Overlapping publications and self-plagiarism

 

General information

Researchers’ re-use of (parts of) their own previous research publications or the material on which these publications are based gives rise to a number of legal and research ethical considerations. In this connection different types of cases can be identified:

  • Publication of the same script more than once (actual duplicate publication)
  • Publication of several scientific products based on (partially) the same material or disseminating (partially) the same results (overlapping publications)
  • Re-use of own text, structure, ideas, interpretations, etc. (self-plagiarism)

While rather firm principles for these phenomena have been formed within some scientific areas, they give rise to doubt in other scientific areas.

The Danish Code of Conduct for Research Integrity (2014) deals with certain aspects of these issues, see details below. In addition, the Committee for Responsible Conduct of Research at Aarhus University has published a note to that effect in 2016, see Note on duplicate publication and self-plagiarism.

Actual duplicate publication

Actual duplicate publication arises if the same manuscript, for example, for a scientific paper, is published more than once. In that case, we are dealing with primary publication and secondary publications. There are many examples of legitimate duplicate publishing, including

  • Publication of an article in several languages
  • Reprint of a previously published scientific article in an anthology that brings together a number of significant scientific contributions in the current area
  • Publishing a contribution to a festschrift, which often have few readers, in a scientific journal
  • Parallel Publishing in an open access repository

Actual duplicate publishing without a clear indication that the manuscript in question has been published previously is generally considered to constitute a breach of good scientific practice. Many journals require that submitted manuscripts have not been previously published (or made public). If they accept secondary publications, there are often specific rules governing these. Some scientific areas have developed more general guidelines for secondary publications, see for example the ICMJE recommendations for secondary publications.

Actual duplicate publishing often requires permission from the publisher/journal that first published the article, etc. If the manuscript has multiple authors, actual duplicate publication will also require the consent of all the authors. This is due to copyright rules, as discussed below.

Overlapping publications

A certain overlap between a scientist’s research publications is not unusual in many scientific fields. For example, several publications may be based on the same material (research data) or disseminate (partly) the same results. In some areas researchers may even be said to have a strategic interest in publishing their research in least publishable units, so-called salami publishing.

General standards

Based on the basic research ethical standards of honesty and transparency (see the article on Good Scientific Practice), the Danish Code of Integrity in Research (2014) sets out the following general guidelines for overlapping publications (bold type added):

i. Research results should be published in an honest, transparent, and accurate manner.

ii. Publishing the same results in more than one publication should only occur under particular, clearly explained and fully disclosed circumstances.

iii. Recycling or re-use of primary materials, data, interpretations or results should be clearly disclosed.

…

v. When using one’s own work … in a publication, appropriate and accurate references to such work should be provided.

Re ii: Re-using results

According to the guidelines, it must always be clear if results are being reused. However, not at all scientific areas have a clear definition of “results”.

Re iii: Re-using primary material and data

The concepts “primary material” and “data” are defined as follows in the Code (page 9):

Primary material is any material (e.g. biological material, notes, interviews, texts and literature, digital raw data, registrations, etc.) that forms the basis for the research.

Data are detailed records of the primary material which forms the basis for the analysis that leads to results.

These definitions may give rise to doubts in some scientific fields.

Section iii of the guidelines does not reflect the established practice in all scientific fields.

Example: A researcher in legal science working with a number of sources of law (including legal texts and rulings), which are considered to be the “primary material” in the sense of the guidelines. Such a researcher will often make records of, for example, the rulings that exist on a particular issue, which is then included in the researcher’s legal analyses. It is certainly not a requirement for legal research that when “re-using” the same sources of law (possibly based on past records above) you “clearly account for this” as stated in the section iii of the guidelines.  

Re v: Use of own work

Section v of the guidelines gives rise to considerable doubt. This guideline may give the impression that you must be pedantic and obsessive about referring to your own previous work (regardless of whether it has been published as a research publication or e.g. as teaching material), even if it reproduces only a single sentence or a paragraph describing generalities. This does certainly not correspond to the established practice in many scientific fields.

