st: bootstrap and conformability error

["sandrine rospabe" <srospstata@hotmail.com>]

Dear statalist,

I am trying to estimate the significance of a few statistics using a 
bootstrap analysis.

I have created a program in which I estimate for 2 different sub-samples 
(respectively 6610 and 1114 observations) a regression with a treatment 
effect (treatreg command) using the MLE and then compute the difference in 
expected earnings between those who receive the treatement and those who 
don't. Having done this for the 2 sub-samples I compute the difference 
between the treatment earnings effect for subsample1 minus the treatment 
earnings effect for subsample2.
Then I use the command bstrap to test for the significance of these 3 
differentials.

My problem is the following: When I set 50 iterations (reps50), the program 
is working. But when I set 200 iterations (which is the number of iterations 
I need), an error message says "conformability error".

I guess the problem comes from the drawing of a  particular sample which 
prevents a certain matrix computation. But I just don't know how to solve 
the problem.

Would someone have an idea of a solution?

Many thanks

Sandrine Rospabe


Dr. Sandrine Rospabé
Université de Rennes I - Faculté des Sciences Economiques
CREREG
7 Place Hoche
35065 Rennes cédex
02-23-23-33-24
06-24-24-07-63

_________________________________________________________________
Découvrez les nouvelles émoticônes animées de  http://g.msn.fr/FR1001/866 
MSN Messenger nouvelle formule

*
*   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/