« ABA Committee Discusses Tenure Requirements and Law School Accreditation | Main | JOTWELL: Tidmarsh on Lemos (and Hensler) on parens patriae »

Tuesday, April 30, 2013

Are Your Students Cheating On Your Take-Home Exam? Would You Know? What Should You Do About It?

After the well-publicized cheating through collaboration scandal at Harvard College (not the law school to be clear) I have been thinking more about whether law students routinely cheat through collaboration, whether I would know if they did in my classes, and what fears about their doing so should cause me to do in terms of my exam format.

In Civil Procedure I give an 8-hour take-home, typically with one long multi-part issue spotter (worth most of the grade), a few true/false (and explain why if false statements), and a theory question.

When I asked some of my former students I trusted to be honest to me if they had encountered cheating at Harvard Law (and specifically on my exam) they told me emphatically no, and in a back-handed compliment told me my exam was hard enough to require so much of the eight hours they would think it would be very risky to try to do this.

I certainly do not want to help people with “how-to”s on cheating, but I told these former students that if one was worried about this drawback, I could imagine telephoning a friend mid-way through, comparing issues spotted on the issue-spotter and the true/falses, and then getting back to work.

I am curious whether others have thought about these issues and what it has or has not motivated them to do. Are your students cheating on take-homes? How would you know? Unlike other kinds of cheating (like copying) this form strikes me as hard to detect: among 82 students I suspect people often cluster on the issues they spot or do not spot) Should I be thinking about moving to an in-class exam (which, I think, is for this course pedagogically less good for my purposes) to avoid it?

-I. Glenn Cohen


Posted by Ivan Cohen on April 30, 2013 at 03:39 PM in Teaching Law | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Are Your Students Cheating On Your Take-Home Exam? Would You Know? What Should You Do About It?:


What "another former HLSer" describes is entirely doable on many exams. I never did it when the professor said it was forbidden, but if the professor just said it was a bad idea (because he/she could tell and it wouldn't help), I sometimes did do it. I didn't (and don't) think there's anything wrong in doing it, but if you want to limit its use, strict word/page limits help.

If you are good enough at predicting what is important enough to be tested and thinking about how it can effectively be tested, though, you can prepare these kinds of prewritten bits of wisdom regardless of how carefully the professor tries to vary tests over the years. If the test is fair (i.e., covers what is important), then it is going to be somewhat predictable. And even blocking the copy/paste mechanism doesn't defeat the underlying practice of formulating an excellent analysis to be dropped in by typing from memory or notes. That's just good test prep practice: preparing 100% for the predictable so that you can spend time and energy focusing on what cannot be predicted.

What I'm coming to is that what makes the copy/paste successful (where it is successful) is actually genuine skill and preparation for a task. It's not unlike having prepared answers to predictable questions in oral argument. For every single exam I took in law school, I had prepared in part by typing out the same basic analysis to the kinds of questions I thought were most likely to arise, and I made sure that my stock approach was well-written and nuanced.

Contrast this skill with collaboration on a test that bans it. That's just cheating, and not in a way that is likely to correlate well with intelligence or any other useful quality in a lawyer.

Posted by: anon | May 6, 2013 9:05:25 PM

I feel compelled to reply to Anon First Year Prof and former HLSer. When I was a 2L at HLS, a fellow student bragged about bending the rules on an in-class exam (the bravado may have reflected self-denial that such conduct constituted cheating). The professor had asked us not to copy and paste pre-written exam answers during the test, explaining that he considered this cheating, but he acknowledged he would have to rely on our honor to comply with his wishes. As with many classes at HLS, old exams for the class were posted online to help students practice and prepare. The student in question had realized he could copy into his final a detailed analysis of a tough and fairly obscure theoretical issue he had written for a practice exam. I presume this passage went beyond bland doctrinal statements and may have benefited the student in an exam that many were unable to finish in the allotted time. The student was on the law review and went on to an elite clerkship.

I like to believe that law students - no matter how prestigious the school or how insecure the student's future employment - are honorable and do not cheat. But law school is too much of a game for too many people, and I think students delude themselves regarding what is or is not cheating. My key take-away from this experience: the student and most of the other students to whom he told this story did not see anything wrong with what he did.

