Bugs on Stata 16 - any advice?

[Rana Zahroh]

Dear All,

I am trying to calculate ratio with nlcom after the svy, over(exposure): proportion outcome. It works really well on Stata 15.1 as attached below:





However, this does not work at all on Stata 16 - and how the Stata 16 generated the table also is different with Stata 15.1 (without list of parameters above the table), attached here:



Does anyone have any experience on this? Any solution on how to calculate nlcom after svy, over(exposure): proportion outcome on Stata 16?

Appreciate your help. Thank you.


Best,
Rana

[Rana Zahroh]

Hi All,

I realized now that this is not a bug, just Stata 16 generates the table differently from Stata 15. I tried to generate the nlcom with the command below, as suggested by the manual:

. svy, subpop(if bidx==1 & b5==1): proportion v401, over(birth_place)
(running proportion on estimation sample)

Survey: Proportion estimation

Number of strata =      67            Number of obs   =     17,836
Number of PSUs   =   1,967            Population size = 17,010.939
                                      Subpop. no. obs =     15,028
                                      Subpop. size    = 14,740.883
                                      Design df       =      1,900

------------------------------------------------------------------
                 |             Linearized            Logit
                 | Proportion   Std. Err.     [95% Conf. Interval]
-----------------+------------------------------------------------
v401@birth_place |
         no pub  |   .7844617   .0079373      .7684879    .7996208
         no pri  |   .7804379   .0073014      .7657834    .7944217
       no other  |          1          .             .           .
        yes pub  |   .2155383   .0079373      .2003792    .2315121
        yes pri  |   .2195621   .0073014      .2055783    .2342166
      yes other  |          0  (no observations)
------------------------------------------------------------------

And then:

nlcom v401#0.birth_place/v401#1.birth_place

But still does not work, anyone has any advice on this?

Thank you.

[Richard Williams]

I don’t really know. But if all worked well in Stata 15.1, try preceding your other commands with

version 15.1

[Rana Zahroh]

Thank you! Unfortunately I do not have the 15.1 package as it is my supervisor's not mine. But I have found a way to calculate the nlcom on Stata 16:

nlcom _b[1.v401#0.birth_place]/_b[1.v401#1.birth_place]

Might be useful for other people encountering the same problem.

Also just got a call from technical staff in Survey Design Australia that if we have a command from older version that does not work in newer version, we can just copy paste the commands from the do-file from older version to the do-file in the newer version.

Cheers,
Rana

[Richard Williams]

You don’t need Stata 15,1. Read the help on the version command. It can keep your commands running correctly as you upgrade to newer versions of Stata.

[Rana Zahroh]

Oh my God, I did not know this. I just tried it and it worked really well! Thank you so much for letting me know this, it is (and will be) really helpful for the future!

version 15.1: svy, subpop(if bidx==1 & b5==1): proportion v401, over(birth_place)
nlcom [yes]pub/[yes]pri

Thank you!

Best,
Rana

[Scott Merryman]

The option -coefleg- shows how specify the coefficients in expressions:

--------------------------------------------------------------------------

. webuse nhanes2f

. svy , over(rural): proportion smsa, coefleg
(running proportion on estimation sample)

Survey: Proportion estimation

Number of strata =      31      Number of obs   =       10,337
Number of PSUs   =      62      Population size =  117,023,659
                                Design df       =           31

------------------------------------------------------------------------------
             | Proportion  Legend
-------------+----------------------------------------------------------------
  smsa@rural |
        1 0  |   .3844816  _b[[email protected]]
        1 1  |   .0016318  _b[[email protected]]
        2 0  |   .4193348  _b[[email protected]]
        2 1  |   .2238066  _b[[email protected]]
        4 0  |   .1961836  _b[[email protected]]
        4 1  |   .7745616  _b[[email protected]]
------------------------------------------------------------------------------

[Rana Zahroh]

Thank you so much! I was figuring out the commands by myself and did not know there is an easy way to find out the expressions! This is really helpful!

Best,
Rana