SoylentNews is people
SoylentNews
Kevin Chen writes a post in his blog about incentives and scaling from his two years as a teaching assistant. Specifically in his current post he addresses plagiarism in computer science and why it has still not stopped.
The most important goal is to keep the course fair for students who do honest work. Instructors must assign grades that accurately reflect performance. A student who grapples with a problem — becoming a stronger programmer in the process — should never receive a lower grade than one who copies and pastes.
Finally, as educators, we also hope that the accused student can learn difficult lessons about ethical behavior in the classroom rather than the workplace.
From his experience, every semester somewhere between 10% to 40% of the students carry out blatant, indisputable cases of plagiarism with an unknown amount of less clear cases left unaddressed. How does this match with soylentil's experiences here, either in computer science or other fields? The perspectives are likely quite different from institution to institution as well as whether you are still studying in college or university, recently graduated, or in a teaching role.
(Score: 4, Interesting) by Virindi on Monday March 26 2018, @11:13AM (15 children)
I did a CS undergrad program in a mid-tier university in the US. In one senior level class, there was a large final project which was worked on during the whole semester by groups of students, who were able to somewhat choose their approach by themselves within the confines of a general type of problem. At the end of the class everyone presented their project.
The vast majority of solutions were a minimal response to the presented problem, and barely successful at that. Two or three groups obviously had a good grasp on the problem (based on their explanation) and went out of their way to solve something more difficult than the minimum task.
One group presented a piece of software that solved the problem as well as you could imagine, with flashy 3d graphics during the process. They described absolutely no detail and essentially just said, "uh yeah, here is our program". During the question and answer part of the presentation, they gave extremely vague non-answers to every question. They could not even describe the basic approach they had taken to arrive at the solution beyond repeating a few keywords that had been discussed in class.
Afterwards I discussed this with the professor. I said something like, "Those guys were pretty obviously cheating right? I mean there is no way they came up with that themselves? What is the deal with that?"
His response was, "Yes, I am pretty sure they didn't write that, but it is too much difficulty to accuse them of plagiarism. So I am just going to let it slide and give them a B."
Maybe it would have been better if the presentations were recorded, so the inability to answer even basic questions about one's "work" was available as evidence.
(Score: 5, Interesting) by BsAtHome on Monday March 26 2018, @12:23PM
As a teacher, I use to say to my students "It works but I don't know why" is failure. If it doesn't work, but you can explain what and why things went the way they did, you still get a good grade. Neither is an encouragement to do half work.
(Score: 1, Insightful) by Anonymous Coward on Monday March 26 2018, @12:36PM (3 children)
Problem with penalizing those guys is that they would be insanely effective in the real world, especially when they have a flashy demo, as long as the product does what it is supposed to that's all the business cares about.
(Score: 5, Insightful) by Virindi on Monday March 26 2018, @12:48PM (1 child)
If cheating is the way to success in business, then we need to do what we can to change that culture. Such a situation is not good in the long run for the advancement of society, technology, or the economies of places that share this culture.
Rewarding it in school packs the workforce with incompetent cheaters (hey, their on-paper qualifications are good, and that's what HR cares about). These are the people who spend all their time trying to game the system and further take advantage of those who are actually good at their job. This disincentivizes quality work and incentivizes office politics. This reduces the competitiveness and productivity of the economy as a whole, and makes everyone poorer.
(Score: 2) by Runaway1956 on Monday March 26 2018, @05:23PM
Yes. Correct. Perfect answer. I agree about 10,000%. Now, how do we convince managers to understand and adopt the attitude?
A couple places, I've wanted to insert something similar into the previous conversations. Something like, "If you think he cheated, but aren't certain, and/or can't prove it, give him a low passing grade, and let him sink or swim later in life." But, that "sink or swim" stuff really doesn't work. It's the networking that works, for so many. No need to really understand what you're doing in your shop/lab/factory if you have a buddy that you went to school with. Or the boss is an alumni of the same college. Or, you go to the same church as the HR. Or, you drink with the right people. Nepotism has no boundaries. If you have friends in the right places, they won't let you sink.
Those of us who grew up believing in a meritocracy have experiences some pretty sad disillusionments.
(Score: 2) by wisnoskij on Monday March 26 2018, @11:11PM
What jobs are available in the real world with a decent salary where everything you are asked to do can just be copy pasted off of Google?
