Main Page | Report this Page
 
Loading..
   
Science Forum Index  »  Statistics - Education Forum  »  Inverse Mill's Ratio
Page 1 of 1    
Author Message
Bob
Posted: Tue Nov 28, 2006 9:34 am
Guest
Hi,

I am trying to estimate a wage regression with the 2-step Heckman
procedure. I don't have a problem with applying the functions with a
statistics software package and I think I understood why a correction
for sample selection bias is needed but I am really lacking intuition
about what the Inverse Mill's Ratio really is and how it corrects the
bias.

Does anybody know a link where this is explained quite figurative with
an example?

Thanks a lot,
Bob
Back to top
Ray Koopman
Posted: Tue Nov 28, 2006 3:31 pm
Guest
Bob wrote:
Quote:
Hi,

I am trying to estimate a wage regression with the 2-step Heckman
procedure. I don't have a problem with applying the functions with a
statistics software package and I think I understood why a correction
for sample selection bias is needed but I am really lacking intuition
about what the Inverse Mill's Ratio really is and how it corrects the
bias.

Does anybody know a link where this is explained quite figurative with
an example?

Thanks a lot,
Bob

InverseMillsRatio(z) = f(z)/(1-F(z)),
where f is the standard normal pdf and F is the standard normal cdf.
It gives the mean of the upper tail of the distribution.
Back to top
 
Page 1 of 1       All times are GMT - 5 Hours
The time now is Wed Dec 03, 2008 3:54 pm
Science Forum Index

Powered by phpBB & ML | Disclaimer



Contact Us