# A Consolidated Homework Policy

A certain theme during the Town Hall Chat was on the question of a homework policy. The idea of such a policy is not new. As one user said in the chat,

I liked Rob john's suggestion towards homework. I quote "I think that hints should first be given to help provoke thought. If they show interest in working the problem, then more of the answer can be shown in an answer. After a time, say a week or so (to give the assignment to be due), a complete answer could be given for completeness of the site." This makes sense and in fact that is what I prefer. It would be great if that is made the policy and posted on the meta discouraging people from posting solutions.

But in the past, we have had such meta questions:

A recurring theme is that many of us agree that giving complete answers to homework questions is poor, but that there is no good way of enforcement. And I don't know of a good way to enforce it either.

I see only a few options that appear reasonable to me.

Those who read the meta might agree to not write up complete solutions. This is a start, but the main problem is that complete solutions are more upvoted. And so there is an incentive to give complete solutions. Just because some of us do not give complete answers does not at all prevent others from giving full solutions.

We might agree to not upvote complete solutions to homework. This would start to remove the incentive a bit.

We might agree to have a pre-written comment, so that when this comment appears on a complete solution we upvote the comment instead of the answer. Maybe the comment would read

Providing an answer that doesn't help a student learn is not in the student's own best interest, and if a solution complete enough to be copied verbatim and handed in is given immediately, it will encourage more people to use the site as a free homework service. In the spirit of creating a lasting resource of mathematical knowledge, please refrain from giving complete solutions to homework questions.

Or maybe not - it's just an idea. Most of that text is taken from the faq.

We might agree to downvote complete solutions. Somehow, this rubs me the wrong way, but it's certain a disincentive. And it's better than systematic 2k+ user editing abuse.

We might agree to bring disputed answers to meta, so that the meta community can decide what to do on a question by question basis. This seems improbable to me, but the idea would be that after an initial burst of attention and flooding of the meta of such questions, more users would realize that posting complete solutions is frowned upon.

We might agree that nothing can be done.

We might agree that something else can be done (This is the cop out - that way, I can say that I gave a complete list of alternatives)

So the question remains: what do we do to consolidate and enforce our homework policy?

-
I can certainly agree not to post complete solutions and not to upvote those who do. Downvoting though seems a bit harsh; I currently only downvote answers which contain significant errors or are completely off-topic. – Alex Becker May 8 '12 at 2:41
@Marvis: Sine there is no consensus, I find the suggestion of moderators deleting answers to be quite absurd. – The Chaz 2.0 May 8 '12 at 3:36
Sometimes I give complete solutions; sometimes I don't. It usually depends on a complex combination of (i) the question; (ii) the effort the student has displayed; (iii) the difficulty of the problem; (iv) how much the problem might be suitable for a more general discussion than a specific one. I can't say I have an algorithm, more a heuristic, very similar to the one I use in person with students when they come to office hours: sometimes I give them hints, sometimes I walk them through solutions. – Arturo Magidin May 8 '12 at 3:51
While hints are often preferable to full solutions, a policy of feeding the solution one hint at a time, and making the recipient ask for more, can be rather cruel. Sometimes a very full solution, far fuller than the one the OP would produce or is expected to produce, can contribute to the OP's mathematical education. – André Nicolas May 8 '12 at 3:57
I don't think that there can be an official policy without some censorship. If there will be censorship I am sure that I will not be the only member missed here. – Asaf Karagila May 8 '12 at 4:58
@Arturo Magidin: For students in office hours, I have often given full solutions, after asking them not to take notes. – André Nicolas May 8 '12 at 5:11
math.stackexchange.com/questions/142008/… I got two down votes for this question since I gave a complete solution. This is why I want to know whether solutions to homework can be written out. As I have said in the comments of that answer, I have become indifferent to homework questions and sometimes I give the entire solution and sometimes I don't. If there is a policy that solution must not be given, then I accept the down votes. But when there is nothing agreed upon, I don't understand the reason for the down votes! – user17762 May 8 '12 at 7:53
@mixedmath: You forgot: We might agree in principle, but disagree which questions are homework questions. and We might just disagree. – Phira May 8 '12 at 10:18
I agree with Arturo that it really depends on the OP and the solution-poster's discretion. That said, when giving (almost) complete solutions I generally have a policy of omitting easy-to-verify details, and using phrases like "you can show that" and "you need to verify this" so that the student still needs to do some work with the problem, but that the main ideas are all there. I personally think that this is not only useful to the OP, it also makes a good reference for others. – Brett Frankel May 8 '12 at 18:44
It sounds like it should be more of a culture rather than a rule. There are websites that specifically help you with your homework problems, like Dr. Math and they have their own rules to assure they're not providing a lazy student with the full solution if they haven't tried enough. On the other hand there are websites like art of problem solving that are mostly student based, and it tends to be the playground for many homework problems but it doesn't really happen, since the "culture" of the forum does not allow this. – Keivan May 16 '12 at 22:59