(Score: 3, Insightful) by theluggage on Monday March 26 2018, @12:46PM (6 children)
Why do you need to prove plagiarism? They should have failed because they gave a poor presentation and didn't demonstrate understanding of the topic. Conversely, even if the code was pure copypasta but they clearly understood it (e.g. how would you modify it to X? why didn't you use method Y?) , who cares? The correct answer to virtually any problem in programming is "someone out there has probably spent several years researching this specific topic - I'm not going to come up with a better algorithm in an hour".
Oh, right, the presentation is a token-bolt on which you can't actually fail, because that would require a qualified person to exhibit human judgement and a management team with the balls to support their unpopular decisions, whereas our entire education system is predicated on memorising shit for unrealistic, but easy and cheap to assess assignments - preferably picked from the same handful that get re-used year after year.
But yes, we should definitely go after lazy plagiarists who just copy and paste - they are an insult to all the other hard-working plagiarists who have the nouse to change the variable names and re-word the comments.
(Score: 3, Interesting) by Virindi on Monday March 26 2018, @12:52PM
It was not prohibited to look up how someone else solved the problem. In fact, my solution personally was based on a paper I found. I mentioned that fact, explained what I had to come up with on my own, and of course I wrote the code myself since the paper only described a general approach and not every detail. I got an A.
(Score: 2) by All Your Lawn Are Belong To Us on Monday March 26 2018, @01:44PM (4 children)
And what do you then do when the problem you have is genuinely unique and Google fails you?
This sig for rent.
(Score: 2) by theluggage on Monday March 26 2018, @04:08PM (3 children)
The same thing as you do when the problem wasn't on the course: hope that you've actually gained some understanding of the subject. If you've bluffed your way through by just blindly copying stuff from Stack Overflow without making an effort to understand it then yeah, you're stuffed - but if you've just bluffed your way through on good exam technique you'll be equally stuffed. Courses which require students to show that they understand their answers, and that include problems you won't find on Google, won't have either problem.
(Score: 1) by nitehawk214 on Monday March 26 2018, @04:46PM (2 children)
So, learn the material and come up with a unique solution to the problem?
That is the most diabolical form of cheating I have heard of!
"Don't you ever miss the days when you used to be nostalgic?" -Loiosh
(Score: 2) by Runaway1956 on Monday March 26 2018, @05:29PM (1 child)
Don't forget the icing on this diabolical cake. Make sure it's over the teacher's head, so that he/she is simply awestruck, and begs to be permitted to prostrate him/herself before your superior intellect.
(Score: 1) by nitehawk214 on Monday March 26 2018, @08:36PM
Oh I thought the top form of cheating would be to use these skills to get a job. Then they give you the money. Then, 20 to 30 years later you just walk out the front door like nothing happened.
http://www.cc.com/video-clips/acaxis/key-and-peele-the-heist [cc.com]
"Don't you ever miss the days when you used to be nostalgic?" -Loiosh
(Score: 5, Interesting) by bradley13 on Monday March 26 2018, @02:41PM (2 children)
We also have a big, semester-long project. As part of the grading, we invite each, individual student to an oral exam. We pick some random piece of code that they claim to have written, and ask them to explain it.
If they truly wrote the code themselves, they are usually out in 1-2 minutes. If they can't explain the code, we pick another piece, and another, and another. Eventually, after 15 or 20 minutes, they run out of chances. We fail them as an individual;'the group as a whole may still pass. Since we put this process in place, the number of students trying to ride along on the coattails of their friends has dropped dramatically.
Critical to any sort of procedure like this: the support of the school administration.
Everyone is somebody else's weirdo.
(Score: 4, Interesting) by Anonymous Coward on Monday March 26 2018, @03:16PM
The best prof I ever had ran his computer classes thusly: people could do the assignments on their own, or as a group. The main thread through the course was taking the same program and modifying/adding features as the class progressed.
Then came the final -- we were each given a new modification to make to the program we'd spent all quarter building, there were multiple modifications assigned, and everyone had 15 minutes to think about the problem and then we were marched to the computer room and we had to sit down and actually modify the problem, within four hours. Those who'd done everything themselves had no problem with the changes. Those who'd worked in groups but had actually done the work, also did well. Those who'd done nothing while relying on the group to do it all, failed utterly.
(Score: 0) by Anonymous Coward on Tuesday March 27 2018, @10:27AM
My first CS class (on the C language) had a two-part exam.
One was "theoretical", where, among other things, we had to write, on paper, a working C code that produced a desired output (just printf, nothing fancy).
For the second part we got random problems (literally picked out of a hat) to solve by writing the code within 30 mins at the lab computers (BorlandC, no internet access).
I loved that exam.