7
$\begingroup$

Some questions on this site are of the basic form: "I ran this model and I don't know how to interpret these results." The individuals who ask these sorts of questions sometimes have a limited prior background in statistics or math more generally (that's why they're asking us to help them!). They may not be able to parse complex equations, or fully grasp some "basic" statistical concepts that we often assume everyone knows, or just have difficulty applying an idea from one context to a different one.

In many cases like this the difficulty the OP is struggling with is just one example of a more general issue that has been discussed many times on this site (e.g. how to interpret interaction terms, confusion about what p values mean, misunderstanding what odds ratios represent, challenges of causal inference). However, the answers that have already been provided on those topics may be too technical for the OP to understand, or the situation those answers are discussing may differ in some way from the OP's situation that makes it difficult for them to know how to apply it.

My approach has always been to try and answer these questions, to help the OP by providing an answer that is tailored to their apparent level of statistical acumen, and which relates specifically to their situation.

But recently I saw one of these sorts of questions being closed because it did not "formulate a distinctly new problem or question." The OP was instead directed at a set of other answers that did broadly touch on the particular issue they were struggling with, but the OP did not understand them because they were too technical or not close enough to their particular situation. The OP begged for additional help in the comments, but the question remained closed.

What do we think is the right approach here? Should we close these sorts of questions because the knowledge needed to answer them exists somewhere else on this site (even if it is not in a form the OP can understand)? Or should we keep them open, and try to answer them in a way that makes sense to the OP (even if this represents a "duplicate" of some previous answer about the same general topic)?

Upon request, here is the original context.

$\endgroup$
7
  • 1
    $\begingroup$ This is a nice question. However, could you provide link to the stated thread that was closed? In that context, perhaps, a better commentary could be made. $\endgroup$ Commented Apr 18 at 14:09
  • 4
    $\begingroup$ +1. But re "The OP begged for additional help in the comments, but the question remained closed," that's exactly how this site is supposed to work. Begging for help is rarely productive. Users are not entitled to answers just because they start a new thread or are desperate or need to get their homework/project/studies done asap. We encourage editing one's post to clarify what is needed in an answer. Sometimes that helps us find a more suitable duplicate and otherwise helps us appreciate what is new or unique about the question. $\endgroup$
    – whuber Mod
    Commented Apr 18 at 15:40
  • 3
    $\begingroup$ @whuber re: "We encourage editing one's post to clarify what is needed in an answer". The "closed as duplicate" message does not seem to mention the possibility to edit the answer. If I were a new user I'd take the current default message as a definitive refusal. Should/can this automatic message be amended? $\endgroup$
    – J-J-J
    Commented Apr 19 at 13:35
  • 2
    $\begingroup$ @J-J-J That's an excellent point. I would suggest raising it in a new Meta thread so others can contribute their suggestions. Thank you for bringing it up. $\endgroup$
    – whuber Mod
    Commented Apr 19 at 14:11
  • 2
    $\begingroup$ Nice catch, J-J-J. Although @whuber, while I agree it's not explicit, once OP tries to reopen, they would get the message that they can (vote to) reopen if the post is improved. Same is with the "Edit" option. $\endgroup$ Commented Apr 19 at 14:23
  • 1
    $\begingroup$ @J-J-J: I agree, great point. $\endgroup$ Commented Apr 19 at 19:04
  • $\begingroup$ I wouldn't describe the particular question as "I ran this model and I don't know how to interpret these results."-type of question. Those q's tend to include actual results, model summaries & diagnostics plots (and oftentimes though not always the OP does add extra background info when requested). In this particular case it was the OP's own choice to try to write an abstract, technical question presumably with the expectation that there is a recipe to be applied to all analyses of this kind. $\endgroup$
    – dipetkov
    Commented Apr 21 at 8:59

2 Answers 2

9
$\begingroup$

If the O.P.'s question has already been asked, but the answers so far are v. technical, one approach is to provide a less technical answer to the prior question. Another is to encourage them to edit their question to differentiate it from the other, perhaps by focusing on what they didn't understand in the technical answers.

$\endgroup$
2
  • 1
    $\begingroup$ (+1). If the already existing question lends itself to less technical answers, it means that quite diverse answers may be valid, hence it might raise the issue of making it a community wiki. Arguably, the question that raised this discussion (stats.stackexchange.com/q/11009/164936) could be a good candidate for this: it is very general, and has currently 18 different answers (though I don't know if we have very specific criteria to decide if a question should be a community wiki). $\endgroup$
    – J-J-J
    Commented Apr 19 at 12:16
  • 2
    $\begingroup$ Another thing: It may be worth noting that when a question is closed as a duplicate, the original poster is not informed that they can edit their question to provide details as to why the existing threads do not answer their question (the message I see is simply "This question already has answers here:"). The terse format of the automatic message may be both intimidating and discouraging to a new user, if they're not subsequently informed of the possibility to detail their request. $\endgroup$
    – J-J-J
    Commented Apr 19 at 12:31
4
$\begingroup$

Both Scortchi and whuber have provided accurate responses.

In the context of the question, whuber again has already provided pointers: firstly, the comment explained there are already comprehensive threads in CV, so the natural expectation is whether OP has gone through those threads and made anything out of those. When OP stated their inability to get any help from the linked thread, whuber further explained what the issue was and sought for clarification.

Finally you responded and when you lamented for the post to be reopened so that you might be able to provide an answer, whuber aptly stated if needed, you could post your answer in the other threads watering down the technicalities.

To sum up what has already been reiterated:

  • If there already exists posts that have received detailed answers, we should cite those posts and close any future queries based on those as duplicate.

  • OP should explain clearly why those queries didn't answer their problem, in fact, they should state explicitly their confusion rather than just saying other posts couldn't provide insight and has possibly spent hours in understanding that. We do have questions that warranted additional clarification of past posts.

$\endgroup$
1
  • 2
    $\begingroup$ (+1). Some nuance though: "OP should explain clearly why those queries didn't answer their problem". But the OP should be informed that they can give such an explanation. (This is in general, I'm not talking about this specific question) . The current message when a question is closed as duplicate is very terse ("This question already has answers here:"), and does not mention the possibility for the OP to improve the question for it to be reopened. I think that's something to keep in mind when interacting with new users, as we can't expect them to know all the ins and outs of the website. $\endgroup$
    – J-J-J
    Commented Apr 19 at 13:03

You must log in to answer this question.

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