Posted by: another former HLSer | May 6, 2013 7:23:25 PM

I can say with certainty I never heard of cheating on exams, take home or otherwise.

First off, and perhaps this is my naivete, but these classes are on a curve (or at least they used to be) -- why on earth would I want to help somebody else once the test has started?

Second, though related to the first, those who would be cheating are not the people getting As (I would suspect), so while it would still bother me a great deal to learn my fellow students were cheating, I would be a little surprised to hear that cheaters ended up getting top scores.

Third, the idea of a closed book take home is stupid. If you want to do a closed book, do it in-class and make it a shorter exam. Have some damn mercy.

Posted by: former HLSer | May 6, 2013 3:15:14 PM

If we are going to give grades that comparatively rank students, I think that we have to make our procedures as robust against cheating as can be reasonably done. My policy is that I don't impose any restrictions on students that I am unable to actually enforce. My grading procedures cannot rely upon compliance with an honors code, because then I would be really grading the inverse of compliance with the honors code, a behavior that is likely not correlated with the variable I am trying to measure.

I don't see the prima facie pedagogical or assessment value of allowing students to work at home. But, assuming there is some reason, then it is necessary to allow collaboration, and of course make it open-book and open-internet. Perhaps a collaborative exam has more ecological validity for measuring the capacity to succeed as a lawyer, in a collaborative profession.

If collaboration is unacceptable to the prof or the College (likely), then enforce solitude, rather than merely request it. If an 8-hour non-collaborative exam is necessary (which I doubt), then do it where it can be observed. The bar exam takes all day (actually multiple days) without letting people work at home unobserved.

BTW: The students we trust (i.e., the nerdiest ones who are least likely to cheat) are likely not aware of whatever cheating may have occurred. Indeed, they are the unwitting victims of cheating, by those who are trying to displace them in the relative ranking. Cheaters are unlikely to disclose their tactics to the victim or to anyone else.

Posted by: ctr | May 5, 2013 6:59:19 PM

I suspect that cheating is more rampant at lower-tier schools than at elite schools, simply because the stakes are so much higher. Even in this economy, all but the very bottom of the class at elite schools are able to secure good jobs...of course, at the lower tier schools things are very different.

I don't mean to imply that this excuses cheating in any way. In fact, I think that the strong temptations to cheat must be deterred with strong penalties for getting caught - e.g., getting suspended/expelled, not being able to sit for the bar, etc.

Posted by: Anon First Year Prof | May 2, 2013 9:27:30 PM

At my school, a contracts professor used old exams (s)he had previously used at another law school in the same city, which kept the exams on a secured webserver for its students' use. Naturally, some of our students had friends who shared the password to the exam bank. You can imagine the result.

Posted by: anon | May 2, 2013 5:32:25 PM

Anon Prof, thanks. Your approach sounds right for your objectives and for your kind of course. I have a number of reactions -- including on the accommodations fairness question you pose, and about the fairness of longer take home exams for students -- but these are separate topics in their own right, and don't take away from the point that a diversity of choices is possible and may be desirable.

Maybe it would have been best to ask commenters to differentiate between first year, other required courses, and elective courses, as well as between curved and non-curved environments. My guess is that the prevalence and impact of cheating varies substantially among these contexts, and among institutions, but I really don't know, which I think was the point of the original inquiry.

Posted by: Ed | May 2, 2013 9:35:13 AM

Wow, Anon Prof must be reading my mind. To that answer, I'll add that I give the take home in patent law, a course where I provide an invention and prior art. I don't want a student's score to reflect differential time spent figuring out the invention (no matter how simple I thought it was) and how the prior art might apply. I promise my students that they don't need any kind of technical background to take the course, and this is one way I fulfill that promise.

Posted by: Michael Risch | May 2, 2013 7:45:36 AM

Answering Ed: Part of my calculus has to do with my style of exam, part has to do with the students I teach, and part has to do with my resignation that there is cheating in the in-class context too.

