Retraction Watch

Tracking retractions as a window into the scientific process

How much text recycling is okay?

with 14 comments

Are there a limited number of ways to describe the the background and methods of an experiment? Once something has been written well, and vetted by editors, is it a waste of time to rewrite it ? And if text has been reused, how should that be indicated — if at all?

These are questions we’ve asked before — and are revisiting after reviewing a pair of commentaries published earlier this year in Research Integrity and Peer Review. We’ve certainly seen our fair share of retractions due to duplication (so many we can’t cover them all) — but in one commentary, Cary Moskovitz — the Director of Writing in the Disciplines at the Thompson Writing Program at Duke University — argues that some text recycling — aka, “self-plagiarism” or duplication — is sometimes unavoidable, and, in some situations, even preferable. He told us:

Some amount of recycling is appropriate in some situations. Of course, it is also true that recycling text in some situations is not appropriate.

Examples of the former, he told us, include when scientists are writing background or methodology for the same audience and context as the earlier work  — such as readers of a specialist journal:

In those situations,  it seems disingenuous to suggest that authors can always improve their prose by rewriting it.

The problem, said Moskovitz, is that there are no guidelines for authors to let them know how much recycling will be acceptable to journals. He suggested establishing a “threshold” — say, 5 or 10% — at which the community could conclude that any duplication below this amount “doesn’t really impact the scientific legitimacy or the scientific record.”

If papers contain duplicated material, Moskovitz argued it shouldn’t be flagged by quotation marks: “Using quotation marks for this purpose seems kind of preposterous for a scientific research article.” But authors should, somehow, indicate which text is recycled to editors, so they can adjudicate whether the kind and amount is appropriate, he said:

If the norm in some field is to recycle text for some kinds of things, and readers expect that, then I don’t think that readers need to – and don’t want to – be shown which specific text was recycled.


Miguel Roig, a professor of psychology at St. John’s University who wrote a response to Moskovitz’s commentary, disagrees. Readers deserve to know which text authors have lifted from their previous work, he told us — although authors don’t have to use quotation marks, they can add a phrase like “as I described in my previous article, verbatim,” Roig said:

Let the reader know you’re reusing stuff. What’s the harm?…The English language is really rich. There has to be a creative way of saying ‘I’m reusing this.’

Roig — a member of the board of directors of our parent non-profit organization, who has written extensively about plagiarism and duplication in academic writing — conceded that there are rare times when it’s okay to use similar text:

When you’re dealing with extremely complicated methodology – and really, that’s the only instance in which I think that would be acceptable.

But Roig said he disagrees with the often-used argument that revising text just to avoid duplication could worsen the writing unnecessarily:

Just because something is published doesn’t mean it can’t benefit from revisions…There’s always room for improvement. I don’t care how good a writer you are.

And that’s a practice he tries to follow in his own writing, Roig added:

I go through great lengths to not [reuse text]. However, we all have a fingerprint. And I am sure that if you were to run all of my articles through some plagiarism detection software you would find snippets I would repeat. So it’s not that I’m consciously reusing them. I start all my articles from scratch.

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 daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

Written by Alison McCook

July 6th, 2017 at 8:00 am

  • Erin Jonaitis July 6, 2017 at 10:05 am

    I work on an ongoing longitudinal study, so the recruitment strategy, methods, etc are more or less fixed at this point. Given that, from a reproducibility standpoint, isn’t it better to describe what was done once, unless it changes? Perhaps a pointer is a better practice than reusing text, e.g. “for a full description of our methods, see Lannister et al (2014).”

    My subjective sense is that the “real” reason for complaining about self-plagiarism is that academics are evaluated based on the quantity of their output, so when you see someone do something that artificially inflates their productivity, it’s going to feel like you’re watching someone cross the picket line. If I’m right, I think that’s a reasonable thing to be annoyed about, but I’d rather we put blame for that on the right doorstep.

    • Debora Weber-Wulff July 6, 2017 at 4:09 pm

      But if you are reusing the text, let the READER know where you first published it. Then they can skip it if they already know the older paper. But if you make even the slightest change: make this clear! Don’t just set a “pointer” to the older article, this one may not be in the collection of the library of your readers.

      Don’t focus on you the writer in your papers, but on your readers, make it easy for them to know when you are resuing what.

    • Erin Jonaitis July 6, 2017 at 2:51 pm

      Oh, that’s interesting. I love the idea, but it seems they extend it to the Background, not just Methods. I’m not sure I like that. (Maybe they mean something different by Background – to me that’s the introductory lit review, which I’d think should change as the scientific literature grows.)

  • Lee Rudolph July 6, 2017 at 11:08 am

    When you’re dealing with extremely complicated methodology – and really, that’s the only instance in which I think that would be acceptable.