I am opposed to such a policy, as I believe it would have a negative long-term impact on the site as a whole.

Here are just a handful of the many problem associated with such a policy:

1. One purpose of stack exchange sites is to have long lasting questions and answers, this is why we close as duplicates, because the original is meant to be found on search engines, and be a reference for future question askers. Having an inordinate number of questions with only incomplete hint-answers, and where posting a complete answer is not allowed, nearly defeats this purpose. I am not saying hints are bad, they have their place, but having a policy which only allows hints on a whole class of questions is ludicrous.
2. I can imagine situations where people make accusations about what is and what is not homework, re tag questions, and refuse to give full answers. In the worse case, things may degrade into a witch hunt scaring away new users.
3. How much of a hint is too much? What may be a complete solution for one user would not suffice for another. How can we standardize this? It seems to just create a whole host of reasons based on ones opinion of "what is a sufficient hint" to downvote/not vote up otherwise good answers.

There are a plethora of ways for students to get help on their homework, and different standards exist around the world for what is and is not allowed. Who is to say your standard is the correct one? I don't think that we should compromise the functionality of the website, the happiness of our answer writers, and the health of the community to try and be some kind of academic police. Hints can be very helpful for students learning, but there is nothing wrong with posting a complete solution.

Edit: Also see this related question regarding downvoting complete solutions.

Edit: Removed point number $4$, as it is not consistent with closing questions as duplicates. Hurkyl also makes a good point in the comments below.

-
I can't agree with you more. I think the main reason that people are objecting to giving a full answer to homework is that they don't like the idea of themselves being used by the asker. But IMO, this is narrow minded. Regards, – Makoto Kato Aug 11 '12 at 23:54
I think should compromise the happiness of our answer writers. After all, while some people are happy to write complete answers to homework questions, other people are quite unhappy to see such answers. The alternative to compromise is to completely disregard the happiness of one demographic or the other. And by admitting a compromise is necessary, it is easier to avoid offensive statements like "there is nothing wrong with posting a complete solution" when that's not really what you mean. (I'm assuming that's not really what you mean....) – Hurkyl Oct 31 '12 at 2:00
"...compromise..." The wording of this sentence is steeped in your opinion on the subject, but that is not a bad thing, and may very well reflect majority opinions on this site. I'm more comfortable with another site wherein I've written many answers, and in the culture of that site I'd say something more like "I don't think we should compromise the integrity of our website, the ethical standards of our answer writers, and the academic growth of our question writers, by not trying to be some kind of police". But I can adapt to the way this site works, especially due to points 1 and 4. – Mark S. Feb 8 at 5:26
@Eric By making a major edit (deleting point 4), you essentially invalidate all of the votes (currently +42,-2). Better would be to add a note saying that your views have since evolved, and link to a different post explaining your current views (which can be voted on independent of the old post). – user23500 Apr 24 at 16:18
@Math Gems, you realize that people can accordingly undo their votes after Eric's edit if they so desire, yes? – J. M. Apr 26 at 12:11
@J.M. Yes, but many prior voters may not even see the edit. Without such a policy there may be much abuse. E.g. start with a popular position, then edit it gradually it into something very different. For the votes on a meta answer to be meaningful it needs to contain a unique static voting target. Otherwise the votes can be rightfully criticized as ambiguous by opponents of the position. It is important to keep that in mind when designing answers that poll community opinion (which is, no doubt, a difficult task given the limited tools we have to work with on the SE platform). – user23500 Apr 26 at 13:35

