Introduction

I am on record as saying that I regard plagiarism as an ethical and copyright problem, but that other forms of integrity violations such as data manipulation, image manipulation, or data falsification are more serious because they undermine the reliability of data and our ability to build on the results. (Seadle, 2019) Plagiarism hunters often emphasize the ethical damage or the putative loss to copyright holders. The goal of this column is to examine the damage done specifically by scholarly plagiarism to attempt to measure its consequences.

Personal damage

The number of people accused of plagiarism is growing. In Germany, for example, VroniPlag has accused 203 works, most with single authors, of plagiarism in amounts that, according to VroniPlag, range from 10.9% to 100%. (VroniPlag, 2019) Academics lose their jobs, their careers, sometimes even their pensions, especially if they are professors and a university takes away the doctorates. Government ministers can lose their jobs as well. Strictly speaking all of these people suffer because of the accusation, regardless of whether they intended to commit plagiarism, or were merely careless about their references. A few admit their guilt. Many do not. Those making the accusations appear to feel that those accused get the punishment they deserved. Even if all accusations were true, destroying a life and a future because of text-based copying arguably resembles the 19th century British law that sent people to prison for failing to pay a debt: it seems out of proportion to the crime and does not solve the problem.

The moral damage that results from intentional plagiarism is a form of self-harm. Cut-and-paste tools make plagiarism easy, and there is some evidence that cut-and-paste is a serious enabler. For example Kauffman and Young (2015) did an "empirical study of digital plagiarism" using undergraduates, and their "[r]esults indicated that overall 79.5% of the writers engaged in digital plagiarism. There was a significant interaction in which instructional performance goals lead to more plagiarism when copy-paste is available." There is certainly no doubt that cut-and-paste makes plagiarism disproportionately easy compared to putting together a correct reference. When copying is easy and the likelihood of being caught is small, plagiarism becomes almost a rational economic choice. In purely ethical terms, intentional plagiarism could potentially damage a person's character and lead to future bad decisions, but that kind of damage is hard to measure. It could be, for example, that plagiarists cheat more than others, but there is currently no clear evidence for that and it could be hard to prove. This column will focus on more measurable forms of damage.

Economic Damage

One way of changing the scenario is to ensure that people are caught, and a variety of modern text-comparison tools such as Turnitin or iThenticate using CrossRef make the likelihood of catching plagiarism much higher than in past eras when the comparison had to be done by hand. In monetary terms, the big losers are the institutions that have to pay for the software to discover whether plagiarism occurred. The institutions are acting to protect their reputations, which has an economic value. They might not pay so much if only morality were at stake.

These institutions also need people to run the tests and people with enough experience and expertise to recognize false positives. Some institutions push the effort back on the students. King's College London, for example, requires students to submit their essays and theses to Turnitin until it comes back with a low enough similarity report to be acceptable. Some publishers do not actually routinely check for plagiarism themselves, but rely on the fact that other publishers do check to be a sufficient disincentive. Students who are caught plagiarising may fail and scholars may get an embarrassing rejection or retraction, or they may lose their job as noted above. The economic damage varies greatly and is largely personal.  

Copyright holders could suffer a loss of revenue if the whole of a best-stelling work is taken, but economic success is rare for scholarly works. Charles Dickens famously complained about the injustice of US publishers publishing his works and not paying him. (Hawksley, 2015). That was an extreme case, however, and the Dickens works were not scholarly, even though many scholars now study them. The economic losses from the publication of most genuinely scholarly works is scarcely measurable, especially when the amount copied is too small to reduce the value of buying the original.

Deadly Damage

A friend and colleague recently brought the following case to my attention in reference to the opioid crisis with the comment that plagiarism kills. This is the passage in question:

Reference "4" leads to this citation, which is just a letter to the editor: "Porter J, Jick H. Addiction rare in patients treated with narcotics. N EnglJ Med 1980;302(2):123". The letter has two references, one of which leads to this 1978 article:

In the end it is impossible to tell what the reference really means because the original is so deeply buried. Whether this represents classic plagiarism (copying without a proper reference) or a form of source-obfuscation that itself counts as data falsification is perhaps a matter of definition. My friend is right that this kind of misconduct can have deadly consequences in medical research.

Indirect damage

A lot of resources today focus on combating plagiarism, but what if those resources were applied to image manipulation, which can lead to false information in medical and biological research, or data falsification, which undermines the reliability of natural science as well as other forms of scholarship? One of the reasons that fewer resources are spent on other forms of integrity violation is that they are just much harder to detect reliably. Some work on intelligent algorithms to detect image manipulation is underway, and scholars have developed some statistical tests aimed at certain kinds of data fraud, but images and data are far more various and complex than text. Text-to-text comparisons on topics where major databases exist works because the algorithms need only compare combinations of 26 letters, not variations running into the thousands.

Funding matters, of course. Potentially outraged copyright holders believe that they have an economic interest in combating plagiarism, but image and data fraud has no well-organised lobby to fund research or to persuade public institutions that they must combat those forms of fraud.

Conclusion

With the possible exception of the opioid case, plagiarism does not in general intentionally propagate false information. Misquoting results can also be harmful, but the harm comes from the falsification, not from a missing reference. Most people who knowingly (or unknowingly) plagiarise seem to prefer reliable sources, and in any case the issue with plagiarism is not whether the source is true, but whether proper attribution was made.  A flawed reference does not in and of itself undermine future research, and in the end, that is what matters. Ethical and copyright violations are bad, but do less actual damage to research than other forms of scholarly malpractice. Perhaps it is time for the world to rethink the emphasis on plagiarism and consider what other integrity issues need greater emphasis.

References

Hawksley, Lucinda. 24 June 2015. "Charles Dickens, Copyright Pioneer" in ALCS News. Available: https://www.alcs.co.uk/news/charles-dickens-copyright-pioneer.

Rich, Ben A. 2001. “Prioritizing Pain Management in Patient Care: Has the Time Come for a New Approach?” Postgraduate Medicine 110 (3): 15–17.

Kauffman, Yashu, and Michael F. Young. 2015. “Digital Plagiarism: An Experimental Study of the Effect of Instructional Goals and Copy-and-Paste Affordance.” Computers and Education. https://doi.org/10.1016/j.compedu.2014.12.016.

VroniPlag Wiki. 2019. Übersicht. Available: http://de.vroniplag.wikia.com/wiki/%C3%9Cbersicht.

Return to Top