8
$\begingroup$

We have an ambiguous tag that says right in the tag excerpt that it's ambiguous and should better be avoided. @gung moved most of the questions out of it several years ago, but did not empty it completely so it got lots of new threads in the meantime (75 currently). I think this should still be done.

More specific tags are , , , , , and -- all fine tags with clear scopes and excerpts.

Update: We now also have the following synonyms:

I am posting it here hoping that maybe somebody will want to do this task or join the effort.

$\endgroup$
17
  • 1
    $\begingroup$ @gung I retagged some Qs about [discriminant-analysis] but it might make sense to create a synonym [fisher-linear-discriminant]->[discriminant-analysis] for the future. $\endgroup$
    – amoeba
    Nov 5 '17 at 13:50
  • 1
    $\begingroup$ Similarly, I retagged some Qs to [combining-p-values] but it might make sense to create synonyms [fisher-method] and [stouffer-method] (or fishers- and stouffers-?) pointing to [combining-p-values]. $\endgroup$
    – amoeba
    Nov 5 '17 at 13:54
  • 2
    $\begingroup$ I've done 6 now of the ones since I last did it (July '13). I'm thinking we should remove the tag from everything, though. It's best for the tag to disappear, this isn't something people are going to be able to use correctly. $\endgroup$
    – gung - Reinstate Monica
    Nov 5 '17 at 14:33
  • $\begingroup$ @gung Yes, I agree; that is what I meant. $\endgroup$
    – amoeba
    Nov 5 '17 at 16:12
  • 1
    $\begingroup$ @gung What do you think of the synonym suggestions above? $\endgroup$
    – amoeba
    Nov 5 '17 at 16:14
  • $\begingroup$ They're fine w/ me, @amoeba. I don't really know these topics well enough. It seems there were several FLA threads. I don't know FLA vs L / Q DA. Is it worth its own tag? We could also make it a synonym, but maybe there's enough interest on its own. $\endgroup$
    – gung - Reinstate Monica
    Nov 5 '17 at 20:36
  • 2
    $\begingroup$ @gung Nah, it's the same thing. "FLA" is LDA for two classes. We definitely don't need a separate tag for it (but the synonym would be handy). I have some authority as the highest-rep user in discriminant-analysis tag :-) $\endgroup$
    – amoeba
    Nov 5 '17 at 21:01
  • 1
    $\begingroup$ I made a synonym. $\endgroup$
    – gung - Reinstate Monica
    Nov 6 '17 at 1:55
  • $\begingroup$ Great, thanks. I still think that en.wikipedia.org/wiki/Fisher%27s_method deserves a tag (mapped to combining-p-values) as well (see my 2nd comment above). $\endgroup$
    – amoeba
    Nov 6 '17 at 8:34
  • $\begingroup$ There are of course other methods for combining p values which have eponyms so we may be starting down a slippery path if we create tags for any more of them although I suppose if they are synonyms that does not matter too much. $\endgroup$
    – mdewey
    Nov 6 '17 at 9:36
  • $\begingroup$ @mdewey I think it's a question of what eponyms are popular enough to warrant a tag synonym. This is an empirical question that should be easy to answer using our website: a search for "Fisher's method" yields 77 results, for "Stoufer's method" - 28 results. Anything else with >10 mentions? $\endgroup$
    – amoeba
    Nov 6 '17 at 9:53
  • 1
    $\begingroup$ @amoeba, [fishers-method] is a synonym here. $\endgroup$
    – gung - Reinstate Monica
    Nov 6 '17 at 12:43
  • $\begingroup$ Are we trying to eliminate the tag? There are four left now for none of which I can find a good replacement. $\endgroup$
    – mdewey
    Nov 8 '17 at 17:05
  • 1
    $\begingroup$ I noticed that a lot of the [fisher] questions were about fishers-lsd, which is legitimate, so I created that & retagged them, @amoeba. I'll post an answer. LSD is mentioned in various threads on CV, but I found 9 that were sufficiently about LSD that I added the tag. $\endgroup$
    – gung - Reinstate Monica
    Nov 8 '17 at 19:59
  • 1
    $\begingroup$ @gung It looks like almost all of these threads are not really about Fisher's scoring. I noticed one that clearly is and put the tag on it now, but overall I don't think we are missing a lot of questions. $\endgroup$
    – amoeba
    Nov 8 '17 at 20:17
2
$\begingroup$

Years ago I reorganized the [fisher] tag. I created several new tags (e.g., [f-e], [f-i], [f-t], and [r-f]). But I left the tag on ~25 threads that seemed to merit some connection to "fisher", but where idiosyncratic. I edited the excerpt to put a rather specific message that the tag should only rarely be used, and the other tags should be preferred. I think now that that strategy is unworkable and was a mistake. Not surprisingly, perhaps, many people over the intervening years ignored that usage guidance, necessitating this thread.

Over the past couple days, @mdewey, @amoeba (thanks, guys), and I have cleaned up the remaining [fisher] tagged threads. In doing so, we have created and . I also created two synonyms (noted in the body of the question above). After 24 hours, [fisher] should disappear. Let us hope it won't be recreated.

$\endgroup$

You must log in to answer this question.

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