If one is to have any hope of tackling these complex issues then one will need to be completely informed of the diverse opinions on such matters. For an example of a different viewpoint, see this answer of JDH (a professor) which currently has (+22/-5) votes and which concludes as follows:

Finally, let me say that the policy of encouraging weasily half-answers to questions that have been deemed to be homework, consisting of obscure hints only, amounts to an annoying policy of encouraging bad answers here at math.SE, and I am completely opposed to it. For this reason, I think we should abandon or ignore the homework tag. If we are to answer mathematics questions, then let us answer them well, with solutions exhibiting such clarity and elegance as we can muster.

Obviously it will be a difficult task to devise a compromise between that and Rob's suggestion. But I think we should make a sincere attempt. For even if we do not completely succeed in this endeavor, we may go far enough to eliminate a great deal of tension on the main site.

-
Plus one! ${}{}{}$ – Matt N. May 8 '12 at 6:27
To JDH's answer (which, for full disclosure, I downvoted) I say: O Americano, Outra Vez! – user16299 May 8 '12 at 9:47
+1 to this particular answer, nevertheless – user16299 May 8 '12 at 9:50
JDH wants to throw the baby out with the bath. The obvious alternative is to encourage good hints, and perhaps to encourage people to go back after a few days to flesh out the answer. – Brian M. Scott May 8 '12 at 21:59
@Downvoter What do you disagree with? That we should strive to be aware of the diverse opinions in the community? That we should attempt to find a compromise so to help eliminate tension on the main site? I'm surprised either might prove controversial. But I am genuinely interested in every point of view, so please do speak up. Please note that the point of including the quote has nothing to do with endorsing it. Rather, it was included merely to give an example of a viewpoint very different from the OP's. – Gone May 9 '12 at 16:47

I don't think that we will ever have a consensus about how to deal with homework. I don't think that we ever should. Therefore enforcing a policy sounds silly to me. A few sporadic thoughts:

1. I feel that the acceptability of giving a full answer depends on: A) who's asking? B) who's answering? I don't have serious qualms about a bright freshman (or a high schooler) helping out a classmate (or someone a year or two behind). The said bright person will have scant opportunity to earn reputation here otherwise. OTOH, someone with a PhD should IMHO show some restraint, unless answering a colleague or a (post)graduate student. So it is not always clear cut.
2. My own policy is actually very selfish in a way. If I derive pleasure from solving the question, then I am inclined to write up an answer. If the question is total turkey shoot I will only give hints and/or break the question up into smaller steps, and seek to derive pleasure from teaching, and trying to accurately judge what kind of hints would serve that end. Not altruistic at all! Consequently my policy depends on how familiar I am with the topic. Hmm. I'm actually not sure that this is a defensible policy? Doesn't matter much, because I usually steer clear from the topics that don't interest me, but even so?
3. If (or when?) many teachers' follow a somewhat similar policy, and all give hints in comments only, then we occasionally end up with the question with several good hints given in comments, and no answers. This is a bit inconvenient, if the OP does not see the light, because the forum software will then leave the question hanging in the unanswered files. At some point somebody has to type up an answer. I have mixed feelings about this. I would not step up to the plate, unless I can add a different point of view. I guess it depends on how many days the question has remained unanswered. I guess somebody may know (from past dealings with the OP) that it is pointless to wait. In a way an unanswered question is fair game. But don't expect me to upvote such answers.
4. May I rephrase my earlier enforcing a policy sounds silly to me to read: it is silly wrong to enforce a policy that befits some posters but not all of them.
5. I should not try and pontificate on this (or any other) matter. Up/downvoting is the perfect mechanism for dealing with large deviations from the community norm (whatever that may be).

It is quite possible that I have completely overlooked some aspects of this question. All criticism obviously welcome.

-
+1, I agree that having standard against providing full answers would have a negative impact. To add to the confusion, there is also the subjective nature of what is a full solution. What one might find to be a complete solution, another might feel misses many details. One might believe that explaining a particular idea gives away too much, while another might think it doesn't say enough. It makes having a standard even more difficult. – Eric May 8 '12 at 7:49

I think there is more than one "homework" issue, and the discussion is confused when they are conflated.

