Please post your tag synonym suggestions as new answers in this thread, one answer per suggestion. Upvote answers where you believe that the suggested tags should be made synonyms, and downvote answers where you believe the tags should remain separate. Well upvoted suggestions will be eventually implemented by the moderators (and then the corresponding answers will be deleted).
-
6@whuber, if this is inappropriate or unhelpful, then I can delete it, but my goal is the opposite: I don't want you (or the other mods) to have to spend time thinking about this & doing a lot of extra research. I'm hoping to bring this to the attention of the community, which, by voting (commenting, etc), will have done all of that for you guys. Nb, w/ respect to lme & multilevel, I'm not sure anyone will have the requisite upvotes--there have only been 3 questions w/ multilevel, according to it's page. – gung♦ Jun 7 '12 at 16:59
-
Comments are not for extended discussion; this conversation has been moved to chat. – whuber♦ Jan 21 '16 at 14:55
estimation (1283) $\leftarrow$ estimators (135)
Perhaps one could argue for some slight difference in meaning, but looking at the threads tagged with these two tags I cannot notice any systematic difference.
This suggestion is triggered by this great thread: Is p-value a point estimate?
-
6At some point I'll have to investigate this. I can see how they could both be valid & useful (which isn't to say that they are in practice). – gung♦ Nov 30 '15 at 0:21
-
Hey @gung, this is a nudge for you to investigate this a little! If you think this is too controversial, I will remove this suggestion and might (or might not) post it separately. I agree that the words have distinct meaning, but looking at the threads in these two tags, I fail to see a consistent difference in usage. But I did not look too long. – amoeba Jan 31 '17 at 22:00
-
I didn't vote on this one but estimators are prescriptions for defining an estimate whereas estimation is the act of making estimates. – Michael Chernick Jan 31 '17 at 22:56
-
@MichaelChernick Sure. But the question is if these two usages are sufficiently distinct to justify two distinct tags, i.e. if it is helpful to have these two tags separate. – amoeba Jan 31 '17 at 23:01
-
@amoeba This one deserves two distinct tags and 4 people voted for it. Nobody voted against ti so far. I have not chosen to vote on this one. – Michael Chernick Jan 31 '17 at 23:05
robust-standard-error $\leftarrow$ neweywest
Strictly speaking, this isn't a synonym. The Newey-West estimator is a specific instance of a robust standard error estimator. That is, this is a set-subset relation. People should feel free to disagree with this suggestion on that basis. However, I question whether they are distinct enough that people are going to want information (i.e., search for) on only one without the other. I further wonder whether there is sufficient interest to justify it: it was created in Sep 2011, and categorizes 27 questions so far, although no doubt there will be more eventually. Robust SE seems to have been created more recently (April 2014), but to have collected nearly twice as many questions.
[robust-standard-error]has 49 threads, with an excerpt and wiki.[neweywest]has 27 threads, at least a couple of which seem to be about robust SEs more generally rather than specifically only about the Newey-West estimator per se. It had neither an excerpt nor a wiki, but I just made an excerpt suggesting people use robust SE instead.
Updated with two more potential synonyms:
robust-standard-error $\leftarrow$ sandwich
robust-standard-error $\leftarrow$ clustered-standard-errors
[sandwich]has 17 threads, with a good excerpt and wiki. One thread has both sandwich and robSE.[clustered-standard-errors]has 79 threads, with a good excerpt and wiki. The excerpt specifies that this is a subset of sandwich estimators. Three questions have both clustSE and robSE; one thread has clustSE and sandwich.
-
This sounds reasonable, but I know almost nothing about these topics and don't think I can judge. In particular, I am wondering if we have other tags that are examples of robust standard error estimators? Or is Newey-West the only such tag? If so, does it correspond to Newey-West being the most popular one? – amoeba Jan 30 '17 at 19:33
-
-
Hmmm, good point @amoeba.
