4 ms·

One problem is that some students cheat. They develop and pass around open-source sets of answers. If the problems don't change, then students don't learn as

One problem is that some students cheat. They develop and pass around open-source sets of answers. If the problems don't change, then students don't learn as well.

Chaebixi 9y ago
> One problem is that some students cheat. They develop and pass around open-source sets of answers. If the problems don't change, then students don't learn as well. The solution to that is to not revise the textbook, it's to pass out different homework handouts each semester. IMHO, the textbook problems sets should be for student practice and all the solutions should be in the back of the book (vs. just the odds).
- cmdrfred 9y ago
  >pass out different homework handouts each semester. That sounds like work, we are talking about academia here.
  - JustSomeNobody 9y ago
    How often does Calculus change? Over the course of a few years, one could develop enough homework problems to start rotating them out each semester.
  - kesselvon 9y ago
    It's not like they write those problems. They have auto generators now, go to Khan Academy or WA and it'll spit out all the calculus problems you want
    - germinalphrase 9y ago
      I believe that auto-generation of homework/assessment elements (or the maintenance of a large corpus of previously developed assessment elements that can be curated at will) should be ubiquitous in all subject areas. Mathematics is clearly the easiest and should be available now to everyone. I suspect, however, that academic publishers would attempt to sue the pants off anyone that built such a service if it relied on user submissions of content (e.g. a professor uploads their homework assignments to the library of content and that content can be utilized by anyone else on the system).
    - wolfgke 9y ago
      > and it'll spit out all the calculus problems you want Better look into old Russian textbooks for interesting (often very hard) calculus problems.
- Jtsummers 9y ago
  A better solution is to reevaluate your evaluation scheme. If homework solutions become available then homework answers can't be trusted as a measure of understanding. You need exams, essays, and projects. Homework should be left as a tool to develop understanding and measure progress, but not for final grades.
  - Chaebixi 9y ago
    That's true. However, in all the university classes with graded homework that I took, the homework only counted for a small fraction of the final grade (usually <=10%). I think it was only "graded" to give the less-motivated students a more straightforward incentive to do it, and encourage more consistent engagement with the course content.
    - Jtsummers 9y ago
      Same for me, which is another reason, though, to not worry about cheating on homework. Cheating on tests, yes, but homework, with it counting 10-15%, only means the difference between one grade level, at most. And a failing student would only move to a C or D.
UnoriginalGuy 9y ago
If you don't want students to cheat or share answers then you wouldn't be getting the questions from a published textbook, regardless of if it was open or proprietary. How does a proprietary textbook solve the cheating problem? Even a new copy yearly doesn't do much of anything, students can still share answers or pay a third party to answer the questions on their behalf.
greglindahl 9y ago
If you ask any professor, they'll tell you that homework can't stop cheaters, and it doesn't really matter if you change the problems every year. Exams are how you catch cheaters, especially if you have subtly different problems to catch copiers at the exam itself.
JustSomeNobody 9y ago
So? They're only hurting themselves. If they happen to get a job, they won't last long. They'll be destined to float from company to company, changing jobs when they're found to be frauds. Or, they'll actually pick up enough knowledge to hide away at some larger company.
- saint_fiasco 9y ago
  That sounds like a pain in the ass for their employers. An incompetent person with a college degree makes life more difficult for recruiters (directly), and for competent people looking for jobs (indirectly). Interviewing a bunch of incompetent applicants to find the competent one is not cheap.
adamnemecek 9y ago
If the homework for your class is easily copyable, you are doing it wrong.
- logfromblammo 9y ago
  Every professor should have a procedural problem generator that takes a seed number input and creates an entire, unique assignment, with a grading guide generator that operates from the same seed to demonstrate the "correct" steps to reach the solutions for the assignment. That is something you could literally write once and continue to use for 100 years, tweaking only when the course requirements change. Even if you're lazy, and write all the questions and solution guides by hand, you could still select from every such question that you had ever written, and by your 20th year on the job, memorizing from the bootleg study guides would be harder than just learning the course material.
AceyMan 9y ago
> One problem is that some students cheat. Do not all colleges have some form of honor code? Where I attended (two schools), it seemed universally understood that the Honor Code had meaning — I may have had blinders on, but I never saw evidence of cheating. I readily admit my school's (while highly selective) were not known as cutthroat, "keep up or get out" institutions, which I'm aware is a de facto state at some places. Sadly, such rank competitiveness fosters a "need" to cheat.
- HillaryBriss 9y ago
  as a math professor once said to me just before he gave an exam to his students at a highly regarded, nationally-known, top-25 university: "they cheat like crazy."
RodericDay 9y ago
Every prof I had in university trotted out this same issue. They all tried to take down answer sheets as soon as possible, got upset when people asked with help solve them past questions, etc. Except one. This guy had been teaching for like 10 years, and put EVERY SINGLE PAST PAPER HE HAD out there. He provided students with a massive, massive stash of questions and answers. For any given problem, you could either solve it, or you could just navigate thousands of PDFs trying to find the solution. And when you wanted to do some solo practice, you had endless options. He reused questions from those papers in midterms, but the chances of you actually picking the right question from all the possibilities was very slim. It's one of those situations where, paradoxically, by being completely gregarious, he achieved better results than by being ultra guarded and secretive.
Cerium 9y ago
The honor society at my college decided to solve the problem of students passing around answer sets. They created an answer database (with support of the dean) to ensure that answers were not unfairly shared. It worked great. Students got a lot more practice material. Some teachers hated it since they had to change tests, but even if they refused, at least the system is fair (thus no more unfair advantages from "cheating").
emptybits 9y ago
> If the problems don't change, then students don't learn as well. Do you mean that students who choose to cheat don't learn as well? (That doesn't garner much sympathy from me.)
dubya 9y ago
I stopped grading homework in mid-level math classes except for completion. If students turn in something, they get credit, and I grade two or three problems, of their choosing, carefully to give them feedback on their proof writing. I still have students hand in work straight from the solutions manual.
kwhitefoot 9y ago
The solution to cheating is to reinstitute the viva voce and make passing it a major part of the score required to get a degree. I'm quite confident that no one at Exeter Uni. in 1977 could have cheated their way to an honours degree in physics because defending your final year experimental report (120 typed pages of data, charts, conclusions, explanations, and procedures) required answering live questions from two very smart faculty members.

