Adjusted Wald test not working on Stata 17

[Lora Tupas]

Hi. I am new to Stata and am doing a DHS dataset analysis (2 appended datasets from the same country) generating a new variable which i named "Period" and assigned values: 0=Pre-pandemic, 1=Pandemic.

I have renamed the variables: for example, v012 (current age, continuous variable) was renamed "age".

I have performed as svyset and proceeded to compute mean age by Period for a subpopulation of women who gave birth in the 2 years prior to the survey:

svy, subpop(if childage_mos<=24): mean age, over(Period)

Hhowever, when I performed an adjusted Wald test using:

test [age]Pre-pandemic = [age]Pandemic

Stata returned the following error message:

Equation [age] not found
r(303);

How can i make adjusted Wald test work in this situation?

Could this be due to mistakes in how i svyset the data?

THank you for your help.

Lora

[Andrew Musau]

You are not dealing with multiple equations, but levels of a categorical variable. Use the -coeflegend- option to see how to refer to the coefficients as below.

sysuse auto, clear
mean mpg, over(rep78)
mean mpg, over(rep78) coeflegend
test _b[[email protected]]=_b[[email protected]]

Res.:

. mean mpg, over(rep78)

Mean estimation                             Number of obs = 69

--------------------------------------------------------------
             |       Mean   Std. err.     [95% conf. interval]
-------------+------------------------------------------------
 c.mpg@rep78 |
          1  |         21          3      15.01359    26.98641
          2  |     19.125   1.328768      16.47348    21.77652
          3  |   19.43333   .7560991      17.92456    20.94211
          4  |   21.66667    1.16316      19.34562    23.98772
          5  |   27.36364   2.632913      22.10974    32.61753
--------------------------------------------------------------

. 
. mean mpg, over(rep78) coeflegend

Mean estimation                             Number of obs = 69

------------------------------------------------------------------------------
             |       Mean   Legend
-------------+----------------------------------------------------------------
 c.mpg@rep78 |
          1  |         21  _b[[email protected]]
          2  |     19.125  _b[[email protected]]
          3  |   19.43333  _b[[email protected]]
          4  |   21.66667  _b[[email protected]]
          5  |   27.36364  _b[[email protected]]
------------------------------------------------------------------------------

. 
. test _b[[email protected]]=_b[[email protected]]

 ( 1)  [email protected] - [email protected] = 0

       F(  1,    68) =    0.33
            Prob > F =    0.5696

.

[Carlo Lazzaro]

Lora:
welcome to this forum.
Being totally unfamiliar with -survey- and survey-related methods, I tried to understand the nitty-gritty made by Stata behind the curtain:

. use https://www.stata-press.com/data/r18/stage5a.dta

. svy, subpop(if ylogit <=0): mean stdize , over( yprobit )
(running mean on estimation sample)

Survey: Mean estimation

Number of strata = 3                  Number of obs   =     11,039
Number of PSUs   = 9                  Population size = 529,810.54
                                      Subpop. no. obs =      7,808
                                      Subpop. size    =  375,429.8
                                      Design df       =          6

------------------------------------------------------------------
                 |             Linearized
                 |       Mean   std. err.     [95% conf. interval]
-----------------+------------------------------------------------
c.stdize@yprobit |
              0  |   2.978111   .0167907      2.937026    3.019197
              1  |   2.973214   .0357534      2.885728    3.060699
------------------------------------------------------------------

. mat list e(b)

e(b)[1,2]
     c.stdize@  c.stdize@
    0.yprobit  1.yprobit
y1   2.978111  2.9732138

. test [email protected][email protected]

Adjusted Wald test

 ( 1)  [email protected] - [email protected] = 0

       F(  1,     6) =    0.01
            Prob > F =    0.9068

.

[Lora Tupas]

Thank you so much for your help!!!
the "mat list e(b)" command works...even for svy:logit commands

Truly grateful,
Lora