Following Cheating Scandals, Harvard Dean of Undergrad Ed Visits CS50 Class and Tells Students Not To Cheat (thecrimson.com)

theodp writes: After a flood of cheating cases roiled Harvard's Computer Science 50: "Introduction to Computer Science I" last year, Dean of Undergraduate Education Jay Harris implored students in the course not to cheat on assignments at an orientation session Wednesday night. Course head David Malan, the Harvard Crimson reports, spent the last five minutes of the orientation session fielding questions from students confused about the course's collaboration policy and whether or not CS50 enrollees are allowed to use code found online. He told them never to Google solutions, and never to borrow a friend's work. Last week, CS50 students were informed via a CS50 FAQ that they are also now "encouraged" to physically attend the course's taped weekly lectures. In an essay last year, Prof. Malan had questioned the value of saying everyone should attend every lecture. Attendance is now also expected at every discussion section until the first mid-semester exam. In case you're curious, the estimated sticker price for attending Harvard College during the 2017-2018 school year is $69,600-$73,600 (health insurance sold separately).

Comments

Well, that solves that!

[aardvarkjoe]

The problem was that nobody told the students not to cheat. Now that that little misunderstanding has been cleared up, the problem is fixed.

Re:Well, that solves that!

[infolation]

He told them never to Google solutions

To me the Google issue is part of a deeper problem with learning and information-retention in a search-engine-driven world.

As a child, when I didn't understand a word or concept, my parents would never provide the verbal definition; we had to fetch the thick dictionary or encyclopedia and look it up ourselves. Later on they told us the effort needed to look things up helped us remember the new words, historical events etc.

Nowadays it seems a person's brain intrinsically knows it doesn't need to remember things or work things out for itself. At a subconcious level, the brain knows it can find the solution to the problem at hand by googling, instead of 'reading around the subject' until it finds the full information needed to work the answer out. I really believe this impedes retention.

No matter how much you want to remember things or work things out yourself, your brain somehow knows it doesn't have to.

Re:Well, that solves that!

[jandersen]

The problem was that nobody told the students not to cheat. Now that that little misunderstanding has been cleared up, the problem is fixed.

The problem with cheating is one of what is perceived as socially acceptable, and it isn't limited to cheating with your education. It is the same mechanism that lies behind, say, binge drinking amongst teenagers in UK, low level tax evasion in Denmark, social benefits fraud, using your mobile while driving, corruption etc etc: it has somehow become socially acceptable - "everybody" is doing it. People have somehow persuaded themselves that it doesn't cause real problems, and to address it, somebody has to start telling people that it is not actually acceptable because it does cause real harm.

In the case of exam fraud, the cheater may benefit in the short term, but if it becomes commonplace, then the profession as a whole suffers, and the industries that rely on employing genuinely competent, highly educated people get into trouble. At the moment, Indian university degrees are considered with some skepticism - but the same can happen to American degrees. Basically, it is about trust: it takes surprisingly little effort to break, and it is very hard work to rebuild it.

How I do it

[XxtraLarGe]

Right now, I'm teaching a Web Programming & Database integration course. I do a flipped classroom model where I record the lecture, and we work on the homework in class. They can do the homework before class, but they have to show me their code & explain it before they hand it in. That way, I can catch any errors they have before they hand it in, and answer questions that they run into if they haven't finished it yet. I also know that they're doing their own work.

Also, if I see a common issue, I can do a 5 minute "mini lecture" to give an example technique in front of the class. If I come across a common issue after things have been submitted, I can do a 5 minute recorded lecture to reinforce what they should do in the future in that situation. Seems to work out well for my students.

Re:How I do it

[XxtraLarGe]

Watching a lecture and then getting one on one time for things you don't understand or follow has got to be the greatest idea ever and would be the greatest improvement to the education system if implemented across the board.

While I'd like to claim credit for it, it's called a flipped classroom, and it's being implemented all over the place. It is also very helpful that I have a small class. It is much more difficult if I have more than a dozen or so students.

Re:Do people go to Harvard to program?

[ShanghaiBill]

you can get a CS degree at any state school and they are far cheaper and just as good.

