A game-theoretic model of plagiarism.
What keeps economists from plagiarizing the works of others? Part of the answer is that we are professionals who take pride in our work. Also, we believe that the penalties of such actions are far greater than the benefits if we are caught. Consequently, we believe that our reputations would suffer irreparable damage if it were known that we had plagiarized. There is also the fear of dismissal from our employing institution for such behavior. But are the penalties that great? Before any penalty can be assessed there has to be agreement that the act constituted plagiarism. Research by Enders and Hoover [2004, 2006] reveal that the profession does not have a clearly defined notion of what does and does not constitute plagiarism. There is also no governing body that has shown a willingness to act as arbitrator when a charge of unethical behavior is made. Should decisions come from a panel formed by the American Economic Association (AEA), journal editors, or employing institutions? None of these issues have been adequately addressed by members of the profession.
In a recent series of articles in the Chronicle of Higher Education, it was revealed that many cases of clear-cut plagiarism that get reported in academia go unexposed or unpunished. (1) If a potential plagiarist's greatest fear is being found out and having a damaged reputation what does it say that so few of these cases are ever publicized and the offender punished? This paper shows how it is possible that a case of plagiarism can occur without any public scrutiny, hence no damage to the offenders' reputation, even if the original author is made aware of the act. This assumes that there was some sanctioning body willing to hear the case, which presently does not exist.
In this simple model, there are two players (see Figure 1). The first mover in the game is the Author who is the originator of the idea or genesis of the paper. The second mover in the game is the Plagiarist.
[FIGURE 1 OMITTED]
The author has two actions available to her. She can choose to create a new paper for publication or not. If the author chooses not to create the paper, the payoffs for both her and the plagiarist would be zero.
Enders and Hoover  and Arce, et al.  show that authors typically take two paths to protect themselves from being plagiarized. Some authors will choose to establish ownership by dispersing the paper broadly through electronic working paper series, sending out multiple copies to others in their area, presenting at conferences and handing out copies, and other means.
Other authors do the opposite by never handing out copies of the paper or posting them where they can be seen by others before publication. As Enders and Hoover  reveal and Arce, et al.  show formally, neither method keeps a paper from being plagiarized. Both methods have their inherent costs. One cost of keeping papers secret until publication are lags in citations. Ellison [2002a, 2002b] shows that publication lags have increased dramatically. The longer it takes for a given paper to be read by other researchers and to be incorporated into ongoing research and then published itself, could cost in terms of citations and delay initiating other strands of research. If, however, the paper is widely and quickly dispersed, there is the danger that it will be seen and quickly plagiarized. Our model will show that despite the original author having clearly defined ownership rights, it is possible that the plagiarism can go unpunished.
Regardless of the dispersion method, if the paper is seen by the plagiarist, there are two options available to him. He can choose not to plagiarize or he can plagiarize the work. The reader should note, however, that this model assumes that the plagiarist has already deemed himself as someone that will take credit for work that is not his own. We do not make any claims as to the number of economists that will engage in this activity, but given current incentives and previous empirical research, it is clear that the number is not zero. We assume that plagiarists do exist and do plagiarize. Enders and Hoover [2004, 2006] work clearly show that some economists do plagiarize.
If the plagiarist does not plagiarize, then the payoff to the author will be P-E where P is the benefit that she would gain from an additional publication. There are several benefits both pecuniary and non-pecuniary that come from having additional publications, such as promotion, increased salary, increased citations, and prestige and recognition in the profession. Here, E is the expenditure of effort in terms of time and energy necessary to create an original work. We conclude that the payoff to the plagiarist from not plagiarizing will be [gamma], which is less than or equal to zero. By not plagiarizing, the least severe payoff to the plagiarist will be no different than if the paper had not been written. In other words, the least severe penalty would be nothing and y would equal zero. However, if the plagiarist is in danger of not being granted tenure or missing out on the other benefits of publication, the payoffs could be more severe.
As Figure 1 illustrates, there are options available to the author in the event that she is made aware of the plagiarist's actions. Enders and Hoover  show that there is a substantially high probability that plagiarism will be detected by someone. This work also reveals that the original author will normally be made aware of the act. This revelation is new and rather disturbing.
It would be assumed that being detected would mean exposure and reputational damage to the offender. As mentioned earlier and will be shown later, this does not have to be the case. Given that she is made aware of the act, the author can choose to contest the validity of the offending work or not fight it and let it go. Complaints concerning plagiarism are typically voiced by the original author, even if made aware of the act by a third party. Once again, we assume that there is an arena available where contesting a questionable work can occur. Enders and Hoover  show that journal editors are hesitant to act as judge when actually confronted with such a dispute by third parties or authors. Arce, et al.  formally model under what conditions editors are more likely to act.
