Interpretation of confidence intervals in both frequentist and Bayesian formalisms (credible intervals) have been a constant and repeated source of queries over the years at CV.

Every now and then, there would be a question by OP having confusion re the interpretations or if not, something related that could be answered by a correct reading of the interpretation of CIs.

Now, over the years, some very well-explained and comprehensive posts have been written as answers to many such queries, which can act as potential faq posts to all other future related queries (trust me, questions on CI ain't going anywhere).

So, would the community be willing to have a faq post on this particular topic? Personally, I feel it's time we have one as that could come handy avoiding the redundant cycle of having the same discussion with minor tweaks.

Anyone interested can check the tag for such posts but kjetil has already linked some in the comment, namely:

Some others being

These are few examples only. There could be others. But the issue is whether we need faqs on this, at the outset.


Thanks to Stephan for recommending to post this at Ten Fold.

  • 4
    $\begingroup$ Mostly, we just need to close Qs as duplicates. You can make a case for a particular one of those threads as the most canonical one. $\endgroup$ Commented Apr 17 at 11:30
  • 1
    $\begingroup$ Sure @gung-ReinstateMonica. But since I am new about how faq works, could we have more than one? Also, I would like to hear what others have in their mind and whether there is a consent on this or not. $\endgroup$ Commented Apr 17 at 11:33
  • 4
    $\begingroup$ Yes, you can have more than one. I created the FAQ tag as a way to facilitate searching for duplicates. The motivation was that some threads are difficult to search for or not highly voted but are repeatedly found to answer new posts. Just as we can list up to five duplicates to any post, we should be able to tag multiple threads as FAQs about the same topic. $\endgroup$
    – whuber Mod
    Commented Apr 17 at 14:52
  • 1
    $\begingroup$ Thanks for the insight @whuber. $\endgroup$ Commented Apr 17 at 15:55

1 Answer 1


This reminds me of a discussion of class imbalance we had two years ago. I like the suggestion I posted there and think something similar could work for confidence intervals: a question that we answer with links to the existing good answers along with some commentary about those answers (think of something like an annotated bibliography).

  • 1
    $\begingroup$ That's not at all a bad idea, Dave. If everyone agrees, we can choose a set of comprehensive threads for the ultimate faq post, replete with additional context. $\endgroup$ Commented Apr 18 at 12:26

You must log in to answer this question.

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