Retraction Watch

Tracking retractions as a window into the scientific process

What types of researchers are most likely to recycle text? The answers might surprise you

with 3 comments

Willem Halffman

Serge Horbach

Historians, economists, biochemists, psychologists: Who reuses their own material most often? Does the rate depend on how many authors a paper has, and how far along a researcher is in his or her career? Serge Horbach and Willem Halffman at Radboud University Nijmegen in the Netherlands tried to answer these questions by reviewing more than 900 papers published by researchers based in The Netherlands. And they were surprised by their findings, published last month by Research Policy.

Retraction Watch: How does the amount of text recycling you identified among researchers at Dutch universities (6.1%) compare to what other studies have shown among other groups of researchers?

Willem Halffman and Serge Horbach: Previous studies found varying degrees of text recycling, ranging from 3% to as much as 60%. Ours is, as far as we know, the biggest study on text recycling so far (N=922). Hence we think our figure of 6% is more realistic. However, the precise degree of text recycling depends very much on the threshold used. We used 10% of the text as our detection limit and then stuck to the Dutch guidelines for acceptable text recycling in our manual check. Judging the acceptability of text reuse depends on conventions that may be implicit, or vary between publication cultures of different research fields or even countries. Hence, we think our main finding is not so much the number of 6%, but the variation behind that number. We had expected to find more recycled text in biochemistry, where you could expect formulaic descriptions of highly standardised methods, but there was hardly any. Among Dutch historians we found virtually no text recycling. However, among Dutch psychologists text recycling is more elevated and among economists it is as much as one in seven publications. Text recycling also occurs more often among productive authors, in papers with fewer co-authors, and in journals that do not specify clear rules. This variation is more informative about the origins of text recycling than the overall degree.

RW: You saw the most recycling in economics (which was dramatically higher than other fields)– why do you think that is?

WH and SH: We think this is caused by a combination of a general evaluation culture that values ‘productivity’ and the specific publication culture in economics, which does not strongly disapprove of text recycling. It seems some forms of ‘gaming the reward system’ have become normalized in response to research assessment procedures that over-emphasize and mechanically apply research output indicators. Some Dutch universities distribute research money based on publication output. In our view, this may incentivize researchers in the short run, but is bound to create undesirable behavior in the longer run: Scientists start to conform to the indicators, not to what the indicators were intended to measure. Perhaps it is also the result of a culture that celebrates competition. For example, Dutch economists have created an annual league table of ‘the best economists,’ based on publication indicators. A more detailed study of Dutch economics could further articulate and test these suppositions.

RW: You discuss at length the case of Peter Nijkamp, a Dutch economist who became embroiled in a duplication scandal. Did you include him in your analysis? The story probably had a large effect on Dutch economists, but do you think rates of text recycling changed among Dutch researchers in general following his case?

WH and SH: With our sample it was not possible to detect the effect, mainly because too little time has passed since the Nijkamp case to sample enough articles written after the case. We would need a more extensive study on the publication culture of Dutch economists – and even then it might still be a little bit too soon. Our main message is that text recycling is not an individual problem, but that it also the product of collective behavior and institutions. Whatever the precise processes in the publication or evaluation culture of economists, it is a collective affair and ‘rotten apple’ approaches are highly misleading. For these reasons, we have also not excluded anyone from our sample – including Nijkamp — and there is no reason in our data to do so.

RW: You thought articles with more authors would contain more recycling, but the results suggested otherwise. What are your thoughts on that?

WH and SH: One could assume that more co-authors would mean more people contributing bits of text, which could make it hard to maintain an overview of what text came from where. However, it appears that social control works better among larger groups of co-authors. Alternatively, there could be the temptation of invited papers, extended more often to already famous researchers. This might explain why text recycling occurs especially among highly productive authors with few co-authors. Such a combination of opportunity and low levels of control is a familiar theme in studies of questionable behavior.

RW: In another surprise, you thought junior, less experienced researchers would recycle more text, but again, the results found it was more common among senior scientists. Why do you think that was?  

WH and SH: Questionable research practices are often presented as a problem of inadequate socialization: of young researchers who have not yet fully absorbed the conventions of academic work. Hence commentators suggest integrity training as a solution. Some suggest young researchers have grown up in a cut-and-paste internet culture, or have a harder time writing. But if this were the full story, then you would not find more text recycling among the most productive, established researchers. An alternative interpretation is that some individuals go off the tracks during their research careers, which then needs to be addressed with clear rules and some policing. However, text recycling clearly is not only a matter of just individuals. We now suspect some degree of cynicism might be at work too, as more jaded researchers respond more quickly to publication opportunities. Rules and information about rules can only be part of the answer.

RW: You mention the debate over text recycling, and how many researchers disagree over how much is okay — something we’ve explored, as well. What’s your personal opinion? Is text recycling ever okay, and if so, how much and in what circumstances?

WH and SH: In principle, academics should have intellectual ownership of what they write and should be able to re-use the fruits of their labor as they see fit, as long as they add a simple reference to the original. The problem arises when recycling starts to affect research resources: when recycling burdens peer review with previously checked material, or when recycled text is used to claim funds and rewards. If we stopped our over-reliance on mechanistic performance indicators, a major driver for text recycling would simply disappear. Why would any scientist invest time in copy/pasting text that was published earlier, if that time could also be invested in the excitement of new research? Therefore more, and more specific, rules about what researchers are allowed to do can only be part of the solution. We really need to consider also how reward and publication institutions provoke problematic behavior.

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

October 6th, 2017 at 8:00 am

Comments
  • Ken Pimple October 6, 2017 at 11:55 am

    As I have often said, all you have to do to avoid plagiarism is to add a couple of “quotation marks” (or, in the United Kingdom, ‘single quotes’ [according to http://www.thepunctuationguide.com/british-versus-american-style.html%5D) and a reference.

    Easy as can be.

  • Derek October 9, 2017 at 8:02 pm

    I wonder how much of the problem might be related to empirical papers. Most of my career, I have done applied economic theory where each paper, by its nature is pretty unique. However, recently I have started a series of related empirical papers with a colleague. Already, we are wondering how we are going to make the literature reviews sound different as we will basically be citing the same papers. Add to this, the fact that the data collection descriptions, etc will be very similar. Given I know of many empirical people who do multiple related papers, I can see how they may have a problem, especially given this study’s 10 threshold.

  • Willem Halffman October 11, 2017 at 12:30 pm

    Derek: as long as you do not pretend that the text is original, there is no problem. “As we have described earlier (Derek 2016)…” fixes it (at least according to the principles articulated by the Dutch academy of sciences). Somehow, economists seem to find this more difficult than researchers in other fields – we found that finding more puzzling than the overall rates.

  • Post a comment

    Threaded commenting powered by interconnect/it code.