The key is that you are presuming that there is anything fair or transparent about the in-class exam or about how grades appear on a transcript. There is no perfectly "fair" assessment, esp not when we're talking about in-class vs. take-home exams. Grades on our transcripts don't indicate what kind of assessment the professor used, so anyone seeing the grade *doesn't* know what kind of exam conditions produced that grade. Grades are inherently non-transparent.

The ability to read and write very quickly (as in in-class 3 hour) is not a skill that I wish to test or reward. I do not believe it has significant value for a practicing attorney, and it's not something I teach. [It does have value on the bar exam, so it is important students practice this - but it need not be when they are ALSO being evaluated in every law school doctrinal course they take.]

At my school (I can't speak for anywhere else), I've noted that my students have vastly different processing speeds for reading and writing that does not correlate with actual knowledge or ability to analyze or solve legal problems. (I suspect the difference is more in the reading than in the writing but I don't have enough data to back up that belief.) At another law school I taught at, the variance was not nearly as profound. I believe that the fairness that comes from removing this advantage/disadvantage affects many more students than the costs imposed on everyone created by the marginal few who might use the switch to a take-home as an opportunity to cheat.

There also have been an increasing number of students who receive accommodations (e.g., time and a half/double time) to complete exams. I believe in reasonable accommodations for demonstrated learning disabilities. However, I am also aware that having a diagnosis (esp one that is long-term/long established pre-law school) can be a privilege of socioeconomics that other students with similar disabilities lack. The fairest solution to me is to use an exam that both reduces the challenges that students with time-related disabilities face, while eliminating any advantage that might be gained over students who really are similarly situated.

In addition, one component (not all) of my take-home exams is often an issue-spotting essay. I grade in a way that rewards students for prioritizing the importance of their insight/answer. I do not add up points for "more issues spotted" vs. fewer, and I do not give heightened credit to a student who spots the most esoterical issue that few students catch while neglecting something - or being flat out wrong - in a way that would constitute actual malpractice. The best answers solve problems in ways that would be useful if the problem was a real one. While students can cheat, this reduces the value or benefit from a crowd-sourced or collaborative because having "more" issues doesn't answer what's most important.

Last, I'm not naive about cheating. At my school, there are also problems with cheating or the fairness of the application of the rules of in-class exams as well. I do not believe that I can block ALL cheating, so I am focused on the marginal trade off from the cheating that can occur in our in-class framework vs. moving to a take-home approach.

Posted by: Anon Prof | May 1, 2013 10:59:41 PM

Michael, thanks. Look, I'm happiest if there is no cheating, and instructors can choose whatever means of examination they think is soundest. My question to Anon Prof, however, concerned some trade-off he was positing, and I was interested in why he thought an advantage in terms of his favored assessment was worth an increase in the prospect of cheating.

Your answer, I think, is that the impact of cheating is more attenuated (because you lack a curve) and that cheating isn't likely to get past you. As to the former, I imagine things are still somewhat relative, but perhaps not under your scheme. As to the latter, let me just say that some students might share the gist of their answers -- so as to ensure that none of them missed a big issue or misunderstood a key point of law -- without raising a substantial risk of detection. I have no idea whether that happens, but I would be realistic about the ability to control it.

Posted by: Ed | May 1, 2013 9:36:12 PM

Ed -

I'll give my answer to your question. Anon Prof could have been me (but it wasn't). First of all, I only do the exam in a non-curve class. Second of all, as I noted above, the answers I get vary widely in quality, and vary in different ways. I have an _extremely_ detailed rubric, and I can tell where students fall down or not, and very rarely do students (especially those who take the exam on the same day, since I allow them to pick any day of the week) fall down in the same way. Additionally, where there are similar responses, they are often not the best responses. That's the point of the assessment, of course - only a few people will get the really hard stuff. If everyone started getting the really hard stuff, or if everybody started making the same mistakes, or, quite frankly, if everyone did just so-so in the same way as each other, I might worry a bit more than I do.

Posted by: Michael Risch | May 1, 2013 7:47:50 PM

At my school, we had a wide-open testing policy (tests generally weren't proctored, and you could take your exams anywhere you pleased--a few students who lived nearby would even pick up their exams and go home to take them).

