More information about text formats
In our article reporting a survey of 495 articles, (1) we state “313 (63%; 95%CI 58 to 68) had evidence of plagiarism: 17% (83) had at least four linked copied or more than six individual copied sentences; 19% (96) had three to six copied sentences; and the remainder had one or two copied sentences.”
Flanagin and Ofori-Adjei (2) dispute these findings, stating that “63% of African medical journals are plagiarized to some degree is a gross overestimate.” They provide no data to support their statement, but challenge the definitions of plagiarism we use, stating that the Committee on Publication Ethics (COPE) define plagiarism as “unattributed use of large portions of text and/or data.” They argue that “identical wording with formal source citations” does not constitute plagiarism.
Flanagin and Ofori-Adjei have cited the correct sources for defining plagiarism, but they need to consider more carefully what an editor should expect of authors in terms of proper attribution. Plagiarism is copying of text without giving appropriate credit - which, if it is taken verbatim, “must be enclosed in quotation marks and be accompanied by a citation to indicate its origin.”(3) This is clearly stated in the COPE resources (4, 5) and those of the US Office for Research Integrity (ORI) (3, 6), cited by Flanagin and Ofori-Adjei, and detailed below.
For COPE, their website defines plagiarism as follows: “When someone takes the work of others (data, words or theories) a...
For COPE, their website defines plagiarism as follows: “When someone takes the work of others (data, words or theories) as if they were his/her own and without proper acknowledgement.”(4) COPE provide a link to the position statement on “Responsible research publication: international standards for authors” (5) that was developed as part of the 2nd World Conference on Research Integrity. Section 4.4 of the document states: “Data, text, figures or ideas originated by other researchers should be properly acknowledged and should not be presented as if they were the authors’ own. Original wording taken directly from publications by other researchers should appear in quotation marks with the appropriate citations.”(5)
For ORI, they define plagiarism as “the appropriation of another person's ideas, processes, results, or words without giving appropriate credit” (6) and provide an explanation of the different types of plagiarism in their online module by Miguel Roig, “Avoiding Plagiarism, Self-plagiarism, and Other Questionable Writing Practices: A Guide to Ethical Writing.”(3) Roig explains plagiarism of text in the following way: “When it comes to using others’ word-for-word (i.e., verbatim) text in our writing the universally accepted rule is to enclose that information in quotations and to indicate the specific source of that text.”(3) In addition, the proposed guideline states that “(a)ny verbatim text taken from another source must be enclosed in quotation marks and be accompanied by a citation to indicate its origin.”(3) Furthermore, on patchwork writing, or so-called “mosaic plagiarism”, Roig explains that “(t)aking portions of text from one or more sources, crediting the author/s, but only making ‘cosmetic’ changes to the borrowed material, such as changing one or two words, simply rearranging the order, voice (i.e., active vs. passive) and/or tense of the sentences is NOT paraphrasing.”(3)
Using these definitions, we derive the measures of the extent of plagiarism in terms of the number of completely or substantially copied sentences, excluding those that were put in quotation marks, with or without references to the original source. This is fully and completely described in the article. (1) We provide, in supplementary file 2, a dummy Turnitin report with examples of how we counted copied sentences, and we are confident in the accuracy of our methods and the results obtained. We accept that “some”, “moderate” and “extensive” used without qualification may be interpreted as value judgements, and we are therefore careful throughout to define these categories in terms of the number of plagiarised sentences in the abstract and the research findings.
Flanagin and Ofori-Adjei question why we reported on an analysis stratified by the African Journal Partnership Program. The reason is that we considered the initiative commendable, and likely to show in the data by influencing policies and by implementing such policies. We therefore carried out a subgroup analysis. The fact that it appears not to have influenced the results is thus a finding that we should be allowed to present without being criticised for doing so.
