A recent answer to this question suggested a Bayesian approach to the t-test. I am the sort who doesn't mind every now and again making more trouble for myself by using what I see as the "better" statistical approach even though it isn't common practice in my area yet. However, this Bayesian approach is far enough outside of my area of expertise I would benefit from some elucidation as to why it might be preferred (or not) over a more standard approach. Such clarification could allow me to make the relevant points in a publication in which I might use the method. Is statistical analysis a reasonable place to ask for those kinds of arguments, clarifications and citations?

  • $\begingroup$ Should these sorts of questions continue within the main body of the question that prompted the answer, or should they be established as separate questions? $\endgroup$ Jul 28, 2010 at 0:38

2 Answers 2


As the author of the "Bayesian" response I think that it is extremely reasonable to ask questions on details that you might not understand. But also, people who make these kind of "alternative" proposals should anticipate misunderstandings and difficulties from potential readers, and therefore provide details, or additional sources of information in their initial post.

to put it short: people should definitely ask, but the proponent should avoid the need of others to ask.


I think it is surely ok to ask here for arguments, clarifications and citations.

But, off course don't copy stuff into your publications without the proper cc-wiki / citations. Also it is not smart to put things in publications that you don't understand yourself, so make really sure you get and understand the methods before getting them into papers.

About the question format: try to keep it interesting questions, show where you are busy with and what you have and I think we are very happy to help you.


You must log in to answer this question.

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