11
$\begingroup$

In R, there are two main packages for estimating mixed effects models: nlme and lme4. They have certain differences. (There are also other packages.)

Now it's 2016, and I'd be interested in whether the situation is still as in 2010/2013. (Specifically, I'm most interested in panel data as per the 2013 question, but I'm also interested in any other changes in the differences between the packages.)

Should I just ask a new question, referring to the earlier ones? Or do something different?

$\endgroup$
  • 2
    $\begingroup$ The problem here, as I see it, is that our standards for what is considered on-topic have gotten much stricter since then. I'm not sure the question your suggesting would be considered on-topic today. $\endgroup$ – gung - Reinstate Monica May 20 '16 at 18:53
  • 5
    $\begingroup$ @gung: I'd take a chance on it. I'd say the question is still on-topic, because "it needs statistical expertise to understand or answer". But I'd now certainly include this little argument to stave off closure, if I do post such a question. $\endgroup$ – S. Kolassa - Reinstate Monica May 20 '16 at 19:55
  • 2
    $\begingroup$ Yeah, I recognize that. I'm OK with grandfathering in the existing questions. I honestly think it's borderline & I'm not sure which way I lean, but my impression is that others have become more hard line than I am, & I suspect people would vote to close. I am actually in the process of formulating a meta.CV Q about our close voting. $\endgroup$ – gung - Reinstate Monica May 20 '16 at 20:03
  • 5
    $\begingroup$ Have you thought putting a bounty on the 2013 question? There is a notice box where you can explain what you are looking for in the newer answers. $\endgroup$ – Andre Silva May 21 '16 at 2:39
  • 4
    $\begingroup$ I'd be quite ok with this type of question, for essentially the reasons that @StephanKolassa cites. Software and programming are an essential part of doing statistics and machine learning, I wish our line was a bit softer. $\endgroup$ – Matthew Drury May 21 '16 at 16:34
  • 3
    $\begingroup$ I think these questions, & others like them, show that software recommendations should be on topic. $\endgroup$ – Scortchi - Reinstate Monica May 21 '16 at 22:43
5
$\begingroup$

Based on the comments, I have set a bounty on the 2013 question, claiming that "The current answer(s) are out-of-date and require revision given recent changes", which is not entirely correct but the closest "bounty reason".

(Now I'd just love for the bounty text to respect the paragraph breaks I included...)

Edit: I just awarded the bounty to Ben Bolker's original answer, although there was not as much updating going on as I had hoped for. However, Ben posted a link to some additional recent thoughts of his, which he may include in his answer at some date.

$\endgroup$
  • 3
    $\begingroup$ It drives me completely nuts when apps destroy my whitespace and end of lines. $\endgroup$ – Matthew Drury May 26 '16 at 19:25

You must log in to answer this question.

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