If authors duplicate portions of their own work in multiple papers — such as descriptions of methods, a boilerplate background to their field, etc. — should that be considered misconduct?
Of course, to many journals, duplication — also known as “self-plagiarism” is a retractable offense. A recent letter to the U.S. Office of Research Integrity (ORI) from the Council on Governmental Relations asked the agency to consider it misconduct, and “include self-plagiarism in the definition of plagiarism.” More specifically, the letter — reported by the Report on Research Compliance — says to new ORI director Kathy Partin:
ORI recognizes self-plagiarism as a questionable practice and a form of academic dishonesty. However, ORI’s definition of plagiarism as “the appropriation of another person’s ideas, processes, results, or words without giving appropriate credit,” makes pursing an allegation of self plagiarism problematic, especially if a university has not specifically included self-plagiarism in their research misconduct definition (many have used the ORI definition to construct their policies).
However, we’ve heard from many researchers over the years who believe that self-plagiarism is decidedly not misconduct. They wrote the material themselves, so why can’t they use it again?
Tell us what you think in our poll, below.
[polldaddy poll=9430303]
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.
I think the poll needs a “depends” option: there are circumstances under which I could consider it unremarkable (journal article included in book; methods sections; etc.) and circumstances under which I would consider it misconduct (portion of one book included in another without notice, substantially similar research published in multiple venues, etc.).
First things first, “self-plagiarism” is a nonsense term and should be abolished.
The question is hopelessly vague as posed. Duplication can be anything from submitting the same paper verbatim to multiple journals at once, to re-using a few key sentences in the methodology part because that is the standard way to describe a particular technique. The former would clearly be misconduct, while the latter probably would not be to most reasonable people.
Agree with TL above. There are just only so many ways to say: “We searched the following databases: Pubmed, Embase, CINAHL and Central. The keywords and subject headings were: blah, blah and blah. (Full search strategy in Appendix A).” Completely copying and pasting an introduction that provides background to your topic is lazy scholarship.
I guess the qualifier ‘self’ arose to indicate that the individual(s) concerned did not take from others but re-used their own material. Whether this is a useful distinction is a moot point. There are sentences and phrases that are used by everyone: after all there are only so many ways to say that proteins are polymers of amino acids. That in my mind does not count as plagiarism.
In contrast, using large section of text without attribution and quotes is deception, regardless of the origin of the text. Similarly, passing off data from one paper in another as new data is deception; doing the same, but taking data from experiment type 1 in the first paper and then passing these off in paper 2 for experiment type 2 is even worse.
TL and Amanda, please comment:
http://link.springer.com/article/10.1007%2Fs10555-016-9627-z
Retraction Note
Cancer and Metastasis Reviews
pp 1-1
First online: 30 May 2016
Retraction Note: Retraction note to: Targeted immune therapy of ovarian cancer
Keith L. Knutson, Lavakumar Karyampudi, Purushottam Lamichhane, Claudia Preston
“This article has been retracted at the request of the authors and in agreement with the Editors in Chief. The article contains large portions of text that have been duplicated from the articles:
Immunity and immune suppression in human ovarian cancer
Claudia C Preston, Ellen L Goode, Lynn C Hartmann, Kimberly R Kalli, and Keith L Knutson
Immunotherapy. 2011 Apr. 3(4): 539–556, DOI 10.2217/imt.11.20
The Immune System in the Pathogenesis of Ovarian Cancer
Bridget Charbonneau, Ellen L. Goode, Kimberly R. Kalli, Keith L. Knutson, and Melissa S. DeRycke
Crit Rev Immunol. 2013; 33(2): 137–164,
DOI: 10.1615/CritRevImmunol.2013006813
The authors would like to express their most sincere apology to the editors and readers of the journal.”
PS:
https://forbetterscience.wordpress.com/2016/04/11/academic-self-plagiarism-misconduct-or-a-literary-art-form/
Copying and pasting parts of your own Methods sections should be considered fine. For instance, there’s only so many times (and ways) that you can say “All procedures were approved by the Ethics Committee of […]”
If you actually used the same methods multiple times, there shouldn’t be any reason you can’t use identical words to describe those identical methods.
Yes, methods are a special case…
Or if running a survey or large scale epidemiology then there will be multiple papers and it doesn’t seem sensible or maybe even possible to have a different methods. For example “N patients were recruited from the x geographical area using sampling method s, etc”. Shuffling the words will still make it self-plagiarism.
Surely it will depend on the field, but if you have methods that are becoming quite common in your work, a proper citation could circumvent this conundrum. “Stuff has been done as reported elsewhere…” is a very common way of doing that.
