7
$\begingroup$

The tags mixed-model and multilevel-analysis seem to be entirely redundant upon a quick glance at the question lists. Should we only have one tag? (I'm not quite sure what the difference between tag synonyms and merging tags are.)

$\endgroup$
  • 1
    $\begingroup$ Maybe we can add mixed-effects-model to that list as well. $\endgroup$ – Andy W Jul 6 '11 at 14:47
  • 2
    $\begingroup$ Let's define first what you or others understand as mixed vs. multilevel models. (There are many definitions and theoretical approaches to such models, but some may also argue this is just reflecting a terminological issue.) If it happens that they refer to different modeling strategies or contexts of use, then it might be best to create synonyms. That being said, I agree that 'mixed-model' and 'mixed-effects-model' should be merged. $\endgroup$ – chl Jul 6 '11 at 19:29
  • $\begingroup$ @chl , I would personally define multilevel models exactly how mixed-models are defined in its tag wiki. I come across the terminology "mixed" far less frequently than "multilevel" or "hierarchical", so I'm less comfortable defining how "mixed" is used in practice. Like I stated previously there doesn't appear to be any difference in the nature of the questions that uses each tag that I can discern from a quick glance of the list. $\endgroup$ – Andy W Jul 6 '11 at 19:37
  • 1
    $\begingroup$ An example of what I had in mind when writing my comment: I would consider that a within-subject design in a psychological experiment (repeated measures are collected on the same statistical unit, with potential crossed factors and/or one or more between-subject factor) falls under the mixed-model approach; in contrast, students nested within schools nested within counties or countries falls under the multilevel or hierarchical approach. I agree this is pure terminology. Doug Bates or JK Lindsey provided more detailed explanations. $\endgroup$ – chl Jul 6 '11 at 19:54
  • $\begingroup$ @chl , I agree that is consistent with what I have encountered (casually I would say I only see mixed used in reference to repeated measures designs like in your example). Perhaps updating the tag wikis would suffice. $\endgroup$ – Andy W Jul 6 '11 at 20:02
  • $\begingroup$ Well, I'm just reading the tag wikis :-) You're right, we need to think of refining them (I can even use a mixed-effect model to estimate an intra-class correlation coefficient :-) $\endgroup$ – chl Jul 6 '11 at 20:07
  • 3
    $\begingroup$ FYI, I merged the 'mixed-model' and 'mixed-effects-model' tags. (/cc @whuber @mbq) $\endgroup$ – chl Jul 7 '11 at 19:18
  • 2
    $\begingroup$ I cannot find the "mixed-effects-model" tag now, and "'mixed-model' currently has no approved synonyms". Maybe we can merge "mixed-model" and "mixed-effect" tags? $\endgroup$ – Randel Aug 28 '13 at 9:52
  • 2
    $\begingroup$ @Randel - that seems reasonable to me, also see the mixed tag that should likely be consolidated. Feel free to either give an answer or edit my question so it gets bumped to the top. If you don't do this I will edit when I get a chance. $\endgroup$ – Andy W Aug 28 '13 at 14:29
  • $\begingroup$ +1. I am also curious about the difference between tag synonyms and merging tags. Or @chl has deleted the "mixed-effects-model" tag? $\endgroup$ – Randel Aug 28 '13 at 14:38
  • 2
    $\begingroup$ @Randel I probably only merged the two tags, without making any synonym. (Usually when merging, an old tag is replaced with a new one; it is possible to make the old tag a synonym when doing this.) I agree with you: the mixed-model and mixed-effect tags should probably be merged. $\endgroup$ – chl Aug 29 '13 at 8:11

Browse other questions tagged .