[sandwich]is also a subset (it might be nearly a synonym, but I suspect there might be some kind of robust SE that isn't a sandwich estimator).[sandwich]is older than [rse], has 17 threads w/ a good excerpt & wiki (better than robSE). I could add it as another piece to this answer, but that might make it harder for people to distinguish the parts of the answer they approve vs disapprove, or could add as a separate answer. I would also probably be fine w/[sandwich]as the master. – gung♦ Jan 30 '17 at 20:14 -
Based on what you said, I think either we should map both, NW and sandwich, to robSE, or alternatively leave all of them alone. I cannot say what makes more sense, we should ask people with high rep in these tags. – amoeba Jan 30 '17 at 20:21
-
Perhaps remove this from here & post this (together with [sandwich]) as a separate Meta question? – amoeba Jan 30 '17 at 20:23
-
@amoeba I think as the info tag is written, robust-standard-error could be an umbrella for both sandwich and neweywest, but they are not synonyms as I see it. Of the three, I do see sandwich and RSE often used interchangeably, but it seems Newey West is specifically intended for time-series so not a general approach, but a tool for a specific kind of data analysis. – AdamO Jan 30 '17 at 21:33
-
@AdamO, do you think it's worth making any of them synonyms? I do acknowledge / am somewhat unsure about the umbrella vs true synonym issue. Would it be best to make sandwich a synonym of RSE, & leave NW apart? I could change the NW excerpt I created & we could copy your sandwich wiki to RSE. Or do you think we might as well merge both under RSE? Do you think we should raise this as its own meta.CV thread for fuller & more visible discussion, as amoeba suggested? – gung♦ Jan 30 '17 at 21:42
-
1@Andy, I don't clearly grasp which tags you think we should retain & which should be made synonyms of which. Would your concern about clustered-SEs be allayed by making it a synonym? (Ie, people would type "clustered-standard-errors", it would be displayed in the tags, but then would be silently changed when they hit submit.) – gung♦ Feb 5 '17 at 17:51
-
1Ok sorry, I might not have expressed this very clearly. I think that [sandwich] and [neweywest] should be merged into [robust-standard-errors]. The [clustered-standard-errors] tag I would like to keep given the high demand for this tag and its wider applicability. Your idea of making it a synonym of [robust-standard-errors] would probably also work well though. – Andy Feb 5 '17 at 18:10
-
1@amoeba thanks for your reply. I think also merging clustered-standard-errors into robust-standard-errors would be okay but personally I would prefer if we didn't do it because there seems to be demand for this tag (79 questions). It depends how narrowly we want to define our tag system. In this sense, we could equally move [regressions], [2sls], [logit], [probit] into [estimation]. So I think we should still differentiate a little where this makes sense. – Andy Feb 5 '17 at 23:53
-
1It would have helped to have this moved to a separate thread. Main considerations: (1) Newey-West standard errors are a special type of "robust" standard errors for time series; (2) as @Andy (I think) reminded, not all of the "robust" standard errors are analytical (e.g., bootstrap), although all analytical "robust" standard errors have the sandwich form (after Huber 1967); (3) clustered standard errors are a special type of a sandwich-type estimator / "robust" standard errors that deal with cluster/spatial/sampling correlation. 1/# – StasK Feb 8 '17 at 18:08
-
1... continued: (0) "robust" standard errors are all designed to fix the inverse information matrix var-cov estimate that is sensitive to the model assumption (such as i.i.d. and normality); most of the corrections are originally offered for the linear regression model, and later eventually extended to GLM-type models; (4) a somewhat restricted use of the "robust" standard errors in econometrics literature is indeed in reference to White (1980) standard errors that correct only for possible heteroskedasticity of residuals in linear regression models. 2/# – StasK Feb 8 '17 at 18:15
-
1There are indeed terminological differences between the different disciplines that utilize these estimators. Econometricians call them "robust" (after
robustoption in Stata and package in R); I hate this term because these standard errors are not robust in the Swiss sense of the word to the outliers at infinity. Biostatisticians seem to be using "sandwich" more often in the GEE tradition, although I am less familiar with that literature. Survey statisticians have also been utilizing the same algebra, but refer to these standard errors as Taylor series linearization (Binder 1984) 3/# – StasK Feb 8 '17 at 18:18 -
2@StasK Thanks a lot for this overview, I personally have learnt a lot. I tried googling but e.g. relevant Wikipedia pages are nowhere close to this clarity. My personal impression, based on what you said, is that all of that should be synonymized together under robSE. If terminology differs and topic intertwine, it's more helpful to have everything under one tag so that people could navigate the whole body of these questions. – amoeba Feb 8 '17 at 18:29
-
1@StasK I think the decision should primarily depend on what will help organizing existing threads better. I.e. I don't see how navigating [sandwich] tag alone will help anybody. But one could probably argue that other three tags are fine as they are... My point is merely that given that we have only 165 Qs with any of these four tags, combining them would not be too excessive. – amoeba Feb 8 '17 at 18:41
nonlinear-regression (593) $\leftarrow$nonlinear (221)
"Nonlinear" is a confusing tag name with unclear scope, but its wiki excerpt says:
models which are not linear, ie models with parameters that are not coefficients. See https://en.wikipedia.org/wiki/Nonlinear_regression
and most threads indeed seem to be about nonlinear regression, or closely related. Hence my suggestion.
