Plagarism or inspiration?

It is an unfortunate reality. It is hard to write good questions and the easiest way is to edit some of the existing ones. If you are trying to put a course together, that's the fastest way to get a big database of high quality questions, esp the verbal ones.

As to legal consequences, I guess not
I have not seen GMAC engage in legal challenges to prep companies. I have seen some of the Veritas Prep questions that were very similar to the official ones from many years ago and they were never questioned. (it was a small number that seems one of their tutors took some shortcuts to make). But it is a disservice to students if the questions are sold as "new" and this is something we have tried to avoid in GMAT Club tests since it defeats the purpose. At the same time, if a company cannot use official questions legally, this is a way for them to have almost-official questions in the course, esp verbal ones, which could be useful and valuable to the student. There is of course no good way explain this to the student.

Hello, ekuseru. I agree with bb and IanStewart above. Plagiarizing questions is not just an act of copyright violation, one that can carry legal consequences, but doing so is just plain dishonest. I chuckled when I read the following:


In all seriousness, though, you might want to ask your e-course provider for a refund. Explain the situation: You were robbed of the right to study official questions—those deemed most helpful by nearly everyone in the industry—for your preparation.

I have decried this practice for some time now, and have been much more vocal about it lately. For instance, here is my first post in the forum in which I used the word plagiarism, from 8 February 2021. At the time, I was uncertain what constituted plagiarism. In graduate school, I recall that in a course in which my professor had us diagramming sentences, he told us that it would be fine to copy the same structure, sentence by sentence, of an entire novel, even a great American novel, and call it a new work, since grammar is not copyrightable. But swapping out one noun for another or one adjective for another is a different story. Where am I going with this? Since, at the time of that early 2021 post, I had recently dabbled in writing some of my own questions, I decided to go directly to the source and ask the GMAC™ legal team for a clearcut response on its stance on plagiarism. After a few weeks, the head of the legal team contacted me to provide a response.

• No official question may be altered slightly and passed off as a new question
• Question phrasing itself is protected by copyright and cannot be reproduced verbatim
• Answer choices are protected by intellectual copyright in the sense that a similar passage asking a similar question cannot present answer choices that adhere to the same logical patterns as those present in an official question

Any violation of the above is a criminal offense and can be brought to suit. However, and this is why I quoted bb above, GMAC™ apparently does not actively seek to identify and prosecute offenders. It might take a whistleblower to set the GMAC™ legal team to action.

I wrote a post just yesterday in which I called out another plagiarized question and advised someone to steer clear of it, but at heart, I am no whistleblower. I am more of a mind to keep to myself, even if I am an active voice on this site.

bb, I think the best thing you can do is to take a clear stance on the issue and post somewhere in the site rules that GMAT Club takes plagiarism seriously, and that creating or knowingly posting plagiarized content on this site will result in moderator action (i.e. pulling the post down) and disciplinary action for further offenses. What prep companies decide to do behind closed doors is up to them, but this site represents something much larger, and for obvious reasons, your voice carries a lot of weight.

- Andrew

While I've found examples of paraphrased RC passages and CR question stems, this one in particular sticks out to me the most:

OG CR question:
Sparrow Airlines is planning to reduce its costs by cleaning its planes' engines once a month, rather than the industry standard of every six months. With cleaner engines, Sparrow can postpone engine overhauls, which take planes out of service for up to 18 months. Furthermore, cleaning an engine reduces its fuel consumption by roughly 1.2 percent.

OG answer:
B. the cost of monthly cleaning of an airplane's engines is not significantly greater in the long run than is the cost of an engine overhaul

Unnamed prep question:
The Great Eastern Hotel is fully covered by wall to wall carpeting which the hotel normally pressure cleans twice a year. With the intention of reducing its costs, the hotel now plans to pressure clean the carpets every two months. With frequent cleaning, the hotel can delay the complete replacement of old carpets by new ones, a process which requires the hotel to close business for a week. Moreover, frequent pressure cleaning of carpets increases their durability and longevity.

UP answer:
Whether the long run costs of more frequent pressure cleanings is not significantly greater than the cost of buying new carpets?

AndrewN do you think that's close enough to warrant some action on my part? I'm not a huge fan of whistleblowing as well, especially considering that the company in question is relatively new, but I do think that some corrective actions are due.

This is a pretty well masked question. I have seen a lot closer versions (I can't find them on the spur of the moment but I have seen copies with only a word modified), so this is by no means a copy (my humble opinion compared to what I have seen).

Thank you Andrew - I have no qualms about doing this besides we will inconvenience users who are searching for those questions. Though I don't think it is a great idea to delete/remove questions.

