svy: how to get mean and standard deviation

[Anshul Anand]

Hey all

I am trying to get the mean and the standard deviation from a variable. I know how to do it without svy:

sum lnhourly    if    immigrant==1

Variable        Obs    Mean    Std. Dev.    Min    Max
                        
lnhourlyw        42,245    3.654568    .4636835    2.620039    5.012237

But i am not getting it, if I want to use svy (pweight) command. I can only get the mean, but not the standard deviation:

 svy: mean lnhourly if immigrant==1
(running mean on estimation sample)

Survey: Mean estimation

Number of strata =       1        Number of obs   =     42,245
Number of PSUs   =  42,245        Population size =  3,184,865
                                  Design df       =     42,244

--------------------------------------------------------------
             |             Linearized
             |       Mean   Std. Err.     [95% Conf. Interval]
-------------+------------------------------------------------
   lnhourlyw |   3.610175   .0025817      3.605115    3.615235

[Richard Williams]

Try running

estat sd

after the svy: mean command.

For more, see the postestimation commands for svy.

[Anshul Anand]

Thank you very much, Sir!

[david pinchao]

Hi everybody, after svy: logit, I write mfx. How STATA compute the mean of the varlist un the output of margins after svy:logit?. Exactly how could I get the means of each one of the variables of the margins output if I want to get the means by myself with pencil and calculador?

[Richard Williams]

David, welcome to Statalist.

Unless you have a horribly antiquated version of Stata, you should NOT use mfx. Use margins instead. See

https://www3.nd.edu/~rwilliam/xsoc73994/Margins01.pdf

It sounds like you want to use the atmeans option of margins. I don't recommend it, but you can do it. Use code like

webuse nhanes2f, clear
svy: logit diabetes height weight i.black
margins black, atmeans

[david pinchao]

Good afternoon,
I type this svy,subpop(Urbano):logistic T_Cuenta i.EscoMS i.EscoS i.Ingreso i.Celular i.RegistroIyG
and the output show:" Note: 51 strata omitted because they contain no subpopulation members."
I would like to get the margins with: margins Ingreso, vce(unconditional) or maybe with margins, vce(unconditional)
but I have this answer : missing predicted values encountered within the estimation sample
How could I get the margins in this situation if Iam working with svy,subpop(Urbano):logistic T_Cuenta i.EscoMS i.EscoS i.Ingreso i.Celular i.RegistroIyG ?

[Felix Garten]

Dear all,

I have a very similar question to Anshul Anand, so I thought it makes most sense to post my question in this group. I am trying to calculate standard deviations using svy (pweight). As suggested in this group I should use "estat sd" after running the "svy: mean command". Here is my code:

 svy, subpop(if numparticipants>0 & internetuser==1): mean divopinions

 estat sd

After running estat sd I get the error message: "estat sd is not allowed with poststratification r(322)". Looking through the stata menu I could not find what command I should use for postratifaction weights. Is it simply not possible to calculate standard deviations for poststratification weights?