7
$\begingroup$

I am subscribed to SEMNET, a list on structural equation and latent variable modeling, and it is undergoing some sort of an existential crisis. The talk there is about moving to a forum/website -- it is, amazingly, still a majordomo-operated mailing list.

I'd like to hear some ideas from this group -- see answers below and up/downvote.

See also Why do we have so few questions on SEM and latent variable modeling?... which is becoming relevant.

$\endgroup$
7
  • 6
    $\begingroup$ The sem tag has over 500 questions; this isn't "few". That said, I would love to know what structural equation modeling actually is... $\endgroup$
    – amoeba
    Commented Mar 29, 2018 at 19:34
  • 5
    $\begingroup$ I used to be on SEMNET but left a couple years ago (I no longer use SEM very often and so don't get much out of the discussions). While the scope of SEMNET is technically within our purview, having them move entirely to our site would be a really big change (for them, not us) because SEMNET tends to host a lot of long, rambling, back-and-forth discussions (not to mention arguments). It's not clear to me how smoothly the adjustment to a purely question-and-answer based format would go down for a lot of the SEMNET regulars. $\endgroup$
    – Jake Westfall
    Commented Mar 30, 2018 at 2:29
  • $\begingroup$ @amoeba that looks like roughly a question per day on CV. SEMNET sees maybe 50 emails per day, with ten or so active threads on a given day. Intensity just isn't comparable. For your reference, a good chunk of SEM has been picked up by the Annals-publishing authors and rebranded as "directed acyclic graphs" which were applied to 0/1 variables, as far as I understand DAGs. Beyond that, there's very interesting nonlinear statistics buried deeper, but most of the applied work is drowning in cookbook recipes based on a simulation or two. $\endgroup$
    – StasK
    Commented Mar 30, 2018 at 14:35
  • 2
    $\begingroup$ Individuals active on SEMNET are more than welcome to post here anything that fits CV guidelines. That includes guidelines on behaviour: CV is quite strict on not allowing anything that seems rude or offensive. (I haven't ever looked at SEMNET to know quite how heated it ever becomes.) $\endgroup$
    – Nick Cox
    Commented Apr 3, 2018 at 14:36
  • 4
    $\begingroup$ A thread in 2010 meta.stackexchange.com/questions/72352/… focused on a hypothetical transfer of Statalist to SO. In 2014 Statalist abandoned its listserver format and became a web forum. I can say as one of the people involved that my experiences on SE helped encourage my advocacy of the move to a web forum (traffic exploded, proper posting of code and graphs was an immediate gain, and access to entire threads was improved) and also my advocacy that moving a list wholesale (with its existing mores) to an SE site was not a good option. $\endgroup$
    – Nick Cox
    Commented Apr 3, 2018 at 14:42
  • 1
    $\begingroup$ @NickCox, getting way overheated has become one of the bigger reason SEMNET as a community is looking at other options. They had to block somebody, a pretty senior figure, because their responses to the list have become prohibitively toxic. SEMNETters are now in the process of figuring out whether StackExchange or reddit voting models would become popularity contests (vs. allowing the most scientific answers to raise to the top). You know my opinion of statalist.org. $\endgroup$
    – StasK
    Commented Apr 4, 2018 at 16:49
  • 1
    $\begingroup$ @StasK I don't really know your opinion of Statalist -- the default explanation for a lack of contribution by experienced people is surely that such people are just too busy to solve other people's problems -- but you're welcome to join and post constructive criticism there. To the point, I know even less about SEMNET but I am a little nervous about any idea that this community should provide the discipline SEMNET has not been able to exercise itself! We do have painful experience on CV of people who don't seem to realise -- over extended periods -- that a abrasive style is just not welcome. $\endgroup$
    – Nick Cox
    Commented Apr 4, 2018 at 17:08

3 Answers 3

Reset to default
17
$\begingroup$

SEMNET should just consider posting questions on stats.stackexchange.com, tag them or or whatever, and their experts should register and answer these questions.

$\endgroup$
2
  • 7
    $\begingroup$ Unless there's something particular about their questions (I don't follow that listserv), such as they are all Mplus code questions, SEM is surely within our mandate. I see no reason they shouldn't come here. $\endgroup$
    – gung - Reinstate Monica Mod
    Commented Mar 30, 2018 at 1:09
  • 1
    $\begingroup$ @gung I think the ratio of code to substance is a bit higher on SEMNET than on CV, but not radically higher. $\endgroup$
    – StasK
    Commented Mar 30, 2018 at 14:26
-5
$\begingroup$

SEMNET should form Area 51 and see how that goes. There are good reasons we don't have very many questions on CV: the overlap of people who are interested in topics typical for CV, including machine learning stuff and mathematical statistics, vs. those interested in SEM, isn't that huge, so the SEM community would be better served by a separate website.

$\endgroup$
0
-5
$\begingroup$

SEMNET should neither be here nor in Area 51... they won't make it anyway.

$\endgroup$
3
  • 1
    $\begingroup$ While not 1:1 related, "not being here" is what Stata community have chosen... not very wisely, in my highly opinionated opinion. @nickcox would probably disagree though. $\endgroup$
    – StasK
    Commented Mar 29, 2018 at 19:21
  • 1
    $\begingroup$ I've reomved the commentary on SEM as a field since it doesn't really seem to bear on where SEM fits in SE. $\endgroup$
    – Sycorax Mod
    Commented Mar 30, 2018 at 13:01
  • 1
    $\begingroup$ Most questions on Statalist would be (very) off-topic here. Regardless of that. people can choose where they post. $\endgroup$
    – Nick Cox
    Commented Apr 1, 2018 at 8:34

You must log in to answer this question.

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