1 There are users who use the homework tag as I imagine has been intended - they have tried to answer the question themselves, are stuck, and need a hint. They may post frequently or infrequently. Their object is not just to answer the question, but to learn how. An answer is not what the user wants, but the site is set up to provide answers. One possible route would be to disable the answer facility on questions tagged homework [a big step, I know].

2 There are users who discover the site and ask homework questions without identifying them as such. This is sometimes, but not always, obvious. Often, in practice, someone suggests the question is homework, and hints are provided. Answers are also given, which are what the user wants. Some such users learn to use the site as at 1 above, while others revert to type 3.

3 There are users who persist in asking for answers to homework questions without doing the work themselves. This, to me, is lazy and an abuse of the site. This behaviour should, in my view, be discouraged. Answering these questions provides unnecessary encouragement.

I think most of the real issues are about identifying and dealing with "type 2" in the most helpful way. Other analysis may be possible.

-
+1 for number 3 – user16299 May 8 '12 at 23:49
+1 for identifying number 3. I don't think we should deal with the other types by assuming they are this type. – Ronald May 9 '12 at 9:32
@Ronald: I couldn't agree more. I think one issue is that the homework category pushes the bounds of a site like this, because it is not straightforwardly about giving answers to questions - and that is how the site is set up. So the point subtly changes from answering the question, to helping the person who asked it - and when you start dealing with people rather than questions you run into issues of behaviour. I know my post is a little provocative - but the question I had in my mind was what would it be like to push the boundaries further ... – Mark Bennet May 9 '12 at 16:16
... (cont) My sense is that the consensus would be to hold close to the Q/A model by fuzzing the boundary a bit for homework. The fuzz will create some lack of clarity of discipline - inevitably. I presume the goal would be to keep this at a tolerable level. If this way of thinking about things doesn't help, feel free to ignore it. – Mark Bennet May 9 '12 at 16:24

I don't agree with the premise, even before some of the other comments made.

• Giving incomplete answers to a homework question is educationally unhelpful and frustrating to students.
• Giving an educationally good 'hint' must be personalised for the user asking the question - in the best case, this requires some two-way communication (Socratic Method)
• Later users may have to ask the same question to get 'their' hint. (which will then be "closed as dup")

The Socratic method is an amazing learning tool, but it depends absolutely on a conversation between student and teacher - it is completely unhelpful on a site of this type. If this is the reason for saying we should give hints, I find it misguided.

It's not clear to me why it's appropriate to give 'hints' or 'partial answers' at all? Every answer on this site is educating and homework questions are no different.

I would relate an incomplete answer strongly to a professor who handwaves something crucial on the chalkboard by claiming "this is trivial". I'm sure most users have experienced this at some level - and I see this as equivalent to what's being proposed.

I also find it disrespectful: Imagine there's a challenging, novel, graduate-level question. If we gave a brief hint for how to start out and then said "now I'm sure you can figure out what to do for yourself", that would be pretty offensive to the questioner - in other words, we are not inclined to talk down to people who ask this type of question.

I don't have any reason that, say, a K-12 homework question should be any different. The questioner has the same motivation - to learn about what they're asking (even if they cannot articulate that as clearly). I believe they should be treated with the same level of respect. The main job of an educator is to respect the student's willingness to learn.

In essence, I couldn't begin to agree with any of the proposed points.

-
I mostly agree, however note that people reaching grad school are generally more inclined to learn compared to the average K-12, or even undergrad student. It is not improbable that a K-12 student which is frustrated and just want to get a solution and move on has come to ask a question; while it is less likely (but still possible) to happen with a grad student. – Asaf Karagila May 9 '12 at 10:19
I don't agree with your point about gradute-level questions at all. I would be equally inclined to just give a hint on a graduate-level question, because I assume that the person asking would enjoy working out the rest for themself. – Tara B May 9 '12 at 11:28
@AsafKaragila This is exactly what I mean - a K-12 student has to equally make the effort to find the site and post the question, the same as a graduate student does, to try and get some help with their problem. The K-12 student may not be able to express themselves as politely or clearly, but why should we start with the assumption their motivation is different? Both are trying to learn based on the current state of their knowledge. – Ronald May 9 '12 at 20:50
@TaraB perhaps you (and others) would like to give a hint equally in these two cases, but I don't think that is what is happening in practice - besides, my understanding of the purpose of this site is to provide Questions and Answers. – Ronald May 9 '12 at 22:10