I graduated just as smartphones were taking off in earnest (2010), so even with ExamSoft, the opportunity to cheat was right in front of us.

That said, I knew of only one student who cheated in all three years. He copy/pasted from a treatise without attribution, and our Honor Court came down on him quite harshly.

Now, did more than just that one man cheat during those three years? Of course. But the fact that I never witnessed it firsthand nor heard about it through the grapevine (in a place where secrets are not kept long and everyone knows everyone else's business) suggests that it may be surprisingly rare.

If any institution is ever silly enough to give me a role in molding the minds of future lawyers, I'll default to trusting them not to cheat, and pray that they don't.

Posted by: Leland Unruh | May 1, 2013 5:45:36 PM

"I think the benefit to the cheaters and harm to everyone else is worth the trade-off of the advantages of the style of exam I give for the specific courses I use them for."

Could you explain why you think that? Or what prospect of cheating would lead you to rethink that calculus? For me, any substantial risk of non-transparent unfairness disqualifies a means of assessment. By "non-transparent," I mean to acknowledge that timed in-class exams have their own fairness issues, but at least those are readily understood by those viewing a transcript. Grades are consequential, and particular in the face of a mandatory curve, the prospect that some students may cheat and secure better grades than those who do not weighs pretty heavily against theoretical advantages in the substance of an assessment.

Posted by: Ed | May 1, 2013 5:00:18 PM

I give take-home exams, and I realize there is no effective method of catching cheating.

One key in limiting the damage of cheating is to give:
- an open world exam (any form of materials or research is allowed *except* collaborating with other people)
- word limit
- credit only for applications of law to new facts/problems posed, not unconnected statements of law or rules

This removes the issue of cut and pasting of material entirely. It just doesn't prevent students from privately working together. It is nearly impossible to detect that, especially because students may be relying on a shared corpus of materials - notes, class outlines, slides posted, etc. so similar language isn't indicative of cheating. The other big issue is that my exams are given during a range of dates, so there is always the issue of students feeding questions to another who hasn't taken the exam yet. I warn them of the severity, but can't prove it doesn't happen.

In one rare instance, I started to get deja vu when grading that an exam was so similar to another that I had graded previously in a large class. When I went back to check, I learned why: there was a duplicate print-out and I was grading an exam I had already read. At least, I noticed.

In general, though, I think the benefit to the cheaters and harm to everyone else is worth the trade-off of the advantages of the style of exam I give for the specific courses I use them for.

Posted by: Anon Prof | May 1, 2013 4:34:36 PM

I give take-home exams, and I realize there is no effective method of catching cheating.

One key in limiting the damage of cheating is to give:
- an open world exam (any form of materials or research is allowed *except* collaborating with other people)
- word limit
- credit only for applications of law to new facts/problems posed, not unconnected statements of law or rules

This removes the issue of cut and pasting of material entirely. It just doesn't prevent students from privately working together. It is nearly impossible to detect that, especially because students may be relying on a shared corpus of materials - notes, class outlines, slides posted, etc. so similar language isn't indicative of cheating. The other big issue is that my exams are given during a range of dates, so there is always the issue of students feeding questions to another who hasn't taken the exam yet. I warn them of the severity, but can't prove it doesn't happen.

In one rare instance, I started to get deja vu when grading that an exam was so similar to another that I had graded previously in a large class. When I went back to check, I learned why: there was a duplicate print-out and I was grading an exam I had already read. At least, I noticed.

In general, though, I think the benefit to the cheaters and harm to everyone else is worth the trade-off of the advantages of the style of exam I give for the specific courses I use them for.

Posted by: Anon Prof | May 1, 2013 4:34:34 PM

As a former student, I'll say that I hated timed take home exams and exams where an outline was banned for this reason.

In some classes I was literally given a packet with the exam and told to spend no longer than 3 hours preparing my answer; I followed the rules, but given the overwhelming incentives to cheat, I would bet a substantial amount of money that some folks were cheating.