What we usually do is we post the question and then we put a note that it is a "copy" of another question not to be confused. That is usually what many moderators choose to do so that it keeps things sane for everyone. Otherwise, users will keep posting the same lame question over and over - it is not productive for the purposes of discussion and keeping a catalog organized. In other words, we choose to label them as such as opposed to eradicating them. It is more shameful at the end

It is clearly a copy.

ekuseru, I imagine you have a reason you do not wish to name the company here?

ekuseru,
It is good you spotted the pattern! It is great matched the questions.
Despite Ian calling it "clearly a copy", it is a well masked question that is a derivative of the OG question. I don't think many people would have recognized it.

Most GMAT questions follow patterns in lower/mid difficulties - that is true for Q and V so catching patterns (this one looks like quite a bit was modified except the concept of cleaning and the parallel logic) but I see from the answer choices that even the question type was changed and turned into a question (there is a ? at the end of the answer choice). It is kudos to your brain. They did try to hide it. This does not seem original and very likely derived from the OG question

P.S. This discussion made me think of an interesting dilemma and the reality of writing original questions. It is very hard to do! On one hand if you write a verbal question - to match the pattern of the OG, nobody can argue that your questions are not-gmat-like since you can whoop out the OG and show examples (we do that in quant where patterns are much clearer and you can demonstrate that a certain rule/property is indeed tested though it may be written in a completely different way), but then you have the issue of question being too similar and frankly better to use the original OG material but since GMAC does not license questions, this is sort of a way around it. At the same time, if you write a completely original question, you have to then prove that it is indeed GMAT-like and that's virtually impossible. There is a fine line... I think in this case, the similarity is too close - the copier did not quite step over the safe line by keeping the concept of cleaning but had they used a slightly different concept, it would be harder to catch/link even if still following the pattern. I have to give credit to the original question writers - it is very hard to do for Verbal!!!

I'm actually in the process of drafting a review for said company (waiting for official results) and would like to be as unbiased and fair as possible. If superficially(?) modifying OG questions is a widespread practice then it doesn't feel right to me to fault them for it.

You are right, bb, that things may look quite different from the moderator end. Even I, as a regular user, though, have come across duplicate questions that have flown under the radar for years. (I keep Bunuel busy, even though my intentions are good.) The two lines at the end of your post intrigue me. For one, I have never come across a moderator-labeled "copy" question. Again, I am sure you know the site better than I do, but in the many times I have exposed plagiarized questions, no moderator has ever gone back and added any sort of note. However, I do think a "Scarlet Letter" treatment could prove effective in any of the following ways (and, once again, you probably know more about this backend stuff):

!	The following question has been identified as deriving from (LINK with text "this official question")

I personally like the third, such is my distaste for the practice of plagiarizing. I would like to hear your thoughts about implementing such a system. I will volunteer in the effort to identify copycat questions.


I agree with both bb and IanStewart about the carpet-cleaning question. From a user standpoint, it is fairly well disguised; from a legal standpoint, GMAC™ would have no trouble using such a question to build a case against the company, and the more questions like it, the stronger the case. Based on just the three bullet points from my earlier post—and I would be willing to bet that GMAC™ would get much more specific if filing suit—the question touches on two.

ekuseru, I would let the prep company know about your disappointment and seek at least a partial refund. Whether you decide to take the matter up in any other way is your decision.

bb, of course I agree that writing high-quality questions, especially Verbal questions, is challenging. But if all these great minds at test prep companies devoted their efforts to creating new material that ran close enough to official questions to capture the right feel, rather than copying official questions and giving them a makeover, the entire community would benefit.

Apologies for the meandering post, everybody. I am dashing this off before my weekly grocery run, and I have to get going.

- Andrew

After some thought, I think students would be better served by a tiered system that can account for both blatant copies (e.g. the manatee question) and "well-disguised" questions. If implemented properly, it may even eliminate the need for "I think this is a [high/low] quality question"- and "Please Bunuel, link me more questions like this"-posts which add little to the conversation.

I have to ask, though, who would have the unenviable job of combing through such questions and deciding whether they belonged to Camp A or Camp B? As with many things, I imagine there is probably a spectrum of questions that are nearer to or further from the source material, but are still identifiable as derivatives. As much as I like specificity, I think the more practical and efficient way to "brand" the questions would be to use one blanket method. In this manner, students could prepare for the exam without being blindsided (at least through this site), and test prep companies might reduce their reliance upon official material for fear of being labeled lackluster in their own question offerings. As I said before, I would be willing to spearhead such a movement in an effort to keep everybody happy: GMAC™, students, and test prep companies. (Yes, you read that right. I am not out to get anybody. I just want to promote honest work.)