Of course “it depends”. I know of examples of papers in which 90-95% of the text is copied – this does warrant a retraction.
On the other hand, if a methodology or a dataset are described (sometimes at great effort in redacting the text for best clarity) in one paper, and then used in a different context (the same methodology in a different situation, or a different methods on the same dataset, then auto-copying should be perfectly OK.
The boundary is related to the “newness” of the rest of the paper.
A separate issue, present in some disciplines is the self-plagiarism between the journal articles and conference papers, again, in my opinion, a more lenient attitude is needed, as the two ways of communication are complementary, allowing the authors to reach different audiences, increasing the reach of the work.
PS: “A separate issue, present in some disciplines is the self-plagiarism between the journal articles and conference papers, again, in my opinion, a more lenient attitude is needed, as the two ways of communication are complementary, allowing the authors to reach different audiences, increasing the reach of the work.”
Disagree. Published in published. Duplicate is duplicate. A submission of a manuscript to any source involves a guarantee of originality. Those who use this logic to try and capture an additional copy are getting an unfair advantage over those who try to be fully original. Moreover, I claim that those who seek such an “excuse”, or exception, are simply getting travel grants for committing self-plagiarism.
A hard-line approach has to be adopted: no leniency, and no exception, although I do agree that methodology can be repeated, if the source is cited and if inverted commas are used, to indicate a direct quote. I do this in my own papers, and the editors never object (they cannot object).
The usual objection to duplication is the assignment of undue credit for duplicate publications (the silly scientific merit badge -hunt). In most fields, there is little credit whatsoever assigned for conference proceedings. It then follows that there is no undue credit received when duplicating conference proceedings (intended for fast communication of new results, not padding your CV) as full journal papers with extended discussion and improved results. This may even improve the scientific record by presenting things more clearly and completely. There is very little reason in my opinion to follow your hardline approach.
Well, like some others said, it depends. It is not possible to copy your own work and pass it off as your own. Some researchers may choose to publish a study as 2 or 3 papers, and one way to do this would be to reference all redundant stuff (methods, etc.) in the other publications to the main publication. However, the introduction should be clearly reframed to point out the difference between past work and current work.
One important thing is that, it is unethical to self- plagiarize the exact same work and publish it in a different journal. That certainly is a retractable offense. Therefore, we need another option that says “depends” instead of a yes or no.
Of course duplication is misconduct and of course there should be cases in which leniency is appropriate (methods sections, conference papers that get converted into journal papers, …). The reason why duplication is misconduct is very simple: Research is supposed to be about generating new knowledge, not publishing existing knowledge again. Otherwise, why don’t we all publish every paper n times, where n>1? There should not even be a debate on this issue.
Agree almost 100% with Geoffrey.
However, even in the conversion of conference papers to journal papers, this should be observed with great caution:
a) the original publication (conference paper) should be fully cited and precise text, data, tables, or figures from that paper should be clearly indicated. Otherwise we enter the realm of selective dishonesty.
b) the editors should be informed about this “duplication” at the moment of submission, so that they can decide it it is worth republishing (editorial independence*).
c) The issues of cronyism and editorial bias need to always be considered**. For example, buddies of editors who wish to republish their work may get more selective freedom that purely random scientists wanting to publish a single copy in the same journal.
* Editorial independence may clash with publisher ethical values.
** Naivety and trust can no longer be accepted as the golden standard in science publishing.
Of course, you are correct. I assumed that it goes without saying that the conference paper should be cited if it gets converted into a journal paper.
I voted yes, but would add the qualifier “Should duplicating your own work WITHOUT ATTRIBUTION be considered misconduct?” Copy-pasting text is never OK; even if it’s just the methods section, why not just cite the original report of the method? The existence of this practice is a symptom of having unlimited space for online supplemental material.
Presenting results/ideas multiple times can be OK as long as you’re appropriately citing previously published material.
If I read a paper that self-cites just on the strength of some boilerplate from the methods section or recycled literature review, I think less of the author – because of excessive self-citation, not because they re-use their own boilerplate.
There is also a legal problem regarding duplication. A lot of us sign off on our copyright to an accepted journal article, and under those circumstances it may or may not be illegal to peruse large chunks of text (even the methods part) without permission from the publisher. Although I don’t think (or hope) any publisher would really go as far as suing original authors, the possibility that it may be inherently illegal may need consideration.
The question glosses over the decisive bit of information: who is meant with ‘self’?
