It is nice that people can ask questions and get answers to whatever question they have. However, we tend to close and deleted questions when they are out of the bounds of the website (not interesting to pursue for the sake of creating a database of questions with related answers).

Currently, we do this for duplicates, broad questions, unclear questions, unfocused questions, and off-topic questions.

Among, the off-topic questions we may have self-study questions that get closed because the OP did not do a lot of research themself.

Among these off-topic questions, we could also regard extremely simple/basic/beginner questions (that show signs of low level or little research effort). But it is not so clear how to deal with them.

I am thinking of this question:

Can a regression formula be Y=A - BX

That question is a bit trivial and simple. But it is a bit strange to close it based on little research effort (it is more like the OP is just at a very early stage of research and can't be blamed for asking a basic question). How should we deal with those?

  • 5
  • 2
    $\begingroup$ In addition, note that adding another close reason like that would either be a system-wide change implemented by the developers, or would require us to eliminate one of our existing site-specific off-topic reasons (ie, homework, data request, not about stats / about programming). $\endgroup$ Commented Feb 11, 2020 at 20:01
  • $\begingroup$ @gung I was not so much aiming for an entirely new close category, and more like trying to see whether there are ways to put it under an existing close category (which would either need a different text or there must be some standard way to inform the user about the reason to close). But, more important than just closing questions (I am actually not so much a fan of closing questions), is to find some ways to curate these type of questions as part of the entire collection of 150k questions. If stats.SE would be a museum would it make sense to just mix all these questions together? $\endgroup$ Commented Feb 12, 2020 at 9:36

1 Answer 1


Your example could be a useful question for others at the same basic level—I'll bet some introductory texts don't explicitly point out that coefficients can be negative & explain how to interpret them in that case (which is a further, implicit, question).

I can't see any reason to proscribe low-level/trivial questions, not least because we'd spend more time trying to define "low-level/trivial" (and arguing with askers who dispute the classification of their question as such) than we would answering them. That's not to say anything goes: "How do I calculate the geometric mean?" isn't a good question because we presume anyone who uses the term "geometric mean" has looked up how to calculate it & has a particular difficulty in understanding the procedure that they're not telling us.

One might almost say that "low-level" can only apply to answers, not to questions. In particular, it's not unusual for solutions to an applied problem to vary greatly in their statistical sophistication; but even theoretical issues can be dealt with in more or less depth. It's perhaps only exercises, with a tightly prescribed scope, that can be classified relatively easily (& that suggests a rationale for Math Overflow & Mathematics SE).

  • 3
    $\begingroup$ Unfortunately asking on a site like this is precisely what many people think of as looking up something. I encounter people daily who think the way to learn something is to watch lots of videos rather than find a good text, or who think "research" means using Wikipedia (and I use Wikipedia daily too). But I agree strongly with your stance. There is no future in "low level" as a criterion. There is scope for simple questions that raise interesting issues even about bar charts or means and medians. $\endgroup$
    – Nick Cox
    Commented Feb 13, 2020 at 12:45
  • 2
    $\begingroup$ Glad you agree. And one might almost say that "low-level" can only apply to answers, not to questions. $\endgroup$ Commented Feb 13, 2020 at 15:54
  • 2
    $\begingroup$ You might almost include your no low-level question only answers in your actual answer. When I used to teach students I remember being told there are no bad questions which is more or less the same. $\endgroup$
    – mdewey
    Commented Feb 14, 2020 at 13:40
  • $\begingroup$ Interesting, if it is convergent, a geometric mean (GM) of random data can be calculated. But, I was not aware that there are solutions for all pdf for which convergent GM for random variates exist. Certain ones, yes, but not generally. $\endgroup$
    – Carl
    Commented Feb 18, 2020 at 0:35
  • 2
    $\begingroup$ ""low-level" can only apply to answers, not to questions." I find this a very strong expression which makes me accept the answer. Regarding my question however, it was mostly motivated by some desire to filter the questions somehow (and not to argue about the validity of questions). To some extend one may wonder about the searchability of questions and some way to seperate different levels of questions would be nice (like MathOverflow is seperate from math.SE, and on MO some questions are less acceptable than on the other site, this ain't bad per se). $\endgroup$ Commented Feb 18, 2020 at 22:20

You must log in to answer this question.

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