There has been some discussion on CV Meta of how to deal with homework questions and also job-related questions. Either can be unreasonable in certain cases, but the latter really annoy me, since they ask users of CV to do the work of the OP. These "do my job for me" questions seem to have two main variations, both of which are problematic:

  • I am programming ML algorithm X. Here is my code; it doesn't work. What went wrong?

  • I want to analyse data on Y. Here is a broad description of the problem/data. What do I do?

In both cases, the question seems to invite CV users to essentially do the professional work of the OP for them. In the first case, there is very specific information given, but it is essentially a request for debugging of code. In the second case, the information given is usually very broad, and it is a request for formulation of a statistical model for an applied project. Unless the questions highlight some useful general principle that is helpful to other users, these kinds of questions do not appear to me to be valuable on the site.

I have sometimes seen the first type of question closed as "off-topic (out-of-scope)" and the second type of question closed as "too broad". However, I wonder if a stronger response is warranted for these questions. Would it would be better to give a custom explanation for closure, along the lines of "off topic (CV is not a site to do your modelling/debugging for you)". Is this overstepping/rude?

What is best-practice for dealing with these questions?

  • 10
    $\begingroup$ I've just read the post on homework questions and feel a related approach is relevant. In most cases the reason people ask these professional questions is they have a main job and are being asked to do something out of their comfort zone. We should encourage guidance type answers rather than solutions, similar to self study, in fact I would advise using the self study tag. That way it helps the newbie learn better. It would also stop people doing the OP's work for them. $\endgroup$ – ReneBt Jun 14 '18 at 4:28
  • $\begingroup$ Thanks for the comment (+1) - maybe convert to an answer? $\endgroup$ – Ben Jun 14 '18 at 4:33
  • $\begingroup$ Now added as an answer $\endgroup$ – ReneBt Jun 14 '18 at 5:41
  • 2
    $\begingroup$ Seems like a great opportunity to provide OP with a cost estimate and proposed statement of work and drum up some business... $\endgroup$ – Sycorax Jun 14 '18 at 14:57
  • 3
    $\begingroup$ Many of the first group of questions will be off topic (if it's primarily related to coding issues rather than requiring expertise) or unclear (if you can't even tell whether it's merely a coding issue or not). The second kind should often close as too broad (indeed I just now closed one exactly with this issue) or unclear (often they don't even state what it is they want to find out!). You can vote to close - you should have all the required privileges to cast such votes...ctd $\endgroup$ – Glen_b Jun 14 '18 at 23:30
  • 5
    $\begingroup$ ctd... Of those that could be turned into a specific question with a clear answer of only a few paragraphs with an edit, encourage such an edit. If it's essentially a homework-style question, treat it as one. We cannot, however, effectively police whether or not a question relates to someones job (most of them will); the main criterion should be "is this a question whose answer will help the next person with a similar issue?". If it's not likely to help anyone but the asker, it's probably not a good question for our site. . $\endgroup$ – Glen_b Jun 14 '18 at 23:36
  • 4
    $\begingroup$ @Glen_b "is this a question whose answer will help the next person with a similar issue?" this criterion makes sense for a Q&A site and I've used a few times myself, but I just noticed it doesn't seem to be explicitly stated anywhere. Neither the tour, the list of topics or the list of inappropriate questions seem to mention it. Should we add it? I may have also missed something obvious - I'm in a hurry today. $\endgroup$ – DeltaIV Jun 15 '18 at 8:51

If a question asks about doing someone's homework and does not meet the requirements of questions described in here, then what you should do is cast a close vote:

Close -> off-topic because... -> Self-study questions [...]

If the question asks about debugging or reviewing someone's code (in many cases it will also be off-topic on http://stackoverflow.com/ and https://codereview.stackexchange.com/ but you could consider voting to migrate if it is relevant), then it is also off-topic, so you should vote:

Close -> off-topic because... -> This question appears to be off-topic
                                 because EITHER it is not about statistics,
                                 machine learning, data analysis, data
                                 mining, or data visualization, OR it
                                 focuses on programming [...]

In many cases the descriptions in the off-topic notices are clear enough to show why the question was closed. If you feel that the question could be edited, so that it becomes on-topic, then you can always comment to encourage the OP to relevant edits. Sometimes I comment also when I have a feeling that an OP does not understand how does the site works, so some brief comment may be helpful. Then I usually give them some of the links: to the tour, on-topic help page, and how to ask? guidelines.


