thanks david for your comments.
I guess in order to treat different subsamples of the data as
separate samples one has to assume that they are random and
independent. however, if one deals with sample selection the
question arises whether randomness of a subsample and random
assignment into that sample are the same thing or not. I've kind of
convinced myself that they are not and that even if the heckman
procedure suggests that a sample bias exists the subsamples can
still be random and independent of each other. If one buys this
story one could indeed keep the subsample size constant and
draw from each subsample rather than the entire sample.
concerning the sample size, you are right, the subsample is rather
small and that might be the reason for rather strange s.e. in the
bootstrap. there are some extreme outliers and while around 90 per
cent of the draws yield sensible results these few outliers lead to
serious problems.
but I'm still convinced that there must be some literature on
bootstrapping and sample selection out there.
cheers
axel
Axel Heitmueller,
Centre for Economic Reform and Transformation, CERT
School of Management
Heriot-Watt University
Edinburgh
EH14 4AS
UK
phone +44(0)131 451 3969
fax +44 (0)131 451 3296
[email protected]
www.som.hw.ac.uk/somah3/
*
* 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/