There are several reasons that the author might choose not to contest the plagiarized paper even though she has full knowledge that it originated from her work(s). Original authors choose not to fight because it is not clear that fighting will be successful. Unlike copyright infringement, which hinges on the non-accredited use of words, plagiarism is more subtle in that it revolves around the non-accredited use of ideas or themes which is more difficult to prove ownership of. In addition, Enders and Hoover  and List et al.  show that a good deal of plagiarism is of the hierarchical nature where the author is a subordinate of the plagiarist such as a graduate student or a junior colleague. In such situations, confrontations could be career threatening.
If the author chooses not to fight, then her payoff will be P - E - [delta], where [delta] is the loss of citations that might be unfairly attributed to the plagiarist. In addition, [delta] could be the mental discomfort of knowing that effort E was shared with the plagiarist without permission. For the plagiarist, if the author chooses not to contest the paper, then his payoff will be P - e. Where P is a publication with its associated marginal benefits. We assume that the amount of effort necessary to plagiarize a work is less than that to create an original work, therefore, his effort is denoted as e, where e < E.
However, the author could choose to contest the work of the plagiarist and fight. If this occurs, the author would receive payoff P - E - F - [rho][delta]. In this case, the author still gets the benefits of the original paper less the efforts necessary to produce it. However, there is also the cost, F, which is the cost to fight the plagiarist. This cost could be opportunity costs of lost time that will be required to gather and document the original efforts put forth on the work. Since plagiarism is subtle, a great deal of effort will be necessary to show where the pilfered material is present in the plagiarized work. These costs could include the cost to retain council in the event that legal actions are taken by either party. As mentioned in Enders and Hoover , legal costs are typically borne by the individuals involved and not the affiliated institutions. It is highly unlikely that these costs will be recovered regardless of outcome.
The term [rho][delta] is meant to capture the fact that success in a fight is not guaranteed. In this case, [rho] is the probability of failure. Where 0 < [rho] < 1.
For the plagiarist, the payoff associated with a fight would be (1 - [rho])(D - d)+ p(P - e - d), where D is the punishment associated with being found to have plagiarized. This punishment may be dismissal from the employing institution. It could also be severe damage to the plagiarist's reputation once it is known and proven that this person is unethical in research endeavors or the punishment could be nothing at all. The term d captures the cost of defense against charges of plagiarism.
We will use standard backward-induction to solve for equilibrium strategies. As such, we begin by determining whether the author would fight against the plagiarist. The author considers the following alternative payoffs:
P - E - [delta] > P - E - F - [rho][delta] (1)
which would imply that the author chooses not to fight when:
[delta](1 - [rho]) < F. (2)
Equation (2) states that if the costs of fighting the plagiarist are greater than the expected gains from victory then the author is better off not fighting. If the counter is true, meaning that the expected gains from fighting are greater than the fight costs, the plagiarist may face damage to his reputation or other penalties, D, if the author can find a sanctioning body (employing institution, governing body, editor) willing to make the action known publicly.
What must be considered is what gain the author actually receives from a successful fight with the plagiarist. At best, a victory will make the author no better off than she was before the encounter, since there are no additional considerations for the original publication. Perhaps the original author will be mentioned in a retraction if the plagiarized paper makes it to publication. If caught earlier in the publishing process, the author does not incur the loss [delta], which leaves her at P - E - F. Even if victory were assured, meaning that [rho] = 0, it still might not be best to fight if fight costs are extremely large meaning F>[delta]. In the event that [rho] = 1, any non-negative fight costs would mean that fighting would not be a worthwhile pursuit.
Given that the author does not fight, the plagiarist considers the following alternative payoffs when considering whether to plagiarize or not:
[gamma] < P - e. (3)
Since the greatest value of [gamma] is zero, the plagiarist will plagiarize as long as e is less than P. This must be the case since e<E, and the author would not have written the paper unless P>E, and the plagiarist would have written his own paper if e>P.
The author's choice regarding producing the original work or not is then:
0 < P - E - [delta] (4)
which implies that as long as the benefits of publication outweigh the energy of creating an original paper along with the knowledge that the paper can be plagiarized, the author will create the work.
In equilibrium it is shown that the author will create an original work. It is also shown that the work created by the author can be stolen by the plagiarist. We assume that the author is made aware of the offense. However, due to the high fight costs incurred by the author, the plagiarist may not be punished or exposed to the profession because no action is taken. In the end, there will be papers produced and there will be some degree of plagiarism. As mentioned earlier, not every economist is a plagiarist, hence, every paper produced in the profession will not be plagiarized, but to the extent that plagiarists exist in the profession, they will engage in this activity with impunity.
This scenario is a textbook example of a negative externality. The plagiarist is not intending to cause harm to the author. In fact, plagiarists prefer that the original authors not know that plagiarism occurred.
