General remarks

If researchers reuse material they that have already published once in new publications, they have, to a certain extent, a research ethics obligation to make this fact known. If they fail to live up to this, there may be a criticisable case of “self-plagiarism” under research ethics. The question, however, is where the line goes.

While fairly solid principles have been established for these phenomena in some scientific fields, doubt may arise in others.

Self-plagiarism may be contrary to responsible research practice without being research misconduct

The Danish Committees on Scientific Dishonesty (DCSD) considered themselves competent to deal with issues of self-plagiarism, and did so frequently. This occurred, for example, in the so-called Klarlund case. In this, a well-known doctor, Bente Klarlund, was found guilty of research misconduct by the DCSD, inter alia, because she had recycled measurement results. The DCSD ruling was subsequently declared invalid by the Eastern High Court. See interviews with the principal persons in the Klarlund case at the UBVA symposium on Research, Ethics and Law, 2014 and the UBVA event on Research Ethics and Legal Security at Folkemødet on Bornholm in 2015.

The successor to DCSD, the Committee on Research Misconduct, cannot, on the other hand, deal with cases of self-plagiarism. This is because section 3 defines misconduct as inter alia encompassing “plagiarism”, which in turn is defined as: “Purloining the ideas, processes, results, text or special concepts of other people without providing proper attribution”. It has been deliberately emphasised at several points in the preparatory work to the Act on Research Misconduct that self-plagiarism may at most contravene responsible research practice, but cannot be considered research misconduct. See the preparatory work for the Act: https://www.retsinformation.dk/eli/ft/201612L00117.

Excerpts from the preparatory work for the Bill on Research Misconduct, etc., of 25 January 2017:

“A minor adjustment of the definition of plagiarism is proposed, in the form of an addition stating that plagiarism covered by the concept of research misconduct relates to the purloining of the ideas, processes, results, text or particular concepts of others. This adjustment is proposed in order to accord with international practice and with the perception of the Danish research environments, in which self-plagiarism, i.e. the reuse of one’s own previously used material without proper attribution, cannot be considered research misconduct. It is remarked in this connection that self-plagiarism may fall under the definition of questionable research practice contained in section 3 (1), no. 5 of the Bill.

[…]

“Self-plagiarism in the form of, for example, re-use of one’s own previously used text passages and the like will […] be treated only as a case of questionable research practice. It is furthermore expected that disputes of authorship will in principle be handled as cases of questionable research practice rather than as research misconduct in the form of plagiarism.”

Research ethics norms regarding self-plagiarism

Although self-plagiarism may not be deemed actual research misconduct, it may still run counter to responsible research practice. Clause 3.1 of the Danish Code of Conduct for Research Integrity states in this connection:

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

2. 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..

…

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

The terms “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, recordings, etc.) that forms the basis of the research.

Data are detailed records of the primary materials that comprise the basis for the analysis that generates the results.”

The Code respects traditions within the various scientific fields

At first glance, when one reads the rules of the Code, one might gain the impression that all forms of recycling of previously published material must always be declared, and that a large number of cases of recycling formerly perceived as natural are now contrary to responsible research practice.

This, however, is not the case. The Code’s general point of departure is that its rules must be understood and administered in a manner that respects the traditions within the various research areas. This is made explicit on p.5 of the Code, which reads:

“The Code embraces all fields of research, while acknowledging the fact that the applicability of the standards for responsible conduct of research may differ between various fields of research. This implies that some recommendations may be more relevant for a specific field of research and at the same time be less applicable to others. The recommendations of the Code should always be understood in accordance with established practices predominant within the individual fields of research.”

This is key to the interpretation of the Code’s rules on self-plagiarism.

Double publication

Double publication is deemed to have occurred if the same manuscript, e.g. a scientific article, is published more than once. In this connection, we speak of primary and secondary publication. There are many examples of legitimate double publication, including:

• Publication of an article in several languages
• A reprint of a previously published scientific article in an anthology that brings together a number of significant scientific contributions in the field in question
• Publication of a contribution to a Festschrift, which often has only a few readers, in a scientific journal
• Parallel publication in an open access repository

Double publication without a clear indication that the manuscript in question has been previously published is generally considered to be a breach of responsible research practice.