One problem is that some students cheat. They develop and pass around open-source sets of answers. If the problems don't change, then students don't learn as well.

Chaebixi 9y ago

> One problem is that some students cheat. They develop and pass around open-source sets of answers. If the problems don't change, then students don't learn as well. The solution to that is to not revise the textbook, it's to pass out different homework handouts each semester. IMHO, the textbook problems sets should be for student practice and all the solutions should be in the back of the book (vs. just the odds).

cmdrfred 9y ago
>pass out different homework handouts each semester. That sounds like work, we are talking about academia here.
- JustSomeNobody 9y ago
  How often does Calculus change? Over the course of a few years, one could develop enough homework problems to start rotating them out each semester.
- kesselvon 9y ago
  It's not like they write those problems. They have auto generators now, go to Khan Academy or WA and it'll spit out all the calculus problems you want
  - germinalphrase 9y ago
    I believe that auto-generation of homework/assessment elements (or the maintenance of a large corpus of previously developed assessment elements that can be curated at will) should be ubiquitous in all subject areas. Mathematics is clearly the easiest and should be available now to everyone. I suspect, however, that academic publishers would attempt to sue the pants off anyone that built such a service if it relied on user submissions of content (e.g. a professor uploads their homework assignments to the library of content and that content can be utilized by anyone else on the system).
  - wolfgke 9y ago
    > and it'll spit out all the calculus problems you want Better look into old Russian textbooks for interesting (often very hard) calculus problems.
Jtsummers 9y ago
A better solution is to reevaluate your evaluation scheme. If homework solutions become available then homework answers can't be trusted as a measure of understanding. You need exams, essays, and projects. Homework should be left as a tool to develop understanding and measure progress, but not for final grades.
- Chaebixi 9y ago
  That's true. However, in all the university classes with graded homework that I took, the homework only counted for a small fraction of the final grade (usually <=10%). I think it was only "graded" to give the less-motivated students a more straightforward incentive to do it, and encourage more consistent engagement with the course content.
  - Jtsummers 9y ago
    Same for me, which is another reason, though, to not worry about cheating on homework. Cheating on tests, yes, but homework, with it counting 10-15%, only means the difference between one grade level, at most. And a failing student would only move to a C or D.

