If you google some obvious search phrases, you will find a number of web sites where students can pay people to take their online exams for them.

If I am convinced that a student has used such a service on one of my online exams, how can I prove it? The disciplinary board at my university has fairly strict standards when it comes to proof. When students have cooperated on exams, or copied from each other, there are often obvious similarities between their answers that you can point out to the board, but in a case like this there is nothing to compare to.

These have been open-book exams where any resources, except other people, are allowed. For example, they are allowed to search and read Q&A sites such as Stack Overflow, but not post their own questions. There are no locked browsers or remote video proctoring.

  • Comments are not for extended discussion; this conversation has been moved to chat. Please read this FAQ before posting another comment.
    – Massimo Ortolano
    Oct 25 '21 at 12:25
  • 2
    Do you have tenure yet? If you don't have tenure yet you need to move on.
    – Wakem
    Oct 25 '21 at 14:57
  • 1
    At least no one has suggested setting up a sting operation! ;-)
    – Ed V
    Oct 25 '21 at 15:11
  • @EdV I did, quite a while ago. For those who missed it then, the idea is to impersonate a student cheat, buy the answers and take it from there. Oct 25 '21 at 19:19

If a student has used such a service on one of my online exams, how can I prove it?

I don't know if this works for all of them, but for chegg specifically: if you find one of your exam problems on their site, have your Dean write chegg a formal letter indicating this and ask for an investigation. Chegg will (usually) send you a list of usernames and IP address of the student who originally posted the problem, as well as any students who viewed the problem. This is detailed in their "honor code": https://www.chegg.com/honorcode

I have used this successfully last year to find out who posted copies of my exam to chegg.

Edit: in response to comments, this was in US. I don't know if this process would work in Europe.

Also, all of the students in question had either used university e-mail addresses, or something like firstname.lastname@gmail.com so I was able to identify all of them.

  • 16
    My department had success contacting Chegg like this. Side note: about 2/3 of the students used their university email address to create their Chegg accounts :/ Oct 23 '21 at 7:53
  • 17
    @Greg Martin: Hopefully 2/3 of the students with a Chegg account and not 2/3 in total?
    – Keba
    Oct 23 '21 at 8:20
  • 6
    For anyone wondering: Chegg's privacy policy explicitly mentions giving away personally identifiable information in cases of "disclosures during academic investigations." It's still unclear to me whether that makes it actually legal and/or ethical.
    – ComFreek
    Oct 23 '21 at 10:52
  • 14
    @ScottSeidman I meant Chegg giving out username+IP may be illegal and/or unethical. In some jurisdictions, privacy policies cannot be arbitrary; thus even if you, as a user, agree to it, its legality may be uncertain. And the fact that users agree to something that they have probably never read (nor might have any alternative options for the site in question) doesn't make it ethical. For example, compare this with Facebook.
    – ComFreek
    Oct 23 '21 at 20:31
  • 4
    "It is in the user agreement" is not synonymous with "sharing personal data is legal", in particular in europe with the GDPR. In particular if they are relying on consent then there must be a mechanism for one to withdraw that consent.
    – Dave
    Oct 25 '21 at 10:51

Having thought about this issue both from purely academic side, and also from cryptographic protocol side (trying to be clever...), I think there is no way to achieve this goal. So, then, the real questions are about the size of the issue (how many students do this?) and about revising our notions of "exams" to make evaluation be such that it's "easier" for students to study than to hire other people to do their work (whatever that work is decided to be).

(One example of reconsideration of what happens in an "exam": years ago, it occurred to me that, if the goal of a course is to get students to learn something, it may be better to give open-book exams, because then at least they may learn something during the exam, even if they didn't study before. :) No, I do not want to take this toooo far, but this and related points do not seem to be part of some traditional testing cultures...)