Many journals require that submitted manuscripts have not previously been published. Journals that accept secondary publication often lay down special rules for this. In some scientific areas, more general guidelines for secondary publication have been developed. See, for example, the ICMJE guidelines for secondary publication.

Double publication therefore requires a statement that the manuscript has already been published – otherwise, it is considered a breach of responsible research practice. In addition, it follows from the copyright rules that double publication may require the permission of the publisher/journal that first published the article in question, etc. If the manuscript in question has several authors, double publication also requires the consent of all of them – otherwise, it may violate their rights under copyright law.

Overlapping publications and other forms of recycling

Double publication means that you publish the same manuscript several times. In addition, in many scientific fields, it is not uncommon for a degree of overlap to exist between a researcher’s research publications, even if they are not necessarily identical. Several publications may, for example, be based on the same material (research data management) or communicate the same results. In some areas, researchers may even have a strategic interest in publishing their research in ‘least publishable units’: so-called ‘salami publishing’.

In practice, recycling typically falls into two groups, each of which actualises the research ethics issues:

First, there is a question of the extent to which researchers have a duty to mention the reuse of fragments of text and wording that they have already used in previous publications.

Second, there is a question of the extent to which researchers have a duty to mention the reuse of material other than fragments of text and formulations, such as research data, primary material and conclusions.

Reuse of previously published fragments of text and wording

As mentioned, it is generally considered contrary to responsible research practice to republish the same material several times without mentioning that this is what is being done.

On the other hand, researchers who deal with the same subject in several publications frequently reuse the formulations they have used in the past, but without informing the reader that the text is being reused. A legal researcher who for the sixth time in his or her career describes the rules governing “employees’ copyrights” may not, for example, be bound by any research ethics obligation to invent completely new formulations in each of the six publications, but must be allowed to repeat himself or herself to a reasonable extent, without it having to be made clear to the reader exactly which words and sentences are being repeated. This is quite common in the legal field, where we often describe the same rules in many different places in the course of our careers. The practice is not contrary to responsible research practice, because p.5 of the Danish Code of Conduct for Research Integrity states that the rules of the Code respect established scientific traditions. The question, however, is what extent the text recycling can have, and whether there comes a point at which it is so extensive that the recycling should be mentioned. Where does the line go? At half a page? Two pages? Five pages? Ten pages? Twenty pages?

The Practice Committee of the University of Copenhagen issued a detailed statement on the matter in its ruling of 29 March 2016. In the case, former Minister of Research Esben Lunde Larsen was accused of, amongst other things, self-plagiarism because at certain points in his PhD dissertation he reused formulations that he had already used in previous publications.

The Practice Committee stated:

“The Practice Committee considers that the rules on good scientific practice must generally be interpreted in the light of the purpose of the rules, which in this context is, in particular, to prevent the reader from being misled about one’s own efforts and results. This means that even when reproducing relatively extensive parts of one’s own previously published text, one usually does not intend to mislead others about one’s own efforts and results, and so the objective condition (see p. 6) for a breach of good scientific practice to be deemed to have taken place will not therefore be met. The condition will probably only be met if the reproduction of one’s own previous text without source attribution and quotation marks is to a certain extent systematic, for example if the reproduction is likely to give the impression that the researcher’s body of original scientific production is significantly larger than is actually the case. If the undisclosed reuse of one’s own text is included in an examination/dissertation that is part of, or leads to, an academic degree, depending on the circumstances, misrepresentation may arise concerning the scope of the original work. As mentioned above, this should give rise to independent deliberations on double merit, which, however, lie outside the competence of the Committee to decide.

The use of text written by others in a research work must […] be judged significantly more severely than the reuse of one’s own text. […]”

One of the things for which Esben Lunde Larsen was criticised was his failure to mention that his PhD dissertation had the same subject as his thesis. In this connection, the Practice Committee stated:

“You have stated in your comments that you regret failing to account for the use that you made in the PhD dissertation of the initial work on the topic that you had performed in your thesis.