No, they are not just as good. A degree from Harvard will open a lot of doors. Not only because of the reputation, but also because of the alumni network.

It's not as if the faculty at Harvard are somehow better

It is not the faculty that is better, or the instruction, but the classmates. They learn a lot from each other. Which is why the restrictions on "collaboration" are so stupid. They are taking away the very thing that makes Harvard special.

Previosu to this, did he tell them to cheat?

[gurps_npc]

I am pretty sure that Harvard has previously told students not to cheat.

Doing it again is not useful. Try using tests (including testing conditions) that make it difficult to cheat, rather than yelling at people that do it.

Re:Previosu to this, did he tell them to cheat?

[tgeek]

Or maybe the dean should look at the course instead of the students. Is the material relevant and useful (i.e. are the students motivated to learn it) Or are they viewing it simply as a "bullshit-requirement-or-prerequisite-that-must-be-endured"? And I won't even get into the whole mixed of message of "Welcome to Intro to Computer science. The exciting field that, among other things, promotes the ability to share information. Now please don't share information or cheat."

Re:$70K? Absurd...

[93 Escort Wagon]

Indeed, most Harvard students do not pay tuition at all, because they are minorities or minority women.

Well, that is a load of crap. Roughly 50% of the Harvard student population is white. An additional 22% are Asian-American, who generally don't get classified as underpriviliged minority.

https://college.harvard.edu/ad...

Also, "minorities or minority women"?! Like the latter group magically doesn't get included in the former?

Reduce the incentive and value to cheat

[Anonymous Coward]

First things first, if the lectures are available online, I wouldn't make attendance mandatory. Instead, I'd explain to students why it benefits them to be in attendance:

1) It simply isn't possible to respond to respond to email or discussion forum questions immediately like what can be done in an in-class lecture.
2) It's far easier for an instructor to help students if the instructor knows who they are. It's not that the instructor is biased against anyone else, but it's hard to directly help students who you don't have any direct contact with.

As for cheating, reduce the incentive and value of doing so. That means eliminating any memorization questions. When I taught, I started creating multiple choice questions that asked the students to apply a particular concept to a very simple and straightforward case. It's easy to grade and yet students can't possibly look the answer up online because the question was unique to my class. There isn't a whole lot of extra work needed to create assignments that are mostly resistant to cheating. The ancillary benefit is that students learn the material better if they're asked to use it rather than simply recall it. For an introductory CS class, one simple approach is providing the students with a bit of code that's unique to the class and ask them to modify it in a relatively minor way. The grading is still simple, but it some basic competence is required even to find the appropriate solution online.

Collaboration is a tougher one, but individual exams are a way to thwart impermissible collaboration. I think the best solution is to tell students that they can discuss assignments with other students, but they need to generate their answer individually and are responsible for understanding the solution. It's completely fair to reference assignments on the exam or even to reuse questions. I've done both, and it places greater value on understanding the answers to the assignments if they will come up again on the exam. I liked this because the assignments directly helped in preparing students for the exams. It also takes away many of the potential complaints about exams being unfair if they are given a very good idea of what's going to be on the exam. My view is that there shouldn't be any deception about exams -- be upfront with students about what's going to be on the exams. However, make the questions require an understanding of the material and the exams will still be challenging. If the assignments directly relate to the exams, there's far less incentive to copy answers on them.

I believe that a lot of cheating can be stopped just by designing the course a bit differently. You'll still have students who try to be lazy and get by, but they'll eventually get exposed by the exams anyway.

Re:Do people go to Harvard to program?

[epyT-R]

No, they are not just as good. A degree from Harvard will open a lot of doors. Not only because of the reputation, but also because of the alumni network.

So it's an expensive popularity contest. He was talking about the quality of education.

test for cheating

[stabiesoft]

I recall one of my CS classes, the prof gave an assignment. You were guaranteed to run into a problem. Way back when, the batch file had a limit on memory usage and the assignment needed more than you had. So you went to the prof and he would put a check by your name and tell you how to up the resource. At the beginning of the assignment he was very clear. Do not cheat and that included asking classmates about anything about this assignment. If you did not go see him, you flunked the class. A clever prof figures out ways to detect cheating.