  • 7
    On open book exams: My parents were both teachers and every test they gave was anything-but-other-people. It made them think more in making test questions but it made for better tests--all of their questions required applying the knowledge of the class, not merely looking it up. Know the material, it's a piece of cake. Forget a name or the like, look it up. Don't know the concept, might as well guess. Oct 23 '21 at 1:34
  • 8
    +1 The main thrust (first sentence) here I support. But the one hail-mary re: "make evaluation be such that it's 'easier' for students to study than to hire other people to do their work" is frankly one of those feel-good wishes that doesn't get any work done without concrete application steps (and I don't think there are any). There's no lower-bound for incapacity in students (incl. such aspects as logging into a test on time, staying awake through it, etc.) Suggest you reflect on whether that line is really useful here, and maybe snip it out. Oct 23 '21 at 1:53
  • 4
    All courses I have taken, except the math ones, has been open book exams for me (where many has had 3+ books with ~700 pages each). My experience is that if you do not know the answer, or at least where to find it quickly, then you won't have time to find it and learn during the exam. During my years I've only managed once to solve a problem I had absolutely no clue about on before hand and it was through sheer luck at finding the correct phrasing in the index. However, I was so stressed to figure it out that I completely blocked out any memories of what I read afterwards and learned nothing.
    – Mrkvička
    Oct 23 '21 at 11:44
  • 1
    @paulgarrett: With respect, I don't agree that for a clear majority universally there's a lower bound. This varies by institution, of course. Consider that half of U.S. college students now attend open-admissions community colleges, and only about 20% graduate from those schools, can't pass 8th-grade reading/writing/math tests, etc. Again, in these "better way" posts I'd like to see a concrete citation and I don't see any, so I remain in disbelief. Oct 23 '21 at 14:00
  • 2
    Data point: in the 1960s, a significant fraction (I'd estimate 1/3...) of high school kids of both genders left high school without finishing it, having taken "metal shop" or "auto-shop" or "book-keeping" and other "vocational" courses, and started making good money immediately. The so-called "smart kids" who were aiming at college could/did only do so if they had sufficient family support, I think. Some sort of "long shot". In any case, people who didn't want to take math classes (and other stuff) could avoid it, and very quickly make a good living. I do wonder about analogues these days. Oct 23 '21 at 23:35

In general, it is always possible to safely cheat via expert in written online exams: The student can just forward whatever web interface you throw at them to the paid expert and then the expert can do everything the student has to do without you noticing¹. For whatever it’s worth, I have been a member of huge groups of people with a strong interest to avoid this kind of cheating and nobody could provide a solution.

As a consequence, you can only detect and prove such cheating if the cheaters or hired experts get sloppy (or confess). And even then it will be difficult to provide sufficiently strong evidence to convince the board. For example, I can think of the following:

  • The exam’s handwriting is strikingly different from the student’s usual handwriting and the student fails to reproduce said handwriting under observation. (But of course, a remotely clever cheater would copy the solutions provided by the hired expert by hand.)

  • The student cannot explain their solution at all. However, if you are probing like this in the first place, just make the interview about the solutions the official actual exam.

  • The student cannot explain how they accessed the exam interface from an IP address located in a different part of the world.

  • The exam question appears on a public platform.

If you think that these probably do not apply to your exam, then that’s more or less my previous point: You can only prove this with help of convenient sloppiness from the other side.

¹ “Locked” browsers and video proctoring can increase the technical hurdle (but not make it insurmountable). Theoretically you can take extreme measures where you only accept solutions that the student handwrote without touching their keyboard under complete video surveillance, but even then you cannot do anything against an expert in the same room giving hints. (Also, I do not condone this extreme proctoring because it is a surefire way to trigger a vast array of anxieties in honest students and disadvantage them.)

