P value refering to the POmean in teffects ipw

[Rodrigo Badilla]

Hi all

Im just looking for some guidance on how to interpret the p-value on the POmean results in a teffects ipw analysis, example

use https://www.stata-press.com/data/r17/cattaneo2, clear

teffects ipw (bweight) (mbsmoke mmarried mage prenatal1 fbaby medu)


Iteration 0:   EE criterion =  4.460e-23  
Iteration 1:   EE criterion =  2.251e-25  

Treatment-effects estimation                    Number of obs     =      4,642
Estimator      : inverse-probability weights
Outcome model  : weighted mean
Treatment model: logit
----------------------------------------------------------------------------------------
                       |               Robust
               bweight | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-----------------------+----------------------------------------------------------------
ATE                    |
               mbsmoke |
(Smoker vs Nonsmoker)  |  -231.3359   24.05699    -9.62   0.000    -278.4867   -184.1851
-----------------------+----------------------------------------------------------------
POmean                 |
               mbsmoke |
            Nonsmoker  |   3402.175   9.584056   354.98   0.000     3383.391     3420.96
----------------------------------------------------------------------------------------

I Would like know what the p value is referring to:

-----------------------+----------------------------------------------------------------
POmean                 |
               mbsmoke |
            Nonsmoker  |   3402.175   9.584056   354.98   p-value= 0.000     3383.391     3420.96
----------------------------------------------------------------------------------------

Thanks in advance.
Rodrigo.

[Andrew Musau]

3402 is the weight in grams that would occur if none of the mothers had smoked. The null hypothesis is that the parameter equals zero, as is the standard case in other contexts. To see this, just subtract the estimated coefficient from birthweight and re-estimate the ATE.

use https://www.stata-press.com/data/r17/cattaneo2, clear
teffects ipw (bweight) (mbsmoke mmarried mage prenatal1 fbaby medu)
replace bw= bw-  _b[POmean:0.mbsmoke]
teffects ipw (bweight) (mbsmoke mmarried mage prenatal1 fbaby medu)

Res.:

. teffects ipw (bweight) (mbsmoke mmarried mage prenatal1 fbaby medu)

Iteration 0:   EE criterion =  4.460e-23  
Iteration 1:   EE criterion =  2.251e-25  

Treatment-effects estimation                    Number of obs     =      4,642
Estimator      : inverse-probability weights
Outcome model  : weighted mean
Treatment model: logit
----------------------------------------------------------------------------------------
                       |               Robust
               bweight |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
ATE                    |
               mbsmoke |
(Smoker vs Nonsmoker)  |  -231.3359   24.05699    -9.62   0.000    -278.4867   -184.1851
-----------------------+----------------------------------------------------------------
POmean                 |
               mbsmoke |
            Nonsmoker  |   3402.175   9.584056   354.98   0.000     3383.391     3420.96
----------------------------------------------------------------------------------------

.
. replace bw= bw-  _b[POmean:0.mbsmoke]
variable bweight was int now float
(4,642 real changes made)

.
. teffects ipw (bweight) (mbsmoke mmarried mage prenatal1 fbaby medu)

Iteration 0:   EE criterion =  4.464e-23  
Iteration 1:   EE criterion =  1.075e-27  

Treatment-effects estimation                    Number of obs     =      4,642
Estimator      : inverse-probability weights
Outcome model  : weighted mean
Treatment model: logit
----------------------------------------------------------------------------------------
                       |               Robust
               bweight |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
ATE                    |
               mbsmoke |
(Smoker vs Nonsmoker)  |  -231.3359   24.05699    -9.62   0.000    -278.4867   -184.1851
-----------------------+----------------------------------------------------------------
POmean                 |
               mbsmoke |
            Nonsmoker  |  -4.92e-06   9.584056    -0.00   1.000    -18.78441     18.7844
----------------------------------------------------------------------------------------

.

[Rodrigo Badilla]

Thanks Andrew for you reply, I was confused, now I am clear.