# Should we have [bayesian-vs-frequentist] tag?

We have huge number of questions about Bayesian vs frequentist approaches (913 in fast search), so maybe we should have separate tag for it..? I know that we already have [bayesian] and [frequentist] tags, but since endless comparisons, maybe single tag would be helpful? I guess it could make searching (including searching for duplicates) easier. I imagine that this may be controversial, but what do you think?

• Your search finds many answers too; bayesian frequentist is:question yields 287 questions. Aug 16, 2016 at 19:15
• Personally, I'm not sure if I'd want such a tag. I tend to have mixed feelings on those questions. A lot (but not all) of them are unproductive spats. Aug 16, 2016 at 20:10
• @gung I agree with you, but still, they seem to get much attention and upvotes so maybe for some users they are important.
– Tim Mod
Aug 16, 2016 at 20:34
• I usually make a tag whenever I get tired of searching for things to close as dupes. Perhaps this is one such case?
– Sycorax Mod
Aug 20, 2016 at 17:31
• I agree that there would be use for such a tag -- I don't really enjoy arguing about the topic, but as a novice in statistics, the difference for me is really confusing, so I have a lot of questions about differentiating the two, and I always feel bad when I tag a question as "Bayesian" when I really mean "can you compare/explain to me what the difference is between Bayesian and frequentist"? I have a lot of difficulty identifying what is supposed to be a frequentist method or what is supposed to be a Bayesian method or why they aren't compatible, and I think such a tag would be useful. Aug 20, 2016 at 17:39
• @William: While it is indeed a question that pops up frequently, I am afraid there is no satisfying answer to what a frequentist method is and even less to the comparison between those and the huge range of Bayesian answers. If there were, the debate would be closed by now! Aug 21, 2016 at 6:58

Rather not.

While I agree that such a tag could be useful, it seems that the combination of and tags serves more or less the same purpose. There are 95 threads tagged with these two tags together, see https://stats.stackexchange.com/questions/tagged/bayesian+frequentist, and looking through them it seems that most could in principle qualify for [bayesian-vs-frequentist].

That's why I think such a tag is not particularly needed: we already have its functionality.

In addition, its very name would imply some amount of confrontation, and I doubt this would be beneficial for our forum.

• I do not agree with the argument about "confrontational" nature of such tag. In general such questions confront the two approaches so such tag would have descriptive meaning.
– Tim Mod
Aug 23, 2016 at 7:30
• But do you agree with my "main" argument, @Tim? Aug 23, 2016 at 13:55
• I neither agree, nor disagree :) On one hand, you are right that it is redundant. On another, it could help with finding the relevant threads.
– Tim Mod
Aug 23, 2016 at 14:00
• There is a difference between "compare and contrast" and "confront". But on balance I'd concur that the intersection of the existing tags probably suffices for now. I can, though, imagine a point in the future at which we want to distinguish "threads where both frequentist and Bayesian methods are mentioned" (which may be great in number) and "threads where the juxtaposition and contradistinction of frequentist and Bayesian methods forms the very heart of the question" (which may be a minority of the aforementioned cases). In that case a separate tag may become desirable. Aug 29, 2016 at 21:11
• @Silverfish That's a good point. I am not strongly against such a tag; if somebody were to create it and curate it thoughtfully (write a wiki excerpt, populate it with relevant questions, etc.), I would not mind at all. (However, I guess that after my answer got 6 upvotes, it would make sense to first post this more explicitly here as an alternative answer; otherwise such an action might seem as going against the popular vote.) Aug 29, 2016 at 23:11