28
$\begingroup$

Recently I have been reading quite a lot on CV about fixed / random / mixed effects models and related terminological controversies. Unfortunately, the CV tags for the mixed models are a mess, which does not help navigating the questions (which is a pity, given that the topic is confusing enough).

Namely, there is tag (689 1214 questions, nice wiki), tag (95 305 questions, worse wiki), and also tag (28 61 questions, no wiki). I do know that terms fixed and random are used differently by different people, but as far as I understand, mixed is mixed. So I think these three tags should be merged.

I think the master tag should be , and the other two can be merged into it.

The two related tags are called and , so originally I wrote here that the best name for the merged tag would be [mixed-effects-model] (which currently does not exist at all); however, on reflection, "mixed model" seems to be a more standard terminology.


This issue was brought up multiple times on Meta, but stays unresolved.

  • Here I suggested the same thing yesterday and @gung commented that it seems reasonable.

  • In 2012 a bunch of related tags was discussed with the same suggestion about "mixed" tags, and I can see no counter-arguments in the discussion (even in the careful reply of @StasK).

  • In 2011 a related question was asked, and @chl reported in the comments that he merged [mixed-model] and [mixed-effects-model] (the latter tag does not exist anymore).

  • In 2013 @Randel left a comment in the above thread suggesting the same thing as I do here, and @AndyW and @chl agreed.

$\endgroup$
9
  • 3
    $\begingroup$ Agree. In a similar vein, suggest that multilevel (even hierarchical) should be lumped in the mixed effects tag. For a beginner I think the former terms make things easier to understand, whereas the fixed/random terminology can be a bit off-putting at first. $\endgroup$
    – dardisco
    Mar 1, 2014 at 7:59
  • 2
    $\begingroup$ @dardisco: Personally, I tend to agree, but note that yours is a much more contentious suggestion, see e.g. StasK's answer here. $\endgroup$
    – amoeba
    Mar 1, 2014 at 13:52
  • 2
    $\begingroup$ Sensible suggestion. I'd vote to merge into 'mixed-effect'. I think this just reflects the different terminology in circulation more widely. $\endgroup$ Mar 2, 2014 at 18:23
  • 5
    $\begingroup$ @whuber: Now that I reached 1k rep, I can see that this question was upvoted 16 times and never downvoted. To me it looks like a consensus, but the tags are still not merged. What is actually the procedure (if any)? Is there a certain number of upvotes that has to be reached? I am sorry to bother you about that, but I noticed that most of the recent tag synonyms were declared by you. $\endgroup$
    – amoeba
    Mar 8, 2014 at 15:51
  • $\begingroup$ Any ideas on merging the tag wikis? mixed-model is good but, oddly, doesn't make any mention of fixed vs random, like mixed-effects does. $\endgroup$ Oct 23, 2015 at 12:30
  • 1
    $\begingroup$ I will try to merge the wikis, @Scortchi. However, I should say that even though I appear here as the main proponent of this tag merge, I am actually not really familiar with mixed models and my knowledge of econometrics is zero; so this terminology and the differences in definitions confuse me. I am sure many other users here could contribute a much more meaningful wiki edit... $\endgroup$
    – amoeba
    Oct 23, 2015 at 14:28
  • 1
    $\begingroup$ Well at any rate it's better not to lose previous efforts & it can always be improved in future. $\endgroup$ Oct 23, 2015 at 14:36
  • 1
    $\begingroup$ I wonder if we can recruit someone to assist with this before we merge the tags. There really is a difference between the way the names are used in econometrics vs more canonical statistics terminology. (cc @Scortchi) $\endgroup$ Oct 23, 2015 at 16:24
  • 2
    $\begingroup$ @gung, I have taken a careful look at the two wikis, and figured that I can merge them reasonably well. I did it now. Almost no text has been lost, I just copy-pasted and rearranged it (and adapted the excerpt). Just in case, I posted both tag wikis as they were before my edits as an answer here, and then deleted my answer; I believe both you and Scortchi can see the deleted answer. The wiki might not be perfect and it might be a good idea to ask somebody to take a look at it, but I think we can go ahead with the tag merge nevertheless. The wiki can always be edited later. $\endgroup$
    – amoeba
    Oct 24, 2015 at 23:46

1 Answer 1

3
$\begingroup$

& are now synonyms of (the master tag).

$\endgroup$

You must log in to answer this question.

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