However, the damages caused by plagiarism are not limited to the author. The profession as a whole is damaged by these actions. There is a notion that the profession would be benefited by plagiarism in that it increases the dispersion of good ideas. However, this logic is flawed. Plagiarism causes damages [delta] to the author by reducing the benefits of producing a good publication, P. Increases in damages will ultimately lead to a decrease in good publications. This will in turn lead to an increased dispersion of papers of less quality, which in no way will benefit the profession.
In addition, if the profession hopes to increase the well-being of others and influence policy makers regarding areas that are impacted by economic modeling, it would be distressing to learn that the work is not original. If increasing damages cause the increased production of lower quality publications, end users might look elsewhere for policy suggestions. If the general public or policy makers lose faith in the validity of economic research, the profession loses. It is not clear whether the threat of being plagiarized would lead authors to only disseminate published papers, however, our profession hinges on the free exchange of ideas and the entire profession is hindered by anything that slows or thwarts that exchange.
Arce  show the impacts of corporate culture on unethical behavior. The author shows that what this paper describes as fight costs are real and substantive, so it may not be worth it to the author to contest the plagiarized work. Previous research has shown that economists take a grim view of plagiarism, but this is not supported by the corporate environment in that there is no governing body that is willing to hear cases of alleged unethical behavior.
It is clear that there is an over production of plagiarized papers. However, there seems to be hesitancy to use any of the ideas that are taught in economics classes to decrease it. The profession could have a sanctioning body designed to hear cases of alleged plagiarism. Technology is such that there are many software programs designed to detect works with similar themes. This type of technology should be freely available to all journals. These measures are not currently in place.
As mentioned in Enders and Hoover  and DeMartino , a code of ethics could be helpful in that it could clearly spell out what is and is not plagiarism and other forms of unethical behavior. In addition, if this code were to come from a sanctioning body, it could also be possible to enumerate possible punishments for such actions. Finally, as mentioned in Arce , the culture of the profession needs to be changed so that subordinates, such as graduate students and junior colleagues, are not hesitant to bring charges against superiors because of possible retaliation.
We show that plagiarism is possible in the profession even if the original author of a paper is made aware of the work of an alleged plagiarist. Since fighting the validity of the plagiarized work might be too costly for the original author then doing nothing, the threat of a damaged reputation is not a sufficient deterrent to such behavior. Plagiarism will ultimately slow the free exchange of ideas in the profession. We also discuss how the production of plagiarized papers can be reduced by using principles of economics, such as reducing the costs to those that contest suspicious work, altering the environment by reducing the expected backlash that can occur when a superior is accused by a subordinate, or making software freely available that can help to detect pilfered work. By taking measures to reduce the amount of plagiarized work in the profession, we strengthen our relationship with the end users of our research such as policy makers, the general public, and other economists and social scientists.
This paper shows that plagiarists will always plagiarize given certain incentives. What this paper does not show is what makes a plagiarist engage in the activity in the first place or why those that could plagiarize do not. We are not implying that all scholars are dishonest or will engage in this type of activity. We do claim that for that subsection of scholars that do want to plagiarize, they can get away with it.
Arce, D. M. The Whistle Blower: Hero or Rat? Rhodes College Working Paper, 2005
--. Plagiarism and Its Impact on the Economics Profession, University of Alabama Working Paper, 2005.
DeMartino, G. "A Professional Ethics Code for Economists," Challenge, 48, 4, 2005, pp. 88-104.
Enders, W.; Hoover, G. A. "Whose Line is it?: Plagiarism in Economics," Journal of Economic Literature, 42, 3, 2004, pp. 487-493.
--. "Plagiarism in the Economics Profession: A Survey of Economists," Challenge, 49, 4, 2006, pp. 1-16.
Ellison, G. "The Slowdown of the Economics Publishing Process," Journal of Political Economy, 110, 5, 2002a, pp. 947-993.
--. "Evolving Standards for Academic Publishing: A q-r Theory," Journal of Political Economy, 110, 5, 2002b, pp. 994-1034.
List, J.; Bailey, C. B.; Euzent, P.; Martin, T. "Academic Economists Behaving Badly? A Survey on Three Areas of Unethical Behavior," Economic Inquiry, 39, 1, 2001, pp. 162-170.
GARY A. HOOVER, University of Alabama--U.S.A. The author wishes to thank Daniel Arce, Sondra Collins, Walter Enders, and Paul Pecorino for helpful comments regarding this work. The usual caveats remain regarding errors.
(1) See Chronicle of Higher Education, 17 December 2004.
|Printer friendly Cite/link Email Feedback|
|Author:||Hoover, Gary A.|
|Publication:||Atlantic Economic Journal|
|Date:||Dec 1, 2006|
|Previous Article:||Nonprofit organization and the division of labor: a theoretical perspective.|
|Next Article:||Economics of crime: panel data analysis of bank robbery in the United States.|