4
$\begingroup$

I have given Bayesian answers to questions that do not have the bayesian tag. Sometimes my Bayesian answer is accepted and there's not tag to lead anyone there. Is it ever okay to add the tag? Do I even have the authority to do that?

| |
$\endgroup$
  • 3
    $\begingroup$ You don't have enough reputation yet to edit other people's questions. So you can only suggest edits and see if they get approved by the community. If such edits repeatedly get approved then you are doing a right thing. In general, adding tags is often a good idea. In some cases the tags can only relate to answers and not to the question itself (e.g. when the author of the question was not aware of the relevant concept and only the answers point it out to him/her). $\endgroup$ – amoeba Mar 3 '16 at 17:06
  • $\begingroup$ @amoeba Thank you. $\endgroup$ – mef Mar 3 '16 at 17:38
  • 1
    $\begingroup$ @amoeba I'd make that an answer. I think it covers all bases. $\endgroup$ – Nick Cox Mar 3 '16 at 18:49
5
$\begingroup$

I would usually avoid adding the tag unless there was something about the question that made it Bayesian or at the very least where a Bayesian approach was notably natural or obvious.

Note that adding the tag carries a degree of implication that the OP sought (or should have sought) a Bayesian answer -- if a non-Bayesian approach is just as reasonable, then I would not add the tag, because it could imply other answers (including answers that have not yet been offered) have mistakenly taken a non-Bayesian approach.

Similar considerations would apply when adding a tag implying a particular methodology when some other methodology might quite reasonably be used.

There can be something of a grey area; if you're not sure, one thing you could consider is to ask the OP if they're happy to have the tag there.

| |
$\endgroup$
  • 2
    $\begingroup$ "Note that adding the tag carries a degree of implication that the OP sought (or should have sought) a Bayesian answer -- if a non-Bayesian approach is just as reasonable, then I would not add the tag, because it could imply other answers (including answers that have not yet been offered) have mistakenly taken a non-Bayesian approach." I'm not sure about this. The Bayesian tag is there to help searchers track down relevant Qs (and more generally, As). If someone wants to see a Bayesian approach to a problem, then putting the tag there may help them even if not all answers to the Q are Bayesian $\endgroup$ – Silverfish Mar 4 '16 at 20:28
  • 1
    $\begingroup$ On the flip side, the absence of a Bayesian tag isn't an indication that a question can't be dealt with in a Bayesian way, nor is it likely to be used as a search-strategy pointer for someone trying to avoid a Bayesian approach. I generally interpret the tags liberally when reading and searching - issues which a thread raises, not necessarily a unique approach to a problem. I do agree with your last paragraph though - asking the OP is often a good idea. $\endgroup$ – Silverfish Mar 4 '16 at 20:30
  • 1
    $\begingroup$ Contrary answers are given in: meta.stats.stackexchange.com/questions/1525/… This thread argues that sometimes questions can be re-tagged to fit the answers topic-scope. $\endgroup$ – Tim Mar 8 '16 at 14:17
  • 1
    $\begingroup$ @Tim, actually this thread seems to be a duplicate of that one! I vote to close as a duplicate (and perhaps you will want to join?), but given that Glen_b seems to have a somewhat different opinion from the accepted answer there, I would suggest that he either posts another answer there or merges the threads so that both answers appear next to each other. $\endgroup$ – amoeba Mar 8 '16 at 20:25

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