In another class, we were told that we could not use an outline, but we took our exams in Word on our personal laptops with no monitoring software and much of the test involved detailed Trust & Estates questions. Again, it strained credibility to the breaking point to believe that a bunch of anxious students concerned about getting good grades to have any hope of paying off their loans (and who, by the way, have good reason to suspect that others are cheating), were not cheating in that circumstance.

My personal view is that when law profs wax on about the integrity and dignity of the legal "profession" in these types of debates it's because they have nothing on the line and aren't that concerned about whether the process is fair. Ask yourself this, though: if getting good grades on these exams was the difference for you between $160k a year and being hired back by your law school for a year with no job in sight after graduation, how would you like the exams to be administered? What procedural safeguards to ensure honesty would you put in place?

Posted by: John | May 1, 2013 3:30:30 PM

Regarding the failure to report -- if someone sees unequivocal evidence of cheating but fails to report it, that should itself be an honor code violation. (It was at my UG institution, and I suspect some other schools have similar policies.) I'm somewhat surprised by the number of persons on this thread who report widespread cheating but who don't seem to have done anything about it, or at least as far as I can tell from their posts. Unproctored exams might be more palatable if administrators and fellow students could trust observers to report what they saw. And I view those who fail to report as only slightly less culpable than the actual cheaters -- you are letting down both your peers and your institution.

While I was in law school, I largely cocooned myself off from the rest of civilization, so I didn't see any cheating. (Drug/alcohol abuse seemed to be a more common issue around exam time.) I would have thought that the competitive nature of law school discourages most cheating -- a student might not want to share his brilliant insights on an exam, since that would just allow others to catch up to him.

In any event, it does seem like some schools are able to successfully implement an honor system, and I'd be curious about what makes those schools different than others. Also, I'm really hoping someone can answer my question about whether exams are proctored at divinity school -- I did a quick Google search of the Bible, but the answer isn't there.

Posted by: andy | May 1, 2013 2:07:55 PM

Why don't students report?

Because actual collaboration is hard to prove in school disciplinary proceedings, because it devolves into a he-said-she-said (or sometimes one student's word against the word of two or more collaborators), because being known as a snitch is not likely to make you many friends, and because schools may sweep the stuff under the rug anyway. It's basic stuff people learn in the schoolyard.

I'm sure practicing lawyers can tell of many times they observed or heard of instances of misconduct by other lawyers at their firms or organizations, even though the misconduct places them at risk (much like cheating hurts other people on the same curve).

There was one incident I remember, where a friend was working in a group take home exam with several other students. It was discovered that one of the students had plagarized (copied) large portions of his assignment, exposing the other members of the group to potential honor code violations. The other students wrestled with the decision to report for some time. Some didn't care, others cared immensely. They ended up reporting, and the student wound up with a slap on the wrist.

Posted by: Anon | May 1, 2013 2:04:36 PM

Ed makes an interesting point. I was going to ask all those former students who saw cheating why they didn't report an honor code violation. The failure to report undercuts andy's view that students should be trusted to be professionals.

I guess I don't view the cutting and pasting of a couple intro sentences cheating, so long as the exam is not closed computer (and how can it be, if you can use word and copy in from another document). I guess I'm just not too worried about that kind of cheating.

First, the type of cut listed by anon would get close to zero points on my exam, since it tells me nothing about the actual facts of the case. Second, it would be a total waste of word limit space to prepaste long arm discussion if it's clear there is one. Third, I'm not convinced that it is easier to think about the rest of the exam while you hunt for the appropriate snippet to copy and paste. You could type out a couple sentences (I could, at least) just as quickly, and typing out the rule helps you think about how the facts might apply to it - something that gets lost in the cut and paste.

Posted by: Michael Risch | May 1, 2013 1:02:09 PM

This is an interesting if sobering exchange; personally, I am not surprised by what's being said, and this may be one of those instances in which anonymous commentary has a relative advantage (given that identifiable students will likely under-report cheating not only by themselves, but also by their peers). I wonder if any of those commenting could relate their experiences to self-scheduled exams -- which might, in theory, operate differently than a take-home exam that all students took more or less simultaneously.

