8 Thoughts on Plagiarism
Today, there was an article on plagiarism in the Swiss commuters' newspaper "20 minuten". Please note that the article did not appear in the online edition. The main scope of the article is plagiarism among students and that Swiss universities do not check incoming student submissions with regard to plagiarism. The article's scope lies on student plagiarism and is common for the entire (recent) debate on the topic but is distracting from the actual problem. The actual problem is plagiarism among researchers. This blog post focuses on 8 points that I find important to consider.
1. Student plagiarism is good
I find the massive interest on student plagiarism somewhat disturbing. I find it disturbing because it focuses on (a) the weakest actors in the universities and (b) it targets the learning process. In fact, I argue that student plagiarism is a good thing because it allows lecturers to address good research practices and ethics.
In the papers I recently read on student plagiarism it appears that student plagiarism is an individual problem that seems to be decoupled from learning to conduct rigerous research. I don't think that student plagiarism is an individual problem but related to the complexity of our trade. The fact that some students create plagiats and others don't is in my opinion for many cases sheer luck. Therefore, it is important to highlight the problems of this tricky part of researching as often as possible in class, ideally based on concrete examples as they occur. Unfortunately, the common organisation of the academic year often povides only very limited space to do so. As this topic is so central to all disciplines, I believe that this is nothing that can be left solitarily to classes on scientific work.
2. Integrating automated plagiarism detection into LMSes is good
In order to be able to address student plagiarism such misconduct needs to get detected in the first place. With this respect it is a good thing to integrate automated plagiarism detectors into learning management systems unless it is used for grading. I am against automated plagiarism detection for grading because I don't think that students are intentionally try to sell the achievements of others as their own work. This might not to be the case for near 100% carbon copies of other works or for sharing of works among students.
I never found this a problem because most plagiats I found in student works were small scale and fairly easy to detect. E.g., the good indicator is a sudden switch in writing style for a paragraph or section. Automated plagiarism detectors can help students to avoid intended or accidental scientific misconduct and enhance their work in those cases where this is detected. Consequently, the result of plagiarism detection needs be available to the students. Maybe there should be markers for lecturers that indicate those sections that were previously identified as plagiarism with a reference to the related sources. This would enable lecturers to identify those smart guys who try to play the system.
3. Automated plagiarism detection is just one element of research quality management
A great deal is written about automated plagiarism detection and so does this morning's article emphasise on the software use vs. the lack of it. While I agree that automated testing is an important element for weeding out the worst cases of plagiarism. However, it is still possible to play these systems and human quality control is still needed.
Particularly for the "weaker" forms of plagiarism a human expert is key to additional quality control. This is needed, for example, if ideas or concepts are replicated without reference to the original work. Such "replicaion" is nearly impossible to detect by automated systems and requires knowledge about the domain and the related works. Typically, the lecturers or, in case of team submissions, the most knowledgable author has the responsibility to remind the students or team members on the state of the art and double check on the hidden "theft of ideas" in the core line of arguments.
4. The actual problem is plagiarism in "original work"
While student plagiarism is a problem to deal with in higher education it is not really the issue. The actual problem of plagiarism is fake "original" research. Carbon copy research is relatively easy to identify and these cases are not only scientific misconduct but are also subject to copyright infringement in many countries. Yet, the more subtile form is replicated work without referring to the original source.
Whether or not replicated work is actual plagiarism is hard to tell. I draw the line towards the "original" claims of a contribution as well as where and when it is published. It is important to differentiate plagiarism from co-development. For authors who submit a paper to a journal or a conference it is the minimum requirement to double check whether the previous issues of the very journal or conference do not include already something that is close enough to match one's own claims. While means extra work for the authors, this definitely can contribute to improve the overall quality of a contribution. On the other hand, leaving this out actually means that one acknowledges to be rightfully called a copycat if there is at least one prior publication on the very subject in the same publication channel. In this case "co-development" is not what anyone should be able to claim on their side.
5. There is no organisational quality control for detecting attempted plagiarism of "original work"
While I found many research papers that focus on controlling student plagiarism, I don't know a university that controls their research staff's output before it is submitted into the various publication channels. It appears that the same rules that are getting enforced on students do not apply to researchers anymore as soon they have graduated.
The present case-by-case pratice of dealing with plagiarism appears to be sufficient for most parts of the scientific community. This very reasoning also sheds a light on the overall scale of the problem. It seems that the actual problem of carbon copies is not such a big issue as the prominent cases might appear to the general public. However, without really analying the problem structurally it is hard to tell whether this is really the case or if the scientific community tries not too spill on their own carpet too much. From reviewing a fair bit myself I find that organisational quality control could not harm, but this is because much work does not even reach the level for even start bothering about scientific misconduct other than wasting the time of fellow researchers.
6. Stop over-emphasizing "self-plagiarism"
The easiest form of "plagiarism" that can be detected is so called self-plagiarism. It is so easy to identify, because one just needs to go through an authors publication list and compare the individual publications. Technically, there is no such thing such as "self-plagiarism" just like it is not possible to rob or kidnap yourself.
True, resubmitting the same article to several journals and conferences is defintely a no go. On the other hand, working with one's data and developing ideas and concepts towards theories is an important element of the scientific discourse. At the same time publications in different channels are crucial means to protect one's work against plagiarism claims. I won't count the act of bringing similar ideas into different communities (e.g., those bound by language) as self-plagiarism, although I know researchers who are taking a strict view point on "self-plagiarism" with regard to other researchers' work. I think it is important to understand the differences between copying one's concepts over and over and disseminating them.
7. There are no reviewer guidelines for handling plagiarism
When it comes to published "original research" the peer review system. While there are many arguments on the limitations of the peer review system, its primary role is quality control for the scientific community. The guidelines for the reviewers, if present, are very broad and the few ones I know do not mention plagiarism. This would suggest that the publishers already took precautions for filtering such work before it enters the peer review process.
From my recent experience on the matter, I must say that even those publishers that claim to do plagiarism checks, actually do not (that is "the one publisher I came across"). It would help the reviewing process a lot if submissions would be quality tested before they enter the peer-review process. Such testing could even provide some additional quality indicators to the reviewers. So probably handling the strong forms of plagiarism is something that should be left better to the editors and publishers.
8. There is no codex for treating plagiarism
One thing with plagiarism is that while detected plagiarism may lead to the loss of a degree if this has been detected in the examination work. However, there is no codex in the scientific community how to treat plagiarising members of the community. While everybody agrees that plagiarising is a really bad thing to do, there appear to be hardly any consequences unless the authors were stupid enough to plagiarise in their graduation works.
The missing codex lets the entire plagiarism debate appear as something that is really targetting students and early stage researchers. But if plagiarism is not really such a big deal for the quality management within the scientific community, why then is there all that fuzz about student plagiarism? I can imagine some draconic punishments such as handing over the related research funding to the plagiarised authors or suspending researchers and reclaiming their degrees if they appear on plagiarising works. However, before thinking about punishments I find it more important to provide more clarification on where to draw the line for plagiarism and identify its relevance within the scientific quality management.