The Practice Committee notes your regret, as the Committee remarks that it is not uncommon for a thesis to be utilised to a greater or lesser degree in a PhD dissertation. In such cases, the Practice Committee considers that this should be stated. The requirements in this respect this will depend on the extent to which the thesis is utilised. In the case of text relating to generally known matters, as is the case here, referring to the thesis at every coincidence of text would usually be superfluous and unnatural. In such situations a reference in a single footnote, in the preface or otherwise, will suffice as appropriate. Given the relatively limited use of the thesis in both quantitative and qualitative terms, the Committee finds that the lack of reference to the thesis is not in itself contrary to good scientific practice.”

The ruling encompasses the following:

• Covert reuse of even relatively large pieces of text is acceptable provided it does not give a misleading impression of the author’s scientific efforts
• Covert reuse of one’s own text may be contrary to good scientific practice if done in a systematic manner that misleads readers about the scope of the author’s scientific efforts
• The character of the work in which the text is recycled also plays a role. It is, for example, significant whether it is a work that leads to an academic degree. If such a work is based on the author’s thesis, this must be mentioned (possibly in the preface).

The principles from the ruling have subsequently been repeated by the Practice Committee of the University of Copenhagen in later cases, including 31/2018 and 34/2018.

The Esben Lunde Larsen ruling is not formally binding outside the University of Copenhagen or on the Committee on Research Misconduct. In the view of the UBVA, however, the principles expressed in the ruling are reasonable and balanced. We therefore recommend that the principles contained in the ruling are also applied outside the University of Copenhagen.

Views similar to the criteria in the Esben Lunde Larsen case are also stated in a memorandum on double publication and self-plagiarism published by the Practice Committee of Aarhus University on 24 May 2016. This states, inter alia, that within some of the social sciences and humanities that do not have the character of exact sciences:

“… it regularly occurs – and not in contravention of generally accepted scientific norms – that an author who has previously analysed, for example, a legal provision or a work of fiction (and published the results of the analysis), will subsequently, in another publication, present the same interpretation or assessment founded on the same source foundation without referring to the fact that he or she has previously presented and argued for the same ‘result’ on the same basis – perhaps even using the same wording in some passages.” No one has ever dreamed of describing this in general as scientifically questionable – on the contrary, depending on the circumstances in peer circles, it might be perceived as superfluous or even unnecessarily assertive if someone were to refer to themselves repeatedly in areas where the argument counts for more than references to previous exact studies. Actual text recycling (self-plagiarism), in which an author copy-pastes from his or her own work without disclosing this fact may also here, depending on the circumstances – but not in general – be characterised as “questionable” in the above sense. “

In many areas, it is common practice for researchers to repeat their own texts in sections where they describe their method. This is not considered contrary to responsible research practice. See, for example, the ruling of 12 November 2018 of the Committee on Research Misconduct. The Committee, which may not address issues of self-plagiarism as such, nonetheless stated that:

“The text coincidence between Publication III and Appendix C mainly concerns the methodological section of Publication III. The Committee finds that this text coincidence does not go beyond what may be expected when the same experiments are reported in publication III and Appendix c. A certain text coincidence between methodological sections in articles reporting the same experiments is widely accepted in the scientific world and cannot generally be deemed self-plagiarism. Therefore, in the assessment of the Committee, this part of the case does not involve plagiarism […]”

Reuse of previously published research data, results, etc.

The previous section deals with whether it is contrary to responsible research practice for researchers to repeat formulations and text passages without mentioning that these have been reused.

However, researchers may also reuse materials other than actual text, such as interpretations, conclusions, research data and primary material. The question is to what extent this is contrary to responsible research practice.

Clause 3.1 of the Code, cited above, may leave the impression that all forms of reuse are problematic unless the reuse is clearly indicated. However, this is not a realistic interpretation. As stated on p. 5 of the Code, its rules in this area must also respect established traditions in the various scientific fields, and in many areas the reuse of research data, conclusions, etc. is common with this being mentioned.

Example 1: A legal researcher works with a number of legal sources (including legal texts and judgments), which would be considered “primary material” in the sense of the guidelines. A legal researcher will often make records of, for example, the judgments that exist on a specific issue, and then include these in the researcher’s legal analyses. It is certainly not a requirement in legal research that such “reuse” of the same sources of law (possibly based on previous records of these) should be “clearly stated” in the sense required by clause iii of the guidelines.