As I discussed with you and Bill, I propose a special feature with the [homework] tag. It currently leads the question counting with $4513$ questions versus $3644$ in calculus. The idea is as follows:

It seems there are certain tags that go in pair with the [homework] tag. Some examples are:

• Algebra pre-calculus
• Calculus
• Trigonometry
• Limits
• Sequences and series
• Most tags that encompass highschool or first year undergraduate maths.

What I propose is that these tags be linked to the homework tag, so that when the question is tagged, the following messages will pop up.

Remember: In an effort to diminish plagiarism, we ask you always make a reference if you retrieved this problem from a book, a webpage or any other source.

Example: "This problem is from Rudin, Principles of Mathematical Analysis, Chapter X, Section Y".

Did you know?: You are likely to get better answers if you provide some context to your question. Some useful questions you might want to answer in your question are:

• In what context are you trying to solve this problem? (Specific topic)
• What have you tried to solve it?
• Where are you stuck?

This comment by Arturo is quite a good one to keep in mind:

Any future idea for this walkthrough pop-ups is welcomed.

-
Note that I tend to tweak that comment a bit; for example, the post in question did not have a title (it just had the standard greyed-out statement that appears before you type out the title), which is why I explicitly and early mentioned the "informative title"; other instances of that comment don't say anything about the title. – Arturo Magidin May 8 '12 at 3:58
@ArturoMagidin It chose this one for the particular throughness. What are your thoughts on this matter? – Peter Tamaroff May 8 '12 at 3:59
I mention it because that comment (and its variants) is very close to the limit; so I tend to remove stuff and add stuff to suit the particular post. That was the first time I really had to say something about the title, because it was such a blatant ommission (it only said what's your math question? be specific.`) Sometimes there is no imperative, so I don't mention that. Sometimes it's just the context that is missing, etc. Though the general framework is the same in all, it does vary from use to use, and I cannot put all into the comment because it won't fit. – Arturo Magidin May 8 '12 at 4:02
@ArturoMagidin You're right. But I think this is a good example of a determined yet respectful comment. – Peter Tamaroff May 8 '12 at 4:06
I like this suggestion, but can't in good conscience support the claim that "You will get much better answers if you provide some context to your question." There are many one-line questions that show no work yet receive multiple full solutions. Sometimes I leave a snarky comment, but I've just started accepting it... more or less :D – The Chaz 2.0 May 8 '12 at 4:11
I like the second message, but not the first. I'd prefer that you get rid of the first and add something like Where did the problem come from to the second. In the second message it would be more accurate to say are likely to get better answers instead of will get much better answers. – Brian M. Scott May 8 '12 at 4:14
@BrianM.Scott Noted. I won't delete, but I'll be glad to modifiy. Tweak, tweak, tweak. – Peter Tamaroff May 8 '12 at 4:14
I'm tempted to suggest to add a message like: "Wait a minute, is it your homework which you're supposed to do alone to find out if you've learned enough on the topic? Then add the (homework) tag and we'll be patient enough to help you solve it by giving hints away. We're concerned about you more than anything else, as kind as a nurse" – Gigili May 8 '12 at 5:23
The proposal of "linking to the homework tag", regardless of whether the community likes it or not, requires recoding on the part of SE. In other words, it is not something that we, ourselves, as a community, can implement without help of SE inc providing such functionality. As such, I think that particular part of the proposal should not be compared directly to other discussions about forming (or not forming) some sort of consensus in the community. – Willie Wong May 8 '12 at 9:01
I am not so worried about plagiarism, but if you think that this is the problem, it would be provably much more effective to let people agree to a short sentence that reads "I agree to the StackExchange honor code on cheating on homework." All your messages are much too long and do not take into account behavioural psychology. My problem with the homework issue is that nothing is learned from copying a complete answer, and I am not interested in wasting my time if it does not involve learning. I don't mind others complete answers unless they come after other people's hints. – Phira May 8 '12 at 10:53
@Gigili: Homework questions are not always given for the student "to think about them on their own", mathematicians do not sit in single rooms, they work in groups and they share ideas. Much like this I would expect my students to sit and study together and share ideas. Much like that I expect that people that have no one to share ideas with them to come here and be able to ask questions to get some ideas. Whether the idea is in the form of a hint, or an overly complicated solution full of information. The important thing for the student is to think about mathematics. – Asaf Karagila May 8 '12 at 11:41
@AsafKaragila: As long as we cannot sit and study with them, we give the idea and they sit and work in groups to work on it. It was an encouragement to make them use the tag. – Gigili May 8 '12 at 11:53
@Gigili: Read what you wrote in those quotation marks. The tone is simply condescending (it means talking down to someone). This is what I was pointing out, I have seen several students here which told me that my explanations are clearer than what they heard in class. I wrote full explanations and not comments telling them that they should think about it on their own. – Asaf Karagila May 8 '12 at 11:56
@AsafKaragila: I thought it was obvious that I'm not recommending to use that message. I was being ironic. – Gigili May 8 '12 at 12:25
@Phira If you can improve the above propositions, do so. – Peter Tamaroff May 8 '12 at 14:37

