing is comparing two people who previously received the *same* number of citations from others and then seeing how differences in self-citation influence their citations from others in the future.

Best,
james

Posted by: James Fowler at Apr 23, 2007 6:04:03 PM

liqingchao 07年8月14日

google排名
 google排名
 wow gold
wow gold
powerleveling
powerleveling
wow gold
wow gold
powerleveling
powerleveling
wow powerleveling
wow powerleveling
wow power leveling
wow power leveling
wow power level
wow power level
world of warcraft powerleveling
world of warcraft powerleveling
world of warcraft power leveling
world of warcraft power leveling