16
$\begingroup$

I have a background in pure mathematics, and I happen to be teaching mostly statistics at the moment.

When there is a question that I can't answer efficiently myself, or when I need a confirmation, I feel like posting a question here (because that's how I feel about the questions: they're math questions).

The first time I did that, the question was massively downvoted and got no constructive answer or even comment (question now deleted, and I got the answer on my own). The second time, the question was migrated to stats.SE, where again it received no comments or answer (although it did get a number of upvotes there).

So my question is:

What makes a statistics question on-topic on math.SE by nowadays standards?

$\endgroup$
14
  • $\begingroup$ @RoddyMacPhee The deleted one was a general question about what theoretical framework allows you to answer a certain type of questions. Not the kind of thing where you would expect working out. Yes it was correctly tagged. It was up for about a year. $\endgroup$ – Arnaud Mortier Feb 10 '20 at 10:16
  • 4
    $\begingroup$ @RoddyMacPhee I've been here for quite a while now, it was not opinion based. Either you're a specialist and you can answer, or you're not and you move on. I personally could answer that question pretty well now. $\endgroup$ – Arnaud Mortier Feb 10 '20 at 10:30
  • 1
    $\begingroup$ It's impossible to say why your first question was so badly received, when we can't see it. Can't say much about your second question without seeing that one either. So maybe you have to look at statistics questions that haven't been downvoted, that have gotten useful comments and answers, and see what features they have in common. $\endgroup$ – Gerry Myerson Feb 10 '20 at 11:25
  • 1
    $\begingroup$ @GerryMyerson My question here is not about why these two particular questions weree treated as they were (I would make a poor statistician if that was my question :-) ) but rather, really, when is a statistics question not going to be automatically migrated. $\endgroup$ – Arnaud Mortier Feb 10 '20 at 11:28
  • 15
    $\begingroup$ In my view, a statistics question would be "on topic" if it concerned the mathematical side of statistics, and "off topic" if it concerned explicit modeling issues or if it was too heavily data based (or if it was coding based). Clearly there's a subjective component as to where the line is drawn. $\endgroup$ – lulu Feb 10 '20 at 12:50
  • 2
    $\begingroup$ How about if you have to observe the data to answer the question, it's a CV question? $\endgroup$ – Sheridan Grant Feb 13 '20 at 4:17
  • 2
    $\begingroup$ I would say that most of the time, statistics questions are on-topic, but might be better posed and more likely to be answered on stats.SE. A similar situation to super high level math questions and MathOverflow. $\endgroup$ – Alexander Gruber♦ Feb 13 '20 at 7:53
  • $\begingroup$ @AlexanderGruber Well, I've asked three questions there so far and none of them received a satisfying answer. You should really be able to use MSE rep to start bounties on CV. $\endgroup$ – Arnaud Mortier Feb 13 '20 at 22:10
  • 1
    $\begingroup$ For the record: stats.stackexchange.com/questions/447625 ; stats.stackexchange.com/questions/448774 ; stats.stackexchange.com/questions/443885 $\endgroup$ – Arnaud Mortier Feb 13 '20 at 22:10
  • $\begingroup$ Personally I'd hope that most statistics questions are considered to be on topic on math.stackexchange, as statistics is a part of math. When a new stackexchange site is created that focuses on some subdiscipline of math, I think it doesn't mean that questions related to that subdiscipline are suddenly off topic on math.stackexchange. $\endgroup$ – littleO Feb 14 '20 at 20:39
  • $\begingroup$ @littleO I agree. I've only started to be actively interested in statistics about a year ago, but the way it was set aside as a separate SE looks really weird to me. Now I really feel like I'm at the wrong place for asking questions, on either site. $\endgroup$ – Arnaud Mortier Feb 15 '20 at 17:10
  • 4
    $\begingroup$ Some statistics questions are math questions. But if you ask what statistical model should be used to analyze a certain data set, mathematics may fail to bear upon that. And sometimes people on stats.stackexchange.com know far more about a particular math problem the people here do. $\endgroup$ – Michael Hardy Feb 17 '20 at 19:43
  • $\begingroup$ @littleO : About "statistics is part of math", see above. $\quad\uparrow\qquad$ $\endgroup$ – Michael Hardy Feb 17 '20 at 19:44
  • $\begingroup$ Statistics is math, and statistical questions are on-topic here. Unfortunately, the statistical questions that I've seen here were of very low quality; for example, OP asks about using hypothesis testing in a particular problem, but it is clear from the question that he does not understand what is hypothesis testing, and strictly speaking the question does not make sense. $\endgroup$ – kludg Feb 20 '20 at 15:23

You must log in to answer this question.

Browse other questions tagged .