    Another instance in which it is clearly not only acceptable but universally accepted is in mathematics, where (often) some “extremely complicated” definitions have to be recalled, in detail, from earlier papers if the paper in question is to be understandable at all.

    Although perhaps this isn’t really “another” instance: a case might be made that, in mathematics, the formulation of technical definitions is a necessary and important part of the “methodology”.

  • Mary Kuhner July 6, 2017 at 3:50 pm

    It does not seem difficult to say “The methods used were described in Smith et al. (2015) and are reprinted here verbatim for reference.” Then if the editors don’t like the repetition it’s a very simple fix, and no one can accuse you of plagiarism.
    In most papers I think “The methods used were described in Smith et al. (2015). Briefly, we–” is better; there is not much point publishing exactly the same methods twice. I might feel differently, however, if the first publication was obscure or paywalled, so that readers would have trouble referring to it.
    I would *not* be happy to see a Background section republished verbatim. Background is supposed to position your work in its scientific context. If the context is exactly the same for two papers in a row, I submit that you are probably salami-slicing them.

    • Marya Lieberman July 10, 2017 at 10:48 pm

      Sensible approach.

  • John H Noble Jr July 6, 2017 at 7:54 pm

    I like Roig’s solution . . . ““as I described in my previous article, verbatim, . . . ” I see no reason why author should object to making known the source of a carefully thought out and constructed statement of methods, nor do I see why the reader would not feel happy to be relieved of the burden of searching the literature to see whether the author has “self-plagiarized.”

  • Harold "Skip" Garner July 7, 2017 at 9:14 am

    Regardless of where you fall in this debate, there is more similarity in methods, and to some extent introductions in biomedical manuscripts….analysis of 72,011 articles in PMC. See: Systematic characterizations of text similarity in full text biomedical publications. PLoS One. 2010 Sep 15;5(9):e12704. doi: 10.1371/journal.pone.0012704.

  • Charon Pierson, COPE Secretary July 7, 2017 at 11:57 am

    COPE’s guideline “Text recycling guidelines for editors”, developed in collaboration with BioMed Central, is a useful resources ( for authors as well as editors. These text recycling guidelines support a transparent approach by authors, (i.e., by carefully citing their previous work) and a common-sense approach by editors (i.e., by carefully considering the amount and location of overlapping text). Unfortunately, there is no fixed threshold for text recycling that is considered acceptable in all disciplines. What is important is transparency and avoidance of duplicate publication of data.

  • ICC July 7, 2017 at 10:06 pm

    “it seems disingenuous to suggest that authors can always improve their prose by rewriting it” yet “Using quotation marks for this purpose seems kind of preposterous”.

    Can’t have it both ways. Likewise, I could argue that if the manuscript is so specialized as to require the exact same methods and introduction prose, then the manuscript is incremental research, merely salami slicing, not fit for the scientific literature. Perhaps, in some fields, the bar for “scientific research” is really that low?

    Outside of low detection thresholds and false positives, the only acceptable amount of of text recycling is 0.0. I make this clear to my graduate students, and they know they will not be publishing (or graduating), if they recycle text, or anything else for that matter.

    • hmm July 10, 2017 at 5:18 am

      It is not uncommon for labs to specialize in specific techniques that then are applied to different problems. Of course the total method section will be different but there may be paragraphs that are very similar to other published articles. For methodological details referring to another article (for just that detail) will become very verbose and difficult to read especially if those details have to be sourced from several previous articles. You could paraphrase and change the wording but what is the point of that? Are reworded methodological descriptions of value to the scientific community?

  • ICare July 8, 2017 at 1:51 am

    If you copy paragraphs or even entire sections from a previous paper – even if you mark it by “copied verbatim” or a similar remark – all you’re doing is duplicating text unnecessarily. You can just as well refer to your previous paper. You may argue that you want to make your paper more self-contained, but it’s perfectly reasonable to expect your readers to check references, it’s simply a part of doing research.

  • Donald Osborne July 12, 2017 at 3:45 am

    Plagiarism is never OK. I’m not talking about a short phrase within a sentence but whole sentences or worse. In the Materials and Methods, if the protocol is unchanged, just state that and cite the paper where it comes from. You can then write, “briefly,…” and rewrite the protocol but it should be written anew. If some parts match, it’s OK because it’s been made clear. If you simply cut and paste, that is not OK and journals should catch those before publication. In the rest the of the manuscript, there should be no plagiarism at all. I write all my papers de novo and I’ve put them through Ithenticate and none of them showed any significant text overlaps with previous publications–even in the Materials and Methods. Basically, it’s really unlikely to have “coincidental plagiarism” if you actually write your manuscript from scratch like you’re supposed to. Journals have the responsibility to retract papers that plagiarized, even self-plagiarism, no matter how old the paper is.

  • Post a comment

    Threaded commenting powered by interconnect/it code.