Example 2: A legal researcher who is writing a textbook touches on the topic of “employees’ copyrights”, which is a topic that he or she has dealt with in five previous publications. The researcher maintains the same point of view in the textbook as in the previous publications, and thereby repeats the conclusions reached. There is no established practice in the field of jurisprudence requiring researchers to explicitly list the publications in which they advance the same point of view. Clearly, therefore, a researcher’s failure to mention in the textbook that he or she has previously reached the same conclusions, or where this has occurred, cannot be considered contrary to responsible research practice.

In the “wet” areas of science, on the other hand – including medical science – the reuse of research data and primary material is traditionally viewed rather more strictly. See, for example, clause 3 of Aarhus University’s “Standards for responsible research practice at the Faculty of Health”:

“Covert redundant publication, i.e. identical or nearly identical publications, including translated into other languages, is never acceptable. On the other hand, secondary publication (ex.: English language article subsequently published in Danish in the Journal of the Danish Medical Association, or vice versa) is permitted when undertaken openly. Use of the same data or subsets hereof in different publications does not constitute double publication, provided any data overlap between a previous and a current work is clearly stated to full disclosure with regard to reviewers and readers.”

The same view lay behind the DCSD ruling in the Klarlund case, in which the DCSD found Bente Klarlund guilty of misconduct because she had used the same measurements in several articles without mentioning this. The DCSD remarked:

“The Committee notes, first, that the requirement for empirical science with biological material to divulge the use of the same material in several scientific studies derives from the fact that the results and conclusions of the individual studies are based on random sampling. Based on these samples, the purpose of the studies is to describe correlations of a general nature, i.e. correlations applicable to all individuals with the characteristics included in the studies. Since individual differences in reaction patterns are common in biological materials, the value and generalisability of study results depends to a large extent on the results of other samples which show results of a similar or supportive nature. If it is not revealed that the survey results are based on the same sample, the reader may be led to believe that these are different samples. Thus, conclusions based on the same biological sample material will be assigned more weight than they are entitled to, and erroneous conclusions are likely to be drawn.

Second, the Committee notes that material drawn from experiments performed on human beings may, of course, be used in several studies and scientific articles. In this connection, the Committee is of the opinion that there may be very good reasons to exploit to the maximum material collected at not insignificant inconvenience to the experimental subject. In the performance of responsible research practice, the researcher must always provide the reader with all the information necessary for an evaluation of the results of the articles in question on a sufficiently informed basis. In many cases, this will require that the reader is explicitly informed about the origin of the experimental material and about any previous use of the same material. This can be done in the form of one or two relevant references or a reference to the known name of a project, a biobank, etc.

On the basis of the arguments of the Complainant and the Respondent, including the Respondent’s reference to a letter from 70 Danish researchers and a number of scientific works in which the same experimental subjects are used in different articles, the Committee can state that no sufficiently unambiguous practice exists for providing general information on the origin of the experimental material, including previous use in articles based on the same experimental subjects. The Committee has therefore conducted an overall review of the articles to which the Respondent refers, cf. section 7. However, as stated above, the Committee considers that it follows from good scientific practice that the specific circumstances of a given scientific work may require that information on the origin of the experimental material is expressly stated in the article in question. This applies, inter alia, in the following instances: i. When the results from two or more articles based on the same subjects are compared and used to support the conclusions of the individual articles. If information on the origin of the experimental material is not provided in such cases, information is withheld from the reader concerning the coherence of the study results in the different articles and the interdependence of the conclusions. Ultimately, this can lead to results from the same study being used in different publications to support the conclusions of the articles, without the reader’s being aware that they concern the same experiments and subjects. In this way, a single study could be used to substantiate a large number of conclusions via subsequent articles, with the conclusions thereby appearing to be independent. ii. When a selection of experimental subjects is kept concealed from the reader through a lack of information about the origin of the experimental material. The reader is thereby denied the opportunity to assess the results of the article on a sufficiently informed basis.”

See the DCSD ruling of 28 August 2014, p. 78 ff.

