7
$\begingroup$

There is a tag which covers Type 1 and Type 2 errors; that is very clear from the excerpt and the Wiki. There is also a tag which seems to deal with Type 2 error specifically (has no Wiki, just an excerpt). I think it is redundant.

So shall we get rid of the tag?

$\endgroup$
10
$\begingroup$

This seems pretty reasonable. I suggest either:

  1. I just make [type-ii-errors] a synonym of [type-i-errors] and merge them, or
  2. make a new tag, [type-i-&-ii-errors], and make both a synonym of that.

Upon reflection and discussion, I think #2 is best.

$\endgroup$
  • $\begingroup$ +1 from the OP. I think [type-i-errors] can be understood to stand for $i=1,2$, so you would not need to create the [type-1-&-2-errors] tag, but perhaps this is not obvious to everyone. $\endgroup$ – Richard Hardy Jun 21 '18 at 11:17
  • 9
    $\begingroup$ @Richard I would find "type-i-errors" to be obscure and confusing. Gung's idea of having two tags (as synonyms of each other and perhaps of yet more) is good, because it exploits the tags already applied by users of the site. Those are the ones most likely to be used by future users in their searching. $\endgroup$ – whuber Jun 21 '18 at 12:02
  • $\begingroup$ @whuber, for clarity, you are arguing for making the new tag i-&-ii, right? $\endgroup$ – gung - Reinstate Monica Jun 21 '18 at 12:42
  • $\begingroup$ To be precise, originally you proposed 1-&-2, not i-&-ii. $\endgroup$ – Richard Hardy Jun 21 '18 at 12:45
  • 1
    $\begingroup$ Gung: I advocate letting tags be determined organically by the proposers of questions, insofar as that is possible. If a tag needs clarification, a good approach is to leave it alone but make it a synonym of another--preferably an existing one. That suggests not creating any new tag in this case. $\endgroup$ – whuber Jun 21 '18 at 12:56
  • $\begingroup$ @whuber I don't understand what suggestion you are endorsing; you wrote that you "would find "type-i-errors" to be obscure and confusing" but that is exactly how the tag is currently called and you said you are against creating new tags in this case. $\endgroup$ – amoeba says Reinstate Monica Jun 21 '18 at 20:56
  • 2
    $\begingroup$ @amoeba Thank you for pointing that out. I had misread some tags (the tag font shows up very small on this monitor for some reason). Without context, "type-i-errors" strikes me as strange because it is reminiscent of a mathematical shorthand for "type 1, 2, 3, ..., whatever errors." But that's me and I have to let the general principle override that: we ought to respect, and try to retain, whatever tags users originally pick. However, there's nothing to prevent us from creating some (or even many) synonymous tags, such as "type-1-errors" or even "type-1-&-2-errors." $\endgroup$ – whuber Jun 21 '18 at 22:26
  • 5
    $\begingroup$ @whuber I see. I must say I do not agree with the general principle you are referring to. The tags are often suboptimal; once a suboptimal tag is created it will get used (remember that it requires some minimal reputation to be able to create a new tag) and there is no mechanism to ensure some sort of competition or survival of the fittest tags. I would agree that we should not delete existing tags (as a "weak form" of your principle). However, in my mind, nothing prevents us to create a new tag X and make an existing tag Y a synonym of X. $\endgroup$ – amoeba says Reinstate Monica Jun 22 '18 at 5:04
  • 1
    $\begingroup$ @amoeba Your thoughts have merit; I find them persuasive. I still would like to hold out for the "weak principle," though, because that's part of what I promised when elected as moderator: to moderate lightly wherever possible. Part of that means accepting the decisions of users whenever they do not clearly violate site policies or norms, even if I might disagree. $\endgroup$ – whuber Jun 22 '18 at 13:14
  • $\begingroup$ Is symbol "&" allowed in a tag name, @gung? I've never seen a tag with this symbol so I suspect that it isn't; if so, my preferred tag name would be [type-i-and-ii-errors]. $\endgroup$ – amoeba says Reinstate Monica Jun 26 '18 at 20:47
  • $\begingroup$ I tried: "&" does not work. $\endgroup$ – amoeba says Reinstate Monica Jun 27 '18 at 11:15
  • $\begingroup$ @amoeba, do you prefer the i's or the Arabic numerals? $\endgroup$ – gung - Reinstate Monica Jun 27 '18 at 14:32
  • $\begingroup$ I prefer i's as in en.wikipedia.org/wiki/Type_I_and_type_II_errors but this is not something that is very important to me. $\endgroup$ – amoeba says Reinstate Monica Jun 27 '18 at 14:38
  • 1
    $\begingroup$ @RichardHardy, are you OK w/ [type-i-and-ii-errors], & making both existing tags synonyms of that? At this point, I think I might just go ahead if you're OK w/ it. $\endgroup$ – gung - Reinstate Monica Jun 27 '18 at 14:47
  • $\begingroup$ @gung, Besides & vs. and in the answer, all is well now (if amoeba is right that & cannot be used). Thank you. $\endgroup$ – Richard Hardy Jun 27 '18 at 15:06

You must log in to answer this question.

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