If ‘self’ is really just one author then fine, copy and paste all you want. Depending on what you’re copying from yourself then it only means that either you are using very standardized methods, or that you’re re-cycling the same discussion of your field over and over again. I’d leave it up to the editor reader to decide whether that reflects poorly on your scholarship (“re-cycling the same argument in a variety of outlets to inflate pubcount”), but if it’s your intellectual property please do with it as you please.
If ‘self’ refers at any point to co-authors then it is plain plagiarism. If your PhD student or supervisor hand you a super convincing paragraph and you keep pasting that in every submission without giving them credit for it, you’re stealing their work.
This is not a yes-no question.
Publishing the same data more than once as “new results” is obviously wrong.
Publishing large chunks of exactly the same text repeatedly is wrong, even if its the introduction or a review article (if you’ve got nothing new to say, don’t agree to write another review!).
Re-using the same phrases when introducing a complicated topic is unavoidable.
I hate methods sections that send me on multi-layered paper trail for, say, a protein prep protocol. My call on that one is to (1) refer directly to the paper that initially described the method in detail and (2) repeat the basic overview in every publication that used that protocol (was the a tag, did it involve denaturation and renaturataion, what buffer it ended up in).
Of course, it depends! It is one thing to pass as new previously disseminated portions of our text and I suppose that whether large amounts of text reuse should be labeled misconduct will depend on the specific circumstances and nature of the reuse and the amount (e.g., one short paragraph vs. an entire literature review). And, yes, there should always be flexibility in allowing reuse of segments or even entire methods sections, especially when these are composed of detailed technical descriptions that are laden with unique terms and expressions. Personally, I believe that the clarity of ANY piece of writing can always be improved, and that is most applicable to a methods section, arguably the most important part of an IMRAD paper. However, it is quite another thing to pass as new data (images, tissue samples, etc.) that have already been disseminated. IMO, reused data that are portrayed as new data is the same as fabricated data: They do not exist and THAT is MAJOR research misconduct. Many cases of self-plagiarism of text are largely matters of poor scholarly etiquette and in some instances, a retraction might be warranted. But, when it comes to data, there must be utmost transparency and clarity as to their provenance and readers should never, ever be confused as to whether any portion of them are original or reused in any way.
Many of us work on a problem over several years. Once you arrive at the best way to describe your problem under study, it is always unsatisfying to try to describe it in new terms simply because it is the Intro in a new paper with new experimental findings you wish to present to your colleagues. But that is what we try to do because of 1) the mere thought of self-plagiarism and 2) copyrights. Similarly with methods, you try to be clear, and once you have the best text for clarity, it is self-disappointing to then need to alter sentences to make it new-even if it isn’t. By the way I am sure I repeat verbatim parts of my lecture when giving seminars several times in different venues.
It should definitely be considered misconduct. Copying ones own text results in a researcher getting credit multiple times for the same work. The amount of citations an individual has can be used to gain benefits. Plagiarizing your own work is basically cheating. Maybe you didn’t get into your field because you love writing. But, writing is what you are doing and you can’t phone it in. Also, once your work is published, generally you don’t own the copyright anymore. If you can’t avoid repeating something, just explain it in the paper and let editors know so they can decide ahead of time rather than retracting it later. It’s not hard. Don’t be lazy and don’t cheat.
Three adjectives I would use to describe self-plagiarism, depending on the extent:
laziness, unfair, unethical (and thus misconduct in the latter)
Those that believe the notion of “not being able to steal from oneself”, then simply substitute the term with duplication (partial, or full), to be able to equate it with an issue related to misconduct.
The terminology can be problematic here: the term “self-plagiarism” is especially ambiguous and also potentially emotive. “Redundant/duplicate publication” seems a clearer term for papers where the same scientific results are published by the same authors multiple times. These cases may lead to retraction. The term “text recycling” seems more precise for cases when the authors present new, original scientific results but re-use parts of their own previous texts within sections of the paper where novelty is not crucial e,g, Introduction/Methods. Editors’ views vary on the seriousness of text recycling but these cases rarely lead to retraction.
In general, I say it is misconduct. Setting aside the copyright issues, specifics may be mitigating. Material published elsewhere is “incorporated” into a work via citation. That applies whether it is someone elses work or my own. It may be that I have developed what is, to me, a partiularly satisfying explanation of some tricky point, which I have published elsewhere. For the readers convenience, I may wish to repeat it as part of an introduction (never methods, data, discussion or conclusions!), rather than reference it . In that case a properly cited abbreviation or paraphrase may be appropriate. If that happens more than once or twice, however, than it has crossed from reader convenience to something you believe the reader should be familar with if they are reading further in the area. In that case a reference to the explanation (i.e., “as described in [citation])”) is appropriate.