-
It would of course be best if somebody went through all [nonlinear] threads, retagging ones that are not about nonlinear regression. Then we could make the merge, and then perhaps even delete the synonym mapping. – amoeba Sep 9 '17 at 18:27
Rather oddly, we have tags for both stationarity and non-stationary. I can't see why we need both. I suggest we make [n-s] a synonym of [s]. (I acknowledge this is rather ironic, since they are clearly antonyms, but I think they can be considered 'synonyms' within the context of how that term applies to tag relationships within the SE system.)
[stationarity]has 537 threads right now, with a good excerpt and full wiki[non-stationary]has 229 threads (of which 95 also have[s]), with good a excerpt and wiki as well
-
1
agreement-statistics (x89) and inter-rater (x187) seems to be identical in meaning and use, so one should be synonym of the other.
-
2The more general term seems to me to be agreement-statistics as it might be agreement between thins other than people so I would make that the main term. – mdewey Oct 12 at 15:01
-
1
[as]has both an excerpt & a wiki, whereas[ir]has only an excerpt. There are 20 threads with both tags. – gung♦ Oct 12 at 15:57
Do we need to have separate mixed-model (1937 current tags), and glmm (344 tags) tags ? I would suggest merging glmm into mixed-model
-
2I would keep them separate.
[mixed-model]is fairly generic, in my view, whereas there can be questions that aren't really about mixed-models vs not, but are specific to GLMMs. – gung♦ Aug 21 '16 at 20:25 -
1Wouldn't it be similar to merging [glm] into [regression], which is something that I think most people would oppose? – amoeba Aug 21 '16 at 20:55
-
2
-
1@amoeba The mixed model world is so much smaller than the general regression world that I don't think a comparison with GLM vs LM is warranted – Robert Long Aug 21 '16 at 21:11
-
I have been reading/editing a lot of [glmm] threads recently, and I hate that every time I feel compelled to tag a question with both [glmm] and [mixed-model]. So I began to see your point. +1. Currently, if a Q is about lmer then only [mixed-model] applies, but if it is about glmer then both apply. Whereas questions about lmer and about glmer tend to be very similar. Copy to @gung. – amoeba Sep 5 '16 at 13:03
-
@amoeba, anything that is a mixed model is also a form of a regression model, etc. Do you feel compelled to add those tags as well? There is a fundamental issue w/ the fact that the topics picked out by tags often constitute hierarchically nested concepts. It is clear that GLMM is not a synonym of mixed model (compare them to imbalanced & unbalanced, interaction & moderation, etc). Some questions are focused on the GLMM aspect; I don't personally feel compelled to add anything more & having only
[mixed-model]& not[glmm]is a suboptimal situation, IMO. – gung♦ Sep 5 '16 at 13:30 -
@gung I see your general point about nested tags, but in this specific case tend to disagree. It's good to have this discussion though. I do not feel compelled to add [regression] or [linear-model] because most of the other Qs in these two tags are not about mixed models and will not be helpful. But questions about GLMM tend to be really close to questions about LMM. As I said, there can be very similar questions about lmer and about glmer. If the first is tagged with [mixed-model] and the second with [glmm] then they do not share a single relevant common tag, which is strange and unhelpful. – amoeba Sep 5 '16 at 13:35
-
1@amoeba most of the other Qs tagged w/ MM are not about GLMMs either. It's true that questions about
lmertend to be similar to questions aboutglmer, but that's a side issue. If a question discussingglmerisn't really about theg*aspect, but is primarily about the*meraspect instead, I would agree that that Q should be tagged w/[mixed-model]instead of[glmm]. – gung♦ Sep 5 '16 at 13:42 -
1Hello Robert. With +2 upvotes and -3 downvotes after half a year, it does not look like this synonym suggestion gets enough traction. Do you think we can consider it rejected? If so, consider deleting this answer (all synonym suggestions in this thread that are either implemented or clearly rejected usually get deleted). Just in case you still think this synonym would be a good idea & you want to push forward, you could post this as a new Meta thread to attract attention. – amoeba Jan 30 '17 at 21:54
-
I doubt (just it, don't insist) that we should have tags like [glmm]. This should go as a combination of [generalized-linaer-model] + [mixed-model]. I will request the feature to extend the maximal number of tags to supply from 5 to 7. – ttnphns Apr 6 '17 at 8:05