Posted by: Ed | May 1, 2013 11:44:53 AM

Of course there is cheating. I saw some of it during law school and heard rumors of more. Students cheated on the law review takehomes and on exams, and it's not inconsequential. Beings able to crowdsource issue-spotting and to compare arguments with another person puts you leagues above someone doing it on their own.

The most blatant example was when a group of about 8/24 students all made the same "translation mistake" on a takehome exam. The students had been doing it in a group. They actually had the gall to complain to the professor.

Posted by: BoredJD | May 1, 2013 10:47:15 AM

"At issue is whether the court has personal jurisdiction over [insert party], such that it may subject the party to a binding judgment. The exercise jurisdiction must comport with both the state's long-arm statute and due process requirements.

The long-arm statute provides [etc.]"

anon - you would get nearly zero points on my exam for that part of the answer, and would have wasted some your limited words.

As for the premise that this helps you think, I think it actually diverts thought to be looking for the appropriate snippet to paste in when you could be typing it nearly as fast while actually thinking about how you might apply the facts of the exam to the rule. But that's just me - I guess that's why I'm just not that worried about the pre-writing type of cheating. Collaboration, on the other hand...

Posted by: Michael Risch | May 1, 2013 10:03:34 AM

I had professors in law school who did *not* outlaw the pasting in of pre-written materials, but who advised against it. They said it would be silly and would only hurt you. I think that position reflects a lack of imagination. The truth is that exams tend to repeat common themes over and over, and you can do a lot to structure and answer and drop in cases, etc.

For example, if you know that there's a good chance that personal jurisdiction is coming on a Civ Pro exam, you can say:

"At issue is whether the court has personal jurisdiction over [insert party], such that it may subject the party to a binding judgment. The exercise jurisdiction must comport with both the state's long-arm statute and due process requirements.

The long-arm statute provides [etc.]"

The bottom line is that there are often predictable, multi-step analyses to engage in, each with nuances and applicable cases. You can put that in an outline form, or you can put it in prose to drop in. If the prose is pre-written, you have more time to think about other stuff.

Posted by: anon | May 1, 2013 9:40:04 AM

AnonProf44 - this was for pre-written stuff? Still seems not terribly helpful for an actual answer (note to my students reading this in the midst of a take-home...)

Posted by: Michael Risch | May 1, 2013 7:51:12 AM

Re fellow: exactly. I'm confident that most of my students are not cheating. Indeed, this semester I have a lovely, small class in which I know the students relatively well and I'm pretty confident none of them will cheat. But "most" and "pretty confident" aren't good enough. So after some experimentation with take-homes, it's back to in-class exams for me.

Posted by: Joey | May 1, 2013 1:46:42 AM

For those of you who ask trusted students whether there is cheating: I believe you that these students don't know of any. But the sort of student who becomes close enough to a professor to be trusted may not be the sort of student who will be plugged into the circles doing the cheating.

Posted by: fellow | May 1, 2013 1:21:37 AM


It was obvious that the exams had been copied. The students met in an apartment, co-wrote the exam on a GoogleDoc, then tried to differentiate, by doing things like changing "fee simple determinable" to "a determinable form of fee simple"; one would spell out a term like "Uniform Real Property Electronic Recording Act" while the other would refer to "URPERA". There were also some identical odd formatting errors, like

"missing from the chain of title, In addition, the sales contract did not provide"

where the comma should be a period, and there are three spaces, not two or one, between the sentences.

Posted by: AnonProf44 | Apr 30, 2013 11:36:45 PM

The stakes are too high and the chance of detection too low to believe that no cheating occurs. There are also probably many self-deluded students who don't see the "complete the exam and compare answers with friend" method as full-fledged cheating. And don't forget that many students have helicopter parents who are also lawyers . . .

Posted by: GU | Apr 30, 2013 11:33:23 PM

AnonProf - how did the cut and pasters do? I can spot outline dumps a mile away and always advise my students not to do it.