The Practice Committee at the University of Copenhagen addressed this question in its decision of 29 March 2016 regarding the case of Esben Lunde Larsen’s PhD thesis.  In the decision the Practice Committee among other things states the following:

“The Practice Committee finds that the rules of good scientific practice in general must be interpreted in the light of the purpose of the rules, in this context in particular to guarantee against misrepresentation of own efforts and results. This means that even reproduction of relatively comprehensive parts of own previously published text will normally not constitute a misrepresentation of own performance and results, and the objective condition (for details see p. 6) for breach of good scientific practice will thus not be satisfied. The condition will probably only be met if the reproduction of own previous text without source references and quotes has a certain systematic character, for example, if the reproduction is likely to give the impression that the person has a significantly larger original scientific production than it is in reality. If the undisclosed re-use of own text is part of an exam/thesis, which is part of/leading to an academic degree, this may, dependent on the circumstances, constitute a misrepresentation of the scope of the original work. As mentioned above, this should give rise to independent considerations about double credit transfer, which however is outside the remit of this committee.”

The Practice Committee noted among other things the following regarding the use of a previous master’s thesis in a PhD dissertation:

“… the Committee notes that it is not uncommon practice that a master’s thesis to a major or minor degree is put to use in a PhD dissertation. In such cases the Practice Committee finds that this should be stated. These requirements depend on the extent to which the thesis is put to use. When it comes to text relating to common knowledge, as is the case here, it will usually be obsessive and unnatural to refer to the thesis every time there is a certain text overlap. In such situations a reference in a single footnote, in the foreword or otherwise, will suffice, according to the circumstances.”

In the decision Practice Committee also stated that “it may be natural” – and that in the case it was in order – that there had been a minor re-use in the PhD dissertation of text passages from own texts for dissemination of the PhD project, and that the draft for PhD dissertation during the PhD project has been used as a basis for research dissemination. Finally the Practice Committee found no reason to criticise that texts from unpublished teaching material that had not been part of rated academic work might be regarded as self-plagiarism.

The subject is also referred to at the end of the above-mentioned Note on duplicate publication and self-plagiarism from Aarhus University’s Practice Committee.

Use of own work can also give rise to legal considerations, which is discussed further below.

Interpretation of the Code of Conduct – and the need for a definition of “best practice”

Here (as elsewhere) it is important to remember that the recommendations of the Code of Conduct must always be understood in accordance with the “established practice in the relevant field”, cf page 5 of the Code:

“The Code of Conduct embraces all research areas while recognising that the standards for responsible research practices must be implemented in different ways within the respective fields of research. This means that some recommendations may be more relevant to a particular field of research while being less useful in other areas. The recommendations in the Code of Conduct should be read in accordance with the established practices in each research area.”

In this connection, the guidelines may give rise to at least two important challenges: First, the “established practice” in a field of research is not necessarily unique. Secondly, it is conceivable that the “established practice” is based on historical convention, which does not reflect the ethical requirements that should be made to research publications today.

Both these challenges may lead to criticism of the researchers for breach of good scientific practice, even if they themselves believe that they just follow the research convention in the area concerned. Therefore we recommended that the individual subject areas discuss and define “best practice”, which then (especially) can serve as a guide for junior researchers.

Special standards for the “wet” sciences

The question of overlapping publications has in particular given rise to challenges in the “wet sciences”, which is why numerous guidelines have been established for this. DCSD’s guides to good scientific practice (2009) which the DCSD now considers a historical document, states the following (bold type added):

 “Hidden duplicate publication, i.e. identical or almost identical publications, possibly in translation, may be as discussed in the Vancouver rules not take place, but secondary publication, for example, in two languages ​​(English and a minor language such as Danish) or to different forums within the same language area, is allowed when done in openness to the editors and according to established rules. The use of the same data or subsets thereof, in different contexts and in different representations does not necessarily constitute duplicate publication, provided that the data correlation between a new work and a previous work has been disclosed to both editors and readers. In this context it should be emphasised that any new publication (a so-called salami publication) should include substantial new information, more than 50% is recommended. In that way it is possible to distinguish the citation of own previous results from a new and original way to process raw data. However, there is no international consensus on the rules in this area as yet. It should be noted that the National Library of Medicine only includes secondary publications for primary work that has already been registered in MEDLINE, provided it is clearly stated in the title that it is a secondary publication.”