I think that it is a flawed premise that we will agree on a clear, strict, enforceable and enforced policy, since there are quite obviously irreconcilable opinions on this matter. But I do think that while opinions are irreconcilabe that doesn't mean that friendly coexistence is impossible.

First of all, I personally would not agree with any general policy, as I agree with Arturo's remark that he decides on a case-by-case-basis whether to give hints or a full answer.

But since it is decided on a case-by-case basis, one has to live with the fact that one's hint is occasionally rendered useless by a complete answer and one has to live with the fact that one's complete answer occasionally gathers some down-votes. It's ok to say that you want the other side to change their behaviour, but it's less ok in my view to want to have someone intervene when there is simply no consensus.

-

How do you solve the following trigonometric equation?

I got two down votes for this question since I gave a complete solution. This is why I want to know whether solutions to homework can be written out. As I have said in the comments of that answer, I have become indifferent to homework questions and sometimes I give the entire solution and sometimes I don't. If there is a policy that solution must not be given, then I accept the down votes. But when there is nothing agreed upon, I don't understand the reason for the down votes!

-
It's important to understand that downvotes do not need to have a concrete or universally-accepted justification. Users should up- or downvote answers that are or are not useful, whatever that may mean for each individual. While I can see how one might infer that the downvotes on your answer there were because it was a complete solution to a homework question, I don't see it explicitly stated there (though, of course, such a statement could have been deleted). – Isaac May 8 '12 at 8:08
@Marvis You have a flawed premise (see also my answer). On SE, every user is free to downvote if they think that an answer is useful, they don't have to have their reason vetted by the community and they don't need you "to accept the down votes". What does this mean? You want the moderators to punish the down-voters for disagreeing with you on the homework policy? There is no right to get no down-votes. And there is no right to have others be as indifferent to effort-free homework questions as you are. – Phira May 8 '12 at 10:44
And in case I am not clear, I similarly don't think that the polarized vote on your answer means that you are obligated not to write full answers. Since noone pointed out errors in your answer, I do find it quite plausible that you got your downvotes for posting a complete answer after a hint (although you might well have been writing simultaneously). – Phira May 8 '12 at 15:21
@Phira We were writing it simultaneously. – user17762 May 8 '12 at 15:27
So, you got some downvotes on a contentious issue probably partially due to a wrong impression. You certainly don't need the rep, so you could just shrug your shoulders. – Phira May 8 '12 at 22:26

The problem is that the educational incentives structure is a broken product of historical chance, more interested in judging students than helping them learn.

In the future there will probably be a split between learning and evaluation, as Salman Khan of Khan Academy has talked about in many of his interviews. Until then it's not our duty to be enforcers for the current system.