[John Mullahy]

Two small requests:

1. It would be handy if the -query- command returned not only current settings but also default settings (at least where practicable).

2. I would be handy if all (instead of just some) -set- commands when executed without an argument would reset the setting to its default. E.g. compare the results of -set cformat- with -set linesize-

An alternative would be to permit -set_defaults- to work on specific settings not just on broad categories (e.g. -set_defaults linesize- rather than -set_defaults output-)

I suspect there are good reasons Stata does not allow these but if so I cannot discern them.

[Richard Williams]

Various articles claim to have developed BIC and AIC and other measures that are legit for complex survey data, e.g.

https://academic.oup.com/jssam/artic...56?login=false

https://www.stat.colostate.edu/grayb...ions/Scott.pdf

There is also the user-written svylogitgof command:

https://www.stata-journal.com/sjpdf....iclenum=st0099

If these commands are statistically valid I would like to see them adopted as post-estimation commands in Stata. (If not valid I'd be curious to hear why.)

[Joseph Coveney]

Currently, bayesmh can call only Stata programs for its llevaluator() and evaluator() options.

I would like to see bayesmh be enabled to directly call Mata functions as likelihood or posterior evaluators in a manner analogous to what ml model is already allowed for its evaluators.

This would allow for at least cleaner code and perhaps for faster execution as well.

It also would be worthwhile to take this one step further, that is, have a Mata class with baysemh functionality, analogous to the relationship between moptimize() and ml model.

[Bruce Weaver]

This is a repeat of something I posted in the Wishlist for Stata 18 thread, but with the SPSS link updated.


As discussed in this thread, there is no direct and easy way to use some of the alternative covariance structures (ACSs) Lesa Hoffman illustrates in Chapter 4 of her 2015 book on Longitudinal Analysis. I would like to see more ACS options for -mixed- similar to the options available in both SAS and SPSS.

• For SAS options, see Table 65.17 here
• For SPSS options, see this page

Cheers,
Bruce

[Clyde Schechter]

It would be convenient if the -label copy- command could copy a value label that exists in a different frame. Currently there are two ways of doing this indirectly that I'm aware of.

1. Copy a variable from that other frame that has the desired label and then just make use of its label, perhaps dropping that copied variable itself if it is not needed. But this only works if the two frames are linkable, which is not always the case.
2. In the other frame, use -label save- to write the label to a temp file, and then come back to the original frame and -do- or -run- the tempfile to create the label here.

Neither of those approaches is particularly onerous, but it would be convenient, and probably more efficient, if the -label copy- command did it directly. Syntactically, I imagine just adding a -from()- option to the existing syntax of -label copy-