  • 1
    Not that I disagree with your answer, but in my experience cheating normally happens with the student sending/posting the exam questions (e.g., with chegg mentioned in another answer), and then regurgitating the answers they got back. So the 1st and 3rd bullet points aren't relevant in most situations I would worry about. For the 2nd, are you proposing to interview every student after the exam or just those you suspect?
    – Kimball
    Oct 22 '21 at 20:34
  • 1
    @Kimball: Regarding 1 and 3: As I said, you do need a convenient sloppiness (also see my edit). Regarding 2: For fairness, you would have to interview every student, in which case you might as well make this an oral exam or a mix (and yes, I know that this is unfeasible in many contexts). — If I would need to propose anything, it’s to either give up written online exams or the illusion that you can avoid cheating in those.
    – Wrzlprmft
    Oct 23 '21 at 6:36
  • @Kimball There is normally "sloppiness" associated with cheating b/c enforcement is lax. If enforcement was stepped up, then some cheaters would stop cheating and other cheaters would step up their cheating game. This answer says that it is not possible to stop all cheating and at some point, we should just stop worrying about it.
    – emory
    Oct 24 '21 at 13:02
  • 2
    @emory: This answer says that it is not possible to stop all cheating and at some point, we should just stop worrying about it. – Mind that there is a crucial difference for remote written exams here: If the cheaters have stepped up their game to some (not very high) point, they are almost perfectly safe – the arms race is capped. This does not apply to cheating in live exams, where all cheating methods come with a considerable risk of detection or little benefit (e.g., cheat sheets, which only help you with the memorising component of an exam, if there is any).
    – Wrzlprmft
    Oct 24 '21 at 13:49

Sometimes you can say nothing and wait for the cheaters to get overconfident (i.e.: sloppier) in future assignments. They may forget to remove the invoice from the cheating service, or may let a friend turn in a copy ... or something else glaring.

For now, all you should/can do is note the suspected students and move on. Maybe save copies of the suspect assignments (the goal being for when a student says on a later assignment "but I just cheated on this one in a moment of weakness").

My other advice is not to worry too much about it. Sure, a school's reputation goes bad if they graduate only incompetent students, and other students will yell at you if they hear how "everyone else" in the class is cheating; but a few is inevitable. It's not like they dock your salary for it. And these students often go on to fail the next class (where they are so far behind they don't even know how to cheat properly).

It's far too easy to turn catching a few cheating students into one of those movies where the student is a serial killer and you're the only one who can catch them before you retire (at the end of the semester). Focus on the good students. Spend a few hours here and there on academic dishonesty, and then put it out of your mind.

  • 1
    Thanks. This is a useful perspective. Suddenly I see myself as the retired, alcoholic cop in The Pledge. Oct 25 '21 at 9:24

As others here have noted, proving cheating via technical means can be difficult, if not impossible in some situations.

I've found that a surprising amount of people buckle under pressure and admit to cheating or dishonesty if you simply (gently, but firmly) ask them if any cheating took place. This is especially true if you ask with a kind of calm confidence that says "I know something happened."

Do not throw accusations, show aggression, or insinuate anything. Simply ask an easy-to-answer question or two and take it from there. If they do not admit to anything, call it a day and move on.


Am not a teacher although I do have some experience in sharing knowledge.

Setting aside the thought that this answer may backfire on me in the future :-) , best way would be to select students randomly (or whoever you suspect of not honouring the 'honorcode') and tell them that you are cross-validating that they did indeed answered it by themselves and have them explain how they arrived at that answer.

If the students answer on the lines of -

I googled it and then chose website-a from where I further navigated to website-b or i took samples from website-a,b and c and made up my answer

...that may be just fine.

On the other hand if they reply on the lines of -

I am unable to find the links now but was able to do it before ..

would be red-flags (although not impossible given that google/duck-duck-go would rely on scoring and that may change in between student searching once and searching again).

In such cases need to increase the number of samples (if unable to recollect how they answered question-a, ask them to explain answer for question-b).

Given the pandemic, it would not be unexpected and may even give students and parents that some amount of care is being taken from the faculty end.

You'll need to figure out how to do the co-ordination bit across students though.

  • 1
    Having students prove they did their own work is awkward in practice. For one thing, you're accusing them of cheating (saying it's done at random, etc... won't matter) but with no proof, and making them do extra work because of that. Another is that non-cheating students won't remember details -- often I can't find the same website twice, and not if I'm nervous. And it's tough to write up for discipline: "student was unable to explain how they did assignment to my satisfaction"? Oct 25 '21 at 21:11
  • Ok. Then may be skip for this exam, re-initiate a 'surprise' one few weeks later and this time around ask them to document the steps used to arrive at the answer or tell them they are being monitored and take screenshots of their screen at random ? Oct 26 '21 at 19:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.