Flanagin and Ofori-Adjei also question our results on journals with policies about plagiarism in relation to journals from their network, stating that three journals, not one, have policies. When we examined all journals in 2016, we were seeking plagiarism policies that had at least one of the three features: a) a definition of plagiarism, b) reference to text matching software or c) consequences of plagiarism described (see Table 2 in Rohwer et al.(1)). The Annals of African Surgery has a clear statement (“All manuscripts that pass the initial review process are screened for plagiarism, which is a ground for immediate rejection”), and this is the journal we report as having a policy in our analysis. The Malawi Medical Journal did not mention the word ‘plagiarism’ in its instructions for authors available on its website in 2016, so at the time of our study did not have a plagiarism policy that was visible to authors. The Ethiopian Journal of Health Sciences in 2016 described scientific misconduct, including a number of practices such as fabrication of data, duplicate publication and plagiarism, with a note on the general consequences of research misconduct, which we did not count as meeting the criteria we were using for a plagiarism policy. Again, we stand by our results.
It may be useful for readers to consider one of the clearest examples of a plagiarism policy that we found in our sample from the South African Medical Journal, which we give below:
“Plagiarism is defined as the use of another's work, words or ideas without attribution or permission, and representation of them as one’s own original work. Plagiarism may take many forms, ranging from major plagiarism (the copy-and-paste of large amounts of text), to minor plagiarism without dishonest intent (e.g. when an author uses parts of an introduction from an earlier paper) and even self-plagiarism (the reuse of significant, identical or near-identical portions of one's own work without citing the original version). The journal subscribes to CrossCheck, an initiative to prevent scholarly and professional plagiarism. All manuscripts submitted to the journal are automatically scanned against the CrossCheck database to verify originality. Manuscripts containing plagiarism will not be considered for publication in the journal. If plagiarism is brought to light after a manuscript has been published, the journal will proceed to conduct a preliminary investigation. Suspected misconduct will be reported to the institutes and funding agencies of the authors concerned. The journal reserves the right to formally retract such manuscripts and publish statements to reference material as plagiarism.”(7)
We would recommend journals adopt similar clearly explained policies that include an explicit definition and explains their policy for detection and actions arising from detection.
Our study aimed to raise awareness about plagiarism of text. We would encourage editors to routinely check manuscripts with text-matching software; to develop and publish criteria for rejection of manuscripts on the basis of plagiarism; and to write editorials to promote these polices and to encourage good writing practice.
1) Rohwer A, Wager E, Young T, Garner P. Plagiarism in research: a survey of African medical journals. BMJ Open 2018;8:e024777.doi:10.1136/bmjopen-2018-024777
2) Flanagin A, Ofori-Adjei D. Biased Study and Misrepresentation of Actual Rates of Plagiarism in African Medical Journals. BMJ Open 27 December 2018 (response to: Rohwer A, Wager E, Young T, Garner P. Plagiarism in research: a survey of African medical journals. BMJ Open 2018;8:e024777.doi:10.1136/bmjopen-2018-024777)
3) Roig M. Avoiding plagiarism, self-plagiarism, and other questionable writing practices: A guide to ethical writing. 2015. Available at: https://ori.hhs.gov/avoiding-plagiarism-self-plagiarism-and-other-questi...
4) Committee on Publication Ethics (COPE). Introduction to Publication Ethics. Publication ethics and misconduct.
5) Wager E & Kleinert S (2011) Responsible research publication: international standards for authors. A position statement developed at the 2nd World Conference on Research Integrity, Singapore, July 22-24, 2010. Chapter 50 in: Mayer T & Steneck N (eds) Promoting Research Integrity in a Global Environment. Imperial College Press / World Scientific Publishing, Singapore (pp 309-16). (ISBN 978-981-4340-97-7). Available at: https://publicationethics.org/node/11184
6) US Department of Health and Human Services. Office of Research Integrity. Definition of Research Misconduct. https://ori.hhs.gov/definition-misconduct
7) The South African Medical Journal. Editorial Policies. http://www.samj.org.za/index.php/samj/about/editorialPolicies#custom-2
We write to express our concern about the prevalence estimate of plagiarism in African medical journals in the study reported by Rohwer et al.(1) The authors’ finding that 63% of African medical journal articles are plagiarized to some degree is a gross overestimate.
The study definitions of “some,” “moderate,” and “extensive” plagiarism are unvalidated and, as the authors admit in the fourth paragraph of their Discussion section, lack inter-rater reliability and precision. Articles were classified as having “some” plagiarism if there were as few as 1-2 sentences that included identical words or sentences in another article by different authors even if the sentences were properly referenced. Numerous publishing organizations, including the Council of Science Editors,(2) the World Association of Medical Editors,(3) and the US Office of Research Integrity,(4) reserve the use of plagiarism for instances when another’s words are used without proper credit or attribution. The authors developed their definition based on suggestions from the Committee on Publication Ethics (COPE), yet even COPE’s Flowchart for managing suspected plagiarism in a submitted manuscript defines plagiarism as “unattributed use of large portions of text and/or data.”(5)
In fairness to the African journals implicated in the study, we request the authors go back to their data, identify all instances in which identical wording with formal source citations were defined as plagiarism, recalcu...
In fairness to the African journals implicated in the study, we request the authors go back to their data, identify all instances in which identical wording with formal source citations were defined as plagiarism, recalculate their plagiarism estimates with instances of proper attribution excluded, and formally correct their article. Only then will the study estimate be credible and useful to the journals included in this study.
Moreover, the singling out of 7 journals from the African Journal Partnership Program (AJPP) is unfair, and some of the comments made about AJPP journals are erroneous. Specifically, the authors claim that only 1 AJPP journal has a stated policy on plagiarism. In fact 3 journals have specific stated policies in their guides for authors,(6-8) and all journals indicate they follow the policies of the International Committee of Medical Journal Editors, which identifies plagiarism as a form of scientific misconduct. When the authors correct their prevalence estimate they should also issue an additional correction for that inaccurate statement about the AJPP journal author guides.
Audits like Rohwers’ et al can be useful for journals considering how to improve their editorial processes and standards. The AJPP journals, with very limited resources, are working to improve their editorial practices and support for and education of authors, and the journal editors are looking for an affordable way to implement software to assess for similarity, duplication, and actual plagiarism as part of their workflows and plans for process improvement. It is unfortunate that the AJPP journals only recently received notification from the study authors with claims of serious plagiarism for 4 articles - at least 6 weeks after publication of the study and after media coverage of the misleading findings. We note that the communication from the study authors to a few AJPP journal editors includes a link to the COPE flowchart that defines clear plagiarism with the term “unattributed.” This further reinforces the problems with the published study and the need to correct the misleading estimate of the prevalence of actual plagiarism and the misrepresentation of AJPP journal policies.
Annette Flanagin, JAMA and the JAMA Network
David Ofori-Adjei, Ghana Medical Journal
Co-Directors, African Journal Partnership Program (AJPP), on behalf of AJPP Journal Editors
1. Rohwer A, Wager E, Young T, et al Plagiarism in research: a survey of African medical journals BMJ Open 2018;8:e024777. doi: 10.1136/bmjopen-2018-024777
2. Council of Science Editors. White Paper on Publication Ethics. 3.1 Description of Research Misconduct. https://www.councilscienceeditors.org/resource-library/editorial-policie...
3. World Association of Medical Editors. Recommendations on Publication Ethics Policies for Medical Journals. Plagiarism. http://wame.org/recommendations-on-publication-ethics-policies-for-medic...
4. US Department of Health and Human Services. Office of Research Integrity. Definition of Research Misconduct. https://ori.hhs.gov/definition-misconduct
5. Committee on Publication Ethics. What to do if you suspect plagiarism.
6. Malawi Medical Journal. Info for Authors. http://www.mmj.mw/?page_id=266
7. Ethiopian Journal of Health Sciences. Information to Authors. https://www.ju.edu.et/ejhs/for-authors
8. Annals of African Surgery. Guide for Authors. https://annalsofafricansurgery.com/guide