Jo:
There is no fundamental objection against bootstrapping a marginal effect (see example), though
the random effects parts makes me worried: With bootstrap you assume that a "new" dataset, created
by drawing with replacement N persons from a dataset of N, is a dataset that could plausibly occur
if you drew a new dataset from the population. Do that many times, estimate your
coefficient/marginal effect/statistic in each of these "samples", and you get the sampling
distribution of that statistic. Note that each individual has the same probability of occuring in
the "new" dataset. This assumes that your dataset is the result of simple random sampling. This is
unlikely to be the case if you are using a random effects model. Now part of the attraction of
bootstrap is that you can take the sampling design into account while drawing a "new" dataset and
thus get correct standard errors in non-standard sampling designs. So I wouldn't be surprised that
this is also possible in Stata, but I have no experience with that.
*-------begin example--------
version 9.1
capture program drop bsmarg
sysuse auto, clear
program define bsmarg, rclass
tobit mpg weight, ll(17)
sum weight, meanonly
local xb = (_b[_cons] + _b[weight]*r(mean)-17)/_b[sigma:_cons]
local lambda = normden(`xb')/norm(`xb')
return scalar meff = _b[weight]*(1-`lambda'^2-`xb'*`lambda')
end
bootstrap meff=r(meff), reps(1000): bsmarg
*-------end example--------
HTH,
Maarten
--- jyotsna puri <[email protected]> wrote:
> Does any one know if its possible to bootstrap marginal effects.
> I want to do this for marginal effects from a Random Effects Tobit
> model.
-----------------------------------------
between 1/2/2006 and 31/3/2006 I will be
visiting the UCLA, during this time the
best way to reach me is by email
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting adress:
Buitenveldertselaan 3 (Metropolitan), room Z214
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
___________________________________________________________
NEW Yahoo! Cars - sell your car and browse thousands of new and used cars online! http://uk.cars.yahoo.com/
*
* 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/