Concerning the latter part of your post, I agree with you in principle, but many people post comments and requests without having read any of the earlier dialogue in the thread—I can speak to this matter as an Expert who often responds to such posts—and I am convinced that such... banter?... simply comes with the territory. People can already scroll to the bottom of the page to a given question and see a list of recommended (similar) questions, but the sorts of posts you commented on keep rolling in.

This conversation seems to be steering into administrative territory. I hope that if you are continuing to study for the exam, all these concerns are not getting in the way.

- Andrew

Thanks for the concern, I'm slowly coming to terms with my score and don't foresee myself attempting a retake in the immediate future. Just trying to give back to the site that helped me with my prep.

If the topic is inappropriate for this forum, perhaps we can move this discussion somewhere else?



As a student, I honestly wouldn't mind it either way but the gmatclub might prefer a more nuanced solution, seeing that they are caught between students and potential advertisers.
In my experience, I was preparing solely based on the course provider's questions (since they advertised themselves as having "GMAT"-like questions) and it wasn't until two weeks before my exam that I found out about the significant gap in quality and style between their questions and the OG's. Prefacing derivative questions with a moderator-approved message linking to the OG question (without passing any value judgement, leaving the final decision to the student) might be a sufficient intermediate solution. Seems like a win-win to me.

I'd ask you to reconsider this. Plagiarizing official questions may be a widespread practice, but it's not a universal practice. Some companies, so they can advertise that they have "2000+ questions!" have copied questions test takers already have, in better versions, in the official books. Of course those companies don't advertise that they have "2000+ questions that are worse versions of ones you already own!" because then no one would buy their product. But then there are other companies that haven't churned out cheap imitation questions (and I'll post later about just how easy it is to write imitation questions like these), and who might only be able to advertise "we have 500 questions", but who have written genuinely original problems that might actually add to a test taker's study. The honest company with fewer questions is at a disadvantage, from an advertising perspective, because 500 < 2000, unless test takers know that the 2000+ question company is not offering useful questions.

I'll be the first to admit that I fell for this marketing gimmick.

You have an excellent point about honest companies getting shafted if they don't cut corners like the rest. The hammer has to fall eventually and I would prefer to reward honest companies. Can't say I've read a review on this site (or anywhere else for that matter) that mentions the quality and originality of questions though. Would definitely be nice if someone with the expertise and experience can give prep courses the Michelin guide treatment.

No, ekuseru, the topic is posted appropriately. I just meant that the content of the discussion is steering into an administrative track. bb and other moderators, perhaps, would have to take up the dialogue from here, since they are the ones who have the ability to edit posts other than their own.


I agree. The note can simply be informative. If some students choose to ruin their chances of assessing an official question legitimately by working through the plagiarized question first, that is their prerogative. But at least it would cut down on the number of students who might get that gut-wrenching feeling of déjà vu upon seeing the official question for the first time.


I agree with Ian. Just because a practice is widespread does not mean one company should get off the hook for doing something dishonest and illegal. Your experience was with this one company, so that may be the only one you are in a position to write about. Also, IanStewart, you are killing me with the honest version of that advertising slogan.


It is hard to find enough people in the industry with enough expertise who would even be able to identify the questions, let alone wish to expose them (for fear of being blackballed). But I agree. It would be fascinating to read the findings of such a person or group.

- Andrew

It might be possible to automate parts of this process.
I can think of some machine learning methods to handle verbal questions: simple TF-IDF can catch manatee questions and text vectorization, a more involved technique, can account for more "disguised" questions.

Might be something that bb might want to look into, has the potential to be a killer feature and improve the whole GMAT prep landscape for good.

How forward-thinking of you, ekuseru. (Congratulations, by the way, on earning a score in the 99th percentile. I see that you have verified it since our last exchange.) I know nothing of these technologies, but I imagine that colleges are turning to them increasingly as such software becomes more widely available (and, most likely, cheaper). I suppose that for the software to work, someone would have to input the database of questions recognized as official; otherwise, official questions themselves would be getting flagged all the time. It is not uncommon to see older questions, especially boldface questions, in a few iterations. (I have seen the same passage with different boldfaced portions appear in four different questions.) Neither is it uncommon to see the same passage used for two different question types, or for the same passage and question type, but with slightly different phrasing or one or two answer choices, to be used for multiple questions. Of course, this discussion is all theoretical unless the administrators want to implement some of the ideas you have touched on. If you know of some free or cheap software that can do what you have described, I would be happy to take a look. If nothing else, at least I could better appreciate how insidious or widespread the problem might be in the industry.

- Andrew