For health sciences particular reference is made to

Copyright limitations on re-use of own work (self-plagiarism)

Research publications are usually protected by copyright as “literary works”, and this copyright rests with the researchers who have created the work cf sections 1 and 6 of the Danish Copyright Act.

Copyright involves, among other things, that the researchers’ consent is required in order to dispose of the work by reproducing copies of it and making it available to the public (e.g. by publishing or making it publicly available). Copyright does not just protect the work as a whole, but also any part of the work, which is characterised by the author’s personal creative efforts. For example, it may constitute an infringement of the researchers’ copyright if others without their consent reproduce specific excerpts of a book (assuming it is not a legal quote) or write a book with the same original structure/make-up. By contrast, copyright does not protect abstract ideas, theories, subjects, methods, main points, etc.  Thus writing a book on the same subject, which is based on the same basic idea and based on same methods and so on, does not constitute an infringement of the researchers’ copyright.  In these cases, doing this may be wrong in terms of research ethics, if done without proper crediting of the researchers’ book.

You can read more about this in Morten Rosenmeier’s guide to copyright, “Ophavsret for begyndere – en bog til ikke-jurister”: The requirements for copyright protection of literary works are described on page 30 ff. The subject matter of the copyright protection is discussed in chapter 5, and the citation rules are mentioned on page 115 ff.  Finally, chapter 8 addresses the question of when something constitutes an infringement, and how copyright and research ethics complement each other in this area.

Joint copyright to earlier work

When several researchers have collectively produced a research publication which is protected by copyright as a literary work, they have joint copyright to the work. This means that it basically requires the consent of all authors (researchers), if someone wants to control the work in a way that requires the consent of the author, cf. the comments above. This also applies where, for example, one researcher wants to reuse copyrighted parts of the joint research publication. If this is done without consent, the researcher will thus be violating the other authors’ copyright to the research publication in question.

Read more about these so-called “joint works” in Morten Rosenmeier’s above-mentioned book, page 62 ff.

When copyright to earlier work is transferred to a publisher etc.

Most research publications are published by book or journal publishers, and in this context researchers assign their copyright to the publication in question to the publisher in whole or in part. The researcher’s position in this respect depends on the specific assessment of the individual publishing agreement, but in many cases the researcher’s publication of the work entails the following:

  1. That the researcher may not publish or make public the research publication (work) in question in any other way, including by another publisher or in an open access repository, without the publisher’s consent.
  2. That the researcher may not reuse parts of the research publication (the work), which in itself reflects the researcher’s personal creative efforts (cf the notes above), in other contexts, including in other research publications, teaching and dissemination, without the consent of the publishers. This is a consequence of the researcher’s transfer of his or her copyright to the publishers.
  3. That the researcher has a duty of loyalty to the publishers who have published the book/article. For book publications this often entails that the researcher cannot publish other books that directly competes with the work which the publisher has released and which undermines the publishers’ freedom to sell it, without the consent of the publishers.

A researcher who has authored a research publication at Publisher X, will in many cases act unlawfully in relation to Publisher X, if the researcher without the consent of Publisher X republishes the same publication at a different publishing house, for example as a book or an article in a journal. If Publisher X consents to the reissue, the researcher does not act unlawfully in relation to Publisher X.

In some cases research publications are published or made public without any further agreement concerning the terms. For example when no actual publishing agreement is entered with a journal. In such cases the assumptions of the parties about the publication must be taken into consideration. In some cases, the journal has set out the terms for publishing on a website which the researcher may not have been aware of at the time of submitting the article – in many cases these terms, however, may have significance anyway. However, this will always be subject a specific evaluation.

When it comes to PhD theses submitted for assessment at a Danish research institution, the institution is, according to the Danish PhD Order, required to ensure that the thesis is made available to the public in connection with the PhD defence. In UBVA’s opinion this does not involve any transfer of copyright to the institution beyond what is necessary to fulfil the obligations under the PhD Order.

Summary

Sometimes copyright may be taken to mean that researchers can reuse parts of their previous research publications only with the consent of the publishers who have published the research publication, and/or with the consent of the other researchers who have helped produce the publication .