My (perhaps incorrect) impression is that few stats.SE regular users specialize in modern research-level statistics and that the website is most appropriate for undergraduate or masters level content, with a largely applied focus.

Is this a widely known and agreed upon assessment? Is it is right, what is the appropriate venue for modern research level statistics? Is it mathoverflow itself?

Thanks in advance.

  • 5
    $\begingroup$ Math pages rather an exception, than the rule in the SE ecosystem, aren't they? Engineering, Physics, Biology, Economy, etc sites don't have sister sites. $\endgroup$
    – Tim Mod
    Aug 17, 2022 at 13:25
  • $\begingroup$ Hi, @Tim! That's a good point. I wonder if math is also an exception in terms of having more users who engage in modern research. $\endgroup$
    – user365540
    Aug 17, 2022 at 22:12
  • 4
    $\begingroup$ There are some questions (many unanswered!) here that are exceedingly technical. $\endgroup$
    – Firebug
    Aug 19, 2022 at 13:31
  • $\begingroup$ Related: physics.meta.stackexchange.com/questions/9822/… mathoverflow was created before math stackexchange $\endgroup$
    – qwr
    Aug 25, 2022 at 4:25
  • $\begingroup$ There's only one other site pair in the network separating research-level Q&A from the rest: Theoretical Computer Science and Computer Science. Though Network Engineering may be another exception, but it doesn't really have an 'amateur' counterpoint; Super User covers more than just networks. $\endgroup$
    – Glorfindel
    Aug 27, 2022 at 19:56

2 Answers 2


There really isn't one. There just hasn't been a perceived need thus far. You could suggest a new site on Area 51, but we are the site for both applied and research-level statistics. Note that there is a lot of overlap amongst SE sites. A question that goes here can often also be on topic on Stack Overflow, Mathematics, Data Science, Artificial Intelligence, Operations Research, Quantitative Finance, Economics, among possibly others. Math Overflow does have an [st.statistics] tag with 1,666 threads, so I guess people could post there, too.

I'm a bit skeptical, though. I don't think the proliferation of sites on SE is the best way to organize information. Also, people often post what they think are research-level questions that really aren't—an example might be the extremely poor use of our tag.

  • $\begingroup$ I also think Data Science could plausibly be merged into Stats and maybe AI be merged into CS. But we are at the whims of what Stack Exchange Inc wants. $\endgroup$
    – qwr
    Aug 25, 2022 at 4:27

So many people study and apply statistics of some kind that it’s only natural that we would get lots of applied questions.

At the same time, there are excellent statisticians among our ranks, and users should not feel any need to hold back when it comes to posting research-level questions. I personally have posted multiple questions about fairly recent literature, for instance.

Challenge an ICML Paper

Does Multinomial Probability Calibration Consider the Probabilities of the Non-Dominant Classes

If there were to be a “Stat Overflow”, it would come from users wanting to separate out the research-level questions from more routine questions, not because Cross Validated lacks the technical acumen to handle research-level questions.

(And “Stat Overflow” would be quite the name for us to use!)


You must log in to answer this question.

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