I downvoted your question because the part about job-related questions doesn't make sense, and it's unrelated to the site policy:

  • this is a Statistics Q&A - nowhere in the list of questions one should avoid asking is it mentioned that questions related to the OP's job must not be asked. The only valid criterion to judge a question is whether it is a good statistics problem or not: whether the OP gets paid to answer them doesn't matter at all - surely William Deming and Karl Pearson could have asked and answered excellent statistical questions, related to their work as consulting statisticians.

  • such a criterion is also illogical: as a lecturer you're being paid from your university to do research and write papers, and your career depends on how many papers of high impact factor you publish, how many grants you get, etc. With this kind of reasoning, we shouldn't answer questions from academicians because we're making the work of reading papers, studying books and so on, for them. It doesn't make sense.

As Stephan Kolassa explains very well here, whether the OP is a doctor, an high school teacher, a researcher or a company employee is none of our business. In closing, downvoting, migrating, etc. questions we should stick to the on-topic/off-topic rules for this site, which don't mention the OP's source of income.

  • 2
    $\begingroup$ Thanks for explaining your down-vote (+1) $\endgroup$ – Ben Jun 14 '18 at 23:40
  • $\begingroup$ I believe that Ben was (much) more nuanced then just being against job-related questions. $\endgroup$ – Martijn Weterings Jun 15 '18 at 21:07
  • $\begingroup$ @MartijnWeterings I believe you didn't understand my answer. $\endgroup$ – DeltaIV Jun 16 '18 at 7:05
  • $\begingroup$ The point made by Ben is that we have a special off-topic closing option for 'self-study' when these questions go out of boundaries. We do not have this for professional work questions (except the general tag). The analogy, allowing nuanced differences, is certainly valid. $\endgroup$ – Martijn Weterings Jun 16 '18 at 11:13
  • $\begingroup$ @DeltaIV I guess I do not understand your first sentence. But I agree with all of the rest (which maybe I did not understand either). It is just that (simplified) first premise that you extract from Ben's question which is entirely different in comparison to what I got from the question. ----- Certainly we should not need to judge questions on whether the OP gets paid for getting an answer. But the same is true for self-study questions where we should not judge the statistics and validity of the question (being on-topic) based on whether the OP is getting a mark for work copied from an answer. $\endgroup$ – Martijn Weterings Jun 16 '18 at 11:16
  • $\begingroup$ The point made by Ben is that we have a special off-topic closing option for 'self-study' when these questions go out of boundaries (off-topic). We do not have this for professional work questions (except the general tag). The analogy is certainly valid. $\endgroup$ – Martijn Weterings Jun 16 '18 at 11:21
  • 2
    $\begingroup$ @MartijnWeterings it is not valid at all. There are obvious reasons why universities have policies against copying one another's work (or from the Internet), while hospitals have no policies against doctors (professionals) asking for suggestions from other doctors (professionals). This is so self-evident that I won't spend my time arguing the point here in comments. If you're convinced the analogy is valid, you're free to propose a new tag - naturally I'll oppose it. $\endgroup$ – DeltaIV Jun 16 '18 at 13:03
  • $\begingroup$ @DeltaIV I am not so much convinced about a different tag. I was just pointing out that the case is much more nuanced. The same is true for the single argument about university policies. There are many more reasons why we have a self-study tag. The least reason would be that students can copy homework, work which btw is not so much university level but more for school (in that case we should delete the self-study questions after the OPs have discovered the solution, to prevent abuse from other students with the same question) $\endgroup$ – Martijn Weterings Jun 16 '18 at 13:57
  • 2
    $\begingroup$ I downvoted the question for similar reasons, but I do want to make it clear to the OP that downvoting on Meta has very different implications and interpretations to downvoting questions on the main site, and should not at all be perceived as a slight! $\endgroup$ – Silverfish Jun 19 '18 at 16:13

I don't think there's a problem with asking questions related to work, and I would go farther than others and say that this is not always because the OP has a "routine" job.