Posted by: Michael Risch | Apr 30, 2013 9:20:37 PM

I was persuaded to move to an in-class closed-book exam after reading BL1Y's confession to cheating in Sam Isaacaroff's civ pro class at NYU ("I cheated in law school. It was easy, it was worth it, and looking back, I should have done it more.") The full piece is disturbing, but worth reading. It is available at: http://www.constitutionaldaily.com/index.php?option=com_content&view=article&id=650:here-is-a-dime-use-it-to-phone-a-friend&catid=38:there-and-never-back-again&Itemid=65

Posted by: CBR | Apr 30, 2013 8:48:29 PM

The law school (top 5) I attended used word for in class exams, and the amount of students that cut and paste pre-written portions of answers was significant.

Posted by: AnonProf | Apr 30, 2013 8:26:03 PM

There is cheating. The data is overwhelming. Here's just one example: according to one large study of more than 50,000 college students on 60 campuses, more than 70% of students surveyed admitted to having engaged in some form of cheating. Admitted. Nearly 80% said that submitting a cut-and-paste job from the internet was not a serious problem. The fact that 70% willingly admit to having cheated suggests not only the obvious conclusion that cheating is routine, but that social opprobrium no longer attaches to it.

Unfortunately, recognizing cheating on take-home exams is difficult. The problem is that collaboration is hard to detect when it results in good answers. Good answers tend to resemble one another, and there are lots of different ways students can arrive at a good response. The people who generally get caught engaging in unauthorized collaboration are the ones who produce similar answers, but the answers are loony. It is much harder for students to explain plausibly how they independently arrived at the same loony answers than how they independently wrote sound answers.

In the last few years, I've taken to giving explicit instructions on my take-home exams. My instruction begin: "You are to take this examination in solitude. This means that during the entire period of the exam you are not to speak to anyone, take the exam in the vicinity of anyone, or take a break or dine with anyone. During the period of the exam, you are not to make or receive a telephone call, text, or instant message, or engage in any other form of communication with any other person...." And that's only the beginning.

I don't think this has been a success, and in the fall I am going to revert to old-fashioned, sit-down, closed-book exams. It isn't the most realistic form of examination, but it's the only way I can see to eliminate the possibility of cheating.

Posted by: Jim Gardner | Apr 30, 2013 7:47:08 PM

Michael, I agree that cheating should be severely punished and should, in ordinary cases, prohibit admission to the bar.

Like you, I attended schools where I was not treated as a potential felon. That undoubtedly shapes my views on this matter. Our students are budding professionals, and I believe that we should treat them that way. But I understand I'm in the minority on this. And, faculty-student respect issues are just one aspect of this. Perhaps it's fairer to the honest students to take measures to prevent cheating, and this justifies proctoring, in-class exams, and so on.

Posted by: andy | Apr 30, 2013 7:36:35 PM

Sadly, I'm now on the side of "I'd rather give a take-home pedagogically, but it seems better to go with in-class just to avoid this problem."

Many of my students are friends with one another. I worry that with a take-home, talking with one another is a form of cheating that will be too tempting for some. Probably not many, but some -- and that's just so unfair to everyone else that I'm going with the (otherwise inferior) in-class approach.

I make my in-class exam as long as possible (4 hours) and have a word limit, so hopefully I avoid the "write frantically till you drop" experience that I see as the worst problem with some in-class exams.

Posted by: Joey | Apr 30, 2013 7:24:21 PM


Sorry for the ambiguity; I should have written "As I was doing a tenure review of another professor, I discovered that that professor had cheated in the process of preparing the work which I was reviewing."

Posted by: AnonProf44 | Apr 30, 2013 7:14:18 PM

How do you cheat in a tenure review?

Posted by: Michael Risch | Apr 30, 2013 7:03:29 PM

How do you cheat in a tenure review?

Posted by: Michael Risch | Apr 30, 2013 7:03:29 PM

I give a takehome, and I require students to re-sign an honor code form. I realize that it won't cause cheaters to not cheat, but maybe it will deter a bit.

I've found that the best way to limit collaboration cheating is to demand really great exams and to limit word counts. Thus, even collaborators will have to make choices about what to say. And I've found (maybe because I only do take-homes in small classes) that students do well and do poorly in different ways. And when they all say the same thing, it is usually the superficial answer that yields little extra points.

In other words, if my students are cheating, they are doing a poor job of it.

