R help - Sep 2009 - Inverse Mills in clustered (multilevel) cross-sectional panel data

Dear R saviors,
kindly address to this problem, I would really appreciate any takers. I am
trying to resolve this issue of IMR in clustered (multilevel)
cross-sectional panel data for more than two months now,.

The characteristics of my dataset are as follows:
-   some 900 000 individuals
-   total of 60 countries
-   cross-sectional time series at the country level max 10 years, not all
countries included every year

For each country, we have a maximum of 10 cross sectional samples (1 per
year) of at least 2000 adult-age individuals (random selection). But,
individuals are not followed over time. Every year a new random sampling is
carried out.
I am interested in analysing individuals' behaviors in a given economic
activity -- entrepreneurship. To do this, I first need to control for the
fact that some individuals self-select to entrepreneurship. This
self-selection may be influenced by individual-level characteristics (such
as age, gender, education etc) as well as country-level factors (e.g.,
taxation). Because both individual- and country-level factors may drive both
self-selection and behavior, once self-selection has occurred, *multi-level
techniques are required for the selection equation. How to do this in R. *The
results of this selection equation would then be used as a control in the
second stage where an OLS is to be run

Thank you for any suggestions




-- 
Dr.Saurav Pathak
PhD, Univ.of.Florida
Mechanical Engineering
Doctoral Student
Innovation and Entrepreneurship
Imperial College Business School
s.pathak08@imperial.ac.uk
0044-7795321121

	[[alternative HTML version deleted]]