Statistics is a big field! No one person is expert in all of it. Now, if I (as a consultant) get a request for which I am simply unqualified (e.g. something with time series) then I will simply turn it down. But there are lots of areas where I might have some knowledge but want confirmation, or have learned something years ago and be interested in anything new.

Asking to do an entire assignment could be closed as "too broad" but, otherwise, I don't see the problem.


To add to Tim's answer: according to how you've described it, your second example should be closed as too broad (cf Top k List of Reasons to Close a Question Immediately). And it's not germane whether it's for work, study, or fun.

These kind of questions differ from self-study ones (exercises, book-work) in that it's only for the latter that it's even possible to give a complete solution within the confines of a CV answer. So @ReneBT's proposal, rather than imposing a restriction, would in fact be a mechanism for enabling the sort of broad, unfocused, forum-style question that we currently close.


CV (and the field of statistics in general) seems to suffer from the problem that basic statistics knowledge is necessary in so many fields and required to solve so many problems, yet professional statisticians are few and far between. This I think is the main reason why there are so many new unanswered questions every day on this site. (As a corollary, I think this stackexchange can never be the 'small numbers of high quality curated answers' site that some people seem to think it should be.)

Speaking as a non-expert who nevertheless uses statistics quite a bit (a chemistry PhD who now does Data Science), I do occasionally depend on CV to fill in gaps in my self-taught stats knowledge. Some of my questions will be ill-formulated because I don't know what I don't know. Rejecting or discouraging amateurish questions related to peoples jobs prima facie I think would be a mistake.

However, I do think effort on the part of the asker needs to be put into making their question as general as possible. As ever, some proof-of-work or proof-of-googling needs to be apparent.

On the third hand, statistics is one of those fields where it may not be at all apparent how carefully you need to think through a problem and make allowances for subtle issues in the data. Many people may come asking for a technique recommendation not knowing that their problem lies elsewhere. The best response, as ever, is patience.

  • 3
    $\begingroup$ FWIW, I think the recommendation that the asker make "their question as general as possible". This belief often leads to questions that have no relevant context, are difficult to interpret & to answer usefully & correctly. Cf, How can we encourage askers to provide context? $\endgroup$ – gung Jun 16 '18 at 18:38

I would suggest that most professional people asking such questions are doing so because they have a routine job and are being asked to do something outside their comfort zone, rather than being lazy. Ultimately almost everyone on the site will be using knowledge gained to provide monetary gain for someone, so we shouldn't discriminate against commercial companies (conflict of interest alert as a freelancer). I feel they should be labelled self study, as they are seeking to learn new skills and gain new understanding. This would flag to answerers that they should not do the work for them but rather provide guidance. This would allow users to learn not what to do, but how to think for themselves, just as for students.


The comments are fair enough, it is indeed a stretch to ask the remit of self study to be extended to such situations. I guess I also didn't initially clarify that I would still expect any question to demonstrate extensive effort to understand the problems and potential solutions prior to posting, along with a very specific description of what their sticking point is so flagging posts that are trying to avoid all effort is a reasonable response.

  • 8
    $\begingroup$ (-1) I agree with the sentiments here but not with the proposed approach. The self-study tag is for exercises/bookwork - artificial problems for training. Other types of questions should be judged according simply to whether they're good or not. We already close questions that ask for code debugging, or are too broad. (I should add my own answer when I have time.) $\endgroup$ – Scortchi Jun 14 '18 at 8:27
  • 7
    $\begingroup$ (-1) I agree with @Scortchi that the self-study tag must not be abused for questions which are not exercise or bookwork. $\endgroup$ – DeltaIV Jun 14 '18 at 12:14
  • 7
    $\begingroup$ Indeed, if the self-study tag is required for any question where someone is "seeking to learn new skills and gain new understanding", then we may as well add it by default to any question, as it will then be hard to imagine a question where it is not required. $\endgroup$ – Christoph Hanck Jun 14 '18 at 12:22
  • $\begingroup$ +1. The meta tag self-study only makes things more confused in this site. We should eliminate it. $\endgroup$ – Andre Silva Jun 16 '18 at 19:05
  • $\begingroup$ @AndreSilva We generally dislike meta-tags and I do think the self-study tag is extremely poorly worded and poorly named, but then again, I do think it is sensible to deal differently with questions of the ilk Scortchi described. I might suggest renaming "self-study" as "bookwork", for a start... $\endgroup$ – Silverfish Jun 19 '18 at 16:15

I personally do not have problems with the 'do my professional work' question an sich. That means, to me the fact that people are getting paid for answers to questions that other people gave is not a problem (and you could extend this type of questions to more than people with real jobs, also researchers, teachers, other goverment employees and students can produce these type of questions).