However, the High Court ruled that Bente Klarlund was not guilty of misconduct. You can read the High Court’s judgment here.

The aforementioned memorandum of the Practice Committee of Aarhus University states:

“Reuse of one’s own research data (in new publications) may […] be both acceptable and desirable, depending on the circumstances, especially when the material is completely or partially reused/re-analysed in order to generate new results and new knowledge. This applies even if the new results to a certain (limited) extent overlap with those previously published […]. The requirement is simply […] that this must be declared openly and honestly, so that there is complete transparency about the fact, wherever this is relevant. Here the Danish Code appears to go very far when, in its definition of “primary material” and “data”, […] which “should be clearly stated”, it also includes personal notes, literature consulted and the like. In the opinion of the Practice Committee, a limitation of relevance and proportionality must be read into this, based on the special characteristics of the individual research area […].

 […]

Although the Danish Code is generally intended to apply in all research areas, it is undoubtedly the case that these standards should not be implemented in exactly the same way in all subject areas – as the Code itself also states (p. 5). […] Since standards, unlike fixed rules, are characterised by being shaped by their environment and context, the different conditions and traditions of the various subject areas both may and should be taken into account.

In the natural sciences and health sciences, where research is largely based on natural laws/natural constants and exact empirical laws, the demand for openness and transparency will typically require that, in the case of repeated use in several studies of the same raw data or the same ‘primary material’, the data connection must be made clear, including in particular any possible overlap in methodology or results between the ‘previous’ and the ‘new’ work. Here, the new work, if it contains or builds upon previous studies, should clearly distinguish between, on the one hand, the previously obtained results and, on the other, the new way of processing the raw data used and the new benefits thereof. […]

Within some of the social sciences and humanities that do not have the character of exact sciences, the situation is different. Here it regularly occurs – and not in contravention of generally accepted scientific norms – that an author who has previously analysed, for example, a legal provision or a work of fiction (and published the results of the analysis), will subsequently, in another publication, present the same interpretation or assessment on the same source foundation without referring to the fact that he or she has previously presented and argued for the same ‘result’ on the same foundation – perhaps even using the same wording in some passages. No one has ever dreamed of describing this in general as scientifically questionable – on the contrary, depending on the circumstances in peer circles, it might be perceived as superfluous or even unnecessarily assertive if someone were to refer to themselves again and again in areas where it is more the argument that counts, rather than references to previous exact studies. Actual text recycling (self-plagiarism), in which an author copy-pastes from himself or herself without disclosing this fact may also here, according to the circumstances – but not in general – be characterised as “questionable” in the above sense. “

Dissertations leading to an academic degree

In the case of dissertations that lead to an academic degree, e.g. PhD dissertations, it should be explicitly stated if the dissertation continues the work contained in the author’s thesis. This is made clear in the ruling of the Practice Committee of the University of Copenhagen in the Esben Lunde Larsen case, quoted above. It is also assumed in the memorandum from the Practice Committee of Aarhus University. The author may, for example, mention in the preface that the dissertation continues the studies that the author began in his or her thesis.

If a dissertation that leads to an academic degree is based on the author’s previously published works, more precise references may be required. That at any rate is the view expressed in the memorandum from the Practice Committee of Aarhus University, which states that:

“A special issue arises in cases where material that has previously been assessed in connection with the award of a Master’s or PhD degree is to some extent reused in a later dissertation that is submitted for the purpose of acquiring a PhD. The issue is accentuated in those subjects where it is the rule rather than the exception that PhD dissertations are written on essentially the same subject as the student’s thesis, or where the curriculum itself (as is generally the case with the 4 + 4 programmes) lends itself to this.

If the previous work in question has been published, the new dissertation should contain precise references to it in accordance with the guidelines that […] apply to all scientific production. Regardless of whether the person’s previously assessed work has been published or not, it should always be possible to demand of a dissertation intended to lead to the award of a scientific degree and thus, amongst other things, must possess a certain value in terms of new knowledge, that any reuse of previously assessed material is made clear to the readers/assessors. If the previously assessed work is unpublished, this can usually be done in general terms by, for example, simply mentioning that this or that section ‘builds upon’ the author’s previous thesis (or other work for which the author has previously been credited). If the previously assessed work has been published, clearer and more precise references to this must be required.