I think this is a good article on collaboration and exams as well: http://www.citywatchla.com/8br-hidden/4982-why-i-let-my-ucla-students-cheat-on-their-exam

Finally, Andy, I agree with you a bit. I went to schools that didn't proctor. I found it odd that the schools where I teach do proctor, but I've gotten used to it. I don't think it says anything about the honor code one way or the other with respect to take home exams, where there is no proctoring. People will cheat or they won't, and the punishment for cheating in law school should be severe.

Posted by: Michael Risch | Apr 30, 2013 7:03:07 PM

Andy, the fact that dishonest lawyers can do a lot of damage does not imply that we should treat all lawyers as honest people. Better to lessen the opportunity to do damage, I think.

Posted by: Orin Kerr | Apr 30, 2013 6:57:19 PM

Also, I know that this isn't exactly a direct comparison, but does anyone know if exams (or their equivalent) are proctored at divinity schools?

Posted by: andy | Apr 30, 2013 6:42:02 PM


Fair enough, but regardless of what any certificate says, we are still training our students to enter a profession where dishonest persons can do a lot of bad things. If we don't trust our students to take an exam, why should we give them the requisite training to become members of a bar?

Posted by: andy | Apr 30, 2013 6:40:16 PM

Andy, my understanding is that the "Dean's Certificate" is mostly a report on whether the law school has ever taken disciplinary action against the student. I think that each state's form is different, but the question about the student's honesty generally just asks about whether the law school's *records* indicate that the person is honest. Thus, the Dean generally is not asked for and does not make a personal assessment of the honesty of each student (which would be hard for the Dean to do).

See, for example:
https://www.ilbaradmissions.org/getpdfform.action?id=19 (Illinois)
http://www.floridabarexam.org/public/main.nsf/form3.pdf/$file/form3.pdf (Florida)
http://www.jud.ct.gov/cbec/July11/Form4.pdf (Connectciut)

Posted by: Orin Kerr | Apr 30, 2013 6:36:06 PM

I caught two law students cheating on a take-home; it was an essay exam and their answers were copied in substantial part. No more take-homes.

I also caught a professor cheating in a tenure review. Both incidents were very, very disappointing.

Posted by: AnonProf44 | Apr 30, 2013 6:27:14 PM

Why on earth would you consider the responses to you questions as honest? Of course there is cheating, only law professors seems unaware of it.

Posted by: MM | Apr 30, 2013 6:25:20 PM

How do anti-cheating measures comport with the school's certification, after graduation, that a student is honest and properly admissible to a bar? (I think most states require some type of "Dean's Certificate", signed by the Dean, stating that the graduate possesses those characteristics.

It seems to me that, if we are going to certify that a student is honest , we should trust him or her to not cheat on an exam. When I received my LLM at Georgetown, our exams were administered without continuous proctoring, and we typed our exams using Microsoft Word, not Examsoft or whatnot. This approach, I think, shows proper respect for students.

More generally, it seems odd for us to certify to a State Bar that our students are honest if we take measures to question their honesty.

One might say that the temptation to cheat and the rewards for cheating are greater in the law school exam context, as compared to the legal practice context, so prophylactic measures are justified at the stage of the game. I'm inclined to think the opposite, though -- if we are going to trust our graduates to handle cases, their client's money, and so on, we should trust them to take an exam without cheating.

Posted by: andy | Apr 30, 2013 6:07:20 PM

I used to do a 24-hour take home; like Glenn, I asked students I trusted if they knew of any cheating and was told no. But I've never been sure. My solution has been to move to a mix--students do an essay portion at home (and have several weeks to do it) and an in-class mix of M/C and short answer, with the latter worth much more of the final grade.

Posted by: Howard Wasserman | Apr 30, 2013 5:11:52 PM

When I was a 1L at HLS, I heard some believable rumors about cheating on take-home exams in another section along the lines you fear. (All of our 1L exams for our section were in-class, so the issue didn't arise.) For that reason, I have never given a take-home exam, and I don't think I ever would: It's just too tempting to collaborate, incredibly easy to do, and pretty much impossible to detect.

Posted by: Orin Kerr | Apr 30, 2013 4:28:44 PM

The comments to this entry are closed.