There is an old joke sometimes attributed to Einstein: "If we knew what it was we were doing, it would not be called research, would it?"

As an example, I am frequently at a loss of how to rigorously define some informal concept. This is because I do not know what I am doing. But, if I knew what I was doing, it would not be called research, would it? OTOH, there are an infinite number of ways to take an informal concept and cast it into a rigorous framework, so this is a very "open-ended" type of question.

It seems to me that some of the StackExchange sites are geared toward research-level questions. For example, imho, MathOverflow is more researcher-oriented than Cross-Validated is more researcher-oriented than StackOverflow. As such, I am wondering how a "research-level" question doesn't involve some degree of open-endedness. Perhaps it is the case that I am taking the phrase "research-level" too literally, and it does not mean "at-a-level-so-high-that-nobody-currently-knows", but merely "very-high-level".

With this in mind, what does "open-ended" mean here on Cross-Validated, in the context of "research-level" questions?

  • 1
    $\begingroup$ "it isn't about the practice or principles of participating in Cross Validated"... The number one rule on the help page for every stackexchange site is "no open-ended questions." So the ability to clearly discern "open-endedness" seems key to any sort of participation in Cross-Validated, or any other StackExchange site. $\endgroup$
    – Him
    Mar 12, 2018 at 19:03
  • 4
    $\begingroup$ We cannot state how "research-oriented" other SE sites may or may not be. With that in mind, can you adapt this Q to be specifically about CV, & what's appropriate here? $\endgroup$ Mar 12, 2018 at 19:15
  • $\begingroup$ @NickCox "I am statistical enough to say what are the data behind that?". It was not an assertion, it was an opinion... note the "imho". In any case, surely you will agree that some stackexchange sites are more amenable to research-level questions than others. For more examples, there is a site for "code review"... there is also a site "mathematics" that specifically says to ask "research-level" questions on mathoverflow instead. It has been my experience that "research-level" questions are generally welcome on Cross Validated, but that this is not the sole purpose of the site. $\endgroup$
    – Him
    Mar 12, 2018 at 19:21
  • $\begingroup$ @gung, My question is, specifically, how the concept of "open-endedness" differs here on Cross-Validated vs other stackexchange sites. The ban on "open-ended" questions seems to have been cookie-cutter applied to many sites on the network, which have different purposes, and that what constitutes an "open-ended" question in one community may not in another. I will slightly reword the question to focus on CV, as requested. $\endgroup$
    – Him
    Mar 12, 2018 at 19:25
  • 7
    $\begingroup$ @Nick It seems to me that if this question were changed only a little to ask "how do and should we interpret 'open-ended' for the purposes of deciding which questions are appropriate" then it would on-topic here in Meta.CV. $\endgroup$
    – whuber Mod
    Mar 12, 2018 at 19:29
  • 2
    $\begingroup$ What "open-ended" means on other sites is not something we can comment on. Nor can we really say how we compare, for the same reason. We can talk about our policies, but that's really it. $\endgroup$ Mar 12, 2018 at 19:35
  • $\begingroup$ @gung "for the same reason". No reason was given... $\endgroup$
    – Him
    Mar 12, 2018 at 19:41
  • 3
    $\begingroup$ CV is a Q&A site, while open-ended questions are rather topics for discussion, so are better suited for discussion forums. I won't common on other SE sites, but I imagine that different disciplines also differ in the open- vs closed-endness (e.g. engineering vs philosophy), but this site has clear scope and discussing research topics is beyond it. $\endgroup$
    – Tim
    Mar 12, 2018 at 20:18
  • 3
    $\begingroup$ I'm inclined to keep Meta posts open, even when there is controversy, because the very purpose of Meta is to have discussions--even somewhat open-ended ones--about our site's policies and practices. The comments so far display differences of opinion that IMHO could be useful and interesting to discuss here. $\endgroup$
    – whuber Mod
    Mar 12, 2018 at 20:23
  • $\begingroup$ Thanks for focusing the q. on CV - I've re-opened it. $\endgroup$ Mar 12, 2018 at 22:00
  • 1
    $\begingroup$ @Scott, thank you for refocusing the Q. If it helps, I can rephrase my comment above: "What "open-ended" means on other sites is not something we can comment on. Nor can we really say how we compare, also because *what "open-ended" means on other sites is not something we can comment on. We can talk about our policies, but that's really it." I hope that clarifies it for you. $\endgroup$ Mar 13, 2018 at 0:25
  • $\begingroup$ Following helpful revision, I will venture a comment, but not an answer. I don't think I consult any personal or public definition of open-ended. Is there scope for definitive answer(s) of usual length (between a paragraph and an order of magnitude more) is a workable criterion. What will statistical science look like in 2100? or What would be suitable Master's or PhD topics in statistics? would fail that criterion for me in different ways. Being unclear what is asked and/or being too broad are common concomitants of being too open-ended. $\endgroup$
    – Nick Cox
    Mar 13, 2018 at 7:03

1 Answer 1


Of course we know what we are doing. We may not know the answer beforehand but that is a different thing from not knowing what an answer should provide (a problem solution).

When we make research questions then we want it to be clearly defined in such a way that a clear answer is possible. Some definite answer is possible. Or at least the question provides (implicitly) conditions on what a satisfying answer would be.

Stackexchange is a question & answer website, and not a website for discussion. Or at least not primarily and if there is discussion then it is meta discussion in order to get better questions and answers.

Open ended questions are vague about what a good answer would be. (e.g. consider this recent question https://stats.stackexchange.com/revisions/332695/2 where it is unclear what is required for an answer to be a good answer)

a question should be clearly defined such that a well formulated answer can be considered as a definite or at least sufficient answer.


  • How do we express the diffusion coefficient, when we assume that it can be expresses by the random thermal motion of particles that behave like spherical particles in a viscous liquid?
  • Can we write a tensor equation including a four dimensional space-time metric such that the postulates of general relativity are satisfied and the weak-gravity and low speed limit provide the classical field equations?
  • When observing the photoelectric effect, is there a relationship between the wavelength of the used light and the maximum of the energy of the ejected photons?

You must log in to answer this question.

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