Help with the eform option of regress, geometric means

[Daniela Estrada]

Hello,

I´m trying to obtain the geometric means and 95% CI with the eform option of regress. I´m working with log-transformed biomarkers and 3 eating groups. I have read the "Stata tip 1: The eform() option of regress" from Newson, 2003 but have problems working with it.

This is my syntax:

regress lnglucose w17age w17sex i.eatinggroup, noconstant eform(GM/Ratio) robust

-----------------------------------------------------------------------------
| Robust
lnglucose | GM/Ratio Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
w17age| 1.001799 .0002353 7.65 0.000 1.001338 1.002261
w17sex | .7707707 .0082473 -24.33 0.000 .7547574 .7871237
|
eatinggroup |
2 | 1.003426 .0122826 0.28 0.780 .9796131 1.027817
3 | 1.004425 .0145342 0.31 0.760 .9763085 1.033351
------------------------------------------------------------------------------

My question is, how do I obtain the geometric means of the 3 eating groups from the regress command? Is there another way?

thanks

[Ly Tran]

Hi Daniela,
Have you found the answer? If yes, would you please share with me? I have the same problem as you: don't know how to calculate back the geometric confidence interval for each group.
Thanks.