Margins command possible after teffects?

[Sara Sakowitz]

Hello,
I am using teffects with pomeans to estimate the potential outcome means of the treatment "elective":

teffects ipwra (died age female hospital, logit) (elective), pomeans

I see this output:

------------------------------------------------------------------------------
             |               Robust
        died | Coefficient  std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
POmeans      |
    elective |
          0  |   .0098819   .0008252    11.98   0.000     .0082646    .0114993
          1  |   .0024719   .0002165    11.42   0.000     .0020475    .0028963
------------------------------------------------------------------------------

I would like to calculate the marginal treatment effect across age. However, the margins post-estimation command does not work after teffects. I was wondering if anyone had any solutions for a workaround?
If I was using a logistic regression, I would do this:

logistic died elective age female hospital, base
margins, by(age elective)

However, I can't find a translation to teffects. Thank you in advance for any ideas!

[Rich Goldstein]

your question is not clear to me; however, the r.pdf manual, under the -margins- entry, shows how, in some circumstances, you can get the same results from -teffects- as you get from -margins-; so, go to the margins manual (type -help margins- and click on the link at the top of the page) and then search for -teffects- using command-F (MAC) or a similar search function if you use another OS

[Sara Sakowitz]

your question is not clear to me; however, the r.pdf manual, under the -margins- entry, shows how, in some circumstances, you can get the same results from -teffects- as you get from -margins-; so, go to the margins manual (type -help margins- and click on the link at the top of the page) and then search for -teffects- using command-F (MAC) or a similar search function if you use another OS

Thank you- to clarify, using teffects I know we can compute the estimated average effect of a specific treatment. But I would like to consider the estimated effect of that treatment, across levels of another variable (i.e. age). Is there a way to do this using options within teffects?