Hi listers,
I have a query related to demographic standardisation after treatreg model.
I want to standardise my dependent variable y on demographic variables age
and sex. One of the explanatory variables, x1, is a binary endogenous
variables, therefore I am using a treatreg model instead of using simple
OLS. I am following the methodology suggested by O'Donnell et al (2008) in
their book 'Analyzing Health Equity using Household Survey Data", though
they only used an OLS model for standardisation. Here is how I am
proceeding:
1. Run the following model:
treatreg y x1 x2 age sex, treat(x1= x2 age sex instrument_1)
2. Get coefficients and then predict yhat by using mean_age and mean_sex
instead of actual age and sex
3. Subtract: yi - yhat + y_mean
My question is: Does the above prediction take into account the Inverse
Mills Ratio (IMR) from stage 1 or should I run a separate -regress- with the
IMR as an additional regressor to get the prediction?
Thank you,
Shehzad
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