Re:test for cheating

[Anonymous Coward]

If You define asking classmates about anything about this assignment as cheating, You had already failed as a teacher. Not only that, it is even a retarded way of detecting cheating -- it does not prevent cross-year cheating, and it is enough for one clever person to be caught cheating to ruin this really stupid and manipulative thing.
A clever teacher knows already that whatever punishments You employ, they only serve as perverse incentives at best and they always end up punishing upright students.

Re:test for cheating

[Anonymous Coward]

Not to mention that the best students - the ones that can identify a problem, and research how to solve it own their own - would be the ones punished.
The ones that "passed" were the drones that did exactly what they were told, in exactly the way they were told, and went back for more instructions at the first sign of trouble.

Re:test for cheating

[ByteSlicer]

So if you knew enough about batch files and memory to fix this on your own (I know I did, back then), you'd be branded a cheater? The problem with "clever" profs is their ego is so big, they'd never consider a student being smarter than them.

Re:test for cheating

[karmatic]

Fuck teachers who do things like that.

I was failed out of a class (at Harvard no less) for doing too well. The instructor didn't think it was possible for me to do as well as I did, in the time I had, and determined that I had cheated on the final project.

Some of us don't need to ask classmates - I was the nerdy autist who (quite literally) asked for DOS manuals for birthday presents as a child. Dealing with extended memory, high memory, upper memory - all of this was common when dealing with video games, and it was often necessary to move things around in memory to deal with (for example) the CDROM driver (or mouse driver) using up memory in a way that caused issues with DOS Protected Mode apps. Or, TSRs would cause conflicts. When teachers assume that students are incapable, they are bad teachers.

If you are talking about a limit on a server (rather than a batch file), then you are likely talking about unix quotas and limits. These are also common limitations, and one I ran into on the FreeBSD nodes that I would use. I was in around 5th grade at the time.

If I ran into a resource limit on someone else's system (for example, a shared CS unix box), I'd check the quota and see if my limit is soft or hard. If it's soft, the command `ulimit` would let the soft limit be raised up to the hard limit, likely fixing the crash. If it's a hard limit, I'm calling IT, not the professor.

It's possible you actually meant batch file, but when you say "needed more than you had", that's an interesting assumption if people are permitted to use their own computers (unless you are talking about the native DOS memory limitations, which can be fixed with LOADHIGH, which gets you out of the DOS conventional memory area.

Re:test for cheating

[karmatic]

It also assumes they don't already know the answer. The class may prohibit checking outside sources, but there are always other ways to solve the problem.

Re:test for cheating

[chihowa]

One of the points of the assignment was to understand how to work within the limitations of practical work.

Was that a specifically communicated point of the assignment or just something that he tacked on when you complained? It sounds like he was just punishing you for bruising his ego. I've been in the same situation that you were in...

Part of teaching somebody how to work within limitations is to use actual limitations. Rewriting his code was obviously not an actual limitation, because you did it, so if he wanted it to be limiting he should have specified that. These some years later, it's clear what the limitations are and you I hope that don't fire people for not working within limitations that you never communicated to them.

Re:Sticker price?

[doctorvo]

And why the fuck is it so god damn expensive?

That's easy: government subsidies and government student loan guarantees.

Re:Do people go to Harvard to program?

[ChromeAeonium]

I have a very hard time believing that. I can't speak to the quality of Harvard's education because I've never stepped foot in the place, but I've met a few people who went to those so-called 'elite schools.' They didn't seem any better or worse than those of us who could not afford the feeder schools, extracurricular activities, and other elements of institutionalized classism. The only difference I could tell was that they had nicer clothes and more expensive hobbies. Pardon me if I am highly skeptical of the claim that there's anything special going on there beyond networking and brand recognition, but I for one would much rather work with an Iowa State grad than someone from Harvard. At least I know which one has the ability to justly earn their position.

Re:Do people go to Harvard to program?

[stealth_finger]

No, they are not just as good. A degree from Harvard will open a lot of doors. Not only because of the reputation, but also because of the alumni network.

So it's an expensive popularity contest. He was talking about the quality of education.

Dunno about most people but when I hear Harvard I think rich, not smart.