However the 'do my professional work' type questions do often have a certain characteristic that makes them on the edge of being off-topic. Stackexchange is not a crowdsourcing website, like Kaggle or InnoCentive, where solutions are given to problems that have a much larger scale, but also much more specific.

The principle mentioned by GlenB

"is this a question whose answer will help the next person with a similar issue?"

is often missing in these questions. It is not always so much off-topic because, too broad, off-topic, or not about statistics, but because the problem is a very specific, esoteric, complex question dealing with a problem of the OP but very difficult to translate into a general issue that many people are gonna have (or at least somebody searching for a solution to a similar problem that depends on the same principle might not find or understand the question)

Like Tim's answer states one way to close this question is using off-topic with the specification 'not about statistic...', or like the OP states also the 'too broad' classification can be used for such questions.

I guess it is not so much problematic to describe how we currently can close these questions, and instead the main point is the question whether we should change this? Regarding this issue it is interesting to remark two current practices which are already in place related to other types of closing.

  • There is a special page to help people write self-study questions.

    In the case of self-study questions we have a specific way to close the question (as well as even a specific tag but that might be another discussion). I imagine the same is possible as well for professional work questions (I am stating it is possible, I am not sure we should do it).

    So we (some) do to some extend care about the intent of the questioner. Although again here not so many are against the good statistical questions involved in course work, and that closing option exists mostly because of the quality. For example: do not simply copy-paste your homework questions on this website, but at least make it into a decent and useful question (you modify this principle analogously as 'do not simply copy-paste your professional work questions on this website, but at least...').

    Self-study questions (including textbook exercises, old exam papers, and homework) that seek to understand the concepts are welcome, but those that demand a solution need to indicate clearly at what step help or advice are needed. For help writing a good self-study question, please visit the meta pages.

    In the meta-pages it is written how to write a good self-study question (actually not so clearly since the refered page is more like a discussion for the community, how should we deal with these questions, rather than brief clear information for a new user that is asking a question). We could have something equivalent for how to write a good professional (applied) question (although I am not very aware whether the issue is such problematic to do all this effort, yet at least some people, including me, are sometimes annoyed, to some extend, by the questions).


  • In the case of questions that are too broad we refer to 'how to ask'. In the case of off-topic (not about statistics) we refer to 'support links'. For people that are asking the problematic questions, often new users of the website, it may not always be as easy to get to the right help to improve the question. The problems with a 'do my professional work for me' type of question may not always be so clearly defined in as a single case of the many options we have to close a question. Because of that some wider support might be useful. Some solution might be that the off-topic close option might also have the 'how to ask' link in it, for instance. What I did before (in the few times I have voted to close) is also use the 'off-topic, other' closing option and specify more precisely what can be changed to improve the question. This would be nice in many more cases, a better explanation of down votes and close votes.

Why should anyone ask a question in CV?
1) Homework
2) Research or application in a professional capacity (PhD candidate, professor, researcher, analyst, consultant)
3) Self study, for whatever reason

Asking for help in a Q & A forum, is always about "do something that I say I can't". Be it homework, research related or work related, it is always about "do part of my job" -with "job" being broadly defined.

So we regulate depending on the case, and on the extend to which help is asked -it is always a matter of degree, it's quantitative.We try to assess whether the cry for help is "do my job" to an "inappropriate" degree -and again, the criteria cannot be formalized.

If it is about homework, we ask from the OP to show that they really cannot, and not that they are lazy and attempt to leech. Asking to "show effort" is a way for the OP to convince us about it.

If it is something research related, or work-related, we expect that it will be some peripheral aspect of something more general.

"My code doesn't work-why?" appears certainly to be an "inappropriate" degree of help asked - but I do not think that there exist a structural category of answers that could be dealt with some general guidance.

We were able to separate and regulate "Homework" because there exists an objective typical criterion related to the environment into which the OP operates.

But outside of that, we are talking about all the other questions in CV.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .