Accidental similarity

The University of Oxford defines plagiarism as: “presenting someone else's work or ideas as your own, with or without their consent, by incorporating it into your work without full acknowledgement.”[1] There are various types of plagiarism, ranging from subtle (e.g. presenting others’ ideas) to the not-so-subtle (e.g. presenting others’ work), but in essence it means passing off someone else’s work or ideas as your own. Most authors will agree that this is wrong, however, understanding the exact nature of plagiarism can often be tricky in environments where core research method knowledge lacks. With the internet of today providing so many more sources to cite from compared to years gone by, keeping track of similarity without a good similarity checker is often impossible.

The vast majority of plagiarism cases I have had to deal with tend to be due to ignorance, rather than purpose (as a result I prefer the term similarity where copying was unintentional and plagiarism where there was clear signs of intent). Pre-publication, this requires a simple notification of the similarity score to the authors so that they can make amendments and correct identified sections of similarity. Within our journal flow, similarity is only checked once on initial submission. I suspect many other journals will reject a paper with high similarity, however, we have found that in most cases authors simply weren’t aware that what they did could be considered plagiarism. Resultantly, we take a more supportive approach and rarely reject a paper based on an initially high similarity score.

In the case below, the first submitted manuscript did not have a high similarity score. The two paragraphs, taken from another paper published in our journal a few years prior, were subsequently added to the discussion section. As similarity is not checked on resubmission, this was not picked up in the editorial workflow and the manuscript made it all the way to publication. The error was first picked up by the first author of the plagiarised paper. We followed the Committee on Publication Ethics (COPE) guidelines to determine the action required.[2] It became clear fairly quickly that ignorance from the lead author (of the subsequent paper) was the main driver behind the error. I liaised with the authors of both papers and it was agreed that a retraction would not be in anyone’s best interest. We decided on an acknowledgment of fault and apology which was approved by both authors.

At the African journal of Emergency Medicine, we acknowledge that emergency medicine is a new and evolving specialty on the continent, that many persons wishing to contribute to the local knowledge economy are clinician researchers and that most have very little to no formal research training. We have always taken a mentoring approach as evidenced by our Author Assist programme. We felt that retracting this paper would have been wrong, given our assessment of similarity, not plagiarism. As a result we elected mediation between the authors of the two papers, which was a much more constructive learning process, not just for them but also for the editors.

I sought the perspectives of those most involved in this case of similarity to better understand their experiences and their recommendations moving forward. Their words, with minor editing for clarity, follow.

The author’s perspective

It is not that I did not believe plagiarism is wrong, I really did. I simply did not realise that what I was doing at the time could be considered plagiarism. I started writing up the research I did as an undergraduate medical student during my first year as a specialty trainee in emergency medicine. I was keen to publish but had never authored a paper before – I recall writing the first draft using only the guidance for authors provided by the journal. I received support from my supervisor but was determined to do a lot of the work myself. As can be expected, it didn’t fare too well on its first submission. But the journal offered me the use of a free author assist service that could help me make my manuscript more publishable for resubmission at a later stage. I was keen to succeed, felt encouraged by the assistance provided and accepted. The assistant was very experienced and recommended many changes that I incorporated into the manuscript over an 18-month period. I felt strongly that he should be included as a co-author and so included him when I eventually resubmitted the manuscript. In hindsight, I should have communicated better with both the assistant and my supervisor. I was just so excited after making the final changes that I neglected to share the final manuscript one last time with my co-authors. I had made a few final changes to the manuscript at the last minute, adding in two more discussion points from a similarly themed paper from Botswana. It was only after publication that the similarity was discovered: two areas in the discussion section. It was the changes I had made prior to resubmission. Naturally, I know a lot about plagiarism now. In retrospect, it seems silly to have made such an error. Although it happened to me, it could have just as well have happened to any of my peers. I am not convinced that the research training we received as undergraduates adequately prepared any of us for the task of publication. This was a hard way to learn a lesson I should have been prepared for during undergraduate training.

Co-author’s perspective

I carry a large supervisor burden. At the time of writing this, I am supervising no less than seven postgraduate students with their masters degree dissertations. Informally, I mentor at least a dozen more students with a whole range of research-related issues. I am also the academic head of department with all the managerial, clinical and educational responsibilities that goes with it. We do not have a large faculty compared to Western universities but we pride ourselves on quality. At no point did I realise that plagiarism occurred. Our institution did not have a subscription to similarity software at the time (I understand that the software is fairly expensive) so there was no way of checking – even if I did see the final draft. I am disappointed that this had happened but accept that given my substantial workload it would have been challenging to avoid without the scrutiny that similarity software can provide. In my view, focusing on preventative strategies is key. I recommend the following to prospective authors:

• 1.Research training, including teaching about plagiarism and its consequences should be provided at an early stage in academic careers (perhaps even at undergraduate level).
• 2.Clear communication pathways between collaborators to ensure that everyone has the opportunity to review every draft. This has to be agreed from the outset for every research degree.
• 3.Drafts should be subject to similarity software scrutiny prior to submission for publication or examination.
• 4.Clear guidance should be issued by the university or facility on the issues surrounding plagiarism and how it should be dealt with.

Assistant’s perspective

During the first part of 2015, I received a request from the African Journal of Emergency Medicine to work with a young author in Ethiopia through the journal’s Author Assist program. The main goal was to improve the article’s written English and to help with its organisational structure and formatting. Over the next 18 months, I recommended numerous changes to the structure of the article, which were adopted in the draft submissions that were shared with me. I was incredibly dismayed to learn of copied material that was included in the final submission to the journal. I feel that the author assist programme is invaluable as we continue to develop emergency care in Africa and hope that this unfortunate event will not dissuade participation in this programme. I have spent many hours contemplating how I could have helped ameliorate this situation and offer the following points for contemplation by the journal’s readership:

• 1.Outlining set guidelines for communication between all parties (primary author, senior author, author assist, journal).
• 2.Using web based authoring software programme (ex: Google Docs, or DropBox Paper), so all authors and author assist administrators can track changes in the article.
• 3.Having all parties that are involved in the creation of the academic work having the ability to review the article prior to each successive submission.
• 4.Using this unfortunate event as a learning experience for improving the author assist programme. I believe it to be a unique and invaluable programme to help less experienced emergency care practitioners publish their research, education and administrative findings to a worldwide audience.