Overall, the obligation to provide clear and reliable information on the possible inclusion of one’s own previous work must thus generally be assumed to be sharpened in relation to other (‘ordinary’) scientific publication when a manuscript is submitted for assessment with a view to obtaining an academic degree. […].”

Copyright restrictions on reuse of one’s own work

Research publications are normally protected by copyright as “literary works”, and this copyright belongs to the researcher or researchers who have produced the work, cf. sections 1 and 6 of the Copyright Act.

Copyright implies, amongst other things, that the researcher’s consent is required if you wish to make use of the work by producing copies of it and making it available to the public (e.g. by publishing it). Copyright protects not just the work as a whole, but also any part of the work that expresses the author’s personal creative efforts. It may, for example, be a violation of the researchers’ copyright if, without their consent, others reproduce specific extracts from a book (other than as a legal citation) or write a book with the same original structure. On the other hand, copyright does not, for example, protect abstract ideas, theories, topics, methods, points, etc. Amongst other things, this means that writing a book on the same subject, based on the same fundamental idea and possibly using the same methods, does not constitute a violation of the researchers’ copyright. In these cases, however, it may be questionable from a research ethics point of view to do so without properly accrediting the researchers’ book.

You can read more about plagiarism and copyright (in Danish) in the UBVA’s information? book on copyright for beginners (Ophavsret for begyndere).

Joint copyright to a previous work

When several researchers have jointly produced a research publication that is protected by copyright as a literary work, they possess joint copyright to that work. This means that, as a starting point, the consent of all authors (researchers) will be required if someone wishes to make use of the work in a manner that requires the author’s consent, cf. the remarks above. This also applies if, for example, one researcher wishes to reuse copyrighted parts of a joint research publication. If this is done without consent, the researcher will thereby infringe the copyright of the other authors of the research publication in question.

When joint copyright to a previous work has been transferred to a publisher, etc.

Most research publications are published by publishers of books or journals, and in this connection, the researchers transfer the copyright for the publication in question to the publisher, in whole or in part. The researcher’s position in this situation will always depend on a concrete assessment of the individual publication agreement, but in many cases the researcher’s publication of the work will involve, inter alia, the following:

1. The researcher may not, without the publisher’s consent, publish the research publication (work) in question in any other way, including via another publisher or in an open access repository.
2. The researcher may not, without the publisher’s consent, reuse in other contexts parts of the research publication (the work) which in themselves are an expression of the researcher’s personal creative efforts (cf. the comments above), including in other research publications, teaching or communication. This is a consequence of the researcher’s having transferred his or her copyright to the publisher.
3. The researcher has a duty of loyalty towards the publisher who has published the book or article. In the case of book publications, this often means that the researcher may not, without the publisher’s consent, publish other books that directly compete with the work that the publisher has published, or that undermine the publisher’s opportunities to sell it.

A researcher who has had a research publication published by Publisher X will thus in a number of instances be acting unlawfully in relation to Publisher X if, without the consent of Publisher X, the researcher republishes the same publication via another publisher, e.g. as a book or as an article in a journal. If Publisher X consents to the republishing, however, the researcher will not be acting unlawfully in relation to Publisher X.

In some cases, research publications are published with no detailed agreement on the terms. For example, cases may arise in which no actual agreement has been entered into in relation to publication in a journal. In such cases, an assessment must be made of what the parties have mutually assumed regarding the publication. In some cases, the journal may have fixed terms for publication on a website of which the researcher may have been unaware when submitting the article – in many cases, however, these terms may nonetheless be accorded significance. However, this will always depend on a concrete assessment.

With regard to PhD dissertations submitted for assessment to a Danish research institution, the institution is obliged under the PhD Executive Order to make the dissertation publicly available in connection with the PhD defence. It is the UBVA’s assessment that this does not imply any transfer of copyright to the institution beyond what is necessary to fulfil the obligation under the PhD Executive Order.

Summary

Copyright may in some cases mean that researchers may only reuse parts of their previous research publications with the consent of the publisher who originally published the research publication in question, and/or with the consent of the other researchers who contributed to the creation of the publication in question.