I recently answered a question that seemed to address the original question's issue (i.e. how to visualize some data). Then, in the comments, there was an implied follow-on question about how the visualization was created.

Is this an appropriate time to create a new question (e.g. how was the visualization created?) and answer it? Or, would the community be better served by adding onto the answer to the original question? I'm particularly curious about cases where answering the second comment/question will be lengthy and probably off-target from the original question.

  • $\begingroup$ Thanks whuber, I wasn't sure if it was appropriate to link to the specific question, or if it should stand on its own in Meta. $\endgroup$
    – dav
    Commented Jul 10, 2012 at 18:38
  • 4
    $\begingroup$ Possibly of interest: meta.stats.stackexchange.com/questions/1151/… $\endgroup$
    – Macro
    Commented Jul 10, 2012 at 18:49
  • $\begingroup$ In your specific case, you could probably just link to a description of the process elsewhere, if available; or give a string to search the help feature of excel with. $\endgroup$
    – naught101
    Commented Jul 11, 2012 at 0:59
  • 1
    $\begingroup$ The solution is somewhat out of the box as far as Excel goes (plus, We're talking Microsoft Help, so I'm not sure that will do anyone much good). I think Macro's link and Andy W's answer frame the argument pretty well, right now I'm wondering if there's enough difference to warrant separating the answers and providing a link like you recommend. $\endgroup$
    – dav
    Commented Jul 11, 2012 at 1:15

2 Answers 2


IMO in your situation I would just add the response into your current answer. While I don't disagree with the answers to the linked question Macro provided in the comment to this meta question, I don't think the discussion has diverged enough from the original topic to justifying splitting the question/answer(s) across two different threads.

I will say "what is diverged enough" to justify splitting into a separate question is quite vague, but it happens infrequently enough I don't see much reason in trying to layout bureaucratic norms for when it should happen. So the interested parties should just use their discretion to determine if such a threshold has been met.

  • $\begingroup$ I appreciate the thoughts and guidance here. Andy W's answer seemed to work well for this instance, and the open-endedness of "diverged enough" I think provides enough leeway to address issues as they arise. Plus, there's always the mods and community to help keep things from going to far off-track. Thanks $\endgroup$
    – dav
    Commented Jul 12, 2012 at 12:23

I think the asker should be encouraged to select the answer that is correct for their original question and then ask another question. Part of participating in this forum is recognizing it as a resource for all users and not just personal consulting help.

Questions that devolve into back-and-forth conversation make it hard for other readers to glean the message, or for new answerers to provide alternative approaches. Searchs can then turn up many false negatives with tangential content.

The rules of this site encourage posts with a "laser focus" on a specific, statistical question, to provide only the salient details, and to upvote/select answers that address their question. If there is a lack of clarity, the readers and answerers will either ask for it or provide a case-by-case strategy. If the question is answerable and on-topic in its current form, it shouldn't be changed. Follow-up questions with segue from previous attempts will often hyperlink those questions to provide some context, but still ask a sufficiently detailed question in their own right.

Think of it this way: if the asker was aware of the full breadth of their problem, would it be justified to split into two questions at the outset? Often times I believe the answer is yes.

  • 2
    $\begingroup$ Agreed that many threads are spoiled by lack of focus; and very strongly that the aim here is building resources for many, not providing a help line for individuals. But it is also true that many people have great difficulty formulating a precise question; that is why they're posting at all. Also, many threads here are valuable because they expose a variety of view points. The Stack Exchange model is dominated by an ideal (sometimes well approximated in programming) of a sharp question with a precise answer. The tacit question here for many threads is fuzzy, how should I analyse my data? $\endgroup$
    – Nick Cox
    Commented Dec 11, 2017 at 17:18

You must log in to answer this question.

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