UnoriginalGuy 9y ago

If you don't want students to cheat or share answers then you wouldn't be getting the questions from a published textbook, regardless of if it was open or proprietary. How does a proprietary textbook solve the cheating problem? Even a new copy yearly doesn't do much of anything, students can still share answers or pay a third party to answer the questions on their behalf.

greglindahl 9y ago

If you ask any professor, they'll tell you that homework can't stop cheaters, and it doesn't really matter if you change the problems every year. Exams are how you catch cheaters, especially if you have subtly different problems to catch copiers at the exam itself.

JustSomeNobody 9y ago

So? They're only hurting themselves. If they happen to get a job, they won't last long. They'll be destined to float from company to company, changing jobs when they're found to be frauds. Or, they'll actually pick up enough knowledge to hide away at some larger company.

saint_fiasco 9y ago
That sounds like a pain in the ass for their employers. An incompetent person with a college degree makes life more difficult for recruiters (directly), and for competent people looking for jobs (indirectly). Interviewing a bunch of incompetent applicants to find the competent one is not cheap.

adamnemecek 9y ago

If the homework for your class is easily copyable, you are doing it wrong.

logfromblammo 9y ago
Every professor should have a procedural problem generator that takes a seed number input and creates an entire, unique assignment, with a grading guide generator that operates from the same seed to demonstrate the "correct" steps to reach the solutions for the assignment. That is something you could literally write once and continue to use for 100 years, tweaking only when the course requirements change. Even if you're lazy, and write all the questions and solution guides by hand, you could still select from every such question that you had ever written, and by your 20th year on the job, memorizing from the bootleg study guides would be harder than just learning the course material.

AceyMan 9y ago

> One problem is that some students cheat. Do not all colleges have some form of honor code? Where I attended (two schools), it seemed universally understood that the Honor Code had meaning — I may have had blinders on, but I never saw evidence of cheating. I readily admit my school's (while highly selective) were not known as cutthroat, "keep up or get out" institutions, which I'm aware is a de facto state at some places. Sadly, such rank competitiveness fosters a "need" to cheat.

HillaryBriss 9y ago
as a math professor once said to me just before he gave an exam to his students at a highly regarded, nationally-known, top-25 university: "they cheat like crazy."

RodericDay 9y ago

Every prof I had in university trotted out this same issue. They all tried to take down answer sheets as soon as possible, got upset when people asked with help solve them past questions, etc. Except one. This guy had been teaching for like 10 years, and put EVERY SINGLE PAST PAPER HE HAD out there. He provided students with a massive, massive stash of questions and answers. For any given problem, you could either solve it, or you could just navigate thousands of PDFs trying to find the solution. And when you wanted to do some solo practice, you had endless options. He reused questions from those papers in midterms, but the chances of you actually picking the right question from all the possibilities was very slim. It's one of those situations where, paradoxically, by being completely gregarious, he achieved better results than by being ultra guarded and secretive.

Cerium 9y ago

The honor society at my college decided to solve the problem of students passing around answer sets. They created an answer database (with support of the dean) to ensure that answers were not unfairly shared. It worked great. Students got a lot more practice material. Some teachers hated it since they had to change tests, but even if they refused, at least the system is fair (thus no more unfair advantages from "cheating").

emptybits 9y ago

> If the problems don't change, then students don't learn as well. Do you mean that students who choose to cheat don't learn as well? (That doesn't garner much sympathy from me.)

dubya 9y ago

I stopped grading homework in mid-level math classes except for completion. If students turn in something, they get credit, and I grade two or three problems, of their choosing, carefully to give them feedback on their proof writing. I still have students hand in work straight from the solutions manual.

kwhitefoot 9y ago

The solution to cheating is to reinstitute the viva voce and make passing it a major part of the score required to get a degree. I'm quite confident that no one at Exeter Uni. in 1977 could have cheated their way to an honours degree in physics because defending your final year experimental report (120 typed pages of data, charts, conclusions, explanations, and procedures) required answering live questions from two very smart faculty members.